Protein structures you can bank on

摘要

With more than 60,000 structures now in the Protein Data Bank, it is often possible to create homology-based models to help solve the X-ray crystal structure of a protein with an unknown three-dimensional structure. But current techniques usually fail when the protein has less than 30% sequence identity to known structures. A method to overcome this limitation has been developed and used successfully in eight of thirteen X-ray diffraction data sets that could not be solved by conventional means (final structure iterations are yellow in image). The method should allow rapid structure determination without experimental phase information for more than half the cases in which current methods fail, as long as the resolution is 3.2 A or better, with no more than four copies in the asymmetrical unit and the availability of structures of homologous proteins with more than 20% sequence identity.