Intramolecular cavities in globular proteins.

An analysis of internal cavities in 121 protein chains has been undertaken to improve the characterization of their occurrence, morphology and role in protein tertiary structure, including an analysis of the optimal probe size for use in their detection. A number of basic cavity characteristics were elucidated. Cavities are non-artefactual and apparently independent of the method of structure determination, resolution and refinement of the data. Overall cavity volume increases with protein size and yet constitutes only a small fraction of the total protein volume but cavities are nearly always present in proteins > 100 residues in size. They are most commonly found in the protein core. 'Empty' and solvent-containing cavities have been compared and solvated cavities found to possess a more polar surface; the two classes are also seen to exhibit different amino acid type and secondary structural preferences. In general, residues that enclose cavities do not display any extra local mobility relative to their surrounding environments. Water-containing cavities do not impose volume restrictions upon their internal solvent beyond that of bulk solvent and permit good hydrogen bonding. These results should prove useful in protein modelling and design.