Angular disorder in actin: is it consistent with general principles of protein structure?

Harold Erickson has recently provided a useful analysis of helical structures having one class versus two classes of intersubunit bonds. His analysis is based upon an assumption that the subunits themselves are essentially unchanged upon bond formation (polymerization). He shows that such a structure having two classes of bonds (i.e. one in which each subunit interacts with four of its neighbors rather than two) can explain some of the features of actin. While he acknowledges that for actin there could be a conformational change and that, in principle, it could explain such features, he argues that the allowed magnitude of such a conformational change is inadequate. Since kinetics and thermodynamics cannot distinguish between the energy derived from the formation of a bond from that due to a conformational change, the question of whether the features of F-actin are derived from a conformational change or a system of two classes of bonds or both must be answered with high-resolution structural information. Recent studies by K. C. Holmes and others suggest that the second possibility might be closest to the truth. The heart of our disagreement is not whether Erickson's thermodynamic analysis is correct, given rigid subunits, but whether all protein polymers are characterized by rigid subunits with rigid intersubunit contacts. Erickson maintains that the observation of an angular disorder of 12 degrees per subunit within the actin filament conflicts with his formalism of rigid subunit interfaces and must therefore result from the erroneous interpretation of measurements. He presents an alternative model to explain the observations. His model, however, does not account for the observations and we will argue that, ultimately, like the resolution of the matter of the number of classes of bonds and the extent of their contact, the amount of angular disorder will require higher-resolution structural studies.