A comparison of methods for calculating NMR cross-relaxation rates (NOESY and ROESY intensities) in small peptides

Three methods for calculating nuclear magnetic resonance cross-relaxation rates from molecular dynamics simulations of small flexible molecules have been compared in terms of their ability to reproduce relaxation data obtained experimentally and to produce consistent descriptions of the system. The importance of the accuracy of the simulation versus the amount of sampling of phase space has also been assessed by comparing different length simulations performed with different time step schemes. A nine-residue peptide from the protein HPr of E. coli was used as a test system. The work shows that, in this case, single conformations or a limited ensemble of configurations are insufficient to properly describe the behavior of the peptide and that different approaches to incorporate molecular motions lead to significant differences in the cross-relaxation rates calculated. The correlation between the cross-relaxation rates calculated from simulations performed with different time step schemes was high and increased with increasing simulation length indicating that the extent of sampling is more important than the details of the atomic motion