Sample preparation for cryo-electron microscopy (cryo-EM) is a significant bottleneck in the structure determination workflow of this method. Here, we provide detailed methods for using an easy-to-use, three-dimensionally printed block for the preparation of support films to stabilize samples for transmission EM studies.
Structure determination by cryo-electron microscopy (cryo-EM) has rapidly grown in the last decade; however, sample preparation remains a significant bottleneck. Macromolecular samples are ideally imaged directly from random orientations in a thin layer of vitreous ice. However, many samples are refractory to this, and protein denaturation at the air-water interface is a common problem. To overcome such issues, support films-including amorphous carbon, graphene, and graphene oxide-can be applied to the grid to provide a surface which samples can populate, reducing the probability of particles experiencing the deleterious effects of the air-water interface. The application of these delicate supports to grids, however, requires careful handling to prevent breakage, airborne contamination, or extensive washing and cleaning steps. A recent report describes the development of an easy-to-use floatation block that facilitates wetted transfer of support films directly to the sample. Use of the block minimizes the number of manual handling steps required, preserving the physical integrity of the support film, and the time over which hydrophobic contamination can accrue, ensuring that a thin film of ice can still be generated. This paper provides step-by-step protocols for the preparation of carbon, graphene, and graphene oxide supports for EM studies.
Over the last decade, breakthroughs, principally in detector technology, but also in other technical fields, have facilitated a succession of substantial increases in the resolution at which biologically relevant systems can be imaged by transmission electron microscopy (TEM)1,2. Despite the fact that cryo-EM already allows the resolution of high-resolution structures from as little as 50 µg of protein through single-particle analysis (SPA), cryo-EM sample and grid preparation remain major bottlenecks3,4,5. SPA samples consist of macromolecules distributed approximately randomly within a layer of vitreous ice. The ice must be as thin as possible to maximize the contrast difference between the particles and the solvent. Biological macromolecules are more stable (i.e., less likely to lose their native structure) in thicker ice, because they remain better solvated. Moreover, particles are often found to be much better distributed over the field of view in ice much thicker than the particle size6 and frequently may not be found within holes in the carbon films at all.
Additionally, thicker layers of ice decrease the probability of molecules being close to the air-water interface due to the high surface-to-volume ratio, and it has been estimated that using standard plunge-freezing methods for cryo-EM studies results in the adsorption of ~90% of particles to the air-water interface7. Thicker ice results in undesirably high background due to increased scattering events within the solvent and concomitant attenuation of the signal6,7. It is therefore necessary to achieve as thin a layer of vitreous ice as possible; ideally, the layer would be only slightly thicker than the particle. The challenge for the researcher, which must be overcome for every different sample applied to a grid, is to prepare specimens thin enough for high-contrast imaging whilst maintaining the structural integrity of the particles within their sample. Protein adsorption to the air-water interface is accompanied by several, usually deleterious, effects.
First, binding of proteins to this hydrophobic interface often induces denaturation of the protein, which proceeds rapidly and is typically irreversible8,9. A study conducted using yeast fatty-acid synthase showed that up to 90% of adsorbed particles are denatured10. Second, evidence from a study comparing the orientation distribution of 80S ribosome datasets collected either on amorphous carbon11 or without support12 showed that the air-water interface can cause severe preferential orientation compromising 3D reconstruction of the volume13. Methods to reduce particle interaction with the air-water interface include supplementation of the freezing buffer with surfactants (such as detergents), the use of support films, affinity-capture or scaffolding of substrates, and accelerated plunging times. The use of surfactants is associated with its own problems, as some protein samples may behave non-ideally in their presence, whilst affinity-capturing and scaffolding substrates generally require engineering bespoke grid surfaces and capture strategies. Finally, although there is a lot of research on the development of rapid-plunging devices14,15,16, these require apparatus that is generally not widely available.
