An Instrument for Physical Vapor Deposition onto Cryo-EM Samples for Microsecond Time-Resolved Cryo-EM

This paper presents the design and operation of a physical vapor deposition apparatus for cryo-EM samples that enables microsecond time-resolved experiments by depositing compounds onto frozen grids for subsequent laser flash melting, demonstrating its utility in optimizing sealing membranes and initiating protein dynamics.

The Gist

Imagine you are trying to take a high-speed photograph of a protein, the tiny machine inside our cells that does all the work. Usually, to take these pictures with an electron microscope, scientists have to freeze the protein instantly, like flash-freezing a fly in mid-air. This is called Cryo-EM.

However, there's a problem: once frozen, the protein is stuck. It can't move, so you can't see how it works. Recently, scientists figured out how to "flash-melt" these frozen samples for just a tiny fraction of a second (microseconds) and then freeze them again instantly. This lets them catch the protein in the middle of a motion, like taking a photo of a dancer mid-jump.

But there was a catch. When you melt the ice, the protein floats in a tiny drop of water. If that drop touches the air, the protein gets stuck to the surface and can't spin around freely, making it hard to see from all angles. Also, if you want to study how a protein reacts to a new chemical (like a drug), you can't just pour the chemical on top of a frozen sample; it won't mix.

The Solution: A "Vacuum Paint Sprayer" for Frozen Samples

This paper describes a new machine built to solve these problems. Think of it as a high-tech, vacuum-sealed spray booth designed specifically for frozen microscope slides.

Here is how it works, using simple analogies:

1. The "Sandwich" Technique (Sealing the Sample)
Imagine your frozen protein is a delicate sandwich. Usually, the top and bottom are open to the air. The new machine can spray an incredibly thin layer of "glass" (silicon dioxide) onto the top and bottom of the sandwich while it is still frozen.

• Why do this? It seals the water inside so it doesn't evaporate when melted. It also pushes the protein away from the air, allowing it to spin freely when the laser melts the ice.
• The Discovery: The scientists tested how thin this "glass" could be. They found that if the glass is too thin (less than two layers of atoms), it has tiny holes, and the water leaks out. But if it is just over two layers thick, it holds perfectly. This is the thinnest possible "seal" they can use.

2. The "Magic Dust" Technique (Mixing Chemicals)
Imagine you have a frozen cake, and you want to see what happens when you add chocolate chips, but you can't melt the cake first.

• The Old Way: You couldn't really do this with frozen samples.
• The New Way: This machine can spray a fine dust of chemicals (like calcium salts) onto the frozen sample. The dust sits on top, waiting.
• The Trigger: When the scientists hit the sample with a laser pulse to melt it for a split second, the ice turns to water, and the "dust" instantly dissolves and mixes with the protein.
• The Proof: The scientists tested this by spraying calcium dust onto a sample containing a special dye that glows red. When the laser melted the ice, the calcium mixed with the dye, and the glow dimmed. This proved that the chemicals mixed perfectly in the blink of an eye.

Why This Matters
This machine is like a universal remote control for frozen biology. It allows scientists to:

1. Protect the sample so it doesn't evaporate or get stuck.
2. Add ingredients (like drugs or chemicals) to the sample after it is frozen but before the experiment starts.
3. Mix everything instantly by using a laser to melt the ice, triggering the reaction exactly when they want to watch it.

The authors suggest that this machine could become the standard "kitchen" for future experiments, where scientists can build complex, multi-step experiments by alternating between adding ingredients and flash-melting the sample, all without ever taking the sample out of the machine.

In Summary
The paper introduces a tool that lets scientists "paint" frozen samples with protective glass or chemical dust. When they zap the sample with a laser, the ice melts, the paint turns into a liquid, and the chemicals mix instantly. This allows them to watch proteins move and react in real-time, solving the problem of how to get ingredients to mix with a frozen sample in a split second.

