Pnina Dauber-Osguthorpe

Partitioning the motion in molecular dynamics simulations into characteristic modes of motion

Molecular dynamics (MD) simulations result in a comprehensive description of molecular motion. However, to gain insight into the dynamic behavior of molecules it is important to be able to identify different types of motions and characterize them. We have developed a novel technique aimed at characterizing the motion of the system using digital signal processing techniques. The amplitudes and phases of the Fourier transform of the atomic fluctuations are used to define the characteristic modes of motion in the MD trajectory. This yields a pictorial description of the oscillatory motions in a manor analogous to normal‐mode (NM) analysis. The validity of this method has been tested on small molecules such as water, acetamide, and a blocked polyalanine in a helical conformation. The NMs obtained by diagonalizing the mass‐weighted second derivative matrix were combined to generate “NM trajectories” that served as well‐characterized test cases. Distinct characteristic modes can be extracted from both NM and MD trajectories. The modes extracted from the NM trajectories were identical to the original NMs. The modes extracted from the MD trajectories were in most cases highly correlated to the corresponding NM. However, intermixing of some of the modes occurred, particularly when conformational changes took place. This technique is flexible and can be applied to the molecular system as a whole or to a subset of atoms of interest. Fourier transform calculations are fast and therefore the analysis stage is not demanding in computational resources. Anharmonicity is included explicitly in the simulations and solvent can be included as well.

MOLECULAR DYNAMICS: DECIPHERING THE DATA

SummaryThe dynamic behaviour of molecules is important in determining their activity. Molecular dynamics (MD) simulations give a detailed description of motion, from small fluctuations to conformational transitions, and can include solvent effects. However, extracting useful information about conformational motion from a trajectory is not trivial. We have used digital signal-processing techniques to characterise the motion in MD simulations, including: calculating the frequency distribution, applying filtering functions, and extraction of vectors defining the characteristic motion for each frequency in an MD simulation. We describe here some typical results obtained for peptides and proteins. The nature of the low-frequency modes of motion, as obtained from MD and normal mode (NM) analysis, of Ace-(Ala)31-Nma and of a proline mutant is discussed. Low-frequency modes extracted from the MD trajectories of Rop protein and phospholipase A2 reveal characteristic motions of secondary structure elements, as well as concerted motions that are of significance to the proteins biological activity. MD simulations are also used frequently as a tool for conformational searches and for investigating protein folding/unfolding. We have developed a novel method that uses time-domain filtering to channel energy into conformational motion and thus enhance conformational transitions. The selectively enhanced molecular dynamics method is tested on the small molecule hexane.

A novel β-turn location in an LHRH antagonist: A combined conformational search and molecular dynamics study

A 50 pico-second molecular dynamics simulation on a cyclic LHRH antagonist analogue Ac-D-Phe1-D-Phe2-D-Trp3-Ser4-Glu5-D-Arg6-Leu7-Lys8-Pro9-D-Ala10-NH2 (where the cyclisation is via an amide linkage between the Glu5 and Lys8 side chains), reveals some hitherto unseen conformational features. The LHRH analogue is found to adopt a near β-sheet type of conformation with the reversal in the chain being brought about by a D-Trp3-Ser4-Glu5-D-Arg6 β turn. The N- and C-terminal ends of the peptide come close together and interact through a network of hydrogen bonds. Additional hydrogen bonds expected of a sheet type of conformation stabilise the lowest energy minima. A conformational search of all possible cyclic structures of a model system c(Glu-D-Ala-Ala-Lys) which was used to determine the starting structure for the simulation studies of the cyclic LHRH antagonist analogue is also high-lighted. The influence of the cyclic part on the conformation of this LHRH analogue is discussed.

Reduced and retro-reduced peptide analogues-conformations and energies

Abstract A complete search of the conformational space available to residues with modified amide links was carried out by flexible geometry minimismions of a grid of values of all rotatable bonds. The implications for incorporating these residues in secondary structures were studied.

SUBSTRATE AND INHIBITOR BINDING TO PHOSPHOLIPASE-A(2)- STRUCTURE, ENERGETICS AND DYNAMICS

Molecular mechanics and dynamics techniques have been used to study the mode of ligand binding and mechanism of action of the enzyme Phospholipase A 2 .