Reoptimization of the AMBER Force Field Parameters for Peptide Bond (Omega) Torsions Using Accelerated Molecular Dynamics

Abstract

Improving the accuracy of molecular mechanics force field parameters for atomistic simulations of proteins and nucleic acids has been an ongoing effort. The availability of computer power and improved methodologies for conformational sampling has allowed the assessment of these parameters by comparing the free energies calculated from molecular dynamic (MD) simulations and those measured from thermodynamic experiments. Here, we focus on testing and optimizing the AMBER force field parameters for the omega dihedral, which represents rotation around the peptide bond of proteins. Due to the very slow isomerization rate of the peptide bond, it is not possible to sample the phase space with standard MD simulations. We therefore employed an accelerated MD method in explicit water in which the original Hamiltonian is modified to speed up conformational sampling and the correct canonical distribution is recaptured. Using well-studied model systems for the peptide and peptidyl prolyl bonds, we discovered that the AMBER omega dihedral parameters underestimated experimentally measured activation free energy barriers for cis/trans conversion as well as failed to reproduce the free energy difference between the two isomers. We reoptimized the original AMBER omega dihedral parameters and further validated their transferability on several experimentally studied dipeptides. The revised set of parameters successfully reproduced the cis/trans equilibria and free energy barriers within experimental and simulation errors. We also investigated the structures of the transition state and cis/trans isomers of prolyl peptide bonds in terms of pyramidality, a measure of the puckering of the prolyl ring. We observed, as expected from quantum mechanical studies, significant bidirectional, out-of-plane motions of prolyl nitrogen in the transition state.

Disordered Structural Ensembles of Vasopressin and Oxytocin and Their Mutants

“…30 The peptides were solvated in explicit solvent represented by the TIP4P-Ew water model. 31 This combination of the adjusted ff99SB force field and TIP4P-Ew water model has been shown to reproduce experimental NMR observables better than other biomolecular simulation force fields.…”

The puckering free-energy surface of proline

“…22 Doshi and Hamelberg have also modified the Amber force field parameters to study proline isomerization so that the simulation results compared well with the experimental values. 23 Sieradzan, Scheraga and Liwo have utilized a coarse-grained UNRES force field to study the isomerization process. 24 As to the QM calculations, Kang and co-workers have thoroughly studied the preferable proline conformations using the different QM calculations, 11,12,14,16,25 and in a recent work Kang and Byun have compared the various QM calculations in studying the conformations of alanine dipeptide and proline dipeptide in the gas phase and in water.…”

Novel third-generation water-soluble noscapine analogs as superior microtubule-interfering agents with enhanced antiproliferative activity

“…AMBER 10 suite of programs [23] were used to carry out the simulations in explicit TIP3P [24] water model in a periodic octahedron box, using the all-atom Cornell et al [25] force field and the reoptimized dihedral parameters for the peptide ω-bond [26]. For ligand molecules, the partial atomic charges are derived using standard two-step RESP method from electrostatic potential calculated using Gaussian03 program [27] at HF/6–31(d, p) level of theory and generalized amber force field (gaff) parameters [28] are used.…”