Using Tomoauto: A Protocol for High-throughput Automated Cryo-electron Tomography

The overall goal of this procedure is to help users establish a high-throughput collection and processing pipeline of cryo-electron tomography. This method can help answer key questions in the cryo-electron tomorgraphy field, such as:how do I handle data collected with new generation detectors, and process a huge amount of data produced by automated tilt-series collection software? The main advantage with this technique is that the user can configure multiple software packages in a unifying manner, to finding an efficient pipeline for automated tilt-series processing, allowing for more time to be spent on tomogram analysis.

Generally, individuals new to cryo-ET will struggle because many tedious steps are required to handle massive data generated from the electron microscope. We first had the idea for this development five years ago when we had to manually go through hundreds of steps to get a 3D reconstruction of one bacteria. In advance, consult the text protocol on preparing the Shigella flexneri minicells.

Then, prepare the electron microscopy grids as explained by an existing JoVE publication and also documented in the text protocol. After having loaded the EM stage with the collection of minicells embedded in vitrified ice, start collecting low-magnification maps for the tilt-series. While viewing the fluorescent screen at low magnification, of about 2300x, find grid squares that are acceptable to image.

This location should contain thin ice and the subject of interest with no contaminants. Next, adjust the view to the center of the grid square, and click on the Add Stage Position button to save the location. Then, repeat the process until all the acceptable grid squares have had their location saved.

The next step begins with opening a new montage file. In the setup dialog, set the X and Y so that the entire grid square is captured. For a standard, 200 mesh grid, make a 10 by 10 space with a high Binning value, like eight.

Also toggle the move stage instead of shifting image option and the skip correlations used to align pieces option. Now, in the Navigator window, select the first stage position to be acquired using the Acquire toggle. Continue doing this for each stage position.

Next, under the Navigator menu, click on Acquire at Points. In the dialog box that opens up, toggle acquire map image and rough eucentricity. Untoggle all the other options.

Then, click proceed, and a montage will be made for each stage position. Continue working with the SerialEM software. In the Navigator, select the acquired maps, and click on Load Map.

Then, click on Add Points, and select the points on the map where the tilt-series should be acquired. After making the selections, click Stop Adding Points and repeat the process with each collected map. Next, open the parameters under the Camera menu, and set the Focus, Trial, and Record modes.

The option to record dose-fractionated data may be of special interest, and that can be specified under the Record options. To proceed, select one of the points in the currently viewed map, and toggle the tilt-series option in the Navigator window. In the dialog box that opens, choose the parameters for the tilt-series.

Then, repeat the process of toggling tilt-series option for each of the selected points in the map. Next, from the Navigator start the Acquire at Points option. Under the dialog box, toggle the Preliminary Tasks, Realign to item, Autofocus, and rough eucentricity.

Then, set the Primary Task to acquire tilt-series, and choose close column vals at end to close the column when all the points have been collected. When the process is executed, a tilt-series will be collected at each point toggled for tilt-series collection in each map. With the original tilt-series, the output log from SerialEM, and the individual dose-fractionated images all in the current working directory, in a terminal, execute the following command to remove the beam-induced motion artifact.

Proceed with aligning and reconstructing the tilt-series. Type in tomoauto command in the terminal as follows. To the command, append the filename of the tilt-series to process, followed by the fiducial diameter in nanometers.

If CTF processing is not desired, simply omit the option from the command. This command is the main use of tomoauto, allowing users to execute all the processes involved in aligning a tilt-series, which originally required manual user intervention, in a single command, allowing users to quickly script batch processing. Now, using the 3dmod command, look over the aligned tilt-series for any glaring errors.

Here filename is the name of the tilt-series without suffix. Then, inspect the estimated CTF with the command shown here. Again, filename is the name of the tilt-series without suffix.

Also, in the output log produced by the tomoauto command, check the residual error of the alignment. This is a quantitative measure of the alignment's overall quality. If the alignment is acceptable, then proceed with computing the reconstruction.

This command is set to correct the CTF, and erase the fiducial markers used in alignment from the tilt-series. To skip the visual inspection, and go directly to the reconstruction, use this command. Tomoauto has many accessible options.

One in particular is the ability to include local configurations described by an auxiliary file, according to the tomoauto documentation. Include this file's name in the commandline dialog using the local configuration option before appending the tilt-series filename. The ability to create local configuration files is the most powerful feature of tomoauto, allowing users to fully configure every command used, tuning execution to specific or difficult data sets while maintaining the automated processing.

Within the test protocol are additional instructions on how to execute sub-tomogram averaging. Following the described protocols, two aligned tilt-series were made with tomoauto. The fine alignment is calculated from a model that ideally defines fiducial marker coordinates.

At 50 degrees, one tilt-series tracked the gold particles correctly, indicating an acceptable alignment. While, in the case of the second tilt-series, several model points, shown in red, strayed from their corresponding gold marker, illustrating the model did not produce suitable alignment. When tilt-series are collected, tomoauto successfully aligns 80 to 90%of the collected tilt-series.

After reconstruction, the final tomogram is a 3D volume of the image sample. This can then be used for cellular annotation by segmentation, or sub-tomogram averaging to obtain higher resolution information of the molecular machinery within the sample. The 2.7 nanometers sub-tomogram average of the attached Shigella flexneri deposited in the EMDB shows the large improvement of this technique over viewing the injectisome in a single tomogram.

After watching this video, you should have a good understanding of how to collect and process cryo-electron tomography data using SerialEM and tomoauto in an automated, high-throughput manner. Once mastered, this technique can be used to collect almost 100 tilt-series in a day, if it is performed properly, and align a tilt-series in 20 minutes. While attempting this procedure, it is important to remember to always check the quality of tilt-series alignment.

If a set of tilt-series does not align well, try aligning manually, and transfer the parameters to a local configuration. Following this procedure, other methods like sub-tomogram averaging can be performed in order to answer additional questions, like determining the higher-resolution structure of macro-molecular complexes. After its development, this technique paved the way for researchers in the field of cryo-EM to explore high-resolution in situ structures in bacteria, viruses, and other organisms.