Gramicidin A (gA) channels provide an ideal system to test molecular dynamics (MD) simulations of membrane proteins. The peptide backbone lines a cation-selective pore, and due to the small channel size, the average structure and extent of fluctuations of all atoms in the peptide will influence ion permeation. This raises the question of how well molecular mechanical force fields used in MD simulations and potential of mean force (PMF) calculations can predict structure and dynamics as well as ion permeation. To address this question, we undertook a comparative study of nuclear magnetic resonance (NMR) observables predicted by fully atomistic MD simulations on a gA dimer embedded in a sodium dodecyl sulfate (SDS) micelle with measurements of the gA dimer backbone and tryptophan side chain dynamics using solution-state (15)N NMR on gA dimers in SDS micelles (Vostrikov, V. V.; Gu, H.; Ingólfsson, H. I.; Hinton, J. F.; Andersen, O. S.; Roux, B.; Koeppe, R. E., II. J. Phys. Chem. B2011, DOI 10.1021/jp200906y , accompanying article). This comparison enables us to examine the robustness of the MD simulations done using different force fields as well as their ability to predict important features of the gA channel. We find that MD is able to predict NMR observables, including the generalized order parameters (S(2)), the (15)N spin-lattice (T(1)) and spin-spin (T(2)) relaxation times, and the (1)H-(15)N nuclear Overhauser effect (NOE), with remarkable accuracy. To examine further how differences in the force fields can affect the channel conductance, we calculated the PMF for K(+) and Na(+) permeation through a gA channel in a dimyristoylphosphatidylcholine (DMPC) bilayer. In this case, we find that MD is less successful in quantitatively predicting the single-channel conductance.