Molecular modeling of biological systems : from chitinase A to Z-DNA

Citation

Brameld, Kenneth A. (1999) Molecular modeling of biological systems : from chitinase A to Z-DNA. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/mvyy-4570. https://resolver.caltech.edu/CaltechTHESIS:10192009-095619938

Abstract

Quantum chemical methods and molecular dynamics simulations are used herein to address interesting problems associated with chemical systems of biological relevance. Two such systems are investigated: The mechanisms of family 18 and family 19 chitinases and the development of a force field for simulations of nucleic acids from first principles calculations.

Chitinases catalyze the hydrolysis of chitin, a β(1,4)-linked N-acetyl-glucosamine polymer. Family 18 and family 19 chitinases are glycosyl hydrolases with different structures and mechanisms. Using a combination of quantum chemical and molecular dynamics methods, several interesting and unexpected features of the hydrolysis mechanisms of chitinases were discovered. Family 18 chitinases induce substrate distortion forcing the N-acetyl-glucosamine sugar bound at subsite -1 to adopt a boat conformation. Protonation of the β(1,4)-anomeric oxygen leads to spontaneous bond cleavage and the formation of an oxazoline ion intermediate. The oxazoline ion is stabilized through anchimeric assistance from the neighboring N-acetyl group. In contrast, family 19 chitinases do not induce substrate distortion and utilize an oxocarbenium ion intermediate. The first of two acidic residues in the active site serves to protonate the β(1,4)-anomeric oxygen while the second acidic residue stabilizes the oxocarbenium ion through a conformational change within a flexible loop of the enzyme. The second acidic residue also coordinates with and activates a water molecule for nucleophilic attack at the Cl' anomeric carbon to complete the hydrolysis mechanism.

For molecular dynamics simulations of biomolecules, it is desirable to use accurate potential energy functions (force fields) which are also generic enough to be parameterized for most any conceivable molecule. A hierarchical approach is undertaken herein to achieve this flexibility and accuracy. To begin, a rule based force field (UFF) forms the foundation upon which additional parameters are added so as to reproduce structural and energetic properties important for nucleic acids. The specific substructures within nucleic acids which require additional parameterization are the phosphodiester backbone, sugar ring pseudorotation, glycosidic bond and base pair hydrogen bonding. The potential energy surfaces for each of these substructures are determined from high level quantum mechanical calculations and the force field parameterized to reproduce these results.

Thesis Files