In spite of its enormous successes, structural biology is severely limited by the low success rates of macromolecular crystallization, and by the poor performance of many samples in NMR. Many valuable, biomedically important targets elude crystallization attempts, and overall high costs of structure determination are due primarily to the labor and time-intensive screening of targets at the stages of protein production, crystallization and initial HSQC spectra collection. Our research will address this bottleneck, through development of protein engineering strategies that allow for rational design of target variants with enhanced solubility, stability, and crystallizability or performance in NMR, and implementation of web-based, publicly available servers that facilitate application of these strategies. During the previous phase, we demonstrated that protein crystallization can be rationally induced by surface engineering based on the premise of surface entropy reduction (SER), i.e. mutagenesis of large, polar and solvent exposed amino acids, such as Lys, Glu and Gln, with small residues, e.g. Ala. Further, we designed and implemented the first generation XtalPred and SERp servers, which offer automated evaluation of protein's propensity to crystallize and design of variants with enhanced crystallizability based on the SER strategy. These tools have been used successfully by thousands of investigators world-wide, and helped solve nearly 170 crystal structures, including those of novel globular and membrane proteins, complexes and drug- targets in drug design pipelines. We now propose to pursue further experimental and computational studies of the relationships between physical chemistry of the protein surface and its solution properties. Specifically, we will investigate how surface entropy reduction affects protein solubility and stability, and how protein surface properties impact on the quality of heteronuclear NMR spectra. We will design and implement second generation XtalPred and SERp algorithms, with numerous new features, to achieve higher success rates for prediction of protein crystallizability and for design of variants with higher crystallizability, solubility, and with enhanced performance in NMR spectroscopy. The two second generation servers will be integrated into an interoperable network with the fold-prediction server FFAS03, and we will design a Protein Construct Design Metaserver (PCDM) making it possible to access the entire toolbox from a single GUI, proceeding from prediction to interactive design of primers and/or synthetic genes for expression of targets. Finally, to validate the methods, we will test them using a selection of biologically relevant protein targets.

Public Health Relevance

The determination of protein atomic structures is of fundamental importance for understanding of human molecular physiology and for successful structure-based drug design. However, most proteins are not amenable to crystallographic characterization, because they are recalcitrant to crystallization. This increases the costs and time needed for structure determination projects and often leads to failures of such projects. Our proposal is focused on addressing this problem by developing protein engineering strategies which allow for rational design of proteins amenable to structural characterization. Development of such strategies and providing public access to servers advising individual structural biology groups how to optimizing their projects, would speed up and lower the costs of protein structure determination.

National Institute of Health (NIH)
National Institute of General Medical Sciences (NIGMS)
Research Project (R01)
Project #
Application #
Study Section
Biochemistry and Biophysics of Membranes Study Section (BBM)
Program Officer
Edmonds, Charles G
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
University of Virginia
Schools of Medicine
United States
Zip Code
Jahandideh, Samad; Jaroszewski, Lukasz; Godzik, Adam (2014) Improving the chances of successful protein structure determination with a random forest classifier. Acta Crystallogr D Biol Crystallogr 70:627-35
Goldschmidt, Lukasz; Eisenberg, David; Derewenda, Zygmunt S (2014) Salvage or recovery of failed targets by mutagenesis to reduce surface entropy. Methods Mol Biol 1140:201-9
Casey, Theresa M; Crodian, Jennifer; Erickson, Emily et al. (2014) Tissue-specific changes in molecular clocks during the transition from pregnancy to lactation in mice. Biol Reprod 90:127
Khar, Karen R; Goldschmidt, Lukasz; Karanicolas, John (2013) Fast docking on graphics processing units via Ray-Casting. PLoS One 8:e70661