Granada Crystallisation Box (GCB) is a new crystallisation device designed to perform counter-diffusion experiments. Here we describe the device and its use for protein crystallisation purposes. GCB allows one to explore and exploit the coupling between crystallisation and diffusion. The role of viscous fluids, gels and/or microgravity can be enhanced by using capillary volumes, creating a perfect diffusive mass transport scenario. The use of capillaries also reduces the consumption of macromolecules and avoids the handling of crystals for X-ray diffraction data collection.

Protein crystals crack when they are soaked in a solution with ionic strength suf®ciently different from the environment in which they grew. It is demonstrated for the case of tetragonal lysozyme that the forces involved and the mechanisms that lead to the formation of cracks are different for hypertonic and hypotonic soaking. Tetragonal lysozyme crystals are very sensitive to hypotonic shocks and, after a certain waiting time, cracks always appear with a characteristic pattern perpendicular to the crystallographic c axis. Conversely, a hypertonic shock is better withstood: cracks do not display any deterministic pattern, are only visible at higher differences in ionic strength and after a certain time a phenomenon of crystal reconstruction occurs and the cracks vanish. At the lattice level, the unit-cell volume expands in hypotonic shock and shrinks under hypertonic conditions. However, the compression of the unit cell is anisotropic: the c axis is compressed to a minimum, beyond which it expands despite the unit-cell volume continuing to shrink. This behaviour is a direct consequence of the positive charge that the crystals bear and the existence of channels along the crystallographic c axis. Both features are responsible for the Gibbs±Donnan effect which limits the free exchange of ions and affects the movement of water inside the channels and bound to the protein.

A new method of increasing the success rate in protein crystallization screening experiments by combining microseeding with counter-diffusion crystallization in capillaries (SCD) is presented. We have investigated the number of crystallization hits obtained with and without microseeding with 10 model proteins. For the cases studied, SCD generally increases the number of hits and is particularly useful when only relatively low protein concentration stocks are available, either because the stocks were prepared for, e.g., vapor diffusion experiments, or because the protein is poorly soluble. In either case, the addition of seeds becomes necessary to overcome the nucleation energy barrier so that crystal growth can take place even when the wave of supersaturation that passes along the capillary is insufficient to promote nucleation.

Mexicain is a 23.8 kDa cysteine protease from the tropical plant Jacaratia mexicana. It is isolated as the most abundant product after cation-exchange chromatography of the mix of proteases extracted from the latex of the fruit. The purified enzyme inhibited with E-64 [N-(3-carboxyoxirane-2-carbonyl)-leucyl-amino(4-guanido)butane] was crystallized by sitting-drop vapour diffusion and the structure was solved by molecular replacement at 2.1 A resolution and refined to an R factor of 17.7% (R(free) = 23.8%). The enzyme belongs to the alpha+beta class of proteins and the structure shows the typical papain-like fold composed of two domains, the alpha-helix-rich (L) domain and the beta-barrel-like (R) domain, separated by a groove containing the active site formed by residues Cys25 and His159, one from each domain. The four monomers in the asymmetric unit show one E-64 molecule covalently bound to Cys25 in the active site and differences have been found in the placement of E-64 in each monomer.

We report an efficient screening methodology based on the capillary counter-diffusion technique (CCD), which was evaluated using two different practical approaches. The first consisted of kits prepared with the most successful crystallizing agents (PEG and ammonium sulfate) buffered at different pHs ranging from 4 to 9 and tested on 14 samples, including commercial and research target proteins. The second approach was based on the previously identified and highly effective 24 crystallization cocktails adapted to the counterdiffusion setup. This screening was tested with two target proteins, HbII and HbII-III from the clam Lucina pectinate, and the results compared with those obtained with the vapor-diffusion experiment. The success rate was higher than 60% in both approaches. These results experimentally confirm the usefulness of the CCD technique for the screening of crystallization conditions of biomacromolecules beyond its well-known value for the growth of large and high-quality crystals. We describe a detailed protocol for the laboratory implementation of the capillary counter-diffusion technique.