Radiation dose effects in correlative X-ray / cryo-electron microscopy of frozen hydrated biological samples

This study demonstrates that frozen hydrated biological samples, such as apoferritin, can withstand high X-ray doses typical of synchrotron tomography and still yield high-resolution (3.17–4 Å) 3D reconstructions via subsequent cryo-EM, thereby validating an integrated workflow for multiscale imaging of thick vitrified specimens.

The Gist

The Big Idea: A Two-Camera Strategy for Tiny Worlds

Imagine you are trying to study a tiny, intricate city inside a block of ice. You have two powerful tools to look at it:

1. The X-Ray Camera (Hard X-ray Tomography): This is like a super-powerful flashlight. It can shine through thick blocks of ice to see the whole city layout, finding exactly where the interesting buildings are. However, this flashlight is very hot and bright; it might melt a little bit of the ice or scorch the buildings if you leave it on too long.
2. The Electron Microscope (Cryo-EM): This is like a super-magnifying glass. It can zoom in so close you can see the individual bricks and windows of the buildings. But, it can only look at very thin slices of ice. If the ice is too thick, the light can't get through.

The Problem: Scientists want to use the X-Ray camera first to find the "interesting buildings" in a thick block of ice, and then use the Electron microscope to take a super-close-up photo of them. But they were worried: Will the X-Ray camera damage the buildings so badly that the Electron microscope can't see them clearly anymore?

The Experiment: The "Ferritin" Test

To test this, the scientists didn't use a whole city (which is hard to control). Instead, they used Apoferritin.

• The Analogy: Think of Apoferritin as a perfectly round, hollow soccer ball made of protein. It's a standard "test object" in science because we know exactly what it should look like. If the picture comes out blurry, we know the camera or the process is the problem, not the object.

They put these "soccer balls" on a grid, froze them in ice, and then subjected them to three different levels of X-Ray exposure:

1. No X-Rays: The control group (the "safe" zone).
2. Low Dose X-Rays: A gentle flash.
3. High Dose X-Rays: An intense, prolonged blast (100 MegaGray), which is the kind of dose you'd need to scan a thick piece of tissue.

After the X-Ray treatment, they took the grids to the Electron microscope to see if the "soccer balls" were still visible and clear.

The Results: "Scorched but Still Visible"

The results were surprisingly good news!

• The "No X-Ray" Group: As expected, these looked perfect. The scientists could see the soccer balls with incredible clarity, resolving details down to 3.17 Angstroms (about the width of a single atom).
• The "High Dose" Group: Even after being blasted with a massive amount of X-ray energy, the soccer balls were still there. They were a bit fuzzier, but the scientists could still see the structure clearly, resolving details down to 3.88 Angstroms.

What does this mean?
It means that even though the X-Ray "flashlight" did some damage (like scorching the paint on the soccer ball), it didn't destroy the shape of the ball. You can still recognize it and study its structure.

The "Ice" Problem (The Real Villain)

The paper also found a sneaky side effect. When the samples were exposed to X-rays, they got a bit "icy."

• The Analogy: Imagine you take a frozen popsicle out of the freezer and hold it in your warm hand. A layer of frost or water forms on the outside.
• In the experiment, the X-rays caused a layer of ice crystals to form on the sample. This ice acted like fog on a camera lens. It made the picture a little blurry, which is why the resolution dropped slightly.

However, the scientists realized that in a real-world scenario (where you are looking at a thick piece of tissue), you would eventually slice away the outer, damaged, and icy layers to get to the clean, inner part. So, this "fog" might not be a deal-breaker for future experiments.

The Conclusion: A New Workflow is Possible

Before this study, scientists were afraid that combining X-Ray and Electron microscopy would ruin the sample. They thought, "If we use the X-Ray to find the spot, the Electron microscope will just see a mess."

This paper says: "No, it's okay!"

It proves that you can:

1. Use X-Rays to scan a thick, frozen sample to find the interesting parts.
2. Take that same sample to the Electron microscope.
3. Still get a high-resolution, near-atomic 3D picture of the biology.

The Takeaway:
Think of it like a treasure hunt. You can use a metal detector (X-Ray) to find the buried treasure in a thick field. Even if the metal detector makes a little noise or vibration, you can still dig it up and examine the gold coins (Cryo-EM) without them turning into dust. This opens the door to studying complex, thick biological tissues in a way we couldn't do before.

Technical Summary

1. Problem Statement

The integration of cryo-hard X-ray nano-tomography with cryo-electron microscopy (cryo-EM) offers a promising workflow for multiscale imaging of thick, vitrified biological specimens (up to 100 µm³). X-ray tomography can locate regions of interest (ROIs) in thick samples, which can then be thinned (via FIB or CEMOVIS) for high-resolution cryo-EM analysis.

However, a critical barrier to this workflow remains unaddressed: Radiation Damage and Sample Integrity.

