Preparation of protein nanocrystals and their characterization by solid state NMR.

Preparation of proteins in their crystalline state has been found to be important in producing stable therapeutic protein formulations, cross-linked enzyme crystals for application in industrial processes, generating novel porous media for separations, and of course in structure elucidation. Of these applications only X-ray crystallography requires large crystals, defined here as being crystals 100s of microns or greater in size. Smaller crystals have attractive attributes in many instances, and are just as useful in structure determination by solid state NMR (ssNMR) as are large crystals. In this paper we outline a simple set of procedures for preparing nanocrystalline protein samples for ssNMR or other applications and describe the characterization of their crystallinity by ssNMR and X-ray powder diffraction. The approach is demonstrated in application to five different proteins: ubiquitin, lysozyme, ribonuclease A, streptavidin, and cytochrome c. In all instances the nanocrystals produced are found to be highly crystalline as judged by natural abundance 13C ssNMR and optical and electron microscopy. We show for ubiquitin that nanocrystals prepared by rapid batch crystallization yield equivalent 13C ssNMR spectra to those of larger X-ray diffraction quality crystals. Single crystal and powder X-ray diffraction measurements are made to compare the degree of order present in polycrystalline, nanocrystalline, and lyophilized ubiquitin. Solid state 13C NMR is also used to show that ubiquitin nanocrystals are thermally robust, giving no indication of loss of local order after repeated temperature cycling between liquid nitrogen and room temperature. The methods developed are rapid and should scale well from the tenths of milligram to multi-gram scales, and as such should find wide utility in the preparation of protein nanocrystals for applications in catalysis, separations, and especially in sample preparation for structural studies using ssNMR.