Technical Summary

Technical Summary: An Instrument for Physical Vapor Deposition onto Cryo-EM Samples for Microsecond Time-Resolved Cryo-EM

Problem and Context
Microsecond time-resolved cryo-electron microscopy (cryo-EM) based on laser flash melting has recently enabled the observation of protein motions on the microsecond timescale. This technique involves briefly melting a frozen sample with a laser pulse to initiate dynamics, followed by rapid revitrification to trap transient configurations. While this approach has expanded the temporal resolution of cryo-EM, it has also revealed new opportunities for sample preparation. Specifically, the ability to deposit compounds onto frozen samples prior to flash melting offers a method to alter the sample environment dynamically. Previous work demonstrated that depositing amorphous ice or silicon dioxide membranes could mitigate preferred particle orientation by separating particles from the air-water interface. However, a dedicated instrument to perform physical vapor deposition (PVD) of non-volatile compounds directly onto cryo-EM samples within a vacuum environment was lacking, limiting the ability to introduce reagents or create ultrathin liquid cell geometries for these experiments.

Methodology
The authors describe the design and operation of a custom vacuum apparatus engineered for physical vapor deposition onto cryo-EM samples. The instrument features a six-way stainless steel vacuum chamber evacuated to $3 \times 10^{-8}$ mbar. Cryo-EM samples are loaded via a repurposed load-lock from a JEOL 2010 transmission electron microscope, integrated into a motorized rotation stage. This setup allows the sample holder to be rotated, enabling deposition on either or both sides of the sample.

The system supports two deposition modes:

1. Gas Dosing: Volatile compounds are introduced via a gas dosing valve, allowing simultaneous deposition on both sides of the sample.
2. Electron Beam Evaporation: Low vapor pressure compounds are deposited using an electron beam evaporator. The system utilizes a tantalum crucible with a graphite liner, heated by an electron beam generated from a resistively heated tungsten coil. An effusive beam is formed through a 1 mm aperture.

To ensure precise control, the deposition rate is actively stabilized using a feedback loop based on an ion current generated by ionizing a fraction of the effusive beam. This signal is calibrated against a quartz crystal microbalance (QCM). The authors note that while the QCM reading is approximately 40% higher than the actual rate due to radiative heating, the linear relationship between ion current and deposition rate allows for rapid setting of desired rates. The sample is maintained at cryogenic temperatures (~100 K) throughout the process to prevent devitrification.

Key Contributions and Results
The paper provides a detailed characterization of the instrument through two primary demonstrations:

1. Determination of Minimum Membrane Thickness: The authors used the instrument to deposit silicon dioxide ($SiO_2$) membranes onto cryo-EM samples to create ultrathin liquid cells. They investigated the minimum thickness required for these membranes to withstand the flash melting and revitrification process without failing.

   • Membranes of approximately 0.75 nm (about 2 monolayers) successfully sealed the sample, preventing evaporation and maintaining a uniform thin liquid film after a 30 µs laser pulse.
   • Membranes reduced to ~0.5 nm (less than 2 monolayers) failed, exhibiting contrast variations indicative of sample evaporation through pores.
   • This establishes that the minimum functional thickness for sealing membranes in this geometry is just over two monolayers.

2. Microsecond Mixing Experiments: The authors demonstrated the feasibility of initiating protein dynamics by depositing reagents onto the sample.

   • A cryo-EM sample containing the calcium-sensing dye Fura Red was prepared.
   • A 0.5 nm layer of anhydrous calcium chloride ($CaCl_2$) was vapor-deposited onto one side of the sample.
   • Upon flash melting with a 50 µs laser pulse, the calcium ions mixed with the sample.
   • Fluorescence microscopy confirmed that the fluorescence intensity of Fura Red decreased in the revitrified area, indicating the binding of calcium ions to the dye. Control experiments without deposited $CaCl_2$ showed no such contrast change.

Significance and Claims
The authors claim that this instrument forms the basis for an integrated platform for microsecond time-resolved cryo-EM. By enabling the physical vapor deposition of compounds onto frozen samples, the system allows for:

• The creation of ultrathin liquid cells that stabilize thin films and extend the observation window from tens to hundreds of microseconds.
• A new strategy for initiating protein dynamics where reagents (such as small molecules or ligands) are deposited and rapidly mixed with the sample upon flash melting, bypassing the need for photocaged compounds.
• The potential for future integration with soft-landing mass spectrometry approaches to deposit larger, non-volatile biomolecules like peptides or proteins.

The paper concludes that this integrated approach reduces sample contamination by minimizing transfers and enables complex experimental sequences alternating between deposition and laser revitrification, thereby broadening the range of accessible protein dynamics.