Multiple Gaussian network modes alignment reveals dynamically variable regions: the hemoglobin case.

ABSTRACT

Gaussian network model (GNM) modes of motion are calculated to a dataset of hemoglobin (Hb) structures and modes with dynamics similarity to the T state are multiply aligned. The sole criterion for the alignment is the mode shape itself and not sequence or structural similarity. Standard deviation (SD) of the GNM value score along the alignment is calculated, regions with high SD are defined as dynamically variable. The analysis shows that the α1ß1/α2ß2 interface is a dynamically variable region but not the α1ß2/α2ß1 and the α1α2/ß1ß2 interfaces. The results are in accordance with the T→R2 transition of Hb. We suggest that dynamically variable regions are regions that are likely to undergo structural change in the protein upon binding, conformational transition, or any other relevant chemical event. The represented technique of multiple dynamics-based alignment of modes is novel and may offer a new insight in proteins' dynamics to function relation.