Why protein conformers in molecular dynamics simulations differ from their crystal structures: a thermodynamic insight

Pullara F., Wenzhi M., Gur M.

TURKISH JOURNAL OF CHEMISTRY, vol.43, no.2, pp.394-408, 2019 (SCI-Expanded) identifier

  • Publication Type: Article / Article
  • Volume: 43 Issue: 2
  • Publication Date: 2019
  • Doi Number: 10.3906/kim-1808-1
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, TR DİZİN (ULAKBİM)
  • Page Numbers: pp.394-408
  • Istanbul Technical University Affiliated: Yes


Conformers generally deviate structurally from their starting X-ray crystal structures early in molecular dynamics (MD) simulations. Studies have recognized such structural differences and attempted to provide an explanation for and justify the necessity of MD equilibrations. However, a detailed explanation based on fundamental physics and validation on a large ensemble of protein structures is still missing. Here we provide the first thermodynamic insights into the radically different thermodynamic conditions of crystallization solutions and conventional MD simulations. Crystallization solution conditions can lead to nonphysiologically high ion concentrations, low temperatures, and crystal packing with strong specific protein-protein interactions, not present under physiological conditions. These differences affect protein conformations and functions, and MD structures equilibrated or simulated under physiological conditions are usually expected to differ from their X-ray structures at a local scale, while the global fold is usually maintained. To quantify this property, we performed conventional MD simulations for over 70 different proteins spanning a broad range of molecular size and structural and functional families. Our analysis shows that crystal structures are good starting points; however, they do not represent structures in their physiological environment. This fact has to be taken into consideration when computational methods dependent on atomic coordinates, such as substrate/ligand docking, are used to guide experimental analyses.