Streamlined structure determination by cryo-electron tomography and subtomogram averaging using TomoBEAR

Abstract

Structures of macromolecules in their native state provide unique unambiguous insights into their functions. Cryo-electron tomography combined with subtomogram averaging demonstrated the power to solve such structures in situ at resolutions in the range of 3 Angstrom for some macromolecules. In order to be applicable to the structural determination of the majority of macromolecules observable in cells in limited amounts, processing of tomographic data has to be performed in a high-throughput manner. Here we present TomoBEAR-a modular configurable workflow engine for streamlined processing of cryo-electron tomographic data for subtomogram averaging. TomoBEAR combines commonly used cryo-EM packages with reasonable presets to provide a transparent ("white box") approach for data management and processing. We demonstrate applications of TomoBEAR to two data sets of purified macromolecular targets, to an ion channel RyR1 in a membrane, and the tomograms of plasma FIB-milled lamellae and demonstrate the ability to produce high-resolution structures. TomoBEAR speeds up data processing, minimizes human interventions, and will help accelerate the adoption of in situ structural biology by cryo-ET. The source code and the documentation are freely available.

Fig. 1. A standard workflow for data processing in TomoBEAR.

a A flow diagram of data processing with TomoBEAR. Blue boxes outline the steps that are performed fully automatically, orange boxes may require human intervention. The steps highlighted in magenta represent the functionality of live data processing. b, c an example of a.json configuration file that was used for processing the EMPIAR-10064 data set (results below). Panel B contains the “general” section and the steps for processing up to the refinement of gold fiducials. c contains modules for CTF estimation and correction, tomographic reconstruction, and template matching.

Fig. 2. Example of particle picking and subtomogram classification from a tomogram of an ion channel RyR1 imaged in native membranes (EMPIAR-10452).

a A slice through a tomogram with particles of RyR1 marked with red circles. b A result of the Dynamo-style cross-correlation search of a low-resolution template (shown in an inset) in a tomogram from a with the 15-degree angular search and C4 symmetry. c Selecting an offset for the CC values of the extracted peaks based on the histogram of the CC values in the map in b. d A TomoBEAR description of a classification project to run multi-reference alignment for 3 iterations with 3 classes, two of which are “noise traps”. e Classification of particles extracted from the positions corresponding to the top values of the cross-correlation map starting with 3 classes, two of which are noise classes. The resulting first class contains particles that can be further processed, classes 2 and 3 - false positives.

Fig. 3. Benchmarking performance of TomoBEAR.

a Processing of the tomographic data sets of purified ribosomes (EMPIAR 10064, mixed defocus): A structure at resolution 11.0 Å sliced in the middle of the reconstruction and colored according to local resolution. Black scale bars: 10 nm. Lower panels: detection of resolution based on Fourier Shell Correlation between independently refined half-sets. b Structure of purified human apoferritin imaged by cryo-ET in this study at the global resolution of 2.8 Å. The map is sliced in the middle of the reconstruction and colored according to local resolution. c Processing of tomograms of an ion channel RyR1 imaged in native membranes purified from rabbit muscle (EMPIAR-10452). The structure of RyR1 at a resolution of 8.9 Å sliced in the middle of the reconstruction and colored according to local resolution. Estimation of resolution based on Fourier Shell Correlation: the curves for the original processing of this data set from ref. are in orange and blue, magenta - from TomoBEAR+RELION4.

Fig. 4. Processing of FIB-milled lamellae with TomoBEAR.

a, b A tomographic reconstruction of the plasma-FIB milled HeLa cells data set containing 80 S ribosomes (Berger et al, 2023), EMPIAR-11306. The tomogram in a is reconstructed from a tilt stack aligned by patch tracking in IMOD a, and post-processed with IsoNet b. c The resulting 80 S ribosome structure at the resolution 6.2 Å sliced in the middle of the reconstruction and colored according to local resolution. Black scale bar: 10 nm. Lower panel: estimation of the resolution based on Fourier Shell Correlation between independently refined half-sets.