Protein Mechanics Through X-Ray Crystallography
Proteins are dynamic entities which often cycle through a variety of conformational states as they carry out their functions. Despite the success of biophysical methods in determining the physical structures of proteins--that is, the precise three-dimensional configuration of all of their constituent atoms--no comprehensive physical models exist which accurately describe or predict the conformational cycling of proteins. For such models to be built, comprehensive knowledge of the energetics of intramolecular interactions in model proteins is essential. Here, we present three approaches which begin to address this problem, each through a different form of perturbation to a series of members of the PDZ protein family. First, we show that cycles of mutagenesis coupled with X-ray crystallography can reveal an anisotropic, distributed pattern of physical interactions in a PDZ domain, PSD-95 PDZ3. This pattern is functionally important and deeply connected to the evolution of PDZ domains in general. Second, we present a new approach for identifying essential dynamical features of proteins from relatively conventional X-ray diffraction data. Through combined analysis of nine different PDZ domain diffraction data sets, we show that collective features can be extracted and averaged, yielding a consensus picture of dynamics in the PDZ domain family. Finally, we report the development of a novel pump-probe method for directly inducing and reading out motions in proteins through the combined use of strong electric fields and time-resolved X-ray crystallography. We show that the method can be used to drive functionally relevant motions in a PDZ domain, LNX2 PDZ2, and provide a foundation for future efforts designed to directly probe the energetic architecture of proteins.