| Abstract: | This article describes a new method for peptide structure optimization within a Monte Carlo Simulated Annealing (MCSA) framework, namely the cominimization of potential energies under the constraint of a calculated Circular Dichroism (CD) spectrum. We compute potential energy as well as the CD spectrum of every structure generated within the course of the MCSA run, and cooptimize the mean deviation of this calculated spectrum to a corresponding experimental spectrum together with the potential energies of the respective structures using a modified Metropolis Criterion within the MCSA scheme. We compare the performance of this technique with two other MCSA optimization variants—first, a cominimization of potential energies and free energies of solvation; and second, the standard minimization of potential energy alone. We use homoalanines in lengths of 10 and 15, whose optimized structures are highly α‐helical, and correspondingly give α‐helix characteristic CD signals as the test peptides. This circular dichroic constrained optimization of potential energies is, compared to the other methods, highly efficient in locating α‐helical structures with lowest potential energies in both the 10 alanine and the 15 alanine peptide system, within short MCSA runs. The overall structural information embedded in the CD spectrum efficiently guides the optimization towards the native peptide structure. © 2000 John Wiley & Sons, Inc. J Comput Chem 21: 270–281, 2000 |
