“Structure defines function” is a basic idea in biology. However, to reveal the fine architecture of a macromolecule whose diameter can be as small as several nanometers is challenging and requires cutting-edge microscopy. The cryogenic electron microscopy (cryo-EM) technique is firstly implemented about half a century ago and is now a powerful tool to determine the high-resolution structures of biological samples. As the name indicates, cryo-EM uses a high-energy electron beam to visualize macromolecules embedded within a thin layer of ice at liquid-nitrogen temperature (-196℃), which helps to preserve the macromolecules at the close-to-native hydrated state and reduce irradiation damages. Although cryo-EM seems to be a first-to-try approach in the structural biology field nowadays, for many macromolecules, especially those of high fragility, tremendous efforts are still needed to solve their structures with this technique. One of the major challenges lies in high-quality cryo-EM specimen preparation.

The most widely used cryo-EM specimen preparation method has not changed much from its initial form¹. In such a procedure, a small volume (3~5 μL) of solution containing target macromolecules is applied onto a 3-mm diameter EM grid (Figure 1a). The EM grid is made of metal alloy of copper or gold, and coated with a perforated carbon film. After being plunge-frozen, macromolecules are expected to be randomly staying at the same plane within a thin layer of vitreous ice, which enables the generation of rich projection views after electron beam illumination. Afterwards, thousands of projection images are collected and then used for three-dimensional reconstruction. Several critical factors, including the homogeneity of macromolecules, the orientation distribution of these molecules, and the signal-to-noise ratio of the images, are not only crucial for the successful reconstruction, but also the time-limiting factors among different projects.

Macromolecules frozen by the conventional cryo-EM specimen preparation procedure are often found to gather at the air-water interface (Figure 1b), posing risks of preferential orientation, sample denaturation, uncontrollable ice thickness, and uneven particle distribution². To solve these problems, graphene membrane is introduced into the cryo-EM field as the supporting film for cryo-EM specimen preparation. Graphene is a single-atom-thick membrane and possesses strong mechanical strength and superior electronic conductivity. Previous works have demonstrated that graphene features better particle-adsorbing capability and negligible background noise. It also reduces beam-induced particle displacement during cryo-EM imaging^3-5, making it an ideal supporting material for biological cryo-EM analysis.

The surface of graphene membrane synthesized by chemical vapor deposition (CVD) on copper foil, however, has dense static wrinkles, giving rise to a height variation of up to dozens of nanometers. Rough graphene surface will render the ice layer non-uniform and lead to a dispersed vertical distribution of macromolecules in the ice (Figure 1c). This is particularly an issue when collecting tilt-series datasets. The varied vertical allocation of individual particles will also lead to different defocus values (as an analogy, one can imagine the species distribution varies with the altitude gradient of the mountain). Such heterogeneity can be a problem in following cryo-EM data processing procedure, especially for small macromolecules whose accurate per-particle defocus value estimation is challenging, and lead to impairment of high-frequency alignment.

By using the Cu (111)/sapphire wafer as the growth substrate, we synthesized ultraflat graphene (UFG) membrane of atomic-scale flatness. The UFG membrane was then transferred onto EM grid via a face-to-face transfer method. The graphene coverage rate on EM grid was as high as 98% and the averaged surface roughness was reduced to 0.7nm. EM grid coated by UFG membrane successfully shaped a uniform thin layer of ice with the majority of macromolecules located at the graphene surface (Figure 1d), as characterized by cryo-EM tomography analysis. The UFG was pre-tensioned and therefore more resistant to deformation compared to rough graphene, which also helps to reduce the particle motion. A smaller B factor value on UFG support obtained by the three-dimensional alignment suggests improved image quality.

Figure 1. Illustration of macromolecule distribution on EM grid. (a) Schematic drawing of a 3-mm diameter EM grid. (b) Macromolecules tend to be adsorbed at the air-water interface during conventional cryo-EM specimen preparation, posing risks of preferred orientation (up) or denaturation (down) problems. (c-d) Macromolecule distribution on rough graphene (c) and UFG (d) supported ice.

The structures of macromolecules with small molecular weight (<100 kDa) are difficult to be solved to high resolution by cryo-EM due to their low signal-to-noise ratio when embedded in the ice layer. To further prove UFG’s structure-solving capability of small macromolecules, we applied several small macromolecules onto the UFG grid, including hemoglobin (64 kDa), alpha-fetoprotein (67 kDa), and streptavidin (52kDa), and found all three samples are mono-dispersedly distributed with good contrast. We then successfully reconstructed their cryo-EM structures at 3.5-Å, 2.6-Å, and 2.2-Å resolution, respectively, allowing us to recognize numerous fine structural details.

In summary, our work reveals that the surface flatness of supporting film can greatly influence the uniformity of ice thickness and determine the spatial distribution of target macromolecules in ice. By using UFG, macromolecules are located on its surface at the same plane, away from the air-water interface. In addition to more controllable ice thickness, UFG is able to improve cryo-EM image quality by reducing the beam-induced particle motion and background noise, thus especially useful for solving the high-resolution structures of small macromolecules.

Reference

1          Dubochet, J., Mcdowall, A. W. & Lepault, J. Frozen-Hydrated Specimens for High-Resolution Electron-Microscopy. Biol Cell 45, 456-456 (1982).

2          Glaeser, R. M. Proteins, Interfaces, and Cryo-Em Grids. Curr Opin Colloid Interface Sci 34, 1-8, doi:10.1016/j.cocis.2017.12.009 (2018).

3          Han, Y. et al. High-yield monolayer graphene grids for near-atomic resolution cryoelectron microscopy. Proc Natl Acad Sci U S A 117, 1009-1014, doi:10.1073/pnas.1919114117 (2020).

4          Liu, N. et al. Bioactive Functionalized Monolayer Graphene for High-Resolution Cryo-Electron Microscopy. J Am Chem Soc 141, 4016-4025, doi:10.1021/jacs.8b13038 (2019).

5          Russo, C. J. & Passmore, L. A. Controlling protein adsorption on graphene for cryo-EM using low-energy hydrogen plasmas. Nature Methods11, 649-+, doi:10.1038/Nmeth.2931 (2014).