Lecture 8 Video 6

Protein structure

❄️ Cryo-Electron Microscopy (Cryo-EM) — Full Workflow Summary

Cryo-EM is one of the most powerful modern structural biology methods. It allows scientists to determine 3D structures of proteins and complexes — often at near-atomic resolution — without crystallization.

This lecture walks step-by-step through:

Sample preparation
Grid freezing (vitrification)
Data collection in the microscope
Image processing and 3D reconstruction
Comparison to other EM methods

Let’s go through everything in a logical and fun way.

🧪 Step 1 — The Most Important Rule: Make a Perfect Sample

Before anything else, you must have:

✅ Highly purified protein ✅ Homogeneous sample (no aggregation) ✅ Fresh preparation

Structural biology rule number one:

The quality of the structure is limited by the quality of the sample.

Typical purification workflow:

Ion-exchange chromatography → elution peak
SDS-PAGE → purity increases step-by-step
Final polishing: Size-exclusion chromatography just before freezing

This is done because:

Proteins can aggregate or degrade if left on ice
Sensitive samples (like membrane proteins) must be extremely fresh

🧫 Step 2 — Applying Sample to the EM Grid

You apply ~3 µL of protein solution onto an EM grid.

What is an EM grid?

Diameter: ~3.05 mm
Metal mesh with square holes
Mesh density: 200–400 lines per inch
Square openings: ~65–130 µm
On top sits a carbon or gold foil with tiny holes (~1–2 µm)

These holes are crucial — they will contain the thin ice film holding the proteins.

💧 Step 3 — Blotting & Vitrification (Flash Freezing)

This happens in a plunge freezer chamber:

Conditions:

~100% humidity
~4–5 °C
Prevents evaporation during preparation

Process:

Make grid surface hydrophilic
Apply sample
Blot away 99.99% of liquid
Immediately plunge into liquid ethane (~−184 °C)

Why liquid ethane?

Very high heat capacity → ultra-fast cooling
Prevents formation of crystalline ice
Produces vitreous (glass-like) ice

Result:

Thin ice film (~30 nm)
Proteins are randomly oriented and trapped in native buffer conditions

Then grids are stored in:

👉 Liquid nitrogen (~−196 °C)

🔬 Step 4 — Imaging in the Cryo-Electron Microscope

Example instrument:

Titan Krios
300 kV transmission electron microscope
Operated at cryogenic temperatures

Modern microscopes use:

🎥 Movie mode data collection

Instead of one long exposure:

Multiple frames are collected
Motion correction is applied
This removes beam-induced drift

Without this:

❌ Images would be blurred With correction:

✅ Sharp individual particle images → high resolution possible

Typical total electron dose:

~60 electrons / Å²

Low dose is necessary to:

👉 Avoid radiation damage

But it also causes:

⚠️ Very low contrast images — a key Cryo-EM challenge.

🧩 Step 5 — Particle Picking

Each micrograph contains:

Hundreds to thousands of particles

Scientists:

Identify x-y coordinates of each particle
Cut out small image boxes
Build large datasets (100,000 → 1,000,000 particles)

Each particle image is:

👉 A 2D projection of the 3D protein

Because proteins are randomly oriented in the ice.

📊 Step 6 — 2D Classification & Averaging

Particles are grouped based on:

Similar projection views

Within each class:

Images are averaged

Why?

Signal is consistent
Noise is random

Therefore:

✨ Signal-to-noise dramatically improves

This also helps remove:

🗑️ Junk particles 🦠 Contaminants ⚙️ Damaged proteins

This cleaning step is essential.

🧠 Step 7 — Determining Angular Relationships (Fourier Space Magic)

Key principle:

Fourier transform of a 2D projection = slice through the 3D Fourier transform

All projections share common lines in Fourier space.

Using this:

Angular relationships between views can be determined
3D Fourier space is gradually filled
Inverse Fourier transform reconstructs the 3D object

This is one of the most elegant mathematical ideas in structural biology.

🧊 Step 8 — 3D Reconstruction (Electrostatic Potential Map)

Final result:

👉 A 3D electrostatic potential map

Important distinction:

Method	Map type
X-ray crystallography	Electron density
Cryo-EM	Electrostatic potential

Electrons interact with:

Atomic nuclei electrostatic potential

X-rays interact with:

Electron clouds

But visually:

👉 Maps look very similar

So:

Atomic models can be built into Cryo-EM maps just like X-ray maps.

🔁 Step 9 — Iterative Refinement Cycle

Processing is iterative:

Initial 3D model built
Generate projections
Align particles again
Remove bad classes
Reconstruct new map

This continues until:

👉 Resolution cannot improve further

Then:

Resolution is estimated (e.g., correlation coefficients)
Final map + atomic model are deposited in databases.

⚖️ Negative Stain vs Cryo-EM (Two Single-Particle Methods)

🧪 Negative Stain EM

Uses:

Heavy metal salts (e.g., uranyl acetate)

Advantages:

✅ Very high contrast ✅ Cheap ✅ Fast ✅ Good for small proteins (<40 kDa) ✅ Great for checking sample quality

Disadvantages:

❌ Low resolution (~20–40 Å) ❌ Drying artifacts possible

Good for:

👉 Initial screening and heterogeneity assessment.

