A toolkit to characterize protein polymerization from cryo-electron tomography data

Abstract

BackgroundCryogenic electron tomography (cryo-ET) is a powerful method to study protein structures and macromolecular complexes. These studies can provide structural information at the nanometer scale, allowing for the visualization of ultrastructures and access to subnanometer information through localized, subtomogram averaging (STA). STA alignments can provide analysts with an opportunity to quantify relationships between particles; however, the analytical tools to accomplish this are often lab or system specific. We offer a robust MATLAB script package that can be applied to a wide array of systems. MethodsMany tomographic analyses require extensive segmentation and image classification, which though useful can have poorly characterized error rates and user biases. We present a method for unbiased, numerical classification of head-to-tail polymerization in STA datasets with no outside information or user influence required. We provide this code in a modular MATLAB script package for ease of adaptation to other projects, with analyses including volumes, concentrations, binding, and fibril bundling. ResultsWe demonstrate the robust analysis possible with this script package using a model system of Rubisco in -carboxysomes (-CBs), showcasing the codes ability to evaluate global data such as volume and overall organization, polymerization data such as twist and bend, and lattice data such as lateral fibril distances and angles. DiscussionOur script package offers structural biologists a toolkit to conduct a robust biophysical analysis on STA data in an unbiased manner. The information generated will provide new insights into protein-protein interactions and the conditions favorable for larger ultrastructures. Particles can also be classified for further STA processing. This script package can be used for scientists studying proteins within isolated compartments or with clearly defined regions of interest.