Although the standard TEM grid for biological cryo-EM already features a perforated amorphous carbon foil17, there are a number of protocols available for the generation of additional support films and their transfer to TEM grids. The use of these films is a long-established method for sample stabilization18. Amorphous carbon supports are generated by evaporation and deposition on crystalline mica sheets19, from which the layers can be floated onto grids, with the utility of floatation supports as useful tools established in prior reports20. Graphene oxide flakes, typically prepared using a modified version of the Hummers method21, have been used as a preferable support structure to amorphous carbon for their decreased background signal as well as the ability to immobilize and stabilize macromolecules22. More recently, there has been a resurging interest in the use of graphene as a TEM support film due to its mechanical stability, high conductivity, extremely low contribution to background noise23, as well as the emergence of reproducible methods for generating macroscopically large areas of monolayer graphene24 and transferring it to TEM grids25. When compared to amorphous carbon, which undergoes beam-induced motions similarly to, or worse, than ice lacking a support film11,12,17, graphene showed a significant reduction in beam-induced motion of cryo-EM images12.
However, while hydrophilized graphene protected fatty acid synthase from air-water interfacial denaturation, the authors of this study noted that the graphene became contaminated during specimen preparation, likely due to a combination of atmospheric hydrocarbon contamination and from the reagent used to hydrophilize the grids10. Indeed, despite many of the superior qualities of graphene, its widespread use is still hindered by the derivatization required to decrease its hydrophobicity12, which ultimately is chemically difficult and requires specialist equipment. This paper reports protocols for the preparation of amorphous carbon, graphene oxide, and graphene sample supports using a three dimensionally (3D) printed sample floatation block27 to directly transfer support films from the substrates on which they were generated to TEM grids (Figure 1). A key advantage of using such a device is the wetted transfer of films, minimizing hydrophobic contamination of the supports and consequently the need for further treatment, and reducing the number of potentially damaging manual handling steps. These approaches are inexpensive to implement and therefore widely accessible and applicable for cryo-EM studies where sample supports are necessary.
1. General preparation of TEM grids pre-support transfer
2. General preparation of reagent solutions
3. Buffer exchange for carbon support films on mica to prepare negatively stained samples using the support floatation block
4. Application of the support floatation block to prepare graphene oxide-coated TEM grids
5. Application of the support floatation block for the preparation of samples on monolayer-graphene films
TEM grids prepared with amorphous carbon supports are typically covered across the entire grid surface. Although breakage of the carbon film occurs in some instances along with some ruffling (Figure 2A), a large number of grid squares are pristine and thus widely applicable for negative staining purposes. The major factor affecting the integrity of the support is the carbon thickness, which is determined during carbon evaporation. Similarly, with this GrOx protocol, good coverage is routinely achieved across the entire grid (Figure 2B). A single application of GrOx suspension for 1 min is sufficient to ensure few areas with multiple layers, which are easy to see due to flake edges. GrOx grids can be prepared quickly from raw materials and are highly protective of the sample. However, flake edges, incomplete coverage, and ruffling are more frequently visible with GrOx grids than for the other techniques because of the nature of the GrOx flakes.
Although the integrity of the graphene support film, like the amorphous carbon, depends on the deposition process, areas that are well-covered display the characteristic diffraction pattern of single-layer graphene. Importantly, by keeping graphene support films wetted, samples can be recovered from the floatation block after an incubation period and data collected in a manner amenable for single-particle analysis. This method does not require any other treatment of the graphene for wetting, thereby removing the requirement for expensive equipment to render graphene hydrophilic, and it is best to prepare support films shortly prior to sample preparation and grid freezing (Figure 2C).
Figure 1: Sample floatation block design and application during support film preparation. (A) Schematic of top, well, and side views of the floatation block including measurements of the shape, depth, and incline. The groove for tweezer tips to rest, as well as channels to insert needles, are indicated. (B) Amorphous carbon layers can easily be floated onto the surface of buffer contained within the wells of the floatation block using the ramp, i.e., during the preparation of negatively stained TEM grids. (C) The width of the wells is suited to accommodate one TEM grid, whilst the tweezer grooves reduce the need to release and pick up grids unnecessarily during preparation steps, but offer a defined path to recover grids without risk of bending if grids are released. Images in B are modified from 27. Abbreviation: TEM = transmission electron microscopy. Please click here to view a larger version of this figure.
