Achieving High-Quality Tomographic Data Collection Using the 200 kV Transmission Electron Microscope with Dual Condenser Lens System

Summary

Electron tomography enables near-atomic-level imaging of organelles, cells, and macromolecules within tissues in cryo-biological samples. Achieving such resolution requires strictly controlled imaging conditions. Here, using a 200 kV transmission electron microscope with a two-stage condenser lens system as an example, we describe how to perform high-quality electron tomography data collection.

Abstract

Cryo-electron tomography (cryo-ET) has emerged as a powerful technique for elucidating the structures of cellular organelles and large protein complexes in their native environments. Unlike conventional room-temperature electron microscopy, cryo-ET minimizes radiation-induced and preparation-induced damage to internal structures. Moreover, in contrast to single-particle analysis (SPA), cryo-ET reduces the need for repetitive purification, thereby preserving the most authentic structural information of the samples. To achieve high-resolution imaging, cryo-ET also imposes stringent requirements on the imaging conditions of the TEM. Despite the continuous improvement in cryo-ET data acquisition efficiency and the emergence of numerous software platforms, most solutions are proprietary, vendor-specific, and tightly coupled to certain electron microscope models, leading to high implementation costs. In this study, we installed the open-source software SerialEM on a 200 kV electron microscope (Glacios 2 cryo-TEM) with a dual condenser lens system and performed comprehensive calibrations to enable high-quality tomographic data acquisition. Using apoferritin as a model specimen, we describe the entire workflow-from cryo-sample preparation to automated cryo-ET data collection, followed by tomogram reconstruction and sub-tomogram analysis. This method offers a cost-effective and accessible solution for cryo-ET data collection, particularly for laboratories using Glacios microscopes without access to proprietary software.

Introduction

Cryo-electron tomography (cryo-ET) has emerged as a powerful technique for studying the native structures of macromolecular complexes and cellular organelles in situ. A major advantage of cryo-ET over single-particle analysis (SPA) is its ability to avoid the excessive purification steps, thereby preserving the native state of the target biomolecules within their biological context. At present, a broad spectrum of specimens -- including purified organelles^1,2,3,4, cell membrane sheets, macromolecular complexes⁵, viruses⁶, cellular lamellae⁷, and tissue sections^8,9,10 -- can be investigated by cryo-ET. Given the diversity of sample types and experimental objectives, data collection strategies vary substantially and must be carefully tailored to both the specimen and the specific instrument configuration.

Data collection is a critical step in the cryo-ET workflow, as both data quality and acquisition efficiency directly impact downstream processing and the overall experimental success of the experiment. To meet the evolving experimental demands, several data-collection platforms have been developed. Among the most widely used are commercial Tomography and the open-source software SerialEM^11,12,13. Tomography offers a highly integrated and user-friendly interface with a low learning curve, but it is less flexible and requires a commercial license. In contrast, SerialEM is a free, open-source package supporting both graphical user interface (GUI)-based and script-based workflows, providing greater flexibility and extensibility.

Tomographic data collection from cryo-lamellae poses unique challenges. High-quality lamellae are difficult to prepare, and their mechanical instability and high sensitivity to electron irradiation further complicate data collection. To address these challenges, Fabian Eisenstein developed the PACEtomo workflow¹⁴, an enhancement of SerialEM that introduces beam-image shift to acquire multiple tomograms per tilt series, thereby reducing the frequency of focus and tracking adjustments. This strategy maximizes the amount of usable data from each section while minimizing unnecessary electron exposure. Similar functionality is also available in the commercial Tomography5 software.

In contrast, data collection from cryo-fixed organelles or membrane sheets is generally more straightforward, as these samples are typically prepared on holey carbon grids, offering ample area for focusing and tracking^15,16. In such cases, SerialEM's built-in cryo-ET acquisition workflows -- accessible directly via its GUI -- offer a convenient, scripting-free option, particularly suitable for beginners or smaller laboratories. Nevertheless, from the perspective of efficiency, we still recommend beam image shift-based strategies whenever feasible.

Given the high cost of cryo-electron microscopes, optimizing data collection strategies to maximize instrument utilization is essential. A 200 kV TEM equipped with a two-stage condenser lens system offers slightly lower electron penetration for thick samples than a 300 kV instrument, but it remains fully viable for cryo-ET when properly calibrated. However, its relatively narrow range of parallel beam conditions and fixed spot size across magnifications makes it generally less suitable for high-resolution cryo-ET unless the system is precisely optimized.

