Recent studies show that protein function is not only determined by the static folded three-dimensional structure, but also by the distribution of the populations of its conformational sub-states within different environments. A detailed investigation of interacting protein interfaces requires - beyond the knowledge of their structure - information about their flexibility and the range of accessible conformations of the native state and complexed states. Additionally, successful protein structure prediction has demonstrated the suitability of “building blocks” (prototype fragments) as embodiment of local biological and physicochemical features. It is expected that the overall dynamics of protein structures can similarly be assembled from fragment flexibilities embedded in a known protein fold. Therefore this proposal focuses on developing a fast fragment-based method to predict the dynamics of a native or designed protein structure, which will be used to determine or define function in interaction networks and binding specificity in protein-ligand complexes. Specifically, the dynamics of prototypical fragments from a representative protein set will be sampled by molecular simulations; a flexibility predictor will be designed, implemented using Hidden Markov Model (HMM) and trained on dynamic profiles of the prototypical fragments; by using the flexibility predictor protein-protein and protein-ligand interactions will be investigated. For protein-protein interactions the flexibility of all known high-resolution protein structures will be predicted and made publicly available. To study flexibility as determinant of ligand-binding specificity, comparisons between the conformational sub-space of protein-ligand complexes and the space of all accessible conformations of the unbound components will be performed.
Field of science
- /natural sciences/biological sciences/biochemistry/biomolecules/proteins/proteomics
Call for proposal
See other projects for this call