A Monte Carlo method with different treatments for short- and long-range interactions in conformational statistics of polypeptide chains

Abstract

Separation of the conformational energy into terms associated with short-range or long-range interactions allows us to propose a new approach for using Monte Carlo calculations in statistical mechanics of molecular chains. The present method is based on a simultaneous use of statistical weight matrices calculated with short-range interactions and importance sampling applied only on the part of energy coming from long-range interactions. Distributions of end-to-end distances and probabilities of the chain units being in their different conformational states are determined on molecular models of the peptidic hormones enkephalin and β-casomorphin. Results obtained in this way are compared with values calculated using previous computational methods which are also applied on chain models including only nearest neighbor interdependent units. The relative importances of short-range and long-range interactions for the molecular conformation can thus be distinguished. Effects of residue Pro as well as of charged ends on the chain folding are well verified. It is shown that solvent effects can be introduced on long-range interactions mainly by screening the electrostatic term of the energy; extended conformations are then obtained. The proposed method lowers the computer time and by using more information on the moledular system, improves Monte Carlo calculations.