Cellular regulation or signaling processes are mediated by many proteins which often have one or several intrinsically disordered regions (IDRs). These IDRs generally serve as binders to different proteins with high specificity. In many cases, IDRs undergo a disorder-to-order transition upon binding, following a mechanism between two possible pathways, the induced fit or the conformational selection. Since these mechanisms contribute differently to the kinetics of IDR associations, it is important to investigate them in order to gain insight into the physical factors that determine the biomolecular recognition process. The verprolin homology domain (V) of the Neural Wiskott–Aldrich Syndrome Protein (N-WASP), involved in the regulation of actin polymerization, is a typical example of IDR. It is composed of two WH2 motifs, each being able to bind one actin molecule. In this study, we investigated the early steps of the recognition process of actin by the WH2 motifs of N-WASP domain V. Using docking calculations and molecular dynamics simulations, our study shows that actin is first recognized by the N-WASP domain V regions which have the highest propensity to form transient $\alpha$ -helices. The WH2 motif consensus sequences “LKKV” subsequently bind to actin through large conformational changes of the disordered domain V.

Amyloid fibrils formations are involved in many neurodegenerative diseases such as Alzheimer’s disease, Parkinson disease, Amyotrophic Lateral Sclerosis (ALS) and others. The proteins associated with the formation of amyloid fibrils are Intrinsically Disordered Proteins (IDP) in the unbound state. Nevertheless, this type of proteins can self-aggregate and form cross-β amyloid fibrils structures at physiological condition. Due to the flexibility of these IDPs, no single experimental approach could completely characterize this system, especially in the unbound state. All-atom molecular dynamics (MD) simulations could be used to study the conformational ensemble of IDPs. Unfortunately, force fields (FF) and water models (WM) were developed to simulate one structure of folded proteins. Recently, several FF/WM were improved to properly generate conformational ensembles of IDP. However, it is unknown if the force fields (FF) can properly reproduce the behavior of IDP and also self-aggregate in cross-β amyloid fibrils structures. In this paper, we will focus of the R2 region of the FUS-LC domain (R2-FUS-LC region) which is an Intrinsically Disordered Region (IDR) of 16 residues in the unbound state but forms cross-β fibrils in the bound state. For the R2-FUS-LC region, we benchmarked thirteen commonly used FFs for studying IDPs. We show that CHARMM36m (updated in 2021) with mTIP3P water model performs the best to generate extended structures and cross-β amyloid fibril.

Le travail de thèse consistera à explorer et caractériser l'ensemble conformationnel de protéines intrinsèquement désordonnées (IDPs) en utilisant plusieurs techniques complémentaires, notamment des simulations avancées de dynamique moléculaire et la diffusion des rayons X aux petits angles (SAXS). Les IDPs sont des protéines possédant une ou plusieurs régions n'ayant pas de structures secondaires stables lorsqu'elles sont isolées, mais pouvant en adopter lors de leur association avec de multiples autres protéines. La question, à laquelle ce travail souhaite répondre dans le cas de trois IDPs, est de savoir si ces éléments de structures secondaires, formés à l'interfaces des complexes protéine-protéine, pré-existent de façon transitoire, ou non, à l'état non-lié des IDPs en solution. S'il est possible d'identifier et de caractériser ces éléments de reconnaissance moléculaire dans les IDPs isolées, alors les résultats de ce travail permettront de guider par la suite la détermination des structures de complexes protéiques impliquant des IDPs.