Proc. Natl Acad. Sci.http://doi.org/hvw (2012)

Credit: © 2012 PNAS

Obtaining crystals of a protein is essential to determining its three-dimensional structure using X-ray diffraction. This entails laborious screening and optimization work to find the conditions that favour crystallization and the acquisition of quality diffraction data. Computational protein design has therefore become an alternative, and in particular a more powerful strategy for designing protein folds. Now, Christopher Lanci and colleagues have succeeded at designing the folded structure of a protein (a three-helix coiled coil) and also the crystal structure it self-assembles into (a layered honeycomb-like lattice with polar orientation of the proteins). Their de novo approach involves the design of protein sequences for candidate crystalline structures within a set of physically accessible crystalline arrangements of the protein backbone compatible with the target crystalline structure. It also involves experimental synthesis, searching for crystallization conditions, and characterization of the candidate proteins. The researchers show that the X-ray structure of their designed coiled-coil protein and its computational model agree down to the sub-ångström scale.