Our electron microscopy setup is equipped with a Falcon 4i direct electron detection camera, which provides substantial advantages in single-electron counting capability and high-speed continuous imaging. To fully exploit the performance of this detector, the TEM must be operated under well-optimized imaging conditions. For TEMs employing a two-stage condenser system, precise calibration of the illumination conditions for parallel light is an important factor in improving imaging quality^17,18. In modern two-stage condenser systems, the focal point of the second-stage condenser lens must coincide with the front focal plane of the upper objective lens to ensure that the optical path intersecting with the sample plane is parallel. Accurately established parallel light conditions are crucial for ensuring that the under-focus values and magnification remain consistent at local positions in the collected images.

In this study, we focused on establishing a high-quality cryo-ET workflow using apoferritin as a model sample. Since the 200 kV transmission electron microscope system is not licensed for the commercial tomography software, we installed and calibrated the open-source SerialEM platform instead. By carefully optimizing the parallel beam parameters for the microscope's two-stage condenser lens system, we achieved automated tomographic data collection through the SerialEM graphical interface-without the need for Python scripts or complex configuration. This approach provides a user-friendly and cost-effective solution for efficient cryo-ET data acquisition on 200 kV instruments such as Glacios.

Protocol

1. Preparation and loading of cryo-samples

1. To prepare the Vitrobot, turn it on and set the temperature to 10 °C. Set the humidity to 100%. Replace the Filter Paper.
2. Place a 300-mesh R1.2/1.3 holey carbon films quantifoil Cu grid into the glow discharge system. Set the glow discharge conditions to 15 mA for 60 s with a vacuum pressure of 0.39 mBar, Negative charge. Perform the glow discharge to hydrophilize the grid surface.
3. Set the Vitrobot parameters as follows: Blot time: 3 s, Wait time: 0 s, Blot force: 2, Temperature: 8 °C, Humidity: 90%.
4. Prepare liquid ethane in an ethane cup wrapped in liquid nitrogen by cooling with liquid nitrogen.
5. Apply 3 µL of apoferritin sample onto the glow-discharged grid. Immediately vitrify the sample in liquid ethane
6. Transfer the sample from ethane to a storage box pre-cooled in liquid nitrogen. The cryo-sample preparation is now complete.
   CAUTION: The following safety protocols must be strictly followed when handling liquid nitrogen and ethane: Personal protective equipment (e.g., cryogenic gloves and safety goggles) is required to prevent frostbite during operations; prior to ethane condensation, ensure the receiving cup is free of liquid nitrogen to avoid splashing caused by rapid boiling; after use, securely close the main and pressure-regulating valves of the ethane gas system to prevent leakage; residual ethane and liquid nitrogen must be disposed of in designated ventilation equipment within the preparation room after sample preparation.

2. Determining the parallel light conditions for electron microscopy

1. Before performing parallel beam alignment, first adjust the following parameters of the scope: stage at eucentric height, objective lens at focus, nP beam tilt ppX/ppY, coma-free pivot point X/Y, coma, and stigma.
2. Using a cross-grating grid, confirm that the specimen is at eucentric height and proper focus at a magnification of SA 73,000× (without energy filter) in Nano probe mode.
3. Switch to Diffraction Mode (camera length: D750 mm) and insert the 100 µm objective aperture (Figure 1A, 1B).
4. Click on the Adjust tab for the objective aperture and use the multifunction X and Y knobs to move the objective aperture off-center so that the edge of the aperture overlaps the gold diffraction rings and partially occludes the central spot (Figure 1C).
5. Bring the back focal plane of the objective lens into view by adjusting the diffraction focus using the Focus knob until the edge of the objective aperture is as sharp as possible (Figure 1D).
6. Adjust the beam intensity (C2 strength) until the width of the gold diffraction rings is minimized.
7. Center the objective aperture. Increase the diffraction camera length until just the central spot is visible (Figure 1E).
8. Click on the Diffraction tab in the Stigmator panel and use the multifunction X and Y knobs to adjust the shape of the central spot while cycling above and below the crossover point to make the beam circular (Figure 1F).
9. Adjust the beam intensity until the central spot is minimized. This will establish parallel illumination conditions. Since parallel illumination conditions may vary with different spot sizes, document the corresponding C2 settings for each commonly used spot size to facilitate rapid configuration during future experiments.
   NOTE: The specific spot size used during data acquisition depends on the magnification and the C2 aperture setting; we need to ensure imaging is performed at the optimal dose rate for the camera.

