X-ray crystallography primarily consists of three stages: crystallization trials, data collection, and structure determination & refinement. A successful x-ray crystallography run results in three-dimensional coordinates of a novel target protein that can be used for lead discovery and/or three-dimensional coordinates of lead compounds bound to validated targets to serve as a basis for lead optimization.

Above: Crystals of Focal Adhesion Kinase protein.

Crystallization Trials
Purified protein is subject to various crystallization conditions over a range of pH. Typically crystallization is achieved by vapor diffusion under one or two sets of conditions. These crystallization conditions can then be optimized to achieve the highest quality crystal growth through adjustment of the precipitant concentration and pH.

Data Collection
Once crystal growth has been achieved, data collection can proceed. A high-intensity x-ray beam is targeted on the sample crystal. X-rays diffract off the electrons in the atoms of each protein within the crystal; the periodicity of the crystal amplifies the resulting signal to measurable levels and the signal is recorded by a special detector. The resulting data is a set of reflections, or spots, in a series of images. Collecting structural information about a novel protein from crystals therefore requires a source of x-rays. The structure-based drug design core is affiliated with the National Synchrotron Light Source at the Brookhaven National Laboratory, a Department of Energy site. In particular, we work with Beamline X6A for x-ray crystallography data collection.

Above: A portion of an electron density map (blue) and the fitted model. The structure of the tyrosine side chain is clearly seen in the shape of the map.

Structure Determination & Refinement
Finally, the collected data sets must be computationally analyzed and the three-dimensional structure of the protein determined. This is accomplished through the use of a wide variety of computer software packages. The collection of two-dimensional reflections comprising the data set is transformed into a three-dimensional electron density map, into which a protein model can be fitted. The structure is refined to include solvent molecules and any additional compounds present when the protein was crystallized. As a final step, the resulting model can be submitted to the Protein Data Bank prior to publication.