Computational Drug Discovery and Design

(backadmin) #1

induces inflammatory responses. Various anti-inflammatory drugs
like Bextra, Vioxx, Celebrex, and many more were designed to
inhibit COX-2 activity, but most of them additionally interact
with COX-1 and induces gastrointestinal and cardiovascular side
effects. Therefore, need for development of novel anti-
inflammatory drug molecules with high selectivity toward COX-2
emerges. To accomplish this, computational SBDND approach was
used by Dhanjalet al.[24].
SBDND approach requires structure of the target molecule and
small compound library. Structures of human COX enzymes were
not available in PDB. MODELLER v9.11 was used to model
COX-1 and COX-2 structure based on homology with template
molecule prostaglandin H2 synthase-1 of Ovis aries (PDBID:
1CQE) and Cox-2 ofMus musculus(PDBID: 1PXX). Modeled
protein structures were stabilized by 15 ns long molecular dynamics
simulations (MDS) in aqueous environment using Desmond soft-
ware of Schrodinger suite [61]. An average representative structure
was obtained from stable trajectories of simulation. The next
requirement was to generate small compound library, which were
accomplished using LigBuilder. 3D structures of eleven COX-2
inhibitors were obtained from NCBI-PubChem database to pre-
pare seed structures usingextractprogram of LigBuilder. Thirty-
seven seed structures generated byextractwere then linked and
grown bybuildprogram into 1135 and 1522 potential lead mole-
cules respectively, drug-like properties of which was simultaneously
evaluated by software. Similarities between various derivatives were
found by calculating Tanimoto coefficients, which gives a set of
2657 unique drug-like molecules, out of which 35 top scoring
molecules were selected to check their binding affinity with
COX-2 protein.
Ligand molecules were prepared using LigPrep and protein
structure was prepared using protein preparation wizard of Schro-
dinger suite [62]. Glide XP docking protocol was used to dock
ligand molecules at COX-2 active site and compounds showing
binding efficiency greater than 8 kcal/mol were selected and
docked against COX-1 active site [63]. Two compounds showing
least binding affinity with COX-1 were selected for further analysis,
namely, C_773 (IUPAC name:5,5-dihydrogenio-3-[(1Z)-1-
[4-({3-hydroxy-4-[hydroxy(λ3-oxidanidylidene)methyl]phenyl}
carbamoyl)phenyl]prop-1-en-1-yl]-1H-1,2,4-triazol-2-ium) and
C_997 (IUPAC name: (3R)-3-carbamoyl-5-[(1Z)-1-{4-[(4-car-
boxy-3-hydroxyphenyl)carbamoyl]phenyl}prop-1-en-1-yl]-3H-
1,2,4-triazol-1-ium). Molecular dynamic simulations of both
docked complexes were carried out to get more realistic insight
into the protein–ligand interactions. Both ligands were interacting
with conserved residues that play a key role in reaction mechanism
of COX-2 enzyme, thereby suggesting their potential to specifically
block the function of COX-2 isoform of COX enzyme (Table2).


Fragment-Based Ligand Designing 135
Free download pdf