Diego Caballero Orduna

Despite the abundance of crystallographic and structural data and many recent advances in computational methods for protein design, we still lack a quantitative and predictive understanding of the driving forces that control protein folding and stability.   For example,  we do not know the  relative  magnitudes  of the  side-chain entropy, van der Waals contact interactions, and other enthalpic contributions to the free energy of folded proteins. The inadequacy of current computational approaches to the analysis and design of protein structures  has hampered  the  development  of many novel  therapeutic  and diagnostic  agents, and is arguably one of the  biggest open challenges in biophysics and biochemistry.

My work will build on the pioneering work of Ponder and Richards in the 1980s, who while at Yale demonstrated that steric constraints and packing criteria in protein interiors were the most stringent criteria in determining protein conformations. I will present my work developing and using a sterically centered molecular dynamics force field to study amino acid conformations.  I specifically use this force field to model hydrophobic amino acids and study the fundamental driving forces that determine amino acid side chain conformations.

In my research,  I have employed  a hard-sphere plus  stereo-chemical  constraint molecular dynamics model. I have complemented this approach with other numerical and computational techniques such as approximate Markov state models to predict and rank amino acid conformations in different environments.

This dissertation presents three separate but related computational studies.  In the first one, I present an analysis of the equilibrium backbone conformations that the amino acid Alanine can take.  I also study the inter-conversion mechanisms between these equilibrium states and compare my predictions with crystallographic data.

I then generalize the method to study side-chain dihedral angle equilibrium con- formation states and the different transition mechanisms between them in the amino acids Leucine and Isoleucine. I analyze bond and dihedral angle correlations  and predict  a novel  transition  method, which is  consistent  with  crystallographic data, between different side-chain conformations.

I then employ my hard-sphere plus stereo-chemical constraint model to comprehensively study and predict the side-chain dihedral angle distributions of eight different hydrophobic residues in the context of high resolution protein crystal structures. I  distinguish  between  local and environment  effects  and quantify how well  we can predict the conformations that these amino acids can take in protein environments.

This thesis work will not only lead to a more fundamental understanding of the underlying physical forces behind amino acid conformations.  Future progress built on this  work will  also enable  the  reliable  design of specific protein-ligand motifs, the development of efficient computational methods to rationally re-design protein cores and interfaces with tunable stabilities, specificities and affinities, and numerous applications in biomedicine.