❄️ Cryo-EM

Advantages:

✅ Native hydrated environment ✅ No fixatives ✅ Can reach atomic resolution ✅ Best structural method for large complexes

Disadvantages:

⚠️ Very low contrast ⚠️ Technically demanding ⚠️ Requires expensive instrumentation

⭐ Big Conceptual Takeaways

Cryo-EM reconstructs 3D structures from many 2D projections
Sample quality determines structural success
Rapid vitrification preserves native state
Iterative classification and refinement improve resolution
Fourier space alignment is key to angular determination
Modern detectors + movie mode enabled the Cryo-EM revolution

Quiz

Score: 0/30 (0%)

Q0. What is the primary purpose of plunge freezing the EM grid in liquid ethane during Cryo-EM sample preparation?

To stain the protein with heavy metals

To slowly cool the sample and form crystalline ice

To rapidly vitrify the sample and avoid crystalline ice formation

To remove contaminants from the protein solution

Q1. Why is high humidity (~100%) maintained in the plunge freezer chamber?

To increase electron beam intensity

To prevent evaporation of the thin liquid film during blotting

To increase protein concentration

To enhance radiation damage

Q2. What is the typical diameter of holes in the carbon foil on an EM grid?

10–20 µm

1–2 µm

50–100 nm

5–10 mm

Q3. What is the main advantage of collecting Cryo-EM data in movie mode rather than a single long exposure?

Higher protein concentration

Improved chemical fixation

Ability to correct beam-induced motion

Faster freezing

Q4. What is particle picking in Cryo-EM data processing?

Selecting crystallized proteins

Identifying and boxing individual particle images from micrographs

Choosing microscope magnification

Removing ice contamination

Q5. Why are particles averaged during 2D classification?

To increase radiation damage

To reduce protein flexibility

To improve signal-to-noise ratio

To increase sample thickness

Q6. Which mathematical concept allows determination of angular relationships between 2D projections in Cryo-EM?

Laplace transforms

Common lines in Fourier space

Gaussian elimination

Monte Carlo sampling

Q7. What is the final structural result obtained from Cryo-EM reconstruction?

Diffraction pattern

Electrostatic potential map

Chromatogram

Absorbance spectrum

Q8. Why is low electron dose used in Cryo-EM imaging?

To increase contrast

To avoid radiation damage to the sample

To speed up freezing

To enhance crystallization

Q9. What is a major disadvantage of Cryo-EM compared to negative stain EM?

Higher cost and lower image contrast

Inability to study proteins

Requirement of crystallization

Longer protein purification

Q10. Why is size exclusion chromatography often performed immediately before grid preparation?

To crystallize the protein

To label the protein with fluorophores

To ensure the sample is fresh and monodisperse

To increase electron density

Q11. What is the typical thickness of the vitreous ice layer containing particles?

1–2 mm

500 nm

Around 30 nm

10 µm

Q12. What distinguishes Cryo-EM maps from X-ray crystallography maps?

Cryo-EM maps detect electron clouds while X-ray detects nuclei

Cryo-EM maps represent electrostatic potential while X-ray maps represent electron density

Cryo-EM maps are 2D only

X-ray maps cannot show proteins

Q13. Why are junk classes removed during 2D classification?

To increase protein size

To remove contaminants and incorrectly picked particles

To reduce microscope voltage

To improve freezing speed

Q14. What is the purpose of iterative refinement cycles in Cryo-EM processing?

To repeatedly freeze samples

To gradually improve alignment and resolution of the reconstruction

To increase beam intensity

To reduce protein concentration

Q15. In Cryo-EM, proteins are trapped in a native-like environment within vitreous ice.

True

False

Q16. Negative stain EM generally provides higher resolution than Cryo-EM.

True

False

Q17. During particle picking, each selected particle image represents a 2D projection of a 3D object.

True

False

Q18. Crystalline ice formation is desirable in Cryo-EM sample preparation.

True

False

Q19. Low contrast in Cryo-EM images is mainly due to the need for low electron dose imaging.

True

False

Q20. Movie mode imaging allows correction of beam-induced motion.

True

False

Q21. All particles selected during automated picking are valid protein particles.

True

False

Q22. The EM grid mesh typically has square openings on the order of tens of micrometers.

True

False

Q23. Cryo-EM reconstruction involves filling 3D Fourier space with slices from multiple projections.

True

False

Q24. Liquid nitrogen is used directly for vitrification because it cools faster than liquid ethane.

True

False

Q25. Negative stain EM can be useful for quickly evaluating sample heterogeneity.

True

False

Q26. Cryo-EM always requires protein crystallization before imaging.

True

False

Q27. Iterative refinement cycles may involve removing poorly aligned particle classes.

True

False

Q28. Electrostatic potential maps from Cryo-EM can be used to build atomic models if resolution is sufficient.

True

False

Q29. Beam-induced motion has no impact on Cryo-EM image quality.

True

False