3. High-quality electron tomography data collection for apoferritin samples

NOTE: This section outlines a standardized procedure for high-throughput tomographic data acquisition, suitable for both experienced users and new researchers.

1. To clip the grid and load the prepared cryo-sample into the electron microscope, place the O-ring into the alignment tool, with the prepared grid placed in the center of the O-ring. Use the spring pen to clip the C-ring onto the O-ring, fixing the sample in between them.
2. Insert the Autogrid vertically into the slot of the cassette, transfer it to the nanocup through the workstation, and then load it into the electron microscope.
3. To capture a full montage of the entire grid using SerialEM, launch the SerialEM software. Enable the Low Dose mode. Define the Search mode at 115× magnification.
   NOTE: The search magnification is typically used for acquiring large-area maps. This magnification is set as low as possible to speed up the mapping process, while still ensuring the accuracy of the image stitching.
4. Open the Navigator file. Go to Navigator > Montaging & Grids > Setup Full Montage. In the dialog box, set Use Search Mode; save images in bin4 format. Click Start under Montage Controls.
5. To calibrate the search and view fields, in the Search mode, locate a highly recognizable landmark (Figure 2A). Click Add Marker in the Navigator. Click Go to Marker to move to the marker location.
6. Switch to the View (2600×) mode and locate the same landmark. Click the Landmark. Go to Navigator, click Shift to Marker to align the fields (Figure 2B). Following the prompts in the dialog box, the beam shifts from both magnifications are assigned to all markers in the search map, and the shift values are saved (Figure 2C).
   NOTE: The view magnification is generally set to the lowest magnification of the Selected Area (SA) to ensure that both view and Record are at the same mode, making the centering process during magnification switching easier.
7. To calibrate the view and record fields (Figure 3). In the View mode, locate a highly recognizable landmark. Switch to the Record (120k×) mode. Use the mouse to center the landmark in the Record mode.
   NOTE: The setting of the Record magnification depends on the estimated target resolution (where the target resolution is twice the pixel size) and the size of the sample.
8. Switch back to the View mode and take a picture. Drag the target object under the view magnification to the crosshairs representing the center of the field of view using the right mouse button.
9. Click Set next to Shift in the Defocus section to align the fields.
10. To select the target square and capture a medium-magnification map , 5.1 in the full montage, click Add Polygon to outline the area to be captured. After outlining an area, click Stop Adding to finalize the selection.
11. Repeat step 3.10 for all desired areas. For each polygon in the Navigator, check the Acquire (A) and New File at Item options.
12. In the dialog box that appears after checking New File at Item, select the following: Montaged images; Fit montage to polygon; Close file when the next new file is to be opened. Click OK.
13. In the new dialog box, select the following: Move stage instead of shifting image; Use View parameters in Low Dose mode. Save the image to the specified location.
14. To capture the medium-magnification map, in the Navigator menu, click Acquire at Items. Check the following options: Acquire and save image or montage; Make Navigator map.
15. In the Primary Action-related section, check Use View mode for capturing images. In the Tasks before or after Primary section, check Rough Eucentricity. Click GO to start capturing the medium-magnification map.
16. To select points on the medium-magnification map. Open the medium-magnification map in the Navigator interface. Click Add Point and left-click on the Target Objects in the Map.
17. Repeat for all desired points. After selecting all points, click Stop Add to finalize the selection.
18. For aperture alignment, use direct alignments in the microscope UI to adjust nP Beam tilt ppX/ppY and Coma-free Pivot Point X/Y, minimizing beam spot wobble in the field of view.
19. Perform tune scope using a script from the SerialEM script library (https://serialemscripts.nexperion.net/script/47).
20. To initiate tomography data collection, in the Navigator interface, locate the ROI and check Tilt Series. In the dialog box, select Frames Images; choose the data storage location.
21. In the settings dialog, set a degree range from -50° to 50°. Set the Base increment to 2.5°, start angle to 0°, and enable Dose-symmetric scheme.
    NOTE: The settings for the dose per tilt, tilt angle range, and tilt step generally follow the principles outlined below: The total dose is maintained at approximately 100 e⁻/Ų. The tilt range is selected to ensure that, even at the highest tilt angle, the sample remains largely unobstructed and allows sufficient electron transmission and imaging. If the tilt range is relatively large (e.g., beyond ±60°), the tilt step can be set to 3°. For a smaller tilt range (e.g., around ±50°), the tilt step can be reduced to 2.5° or 2° to ensure the total dose remains around 100 e⁻/Ų. This is particularly important for 200 kV microscopes, where radiation damage is more pronounced compared to 300 kV microscopes, making it essential to avoid excessive total dose.
22. In the Beam Intensity or Exposure Control section, check Keep beam intensity. Set the Defocus target to -3 µm and set the Autofocus frequency to once per tilt.
23. Check Stop if Autoalign shifts more than 60% of the image size, set Track Before & after.
24. To start data collection, in Navigator, click Acquire at Items. In primary action, check Cycle defocus target in autofocus, drift wait & TS.
25. Set defocus range from -3.5 to -4.5 µm, in 4 steps. Under General, enable Close column valves at end. In the Tasks before or after Primary section, enable Realign to item and Fine Eucentricity before Primary.

