How Does Cryo-EM Work? A Closer Look at the Process

Cryo-electron microscopy (cryo-EM) has revolutionized structural biology by allowing researchers to visualize biomolecules at near-atomic resolution. Unlike traditional methods, it does not require crystallization, making it especially useful for studying large and dynamic protein complexes. Its impact spans drug discovery, virology, and molecular medicine.

To understand how cryo-EM achieves such detailed imaging, it’s important to examine its key steps.

Sample Freezing And Preservation

Preserving biological samples in a near-native state is one of the most challenging aspects of cryo-EM. The process begins with plunge freezing, where a thin layer of the sample is applied to a specialized grid and rapidly submerged in liquid ethane cooled by liquid nitrogen. This ultra-fast cooling vitrifies the water, preventing ice crystal formation that could distort molecular structures.

The choice of grid material and surface treatment affects sample stability. Typically, grids are made of copper or gold and coated with a thin carbon or graphene oxide film. Some receive hydrophilic treatments to enhance sample adhesion. Ice thickness is another critical factor—it must be thin enough for electron transmission while maintaining a hydrated environment. Glow discharge or plasma cleaning optimizes the grid surface, ensuring uniform sample distribution and reducing aggregation.

Once vitrified, the sample must remain at cryogenic temperatures to prevent structural degradation. Storage in liquid nitrogen or specialized cryo-storage systems preserves integrity over time. Automated cryo-transfer systems minimize exposure to ambient conditions, reducing contamination risks.

Electron Beam And Microscopy

Once the sample is vitrified, imaging begins with a transmission electron microscope (TEM). Unlike light microscopy, which uses photons, cryo-EM employs high-energy electrons to achieve significantly higher resolution. The electron beam, typically in the 200–300 keV range, must be carefully calibrated—higher voltages risk radiation damage, while lower voltages reduce resolution.

To minimize electron-induced damage, low-dose imaging controls beam intensity and exposure time, preserving structural integrity. Phase contrast, essential for visualizing low-contrast biological structures, is enhanced using a phase plate, which modulates phase shifts between scattered and unscattered electrons. Advances like the Volta phase plate have significantly improved image contrast.

Electromagnetic lenses focus the electron beam, shaping it as it passes through the sample. Aberrations in these lenses can distort the image, so modern cryo-EM systems incorporate correction mechanisms for optical precision. Automated alignment ensures consistency across datasets.

Recent developments in direct electron detectors have transformed cryo-EM by capturing images with exceptional clarity. Unlike traditional charge-coupled devices (CCDs), which rely on intermediate scintillators, these detectors directly record electron interactions, improving the signal-to-noise ratio and enabling more accurate structural reconstructions.

Detecting And Collecting Images

Once the electron beam interacts with the vitrified sample, scattered electrons carry structural information that must be captured with precision. Direct electron detectors (DEDs) record these images without intermediate scintillators, significantly improving sensitivity and reducing noise. This allows researchers to use minimal electron exposure, preserving fine molecular details.

To counteract beam-induced drift, motion correction algorithms align a series of rapid subframes, reconstructing a sharper final image. Technologies like MotionCor2 have improved resolution by eliminating motion artifacts.

Contrast enhancement techniques, such as contrast transfer function (CTF) correction, refine image sharpness. Since unstained biological samples produce low contrast, CTF correction ensures fine structural details are accurately represented. Software like RELION and cryoSPARC streamline these corrections, enabling rapid processing of large datasets.

Building Three-Dimensional Models

Transforming raw cryo-EM images into detailed 3D structures requires computational techniques. Each micrograph represents a 2D projection of a molecule in a random orientation, so researchers sort and classify thousands to millions of individual particle images. Machine learning algorithms assist in particle picking, filtering out noise and contaminants.

Once selected, particle orientations must be determined before reconstruction. Algorithms use reference-based or ab initio methods to align particles and estimate spatial distribution. Maximum likelihood approaches, implemented in tools like RELION, iteratively refine alignment accuracy. This step is crucial for flexible or heterogeneous structures, where variations in conformation complicate reconstruction. Statistical modeling helps distinguish distinct structural states, capturing dynamic changes that might be lost in static models.