• Uncertainty: It is unknown whether cryo-preserved samples can withstand the handling conditions at synchrotron facilities without excessive ice contamination.
• Radiation Limits: It is unclear if the high radiation doses required for X-ray tomography (typically 1–20 MGy, potentially up to 100 MGy for ptychography) will degrade the specimen's structural integrity to the point where subsequent high-resolution cryo-EM reconstruction becomes impossible.
• Knowledge Gap: While radiation damage in crystalline samples is well-studied, its effect on amorphous, frozen-hydrated biological samples (resembling cells/tissues) under combined X-ray and electron exposure has not been evaluated.

2. Methodology

The authors designed an experiment to simulate a correlative workflow using apoferritin (a spherical protein complex) as a model system due to its well-characterized high-resolution structure.

• Sample Preparation:

  • Purified apoferritin was applied to Quantifoil holey carbon EM grids.
  • Samples were plunge-frozen in liquid ethane to create a vitreous ice layer.
  • Grids were mounted in specialized 3D-printed holders compatible with SPINE vials for transport.

• X-ray Exposure (Synchrotron):

  • Location: ID30B beamline, European Synchrotron Radiation Facility (ESRF), Grenoble.
  • Conditions: 100 K temperature, 13.5 keV photon energy, 30 × 30 µm² beam size.
  • Doses: Grids were exposed to three conditions:
    1. 0 MGy (Control, unexposed).
    2. 1 MGy (Typical for standard tomography).
    3. 100 MGy (Upper limit for ptychographic X-ray computed tomography).
  • Exposure Method: Continuous "snake-like" trajectory scanning over 90 × 90 µm² areas.

• Cryo-EM Imaging & Analysis:

  • Instrument: Titan Krios G3i (300 keV) with a K3 Summit direct detection camera.
  • Data Collection: ~1,730 movies recorded across the three dose conditions. Total electron dose per movie stack was 60 e⁻/Ų.
  • Processing:
    • Software: RELION 5.
    • Steps: Motion correction, CTF estimation, particle picking, 2D/3D classification, and 3D refinement.
    • Normalization: To ensure fair comparison, particle counts were equalized to 6,488 particles per dataset before final refinement.

3. Key Contributions

• First Empirical Validation: This study provides the first experimental evidence that biological samples exposed to high-dose X-rays (up to 100 MGy) can still yield high-resolution structures via subsequent cryo-EM.
• Quantitative Damage Assessment: The authors quantified the resolution loss and B-factor degradation specifically caused by the cumulative effect of X-ray pre-exposure and electron imaging.
• Workflow Feasibility: The paper establishes that a correlative X-ray/cryo-EM pipeline is technically feasible, provided mechanical handling and ice contamination are managed.

4. Results

• Resolution Retention:

  • 0 MGy (Control): Achieved 3.17 Å resolution.
  • 1 MGy: Achieved 3.56 Å resolution.
  • 100 MGy: Achieved 3.88 Å resolution.
  • Conclusion: Even at the highest dose (100 MGy), the data supported near-atomic resolution model building (demonstrated by fitting PDB 6rjh).

• Radiation Damage Metrics:

  • Rosenthal-Henderson B-factors increased with dose, indicating reduced particle quality:
    • 0 MGy: 83.99 Ų
    • 1 MGy: 111.90 Ų
    • 100 MGy: 120.18 Ų
  • Ice Contamination: X-ray exposed areas showed significantly higher ice contamination (82–83% of particles excluded during 2D classification vs. 44% for controls). This was attributed to surface ice accumulation and local heating during X-ray exposure.

• Mechanical Challenges:

  • The specialized 3D-printed holders caused issues at cryogenic temperatures (brittleness), leading to grid detachment or damage.
  • Gold grids (Quantifoil Au200) were too fragile and failed; Copper grids (Quantifoil Cu200) survived the handling.

• Dose Equivalence:

  • 100 MGy of X-ray dose is roughly equivalent to 27 e⁻/Ų (using a conversion factor of 3.7).
  • The study confirms that this pre-exposure, combined with a standard cryo-EM dose (~60 e⁻/Ų), does not destroy the structural information necessary for high-resolution reconstruction.

5. Significance and Future Outlook

• Enabling Multiscale Imaging: The findings validate the "integrated imaging workflow" where thick biological specimens are first mapped by X-ray tomography to identify ROIs, then thinned and imaged by cryo-EM.
• Sample Strategy: The authors note that while the current model (thin apoferritin film) suffered from ice contamination, future workflows using thick, high-pressure frozen blocks will likely mitigate this. In such workflows, the X-ray exposure occurs on the bulk sample, and the subsequent FIB-milled lamella (for cryo-EM) is taken from the interior of the block, shielding it from surface ice and direct X-ray damage.
• Structural Biology Impact: This approach allows researchers to study complex, thick biological systems (tissues, organelles) in a near-native state with near-atomic precision, bridging the gap between cellular context and molecular detail.
• Limitations: The study used a thin model system. Thicker tissues may generate more free radicals upon X-ray exposure, potentially causing more severe damage than observed here. Further experimentation with thick tissue samples is required.

Data Availability: The datasets are publicly available in the Electron Microscopy Data Bank (EMDB) under accession numbers 55343 (0 MGy), 55338 (1 MGy), and 55344 (100 MGy).