Biophysical Society Thematic Meeting | Trieste 2024

Emerging Theoretical Approaches to Complement Single-Particle Cryo-EM

Poster Abstracts

7-POS Board 7 EXTRACTING PROTEIN ENSEMBLES FROM CRYO-EM 2D DATA AND RECONSTRUCTED 3D MAPS Samuel Hoff 1,2 ; Massimiliano Bonomi 1,2 ; Pilar Cossio 3,4 ; Eric R Greene 5 ; James Fraser 6 ; 1 Institut Pasteur, Department of Structural Biology and Chemistry, Paris, France 2 Université Paris Cité, CNRS UMR 3528, Paris, France 3 Flatiron Institute, Center for Computational Biology, New York, NY, USA 4 Flatiron Institute, Center for Computational Mathematics, New York, NY, USA 5 San Francisco State University, Department of Chemistry and Biochemistry, San Francisco, CA, USA 6 University of California San Francisco, Department of Bioengineering and Therapeutic Sciences, San Francisco, CA, USA Cryo-EM has proven to be an instrumental technique for determining the structures of biomolecules. As the resolution obtainable through cryo-EM increases, it has become possible to generate high-quality single-structure atomic models of proteins from the reconstructed 3D cryo EM maps to represent proteins and other biomolecules. However, as proteins occupy a variety of functionally important conformation states, they cannot be accurately described by single model representations. Critical information pertaining to protein conformational heterogeneity is present in both reconstructed 3D maps and 2D particle image sets but is often not fully utilized during model generation or is not properly extracted from the particle images and 3D maps. To accurately describe proteins as entities which occupy an ensemble of structural states, new approaches must be developed to accurately extract information on biomolecule conformational heterogeneity present in sets of cryo-EM particle images and reconstructed 3D maps. Here we demonstrate the utility of two Bayesian inference based integrative approaches, BioEM and EMMIVox, to determine conformational ensembles of proteins from both 2D particle images and reconstructed 3D maps. We utilize BioEM with coarse-grain molecular dynamics to gain insights into the dynamic nature of a multidomain bacterial toxin and a myosin protein from 2D particle images. Additionally, we apply the newly developed tool EMMIVox to explore functionally important conformational heterogeneity describing the continuous dynamics of glutamine synthetase using high-resolution (2 Å) 3D reconstructed maps from cryo-EM. These approaches allow us to obtain novel insights into the structure-dynamic-function relationships of these important systems. Further development and application of pipelines which extract and utilize the entirety of the data generated from cryo-EM will be critical to accurately describe structurally complex and dynamic biomolecules. EMMIVox is now available in the Integrative Structural and Dynamical Biology module of the open-source, freely-available PLUMED library (www.plumed.org).

57