Abstract
Proteins are dynamic systems. Their dynamic properties can be divided into three broad classes: individual atomic fluctuations, collective motions of bonded and nonbonded neighbouring atoms, and ligand-induced conformational changes (Ringe and Petsko, 1985). The first two classes represent small-amplitude excursions around the equilibrium conformation of a protein; triggered conformational changes lead to the formation of a new average structure. Although the time-scale of individual and collective fluctuations is relatively short (10−13 to 10−9 sec), they can be studied by a variety of spectroscopic techniques and can be simulated computationally by molecular dynamics calculations (Karplus and McCammon, 1983). It has even proved possible to map the spatial distributions of these motions by X-ray crystallography, because they produce a spreading of the electron density around each atom, which may be modeled by various distribution functions (Petsko and Ringe, 1984; Ringe and Petsko, 1985). Triggered conformational changes have proven much more difficult to study in detail. Their time-scales are too long (10−6 to 101 sec) for simulation by simple molecular dynamics techniques. Moreover, since they produce a change in the equilibrium conformation of the protein, they involve crossing relatively large potential energy barriers. Theoretical methods for simulating barrier crossings are only just being developed, and they require detailed knowledge of both the initial and final states of the molecule, as well as of any intermediate structures that have a lifetime longer than that of a single atomic vibration (10−15 sec).