DOI
https://doi.org/10.25772/RBX5-HP72
Defense Date
2013
Document Type
Dissertation
Degree Name
Doctor of Philosophy
Department
Pharmaceutical Sciences
First Advisor
Glen Kellogg
Abstract
The overall aim of this work was to apply HINT, an empirical scoring function based on the understanding of hydrophobicity, to analyze and predict the binding affinities and biological activities of colchicine-site anticancer agents. The second, concurrent aim was to improve the scoring function by incorporating tautomerism within the modeling process. Our belief is that proper evaluation of tautomeric forms for small molecules will improve performance of virtual screening. The novel pyrrole-based compounds targeting the colchicine site were docked into the receptor using HINT as a rescoring function. Two distinct binding modes dictated by the size and shape of a subpocket were predicted to differentiate the highly active compounds from the weak ones. Of the residues predicted to participate in binding for the active binding mode, Cys241β was revealed to form a weak but critical hydrogen bond with the ligand. A larger collection of colchicine-site agents, biologically tested in the same laboratory including our pyrrole-based compounds were subject to 3D quantitative structure-activity relationship (QSAR) study. Using results on docking the pyrrole compounds as a guide, relative binding poses and QSAR models were built to facilitate ligand design and optimization. A new 3D modeling approach was introduced to visually highlight the unique features of highly active compounds and the commonality of all compounds in the dataset using HINT maps and successfully tested on the colchicine-site agents. These results will provide valuable guidance in the future design and development of new colchicine-site agents. To incorporate tautomerism within HINT, we proposed and developed two workflow approaches: a general search tool using a simple and intuitive algorithm analyzing hydrogen shift patterns to identify and enumerate tautomeric structures, and a database that contains commonly observed tautomeric structures. The first approach was designed for small-scale docking studies and the second approach was designed for large-scale virtual screening. The tautomer module in HINT will give more accurate modeling results when the compound encountered is able to tautomerize.
Rights
© The Author
Is Part Of
VCU University Archives
Is Part Of
VCU Theses and Dissertations
Date of Submission
May 2013