(274c) Prediction of Cryoprotectant Requirements for Flash Cooling of Protein Crystals

Shah, B. N., The University of Toledo
Schall, C. A., The University of Toledo
De Witt, K. J., The University of Toledo

Protein structure at atomic resolution is determined from X-ray diffraction patterns of protein crystals at atomic resolution. Data are collected at cryogenic temperatures in order to reduce thermal motion of atoms and radiation damage within the crystal lattice. Protein crystals contain large channels of solvent which along with surrounding crystallizing solution (mother liquor) may form ice at low temperatures disrupting the crystal lattice as well as giving rise to additional diffuse scattering. To avoid ice nucleation, the solution in the channels and the surrounding mother liquor must be rapidly cooled to a temperature where molecular motion and viscosity is similar to that in a solid state and properties (such as density and thermal expansion) are similar to a liquid state. This metastable state is defined as a glass and the temperature of the transition from supercooled liquid to glass is defined as the glass transition temperature (Tg). The rapid cooling rates for glass formation (vitrification) must also be practically attainable. Addition of cryoprotectants such as glycerol increases the Tg and decreases the nucleation temperature (Th) and the melting temperature (Tm), thereby reducing the ice nucleation range. The cooling rates required for vitrification are then practically attainable using liquid or gaseous cryogens such as nitrogen or helium.

In our earlier studies (Chinte et al., 2005) with commonly used crystallizing solutions Hampton Screen I, it was found that cryoprotectant requirements are a strong function of sample size using gaseous nitrogen (100 K) as a cryogen. It is also found to be true for different size crystals of D-xylose isomerase. Ice formation takes place between Tg and Th with Tm being the upper bound. Thermal properties for Hampton Screen I and solutions of D-xylose isomerase are measured using a differential scanning calorimeter. A semi-empirical Miller/Fox model was applied to Tg for Hampton Screen I solutions and D-xylose isomerase. The experimental values were within approximately ±4% of that predicted by the model(Shah and Schall,2006). This model in combination with heat transfer analysis of the flash cooling process can serve as a starting point for selection of cryoprotectant concentration for different size protein crystals. Based on these two models, cryoprotectant concentration can be predicted for different size protein crystals. This was verified for crystals of Ape FEN 1 using deuterated solutions and for lipoxygenase (from soybean) crystals. The heat transfer model can be further modified with temperature dependent thermal properties.

Chinte, U., Shah, B., DeWitt, K., Kirschbaum, K., Pinkerton, A. A. Schall, C., Sample-size: an important parameter in flash cooling macromolecular crystallization solutions J. Appl. Cryst., (2005) 38(3), 412-419.

Shah, B. Schall, C., Measurement and modeling of the glass transition temperatures of multicomponent solutions Thermochimica Acta, (2006) 443(1), 78-86.