MODELING PROTEIN INTERACTIONS THROUGH STRUCTURE ALIGNMENT

Abstract

Rapid accumulation of the experimental data on protein-protein complexes drives the paradigm shift in protein docking from "traditional" template free approaches to template based techniques. Homology docking algorithms based on sequence similarity between target and template complexes can account for ~ 20% of known protein-protein interactions. When homologous templates for the target complex are not available, but the structure of the target monomers is known, docking through structural alignment may provide an adequate solution. Such an algorithm was developed based on the structural comparison of monomers to co-crystallized interfaces. A library of the interfaces was generated from the biological units. The success of the structure alignment of the interfaces depends on the way the interface is defined in terms of its structural content. We performed a systematic large-scale study to find the optimal definition/size of the interface for the structure alignment-based docking applications. The performance was the best when the interface was defined with a distance cutoff of 12 Å. The structure alignment protocol was validated, for both full and partial alignment, on the DOCKGROUND benchmark sets. Both protocols performed equally for higher-accuracy models (i-RMSD &le 5 Å). Overall, the partial structure alignment yielded more acceptable models than the full structure alignment (86 acceptable models were provided by partial structure alignment only, compared to 31 by full structure alignment only). Most templates identified by the partial structure alignment had very low sequence identity to targets and such templates were hard to detect by sequence-based methods. Detailed analysis of the models obtained for 372 test cases concluded that templates for higher-accuracy models often shared not only local but also global structural similarity with the targets. However, interface similarity even in these cases was more prominent, reflected in more accurate models yielded by partial structure alignment. Conservation of protein-protein interfaces was observed in very diverse proteins. For example, target complexes shared interface structural similarity not only with hetero- and homo-complexes but also, in few cases, with crystal packing interfaces. The results indicate that the structure alignment techniques provide a much needed addition to the docking arsenal, with the combined structure alignment and template free docking success rate significantly surpassing that of the free docking alone.