The influence of simulation conditions in molecular dynamics investigations of model beta-sheet peptides

The influence of simulation methods, cutoff based and particle mesh Ewald (PME) on the accuracy by which experimentally derived nuclear Overhauser effect (NOE) data are reproduced, has been investigated using 500-ns-long molecular dynamics simulations on a model beta-sheet peptide in explicit solvent. The structural and conformational features under the different conditions were evaluated in terms of flexibility, secondary structure content, hydrogen-bonding pattern and percent of native contacts as a function of time. It was found that the different simulation methods strongly influence the dynamics of the peptide, confirming previous observations based on ideal peptide models simulated for much shorter times. Moreover, the results of our simulations prove once more that it is necessary to reach extremely long time scales to obtain enough statistics to accurately reproduce experimental NOE restraints even in the case of the PME method, despite its tendency to the stabilization of conformations which are structurally closely related to the ones derived through experiment. Possible implications regarding the stabilization and folding mechanisms, together with their relationship to the experimental study of peptide models, are discussed