4. Subtomogram averaging of apoferritin

1. Perform motion correction and CTF estimation using Warp¹⁹. Conduct tomographic tilt-series alignment using AreTomo3²⁰.
2. Execute tomographic reconstruction and deconvolution using Warp. Perform template-based particle picking using PyTom²¹.
3. Carry out initial refinement using RELION 5²². Conduct high-resolution refinement using M²³.
4. Determine the final structure by processing eight tilt series, followed by the selection of 5,939 particles and iterative refinement, achieving a final resolution of 2.8 Å (Figure 4).

Results

After precise parallel beam alignment, we conducted tomography data collection for apoferritin using the graphical interface function of SerialEM. Subsequent processing with Warp, AreTomo, PyTom, RELION, and other relevant software yielded a final structure at a resolution of 2.8 Å. This resolution sufficiently demonstrates the high quality of our data, reaching a level currently achievable by subtomogram averaging at relatively high resolution. As shown in Figure 4, the aligned and reconstructed results (Figure 4A) clearly demonstrate the overall architecture and spatial distribution of apoferritin. Subtomogram averaging was performed on particles extracted from eight aligned and reconstructed tilt series, yielding a structure at a resolution of 2.8 Å (Figure 4B). The corresponding Fourier shell correlation (FSC) curve is presented in Figure 4C.

Figure 1: Determination of parallel illumination condition using the second condenser lens system. (A) Diffraction pattern in Diffraction Mode without an aperture. (B) Diffraction pattern after insertion of a 100 µm objective aperture. (C) Aperture offset from diffraction center to better visualize edge sharpness. (D) Focused aperture edge and optimized diffraction rings by adjusting objective and C2 lens currents. (E) Elliptical central diffraction spot requiring correction via the Stigmator panel. (F) Circular and minimized central diffraction spot indicating optimized parallel illumination. Please click here to view a larger version of this figure.

Figure 2: High-to-low magnification alignment process from Search (115×) to View (2600×). (A) Search mode image with the marker (red arrow) positioned at the edge of the hole. (B) View mode image with the marker (red arrow) positioned at the same object location as in (A). The Shift to Marker option is clicked in the navigator. (C) In the pop-up window, the beam shifts from both magnifications are assigned to all markers in the Search map, and the shift values are saved. Please click here to view a larger version of this figure.

Figure 3: High-to-low magnification alignment process from View (2600×) to Record (120k×). (A) In Record mode, the target object is positioned at the center of the field of view (red dashed circle highlights the target object). (B) Image captured in View mode, where the target object does not align with the red crosshair (red dashed circle highlights the target object and the red crosshair). (C) The target object is dragged to the center of the red crosshair by right-clicking and moving it, followed by pressing the Set button located beside the Shift button (red dashed box highlights the button). Please click here to view a larger version of this figure.

Figure 4: Structure determination of apoferritin. (A) Structural imaging.Local resolution map of the final subtomogram average. The texture appears uniform, with closely arranged, repeating nanoscale features across the entire field of view. (B) Spatially resolved quantitative mapping. A false-color, circular map of the same or a related nanostructured region. Colors range from blue/green to yellow/red, corresponding to numerical values shown on the color scale below (approximately 2.5 to 3.5 in arbitrary units). The heterogeneous color distribution indicates spatial variation in a measured property (for example, height, thickness, refractive index, or mechanical/optical response) across the sample. (C) Global spectral or functional characterization. Fourier shell correlation (FSC) curve used for resolution estimation. The final resolution of 2.8 Å is indicated by the dashed line at FSC = 0.143. Please click here to view a larger version of this figure.

