Albany 2015:Book of Abstracts
June 9-13 2015
©Adenine Press (2012)
Exploring conformational dynamics of biomolecular recognition via hierarchical sampling approach
Biomolecular recognition is central to many biologic processes and is mediated by specific, non-covalent interactions. The dynamic interplay occurring between the interacting molecules has attracted the interest of many scientists to understand and predict the mechanism of recognition processes. In this context, Molecular Dynamic simulations have the versatility to explore the binding events at an unprecedented resolution than what is possible with experimental technique. For this purpose, we have proposed an advanced hierarchical sampling approach, a theoretical method following recursive disintegration and aggregation of substates obtained from the free energy landscapes. The molecular conformations are distributed in the essential subspace of dominant eigenvectors obtained from principal component analysis (PCA). For this purpose, PCA based on Cartesian coordinates (cPCA) computed on all-atom molecular system and the dihedral torsion angles (dPCA) of protein and glycan backbone was computed. Initially, the molecular conformations were acquired to account the global fluctuations that are conserved, followed by identifying metastable conformations capable of demonstrating subtle internal oscillations particularly localized at the recognition domain and the interacting counterpart. Boltzmann probability distribution was used to compute the free energy landscapes spanning the eigenvector space of the two dominant eigenvectors to enable clustering of the most invariant sets constituting dynamical conformations. Finally, k-means clustering algorithm was applied to geometrically decompose the configuration space into physically meaningful microstates to obtain metastable ensemble to demonstrate the underlying molecular interactions. In this study, we have used malectin and diglucosylated-N-glycan complex as the model system to demonstrate the specific binding events, which supports an extended conformational selection model as the underlying binding mechanism. Further, this approach has a wide applicability to address the mechanism of molecular recognition by MD simulations in general.
This research has been supported by DBT under Bioinformatics
Mamidi A.S., Surolia A. (2014). Hierarchical sampling for metastable conformers determines biomolecular recognition: the case of malectin and diglucosylated N-glycan interactions. J. Biomol. Struct. Dyn. DOI: 10.1080/07391102.2014.948070. Epub ahead of print Aug 20 2014.
Ashalatha Sreshty Mamidi
Molecular Biophysics Unit