Classification and Prediction of Low-Energy Membrane Protein Helix Configurations by Replica-Exchange Monte Carlo Method

Abstract

The effectiveness of our classification and prediction method for transmembrane helix configurations of membrane proteins by replica-exchange simulations is tested with glycophorin A transmembrane dimer. Replica-exchange simulations can sample wide configurational space without getting trapped in local-minimum free energy states and we can find stable structures at low temperatures. We classify low-energy configurations into clusters of similar structures by the principal component analysis. These clusters are identified as the global-minimum and local-minimum free energy states. Our classifications revealed that there are only two major groups of similar structures in the case of the simulation with the dielectric constant e= 1.0 and five such groups in the case of e= 4.0. The global-minimum free energy state in the case of e= 1.0 is very close to the structure of the NMR experiments and the prediction was successful, while in the case of e= 4.0 not the global-minimum but a local-minimum free energy stat...