Abstract

Short hydrogen bonds, which have the heteroatom distances below 2.7 Å, occur extensively in small organic molecules and biological macromolecules. We have recently conducted a statistical analysis of the Protein Data Bank to elucidate the structural and chemical features of short hydrogen bonds and developed machine learning models to predict their occurrence in proteins. From electronic structure calculations, we show that these compact structures exhibit considerable quantum mechanical characters and share common features in their proton potential energy surfaces. We have further carried out first principles simulations on a set of model molecules that mimic these biological short hydrogen bonds and revealed how electronic and nuclear quantum effects promote the sharing of the proton in the hydrogen bonds and lead to distinctive 1H NMR chemical shifts. These findings will facilitate the investigation of the structure and functional roles of short hydrogen bonds in biological systems.