Self-consistent Analysis of Large Fluorescence Data Sets for Integrative Time-Resolved Models of Biomolecules

Abstract

For a mechanistic understanding of large biomolecular molecular complexes it is imperative to (1) characterize individual components by molecular models including excited and transiently populated states, (2) monitor the time-evolution of the complexes performing their associated biological function, and (3) study the biological assembly in living cells. Here, methods and tools for the self-consistent analysis of large fluorescence datasets for integrative molecular models of are presented to (a) generate molecular models in a dynamic equilibrium, (b) describe time-ordered processes, e.g., oligomerization processes, and (c) characterize large molecular assemblies in living cells. An extendable Bayesian framework integrates these methods for ensemble, single-molecule, and fluorescence image spectroscopy data. Forward modeling the multiple fluorescence intensity decays of a network of FRET pairs resolves sample heterogeneities by coarse-grained structural models. The framework, which considers the dye mobilities, their local environment, and FRET, is applied to a network of FRET pairs using T4 lysozyme as a test case. Moreover, this approach is validated by synthetic data and (a) promises structural models with Angstrom resolution, (b) monitors the oligomerization in time-resolved experiments, and (c) recovers equilibrium constants and structural features from imaging data.