(496c) Understanding Unexpected Hydrogen-Bonding Patterns in Proteins with Wavefunction Theory | AIChE

(496c) Understanding Unexpected Hydrogen-Bonding Patterns in Proteins with Wavefunction Theory


Vennelakanti, V. - Presenter, Massachusetts Institute of Technology
Mehmood, R., Massachusetts Institute of Technology
Qi, H., Massachusetts Institute of Technology
Hydrogen bonds (HBs) play a key role in influencing protein-ligand binding, controlling catalytic outcomes, and determining macromolecule structures, which in turn determine their biochemical functions. Cooperative hydrogen bonding, where the donor concurrently acts as an acceptor, is the strongest HB interaction and has been observed among certain amino-acid sidechains in biomolecules. The transient presence or absence of this particular interaction can influence key outcomes in enzymes, such as the reaction rate or selectivity. In order to understand this, it is important to develop an accurate model of how often single hydrogen bond donors/acceptors are formed versus those that involve multiple hydrogen bonds between amino-acid sidechains. We use accurate coupled cluster theory models extrapolated to the complete basis set limit to build a potential energy surface (PES) of these competing types of hydrogen bonds. We use small model systems of N-H···O or O-H···O HBs and their doubly bonded forms to ensure that we can extensively study the relevant PES. We then compare the observations on these systems (e.g., acetamide-methanol) to the representative protein amino acids. We extract representative amino acids with oxygen-atom-containing sidechains (i.e., tyrosine, serine, or threonine) hydrogen-bonded to the nitrogen/oxygen-atom-containing sidechains of asparagine or glutamine from high-resolution (1.5 Å or better) X-ray crystal structures in the Protein Data Bank. Analysis of these high-resolution protein structures reveals that the maxima of O-H···O and N-H···O HB distance and angle histograms coincide with the interaction energy minima obtained from our calculations of model systems. This comparison suggests that we have achieved quantitative prediction of the relative favorability of the hydrogen bonding motifs, despite the approximate nature of the model systems. Comparison of representative systems revealed distinct behavior for interactions between aromatic tyrosine and sidechains versus those with serine or threonine. This study provides essential insight into understanding how noncovalent interactions in enzyme active sites guide catalytic outcomes.