What can simulation tell us about protein-nucleic acid interactions? Successes, failures, and our approach to overcome the deficiencies

COMP 31

Thomas E. Cheatham III, tec3@utah.edu, Departments of Medicinal Chemistry and of Pharmaceutics and Pharmaceutical Chemistry, University of Utah, 2000 East, 30 South, Skaggs Hall 201, Salt Lake City, UT 82117
Molecular dynamics and free energy calculations have proven to be very powerful tools for giving insight into nucleic acid and protein structure, dynamics and interactions. As computer power has increased, the reliability and predictability has improved, we've been able to approach more biologically relevant time scales, and we've been able to provide detailed atomistic insight into protein-nucleic acid interaction, hydration and dynamics. We have also exposed serious deficiencies in the commonly applied force fields and serious sampling problems. These deficiencies, such as minor deficiencies in the nucleic acid backbone parameters and mis-balance in ionic interactions, lead to melting of characteristic RNA structure, population of anomalous backbone conformational states, and incorrect interaction of drugs or proteins with nucleic acids. To better understand these deficiencies, we have started to explore simulation of both loop and protein-nucleic acid decoy sets, nucleic acid structures for which detailed NMR data is publicly available, and putative drug-nucleic acid interactions with a goal to improve the Amber force fields for nucleic acids.