On the intrinsic flexibility of the µ opioid receptor through multiscale modeling approaches

PhD thesis defended by Mathieu FOSSEPRE (Prof. Daniel VERCAUTEREN) - 20/01/2016
Promoter

Prof. Daniel VERCAUTEREN, UNamur, Laboratory of Physico-Chemical Informatics (PCI)

Jury

Raphaël FREDERICK (UCL), Aatto LAAKSONEN, co-promoteur (Univ. Stockholm), Mathieu SURIN (UMons), Johan WOUTERS, président (UNamur), Renaud LAMBIOTTE (UNamur), Daniel VERCAUTEREN, promoteur (UNamur)

Summary

In our PhD thesis, we proposed an integrated hierarchical and multiscale approach for understanding the intrinsic flexibility properties of the µ opioid receptor (µOR), a key protein in the medical field as the target of most used anesthetic agents (Fig. 1). To do so, we first performed classical all-atom (AA) Molecular Dynamics (MD) simulations of µOR in its apo-form to decipher the important mechanical properties of µOR related to its biological functions. Notably, we highlighted that the various degrees of bendability of the α-helices presented important consequences on the plasticity of the µOR binding site, explaining thereby why µOR can interact with diverse ligands that have very different structural 3D geometries. In the second step, AA MD simulations were used as reference for developing purely mechanical coarse-grained (CG) models of µOR adapted for reproducing the intrinsic flexibility of µOR. In the third stage, optimized CG models were themselves spatially abstracted for designing new and original multi-grained (MG) models, so highlighting modular features of µOR flexibility properties. Therefore, one considered network modularization techniques to design so-called MG models. They represent a novel type of low resolutions models. They are different in nature versus CG models as being true multi-resolution models, i.e., each MG interaction site grouping a different number of residues. MG models, despite their very low resolution, reproduced in a high accurate way the dynamics obtained with the AA MD simulations. MG models are therefore useful for simulating the dynamics of μOR and other proteins at a domain or modular resolution.