Page 110 - 76_02

ÁLVARO CORTÉS CABRERA Y CRISTINA RUEDA PÉREZ AN. R. ACAD. NAC. FARM.

Analyzing the docked conformations, we identified three zones
where hydrogen bonds can be found. The first one corresponds to
the residues of Arg81 and Gln40 where ketone groups of the ligands
tend to interact, the second one is placed near residues Gln111 and
Thr205 and bind with hydroxyacethyl chains. In the last one, Leu33
and Asn34 interact with hydroxyl group in the C ring of training
set molecules (Figure 1 and 2). Hydrophobic interactions in the
pocket are due to Met30, Phe31, Leu33, Val41, Trp70, Met71, Met74,
Met115, Tyr210, Leu219 and Ile 22 that surround the ligand and
form the cavity in which this binds. For the test set, the distance and
nature of the interactions are summarized in the Table 2. Apart from
the residues identified in the training set, new binding modes were
found that implies new hydrogen bonds with Cys202, Leu198 and
Met30 backbones.

3.2. Molecular dynamics

MD trajectories were inspected to check the stability of com-
plexes, using de RMSD values of Ca atoms compared to the initial
conformation. After 60 ps of the production run, the complexes were
found stable in all cases, with values of RMSD ranging from 1.29A
± 0.21 to 1.58A ± 0.22 for the whole simulation. Despite of this short
time needed to stabilize complexes, first 100 ps of the MD run were
discarded as equilibration time in order to achieve even higher levels
of confidence in the position of the ligands and the aminoacids of
the binding pocket.

Hydrogen bonding during MD simulation between ligand and the
surrounding environment were analyzed with g_hbond (GROMACS
Hydrogen bond analysis tool) to check the stability in time of the
docking results. In addition, energy values of complexes and free
ligand forms were extracted and considered for LIE analysis.

230