Modeling Chiral Recognition Between Amino Acids and Vaterite Surfaces

Accurate models of molecular recognition between biomolecules and mineral surfaces help to understand biomineralization processes, treat biomineralization related diseases such as the osteomalacias, and develop novel synthesis techniques for next-generation materials.  Single amino acids have demonstrated chiral recognition on mineral surfaces, and their study provides insight into the behavior of proteins and peptides on those same surfaces.  A better understanding of molecular recognition between amino acid building blocks and prebiotic minerals may also elucidate mechanisms for the coevolution of life and minerals on earth.  Here we demonstrate how computational modeling with the RosettaSurface algorithm, combined with experimental data from in vitro crystal growth experiments, is used to study chiral recognition between amino acids and mineral surfaces.  We computationally docked amino acid enantiomers to an array of vaterite surface models in order to propose and test different mechanisms to explain experimentally observed chiral crystal morphologies.  The energy landscape for an amino acid near an ionic crystal is extremely rugged, necessitating Monte Carlo sampling techniques capable of overcoming energetic barriers.  The RosettaSurface algorithm overcomes these barriers by generating O(10⁵) independent structural models using docking, side chain packing, and energy minimization.  We have also developed a method for reflecting chiral amino acids across crystallographic mirror planes using homogeneous transforms.  Our simulations suggest that orientationally dependent hydrogen bonding between amino acids and mineral surfaces is critical to achieving chiral selectivity.  In addition to providing a potential explanation for the observed chiral morphologies, our docking studies also offer insight into the structure of the vaterite mineral, a topic of heavy debate for half a century due to multiple conflicting interpretations of vaterite’s diffraction data.