Kaelin Wilby Tsai November 9, 2018

Docking of Potential Inhibitors of the Human Y-family DNA Polymerase POLK

Introduction

Objectives
The objective of this study is to perform a local docking experiment with potential inhibitors
as the ligands and Pol kappa as the protein using AutoDock4 and AutoDock Vina. To measure
the effectivity of the drug binding, the dissociation constant is determined, along with the
binding energy. The docking score and location of the binding pocket with its corresponding
receptor residues are also shown. The results obtained from the two methods shall be evaluated
through RMSD values and both will be compared to the given literature as well.

Methodology
Given the 2D structures of the inhibitor molecules, their minimized 3D structures were
rendered using Maestro. Next, AutoDockTools was used to prepare the protein and ligands for
docking. Polar hydrogens were added to PolK while the 4 ligands were subjected to the
automatic addition of Gasteiger charges, rotatable bonds and number of torsional degrees of
freedom, and number of non-polar hydrogens merged. The protein was then set to be a rigid
macromolecule while several map files for the inhibitors were produced. The search space was
set afterwards wherein the number of points in each dimension was 60 and centered at the known
ligand-binding residues (47.65, 19.46, 70.18). Once the rigid receptor, search space, and flexible
molecules were set, AutoGrid was run to generate a log file output. The search method used in
AutoDock4 was genetic algorithm and default parameters, wherein the output was a docking
parameter file with instructions for a Lamarckian GA docking. The results were then determined
and visualized with AutoDockTools again. For AutoDock Vina, the same input files were used
and the program was run in the terminal. The results were presented in the log output file.

Results and Discussion

The binding affinities from AutoDock Vina
Possible Sources of Deviations:
1. Minimization: Only the default minimization function was used in Maestro
2. Grid center may have been different, despite having known the residues surrounding the
binding pocket; number of points of the grid box may have also been different
3. Only Polar hydrogens were added to the molecule/protein
4. Torsions/number of torsions of the ligands may have been modified after the automatic
designations
5. Flexible residues may have been added in the receptor
6. Different settings for the search parameter, docking parameter, and type of genetic
algorithm used
7. AutoDock Vina and AutoDock4 both report vastly different values.