Energy Landscapes from Cryo-EM and Simulations

Proteins are dynamic entities that undergo many structural transitions and fluctuations, which are essential to their biological functions.

We, therefore, need continuous descriptions of protein conformational space in the form of energy landscapes in order to properly understand their mechanisms of action.

This is now becoming possible through the use of hybrid methods, which combine computational biophysics with experimental structural biology and overcome the limitations of either approach alone.

In this proposal, we present a new hybrid methodology that leverages recent innovations in cryo-electron microscopy image analysis to examine continuous dynamics and free energy landscapes of large, multi-domain proteins, which are not achievable with existing methods.

Our novel interdisciplinary pipeline will involve the use of efficient coarse-grained representations of proteins from computational biophysics coupled with sophisticated image processing tools including 3D reconstruction, classification, and dimensionality reduction.

The specific objective is to extract reaction coordinates from 3D class averages and use them to generate conformational landscapes onto which the raw 2D images can be mapped.

The resulting free energy landscapes will reveal all conformational states with physiological relevance and the preferred transition pathways, which can be analysed further using molecular dynamics simulations.

We will apply our pipeline to ionotropic glutamate receptors, which are tetrameric ligand-gated ion channels with large, dynamic, multi-domain architectures that are critical to synaptic transmission and plasticity in the mammalian central nervous system.

We expect our results to be of great benefit to the broad structural biology community and to be instrumental in understanding brain physiology and designing treatments for a wide range of diseases.