“Combining Bioinformatics and Biophysics in the Study of Biological Complexes”

Thursday May 16, 2013

“Combining Bioinformatics and Biophysics in the Study of Biological Complexes”

Jiang Zhu, Assistant professor, Department of Immunology and Microbial Science, Department of Integrative Structural and Computational Biology, The Scripps Research Institute

Macromolecular assemblies play critical roles in many biological processes ranging from protein translation and folding to signal transduction. Such complex functions often come hand in hand with complex structures. Experimental techniques such as x-ray crystallography and NMR spectroscopy can provide atomic details of the subunits and domains but face significant difficulties with multi-component complexes. Rapid advances in cryo-electron microscopy (cryo-EM) technique have made it possible to take 3D images of macromolecular assemblies in their relevant functional states at sub-nanometer and near-atomic resolution. Separately, extensive efforts in methodology development in molecular simulation and structural bioinformatics have produced a large repository of computational techniques, which could potentially fill the gap of missing structural information in the study of macromolecular assemblies. Here we present two examples how bioinformatics and biophysics tools can be combined to provide reliable structure models for large molecular assemblies so as to interpret their biological functions. The modeling method used in both studies, iterative modular optimization (IMO), was developed originally in the context of protein structure prediction for refining local structural segments. In the first case, a modified version of this method, EM-IMO, was developed for building, modifying and refining local structures of protein models using cryo-EM maps as a constraint. A multi-scale refinement strategy that combines EM-IMO- and molecular dynamics-based refinement is applied to build atomic models for the seven conformers of the five capsid proteins in our 5~9Å cryo-EM map of the grass carp reovirus virion (GCRV). In the second case, a two-step strategy that combines IMO for rigid-body docking and replica-exchange molecular dynamics for refinement was used to construct a structure model of TRPP2/PKD1 channel-receptor C-terminal domain complex, which were validated experimentally at domain level and full-length channel level. We expect that such integrated approaches will find many applications in the study of biological complexes.