Heteronuclear NMR and soft docking: An experimental approach for a structural model of the cytochrome c(553)-ferredoxin complex

The combination of docking algorithms with NMR data has been developed exte nsively for the studies of protein-ligand interactions. However, to extend this development for the studies of protein-protein interactions, the inter molecular NOE constraints, which are needed, are more difficult to access. In the present work, we describe a new approach that combines an ab initio docking calculation and the mapping of an interaction site using chemical s hift variation analysis. The cytochrome c(553)-ferredoxin complex is used a s a model of numerous electron-transfer complexes. The N-15-labeling of bot h molecules has been obtained, and the mapping of the interacting site on e ach partner, respectively, has been done using HSQC experiments. H-1 and N- 15 chemical shift analysis defines the area of both molecules involved in t he recognition interface. Models of the complex were generated by an ab ini tio docking software, the BiGGER program (bimolecular complex generation wi th global evaluation and ranking). This program generates a population of p rotein-protein docked geometries ranked by a scoring function, combining re levant stabilization parameters such as geometric complementarity surfaces, electrostatic interactions, desolvation energy, and pairwise affinities of amino acid side chains. We have implemented a new module that includes exp erimental input (here, NMR mapping of the interacting site) as a filter to select. the accurate models. Final structures were energy minimized using t he X-PLOR software and then analyzed. The best solution has an interface ar ea (1037.4 Angstrom(2)) falling close to the range of generally observed re cognition interfaces, with a distance of 10.0 Angstrom between the redox ce nters.