How Cryo-Electron Tomography Captures Molecular Machines Inside Living Cells

Cryo-electron tomography (cryo-ET) lets scientists build three-dimensional images of proteins and cellular machinery in their native, frozen-hydrated state—without pulling molecules out of the cell or chemically fixing them. The technique works by flash-freezing a cell, tilting it inside an electron microscope to photograph it from dozens of angles, and computationally combining those two-dimensional views into a 3D volume called a tomogram. The result is a structural snapshot of molecular machines at work in their functional cellular environment, something no other method delivers.

Introduction to Cryo-Electron Tomography

Cryo-ET is a specialized branch of cryogenic electron microscopy (cryo-EM). As Thermo Fisher Scientific describes it, cryo-ET "provides 3D snapshots of proteins at work within their functional cellular environments," delivering both structural information about individual proteins and their spatial arrangements within the cell. A review published through the National Center for Biotechnology Information frames the technique's core appeal: it combines 3D imaging with "close-to-life preservation" and allows studying biological material at molecular resolution. SLAC National Accelerator Laboratory adds that cryo-EM more broadly lets scientists make detailed 3D images of DNA, RNA, proteins, viruses, and cells, revealing how these molecules "change shape and interact in complex ways while carrying out life's functions."

What distinguishes cryo-ET from conventional cryo-EM is context. Rather than purifying a protein and imaging isolated copies, researchers image the molecule where it actually operates—inside an intact virus, cell, or tissue. The trade-off is resolution: a single tomogram of a protein complex is often too noisy to describe molecular structure in detail. The workflow below shows how the method balances that tension.

Step 1: Sample Vitrification – Freezing Life in Motion

Everything begins with vitrification: rapid freezing so fast that the water inside the sample solidifies as amorphous ice rather than crystallizing. This matters because ordered, hexagonal ice would disrupt native cellular structures such as lipid membranes. Vitrified ice preserves them intact, immobilizing cellular components and molecular machines in their near-native, hydrated state and preventing the structural damage and artifacts that slower freezing or chemical fixation would introduce.

After plunge-freezing, samples are kept at liquid-nitrogen temperatures throughout storage and imaging so the water never warms enough to crystallize. The sample is typically prepared on an electron-microscopy grid, and robotics-assisted preparation has improved specimen reproducibility and helped identify optimal conditions, according to a cryo-EM primer for non-microscopists. From this point on, the specimen never returns to room temperature.

Step 2: Data Acquisition – Capturing Tilt Series Images

The vitrified sample is loaded into a cryo-capable transmission electron microscope. A single electron image is a 2D projection—it flattens everything along the beam direction. To recover depth information, the microscope collects a tilt series, acquiring a series of cryo-EM images at defined, incrementally larger tilts relative to the electron beam.

Before those images can be combined, their relative orientations must be precisely determined. As the tomography specialists W. Wan and J. A. G. Briggs explain, this alignment is most commonly performed with embedded fiducial markers—small reference particles whose positions can be tracked across every image in the series. Accurate alignment is the foundation on which the entire 3D reconstruction rests, because any angular error propagates directly into the final volume.

Step 3: 3D Reconstruction – Building the Tomogram

Once aligned, the 2D projections are computationally merged into a 3D volume. The most common method is weighted back-projection (WBP), alongside alternatives such as simultaneous iterative reconstruction (SIRT) and simultaneous algebraic reconstruction (SART). WBP works in Fourier space: each image undergoes a Fourier transformation and is placed into a 3D Fourier volume, with the recorded tilt angles determining where each Fourier plane belongs.

Each frequency contribution is weighted to prevent oversampling before the volume is back-projected into real space, creating the tomogram. Because WBP is non-iterative, it is often chosen when high-resolution information needs to be retained—for example, when the tomogram will feed into further processing such as subtomogram averaging. The output at this stage is a full 3D view of the cellular interior, though individual molecular complexes within it may still appear blurry.

Step 4: Subtomogram Averaging – Achieving Molecular Resolution

Here is where the resolution problem gets solved. While individual tomograms of protein complexes are often at resolutions "not useful for describing molecular structures," as the cryo-EM primer notes, cells frequently contain many identical copies of the same machine—ribosomes, for instance, or the Sec61 protein-conducting channel. When multiple copies of a structure of interest exist, researchers extract each copy from the tomogram, align them to a common frame, and average them together.

Averaging improves the signal-to-noise ratio and yields higher-resolution 3D structures of the molecular machine. The payoff is detailed visualization of molecular architecture within its native cellular context: researchers can see not only what a protein looks like, but exactly where it sits and which neighbors it touches.

The Cryo-Electron Tomography Workflow

The full pipeline can be traced as a sequence of stages, each transforming the output of the previous step. The table below maps the path from a living cell to a high-resolution 3D structural model.

The cryo-ET workflow, from sample to structure
Workflow stage	Purpose	Key process	Outcome
1. Sample vitrification	Immobilize cellular components in their near-native, hydrated state	Rapid plunge-freezing into amorphous ice; storage at liquid-nitrogen temperature	Preserved cell, protected from crystallization and fixation artifacts
2. Tilt-series acquisition	Capture 2D projections from many angles	Imaging at incremental tilts, typically every 2–3° from about −60° to +60°; alignment via fiducial markers	An aligned series of 2D cryo-EM images
3. 3D reconstruction	Combine projections into a volume	Weighted back-projection: Fourier transform each image, place planes by tilt angle, weight frequencies, back-project to real space	A 3D tomogram of the cellular interior
4. Subtomogram averaging	Resolve molecular detail	Extract and average multiple copies of identical complexes within the tomogram	Higher-resolution 3D structures of molecular machines in situ

Reading the table top to bottom shows the essential logic: freezing buys fidelity, tilting buys depth, back-projection buys a volume, and averaging buys resolution. Each stage depends entirely on the one before it, which is why automation and workflow improvements across the pipeline have had such an outsized effect on data quality and speed.

Unique Insights: Visualizing Molecular Machines In Situ

Why go to all this trouble? Because purification-based methods strip molecules of their context. Cryo-ET's high-resolution 3D images of cell interiors, as Thermo Fisher describes, "provide new insights into cellular function and shed light on the arrangement and structure of native protein complexes." Researchers can map how proteins are distributed across a membrane, how molecular machines are positioned relative to one another, and how complexes cooperate to carry out major cellular processes.

The technique's reach extends across structural and cell biology, virology, and medicine. SLAC notes that stringing thousands of cryo-EM snapshots together into stop-action movies and virtual-reality flythroughs lets scientists watch biology in motion. And because cryo-ET works on intact viruses, cells, and tissues, it can reveal organization that no isolated-molecule experiment could—such as the native architecture of a protein-conducting channel embedded in its membrane environment.

Advancing Structural Biology with Cryo-ET

Researchers should consider cryo-ET when seeking to understand the structural basis of molecular function within the complex, native environment of a cell—situations where knowing a protein's shape is not enough without also knowing its cellular address and neighbors. The initial tomogram may not reach molecular resolution on its own, but the in situ context is the technique's primary benefit, and subtomogram averaging closes the resolution gap when repeated copies of a structure are present. The measurable indicator of progress is the publication of new high-resolution 3D structures of molecular machines or cellular components obtained in situ using cryo-ET—each one a data point on how far the workflow can push into the unperturbed interior of a cell.