Ab-initio-based computation of physico-chemical properties of iron proteins and pKA values of organic molecules

Yutao Yue, Purdue University

Abstract

Quantum mechanical based ab-initio methods have remarkable accuracy and versatility, but are somewhat limited in their application to large (bio)molecular systems due to their computational cost. In this work, towards overcoming this problem, three approaches were taken. The first is to study only the most biochemically relevant part of a metalloprotein, namely the metal-containing active center. In this way, the electronic structure of the iron active center of the enzyme homoprotocatechuate 2,3-dioxygenase was extensively studied using density functional theory. Some connections between electronic structure, geometric configuration, and biochemical reactivity were established. The second is to break a large system into small ones. Based on this idea of "divide-and-conquer" and a program implemented by a previous member of the Rodríguez group, Dr. Teepanis Chachiyo, the method was applied to the NO-bound nitric oxide reductase protein. The optimized geometries obtained with this divide-and-conquer method agreed well with conventional methods of optimization but expended only a fraction of the computational cost. The third is to combine quantum mechanical based methods and empirical methods. To accurately predict molecular pKa values, ab-initio methods were used to handle the solute molecule whereas continuum solvent models were used to describe the solvent. A flexible script that can adapt to various protocols was developed. The dependence of calculated p Ka on various computational parameters and the origin of errors were carefully studied. Some improvements in pKa calculations were proposed and implemented and the remaining errors were analyzed.

Degree

Ph.D.

Advisors

Rodriguez, Purdue University.

Subject Area

Molecular physics|Biophysics

COinS