Conformational dynamics of full-length inducible human Hsp70 derived from microsecond molecular dynamics simulations in explicit solvent

Abstract

Human 70 kDa heat shock protein (hHsp70) is an ATP-dependent chaperone and is currently an important target for developing new drugs in cancer therapy. Knowledge of the conformations of hHsp70 is central to understand the interactions between its nucleotide-binding domain (NBD) and substrate-binding domain (SBD) and is a prerequisite to design inhibitors. The conformations of ADP-bound (or nucleotide-free) hHsp70 and ATP-bound hHsp70 was investigated by using unbiased all-atom molecular dynamics (MD) simulations of homology models of hHsp70 in explicit solvent on a timescale of .5 and 2.7 μs, respectively. The conformational heterogeneity of hHsp70 was analyzed by computing effective free-energy landscapes (FELs) and distance distribution between selected pair of residues. These theoretical data were compared with those extracted from single-molecule Förster resonance energy transfer (FRET) experiments and to small-angle X-ray scattering (SAXS) data of Hsp70 homologs. The distance between a pair of residues in FRET is systematically larger than the distance computed in MD which is interpreted as an effect of the size and of the dynamics of the fluorescent probes. The origin of the conformational heterogeneity of hHsp70 in the ATP-bound state is due to different binding modes of the helix B of the SBD onto the NBD. In the ADP-bound (or nucleotide-free) state, it arises from the different closed conformations of the SBD and from the different positions of the SBD relative to the NBD. In each nucleotide-binding state, Hsp70 is better represented by an ensemble of conformations on a μs timescale corresponding to different local minima of the FEL.

An animated interactive 3D complement (I3DC) is available in Proteopedia at http://proteopedia.org/w/Journal:JBSD:30

Keywords:

• human Hsp70
• molecular dynamics
• cancer
• conformational changes
• FRET

Acknowledgments

This research was conducted by using the computational resources of the “Centre de Ressources Informatiques” (CRI), Université de Bourgogne. This work was granted to the HPC resources of CINES under the allocation 2011-c2011076161 made by GENCI (Grand Equipement National de Calcul Intensif). AN thanks the French Ministry of Research and Education for a PhD fellowship.