Discussion

The procedures for cryo-sample preparation and electron tomography data collection must be adapted to the specific specimen type. The experimental strategy described here is well-suited for biological macromolecules, protein complexes, and small organelles. These specimens can be directly vitrified onto EM grids, and their relatively small thickness ensures efficient electron transmission and the acquisition of high-contrast images. Ideally, sample thickness should not exceed 200 nm.

For plunge-frozen samples, a lateral dimension of less than 10 µm is generally preferred to minimize crystalline ice formation, which can compromise structural integrity. For larger samples, high-pressure freezing or the inclusion of cryoprotectants, such as glycerol, in the buffer can delay ice nucleation²³. However, to satisfy the thickness requirements for TEM, further thinning is often necessary. Among the available thinning approaches, focused ion beam (FIB) milling²⁴is the most widely adopted due to its precision and minimal damage, producing lamellae suitable for high-resolution cryo-ET.

During lamella data collection, it is critical to maximize lamella utilization. The implementation of beam-image shift acquisition can reduce the need for repetitive focus and tracking steps, thereby increasing data collection throughput. Although not described in detail here, this technique has been extensively described in the PACEtomo methodology.

Currently, popular software options for tomographic data collection include Tomography (Thermo Fisher Scientific) and SerialEM. Tomography is a commercial solution with limited flexibility for customization, whereas SerialEM, developed by David Mastronarde at the University of Colorado Boulder, is an open-source software that supports script-based automation. However, fully exploiting SerialEM's capabilities requires a substantial learning curve. In this study, we present a streamlined workflow that leverages SerialEM's GUI, eliminating the need for scripting or Python-based configurations while enabling unattended overnight data acquisition.

A frequent challenge encountered during tomographic data acquisition using this method is inaccurate tracking, which can lead to incomplete tilt-series with missing angles or complicate subsequent alignment. To mitigate these issues, tracking parameters must be optimized based on the sample's contrast and the tilt-stage stability. Effective strategies include adjusting the exposure time at tracking positions to ensure image clarity, physically separating the tracking and focusing positions to reduce cumulative radiation damage, and reducing the tracking magnification when stage stability is poor to facilitate reliable pattern matching. Collectively, these adjustments are essential for achieving robust tracking performance across diverse experimental conditions. In this work, we use apoferritin as a model specimen. Under the premise of testing and correcting parallel light, we utilize the built-in functions of SerialEM to collect high-quality data for cryo-ET and perform subtomogram calculations on the ferritin samples, ultimately obtaining the final structure at 2.8 Å.

This method was developed to extend the applicability of 200 kV electron microscopes for high-resolution cryo-ET. Achieving such a resolution fundamentally depends on obtaining high-quality imaging data, which in turn requires precise illumination conditions. To this end, we performed accurate parallel illumination alignment as a critical preparatory step. The alignment procedure described here is specifically designed for TEMs equipped with a two-condenser lens system, such as the Thermo Fisher Scientific Talos Arctica and Glacios. Unlike instruments with a third condenser lens (C3), these systems restrict the condition for parallel illumination to a single, specific excitation value of the C2 lens. This inherent constraint necessitates exceptionally precise alignment of the electron beam to position the source image at the front focal plane of the upper objective lens-a prerequisite for optimal imaging performance. Our subsequent data processing results demonstrate that, following this alignment, the data acquisition quality achieved with this 200 kV microscope for the tested samples is nearly comparable to that typically expected from a 300 kV instrument.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
ANTCryoTM niti grid	Najingdingxin	M04240922A	used for cryo-sample preparation
apoferritin	Biortus	BP17854-00A	used for cryo-sample preparation
cross-grating grid	Agar Scientific	AGS106	used for determining the parallel light conditions
Glacios2	thermo fisher	9959084	used for tomo data collection
Number 2 filter paper, 55 mm	Whatman	10311807	Blotting paper
PELCO easiGLOW	TED PELLA	91000-01020	used for glow discharge
Quantifoil Cu grid	quantifoil	N1-C14nCu30-01	used for cryo-sample preparation
Vitrobot Mark IV	thermo fisher	230101164	used for cryo-sample preparation