Structure-Based in Silico Study of 6-Gingerol, 6-Ghogaol, and 6-Paradol, Active Compounds of Ginger ( Zingiber officinale ) as COX-2 Inhibitors

Ginger’s ( Zingiber officinale ) phenolic compounds, that are 6-gingerol, 6-shogaol, and 6-paradol, have been proven to show anti-inflammatory activity. The purpose of this paper was to discover whether these compounds are potential to be used as COX-2 inhibitors through structure-based in silico study, which is based on the character of the receptor. Docking was performed to the binding pockets of both COX-1 and COX-2 enzymes, to examine their selective character on COX-2. The binding pockets used in this project were the sites where flurbiprofen and SC-58, crystallized in the enzymes. The scoring value of the interaction of 6-gingerol, 6-shogaol, and 6-paradol with COX-1 were -7.40, -7.27, and -7.20 kcal/mol, while with COX-2 were -7.97, -8.10, and -7.80 kcal/mol, respectively. K i value to COX-1 were 3.78, 4.66, and 5.30 μ M, while to COX-2 were 1.46, 1.16, and 1.93 μ M, respectively. We also calculated the selectivity index value of these compounds to COX-2 and resulted an interval of 0.2 to 0.4, which indicated that all tested compounds could be classified as preferential COX-2 inhibitors. It can be concluded that 6-gingerol, 6-shogaol, and 6-paradol could be developed as COX-2 inhibitors.


Introduction
Cyclooxygenase (COX) enzymes play an important role in inflammatory response, i.e catalyzed the prostaglandins biosynthesis.These enzymes are visualized as homodimers that contain 587 amino acid residues in each chain with molecular weight of 67 230 Daltons.Two isoforms, knows as COX-1 and COX-2, have similar amino acid residues composition and hydrophobic channel as binding pocket (Fabiola et al., 2001).The COX binding pocket contain Val116, Arg120, Val349, Leu352, Tyr355, Leu359, Tyr385, Trp387, Ile523 (for COX-1 or Val523 for COX-2), Gly526, Ser530, and Leu531 (Picot et al., 1994).The most important amino acid residue is Tyr385 that catalyzed the transformation of arachidonic acid to PGG2.COX-2 had larger binding pocket due to the substitution of valine to isoleucine at position 523.COX-1 and COX-2 differ in their distribution and regulatory functions.COX-1 is expressed in cells and normal tissues physiological functions.COX-2 is induced by mediators of inflammation in pathological conditions.Inhibition of both COX-1 and COX-2 with non-selective inhibitors lead to renal and gastrointestinal side effects due to inhibition of COX-1 (Kurumbail et al., 1996).
In this paper, 6-gingerol, 6-shogaol and 6-paradol were investigated whether these compounds are potential to be used as COX-2 inhibitors through structure-based in silico study.Their binding modes were compared with SC-58, a selective COX-2 inhibitor, to determine their selectivity to COX-2.Structure similarity of tested compounds with SC-58 is not considered important due our structure-based method approach which is based on the character of the receptor.

Materials
A Windows XP Professional (2010) computer with Genuine Intel Core TM 2 Duo 2.0 GHz, 250 GB, 800 MHz FSB 2 MB L2 cache and RAM 2.0 GB capacity of memory used in this computational study.The X-ray crystallographic 3D structures of COX-1 (PDB code: 1EQH) and COX-2 (PDB code: 1CX2) were downloaded from online Protein Data Bank (http://www.rcsb.org/pdb).

Macromolecules Preparation
The X-ray crystallographic 3D structures of COX-1 (PDB code: 1EQH) and COX-2 (PDB code: 1CX2) were downloaded from online Protein Data Bank (http://www.rcsb.org/pdb/).Hydrogens were added to all COX enzymes PDB crystal structures followed by calculating their partial charges.SwissPDBViewer v.4.01 (GlaxoSmithKline R&D, downloaded from http://www.expasy.org) was used to separate the monomer of the macromolecules.

Molecular Modeling
All three phenolic compounds of ginger, 6-gingerol, 6-shogaol, and 6-paradol (Figure 1), are hydrophobic (cLog P 3.78 to 4.69), due to their aromatic ring and methoxy group.The similarity of SC-58 and our tested compounds is that they show anti-inflammatory activity as proven by in vitro study.Although, their structures are not similar, they indicate the same hydrophobicity character as shown by their log P values (Table 1).This hydrophobicity is important because COX-2 binding pocket is a hydrophobic channel (Fabiola et al., 2001).Geometry optimization of the ligands was performed by AM1 method because this method is used to predict small molecules with better precision.It is also able to calculate energy generated by hydrogen bonding of the O and N atoms (Marcel Dekker Incorporation, 2004).

Macromolecules Preparation
Flurbiprofen and SC-58 which were co-crystallized in the structure of 1EQH and 1CX2, respectively, were extracted and redocked into their original binding pockets.The RMSD values resulted from these ligands redocking were 1.54 Å and 0.85 Å respectively for flurbiprofen and SC-58, which were less than 2.0 Å, a value typically used in evaluating the success of docking algorithms, indicating the docking methods were valid (Figure 2).There was a shift in the position of the ligands which shown through different amino acid residues in the pocket.The redocking results was categorized as close (Jones et al., 1997).SC-58 was docked on COX-1 and COX-2 and the results were compared to those of the tested compounds.This step was performed to calculate its selectivity index value as a standard comparison for the tested compounds.As we have already recognized that SC-58 is a selective COX-2 inhibitor.The interaction energy (scoring value) of the tested compound with COX-1 and COX-2 are similar, though they showed different Ki values, because the amino acid residues that interacted between the tested compound with both enzymes were different.K i represents binding affinity of the ligand to the enzyme.Smaller K i value shows stronger interaction.The K i values of the ligands interaction with COX-2 are smaller than COX-1 (Table 2), which mean that the ligands better interacted with COX-2.The K i values of 6-gingerol, 6-shogaol, and 6-paradol greater than SC-58 (6.98 x10 -3 μM), indicate that these compounds would probably require larger doses to obtain the same effect as SC-58.

Figure 2 .Figure 4 .
Figure 2. Redocking of (a) flurbiprofen into the binding pocket of COX-1 and (b) SC-58 into the binding pocket of COX-2.Ligands are visualized by stick and ball model.Green lines indicates hydrogen bonds which are formed between ligand and the amino acid residues in the binding pocket of COX-1 and COX-2

Table 1 .
Analysis of ligands (calculated by using Portable HyperChem Release 8.0.7)