In the past three decades, macromolecular crystallography has emerged as one of the most powerful tools in biomedical sciences. However, in spite of dramatic advances in relevant technologies and in molecular biology, structure determination of a protein often remains a challenge and may take years to complete. Two strategies are currently employed: research which focuses on a specific structure(s) chosen owing to the known biological role(s); and structural genomics, where large numbers of targets - not necessarily with known function and fold - are selected to allow for high throughput structure determination and sampling of the global ensemble of folding patterns. Both approaches suffer from shortcomings: targeted crystallography may take a long time before it bears fruits, while structural genomics strategies will probably allow for structure solution of 1 in every 5 to 10 attempted proteins. In both cases, the limiting factor is likely to be the preparation of diffraction-quality crystals. In this application we propose to investigate the feasibility of efficient protein crystallization through 'crystal engineering', i.e. mutagenesis of rationally selected surface residues. Although mutagenesis has been shown in the past to affect the solubility and crystallization modes of proteins, it was never used in a rational fashion, nor was it technically feasible as a high-throughput screening method in a laboratory of any size. The approach tested in our pilot study integrates three protocols: rapid cloning using the Gateway method, an optimized QuikChange mutagenesis protocol, and systematic mutagenesis of residues with high conformational entropy and high probability of occurring on the surface (i.e. Lys and Glu). Preliminary results, obtained with the human protein RhoGDI (Rho Guanine nucleotide Dissociation Inhibitor), the LH domain from PDZ-RhoGEF, and the BH domain from GRAF, show that K2A (Lys to Ala) and E2A (Glu to Ala) single and multiple mutants crystallize much more readily than wild-type protein. This is in agreement with our hypothesis that surface conformational entropy creates a barrier for crystallization. Preliminary conclusions inferred from the pilot experiments suggest that high-throughput crystal engineering may significantly enhance the efficiency of crystal structure determination: we show that within two weeks of obtaining cDNA it is possible to screen for crystallization of up to 10 or more variants of the selected protein. The approach is simple and inexpensive, and feasible in a laboratory of any size. Funds are requested to further test the general applicability and success rate of this method, and to apply the method to a subset of targets selected for the Structural Genomic Initiative.
Showing the most recent 10 out of 11 publications
