Phase knowledge enables rational screens for protein crystallization

We show that knowledge of the phase behavior of a protein allows one to create a rational screen that increases the success rate of crystallizing challenging proteins. The strategy is based on using microfluidics to perform large numbers of protein solubility experiments across many different chemical conditions to identify reagents for crystallization experiments. Phase diagrams were generated for the identified reagents and used to design customized crystallization screens for every protein. This strategy was applied with a 75% success rate to the crystallization of 12 diverse proteins, most of which failed to crystallize when using traditional techniques. The overall diffraction success rate was 33%, about double what was achieved with conventional automation in largescale protein structure consortia. The higher diffraction success rates are achieved by designing customized crystallization screens using the phase behavior information for each target. The identification of reagents based on an understanding of protein solubility and the use of phase diagrams in the design of individualized crystallization screens therefore promotes high crystallization rates and the production of diffraction-quality crystals.