EFFICIENCY OF SIMULATED ANNEALING AND THE MONTE-CARLO MINIMIZATION METHOD FOR GENERATING A SET OF LOW-ENERGY STRUCTURES OF PEPTIDES

Peptides in solution are in most cases random coils but under special solvent conditions can produce in NMR experiments medium-and long-rang e nuclear Overhauser effect intensities (NOEs). However sometimes the latter are only compatible with a molecule that populates several stat es in thermodynamic equilibrium. A new methodology for analysing such data is based on an extensive conformational search for the global ene rgy minimum (GEM) as well as the other minima in a certain energy rang e above the GEM. A convenient and potentially efficient method for suc h a search is simulated annealing (SA) in which the energy of selected conformations is minimized during the simulation. We apply such SA pr ocedure to ECEPP models of the pentapeptides Leu- and Met-enkephalin a nd find that it is significantly less efficient than the Monte Carlo m inimization method of Li and Scheraga. (C) 1997 Elsevier Science B.V.