Nat. Methods, published online 4 August 2013; doi:10.1038/nmeth.2592

The amino acids in proteins are highly dynamic, with their side chains sampling many different conformations every second. Although X-ray crystal structures of proteins are usually presented as a single, static structure, the diffraction data may contain additional information about transient, alternative conformations of the protein. van den Bedem et al. now report a new algorithm, CONTACT, that can be used to identify networks of dynamic amino acids in a protein from high-resolution X-ray crystallography data. In a proof-of-principle study involving dihydrofolate reductase, the algorithm predicted that the NADP cofactor might mediate long-range interactions between the FG loop and the adenosine-binding domain; NMR spectroscopic measurements were shown to support this hypothesis. The authors then solved a high-resolution, room-temperature X-ray crystal structure of a variant of dihydrofolate reductase in which the Met20 loop was destabilized by making two mutations. CONTACT indicated that there was a large increase in the number of dynamic contacts between amino acids in the double mutant, and its active site was significantly more heterogeneous than that of the wild-type protein. This heterogeneity may explain why the rate of hydride transfer for the double mutant is much lower than that of the wild-type enzyme. Algorithms like CONTACT will make it easier to identify and study dynamic regions of proteins, which will help structural biologists relate protein motions to function.