Bioinformatic analysis of protein/DNA complexes

Bohdan Schneider¹, Jiří Černý¹, Jean-Christophe Gelly², Alexandre G. de Brevern²

¹Institute of Biotechnology AS CR, CZ-142 20 Prague, Czech Republic

²INSERM - University Paris Diderot, Paris 7, France

Recognition between proteins and DNA is a key step of essential biological processes. To investigate structural features of the recognition we analyzed the geometry of the protein/DNA interface calculated from more than thousand crystal structures of protein/DNA complexes.

Novelty of the analysis lies in our fine-grained categorization of protein and DNA local conformations involved in the interactions. The conformations of proteins and DNA were categorized into structural classes according to the previously published structural alphabets. Protein structures were sorted into sixteen pentapeptide units called “peptide blocks” [de Brevern et al. Proteins 41, 271 (2000)] and DNA into 14 dinucleotide conformers derived from the original dictionary of conformers [Svozil et al. Nucleic Acids Research 36, 3690 (2008)]. We determined what peptide blocks and dinucleotide conformers formed the protein/DNA interface and which of them were in mutual contact. The resulting matrix showing which peptide block contacted which dinucleotide conformer was called interaction matrix. The interaction matrices were constructed and analyzed for various types of protein/DNA complexes such as DNA complexes of polymerases, transcription factors, or structural proteins. The variability of these interaction matrices characterizes differences of recognition patters of different types proteins. The interaction matrices were constructed and analyzed for direct polar contacts between protein and DNA atoms and for contacts mediated by water molecules and also separately for the DNA atoms forming the minor groove, major grove, and the phosphate group. Water-mediated contacts are almost as frequent as the direct polar ones and the contacts to the minor groove display distinctly different patterns from the major groove and phosphate atoms.

This work was supported by the Czech-France collaboration Barrande (MEB021032) in years 2009-2010 and is supported by the Czech Science Foundation (P305/12/1801). BS and JC are supported by the institutional grant AV0Z50520701.