A scale-spanning integrative modeling strategy to study structure, dynamics, and function of vital molecular machines.

Abstract

Life is motion driven through cellular processes carried out by molecular machines. To understand and manipulate such cellular processes, scale-spanning knowledge of structure, dynamics, and function of molecular machines is needed. This knowledge cannot be obtained by one method alone. Therefore, I developed a strategy that combines ab initio structure prediction, molecular dynamics (MD) simulation, and quantum chemical calculations with data from structure resolving experiments to investigate the assembly and functional cycle of molecular machines from the electron up to the molecular level. First, experimental data is converted into static structural models of different mechanistic states of molecular machines using bioinformatics methods. Second, active sites are refined at sub-Å resolution and intermediate states inaccessible by experiments are identified employing QM/MM simulations. Third, the obtained refined snapshots are merged to map the dynamics of the processes in their native environment using MD simulations. The strategy provides insights into the interplay between local processes like chemical reactions at the active sites and global conformational changes driving cellular function. These insights lead to mechanistic hypotheses and key functional residues that motivate targeted experiments, like mutagenesis studies, to validate the structural dynamic models. This generally applicable strategy was applied e.g. to obtain insights into ATP‑hydrolysis driven protein recycling by the proteasome [Schweitzer et al. (2016) PNAS; Hung et al. (2022) Nature Communications] or the assembly of proteins involved in photosynthesis like photosystem II [Zabret et al. (2021) Nature Plants] or vesicle inducing protein in plastids Vipp1 [Gupta et al. (2021) Cell].