An alternate approach to protein structure determination from a minimal set of NMR input data has been developed, which is applicable to larger proteins representing a wide variety of folds. The method is based on experimental input data comprising 13Ca, 13Cb, 13C, 15N, 1Ha and 1HN NMR chemical shifts, plus sparse NOEs if available, and directly exploits the powerful bioinformatics algorithms previously developed for sequence-based homology modeling. For a protein with assigned chemical shifts, the approach relies on a newly designed chemical shift guided protein structure alignment method, POMONA, to select protein templates with the best matched local structure from the Protein Data Bank. Subsequent chemical shift based Rosetta comparative modeling (CS-RosettaCM) is used to generate full atom models from the selected structural templates. POMONA is shown to identify structural homologues in the absence of significant sequence similarity and, in combination with CS-RosettaCM, the approach then generates full-atom models that are demonstrated to match well to the corresponding structures experimentally derived from X-ray diffraction or NMR data. The method is likely to be applicable to any protein for which NMR resonance assignments can be completed and should impact protein structural work in a wide range of disease areas.
Showing the most recent 10 out of 31 publications