Protein Structure Determination by High-Precision FRET and Molecular Modeling

Abstract

A comprehensive methodology for FRET based modeling of biomolecules is presented. Conformations of six multi-state model proteins are determined based on synthetic FRET data in the given benchmark. The workflow consists of five steps:1. Initial conformational ensemble generation with NMSim geometric simulations utilizing available prior knowledge like homology model or crystal structure of the one of the states.2. Identification of the most informative labeling positions and FRET pairs for fluorescence measurements based on proposed ensemble generated at first step.3. State-representative conformer identification, based on their agreement with experiment, using FRET-screening and Accessible Volume simulations.4. Expansion of the initial conformational ensemble with FRET-guided NMSim Markov Chain Monte Carlo simulations based on FRET data. As a result conformations with better FRET-agreement are identified.5. Precision and accuracy assessment and quality control to outline the capabilities and limitations of the shown methodology.Using these steps, conformations of the model proteins were determined accurately. Confidence levels were determined together with the corresponding confidence levels and accuracy measures. Performance capabilities and limitations of the methodology are rigorously assessed. The software package is presented on site.