A novel way for inferring DNA-protein binding
The thorough characterisation of individual nucleic acid binding proteins is of primary importance for delineating the complex interplay of gene regulatory networks. In transcription and translational regulation, as well as the cell cycle, accumulating evidence is pointing towards a role of proteins which do not assume a fixed conformation in the native state, but become ordered upon binding. Available predictors, however, can only provide information regarding the structure of such protein regions and not their ligand types and mode of interaction. The primary focus of the EU-funded PROTDNABINDSPEC project was to predict the binding pattern of such natively unfolded protein regions. By using structural bioinformatics, scientists designed, implemented and tested a method to predict the DNA-bound conformation of disordered protein regions at atomic resolution. As a first step, the interaction energy between different amino acids and nucleotides was computed and put into the Fragfold method of protein folding prediction. The predicted DNA-binding specificity of various natively unfolded protein regions was subsequently used to identify DNA binding sites in the genome and was validated on an experimental basis. The PROTDNABINDSPEC project deliverables are expected to improve our understanding of the molecular details of DNA-protein macromolecular interactions. Long term, this method will pave the way for the identification of novel targets and the design of regulatory molecules.
