Technical Highlight - November 2009
Short description: The first report of a protein structure obtained without molecular replacement or derivatives from a protein diffracting to 2.0 Å gives hope for direct methods for solving macromolecular structures.
X-ray crystallographic structural models cannot usually be directly calculated solely from experimental data collected from crystals of the native protein because phase information is missing, and only the intensity of diffraction is recorded. Yet phase information is essential for structure determination.
For macromolecules, initial phases are usually worked out either using molecular replacement with a related structure, although that risks introducing a bias towards the already solved structure, or using derivatives of the protein incorporating heavy atoms to produce an anomalous scattering pattern from which phase information can be recovered by comparison with data from the non-treated protein. However, the latter approach requires soaking the crystals to incorporate the heavy atoms and takes additional experimental time.
Direct or ab initio methods of establishing the phases rely on mathematical calculations that are based on the electron density being concentrated at randomly distributed, resolved equal atom positions, and until now this approach has only been possible for protein crystals that diffract to around 1.0 Å resolution.
Now, Rodríguez et al. 1 report the structure of phosphotransferase system regulation domain II (PRD-II) from the transcriptional antiterminator protein GlcT of Bacillus subtilis at 1.95 Å, determined ab initio.
Their approach exploits the fact that macromolecular structures are made up of small fragments of known geometry, such as a-helices and β-sheets in proteins and base pairs in nucleic acids. The authors used the program Phaser to search structural databases to find ideal α-helical polyalanine fragments of 14 residues. Their search initially returned thousands of potential fragments but as they are small they only account for a small proportion of the total scattering mass.
Partial solutions made up of one or two helices did not produce a solution, but three out of the 1,465 top Phaser combinations of three fragments afforded an interpretable electron density map when subjected to density modification with the program SHELXE. By looking at whether the program can trace a number of residues and the correlation coefficient of the partial structure, the successful phasing cases could be identified. Solutions can be further enhanced by B-value refinement of the main-chain atoms.
This approach will make ab initio phase protein structure solution a reality.
D. D. Rodríguez, C. Grosse, S. Himmel, C. Gonzalez, I. M. de Ilarduya et al. Crystallographic ab initio protein structure solution below atomic resolution.
Nature Meth. 6, 651-653 (2009). doi:10.1038/nmeth.1365