Figure 2: Typical examples of sample support films prepared using the floatation block. Grid square (left) and image (right) views are shown for (A) amorphous carbon, (B) graphene oxide, and (C) graphene support films prepared using the floatation block. The amorphous carbon support was used in the preparation of 70S ribosomes for negative staining, whereas the graphene oxide and graphene supports were used in the preparation of 70S ribosomes for cryo-EM. Images in A and C are modified from 27. Scale bar for A grid square = 10 µm; scale bars for B and C grid squares = 5 µm; scale bars for A–C image views = 50 nm. Abbreviation: cryo-EM = cryo-electron microscopy. Please click here to view a larger version of this figure.
This paper presents protocols for handling of both amorphous carbon and graphene films for cryo-EM sample preparation using a sample floatation block27. An STL file for the support block is freely available from the public Thingiverse repository [www.thingiverse.com/thing:3440684], and can be 3D-printed with any suitable stereolithography printer from a suitable resin. The use of carbon films covering a TEM grid usually involves the carbon floatation onto the sample28. This approach to preparing negative stain grids minimizes air exposure during support handling, thus reducing contamination and protein denaturation. The preparation of grids using floating carbon in small wells is advantageous to floating a larger surface area, i.e., in a water bath or Petri dish, in which case mechanical shearing of the carbon occurs much more readily.
UAc may be difficult to purchase due to current health and safety regulations at the time of publication. Many other commonly used, non-radioactive, negative staining reagents are available, and protocols for their preparation have been described previously29. Although alternative stains have not been used with this support floatation block, it is not likely that there would be any differences in these protocols besides the optimization of incubation time with sample (step 3.5), which is already inherently sample-dependent. The key step in this GrOx support preparation protocol is step 4.4, highlighted by the note to prevent the water and GrOx solution from making contact around the grid edge. Inappropriate mixing of the water and GrOx solutions prevents unidirectional settling of the GrOx flakes by capillary action. Having GrOx flakes on both sides of the carbon foil results in thick layers, thus negating the advantages of using GrOx as a near-single layer support, as well as trapping water between the flakes, which causes contamination of useable areas with additional layers of ice. Graphene oxide support preparation is relatively easy to achieve using droplets of solution on flexible polyolefin film. However, when performed in that way, it is easier to accidentally contaminate the copper side of the grid by mishandling errors; the use of the floatation block reduces the likelihood of this eventuality.
Finally, this paper presents a protocol to prepare graphene-covered grids that avoids any kind of graphene pretreatment to render it hydrophilic, thus reducing its cost and increasing its accessibility. Maintaining a wetted film throughout specimen preparation and applying the sample in situ in the block just before freezing is sufficient to allow the generation of suitable ice layers for cryo-EM with a homogeneous sample distribution. Overall, the protocols presented here minimize sample contact with the air-water interface, therefore reducing sample denaturation and support contamination. For the three support films used in these approaches, homogeneous sample distributions could be achieved across the grids along with imaging of intact, well-preserved single particles.
The authors have nothing to disclose.
The authors would like to thank all the members of the Section for Structural & Synthetic Biology at Imperial College London who have helped test these techniques, as well as Harry Barnett at the Imperial College Advanced Hackspace, and Paul Simpson at the Centre for Structural Biology. CHSA is supported by a Sir Henry Dale Fellowship jointly funded by the Wellcome Trust and the Royal Society (206212/Z/17/Z).
Basic Plasma Cleaner (230 V) | Harrick Plasma | PDC-32G-2 | |
Dumont tweezers N5A INOX. | Dumont Swissmade | 0302-N5A-PO | |
Dumont tweezers NGG INOX. | Dumont Swissmade | 0102-NGG-PO | |
Ehtylacetate | Sigma-Aldrich | 270989-250ML | |
Fishing Loops 10 μL | VWR | 612-9353 | |
Graphene Oxide 2 mg/mL | Sigma-Aldrich | 763705-25ML | |
Iron (III) chloride | Sigma-Aldrich | 31232-250MG | |
Mica Sheets 75 mm x 25 mm x 0.15 mm | Agar Scientific | AGG250-1 | We usually coat mica with a target carbon film thickness of 2 nm |
Monolayer Graphene on Cu | Graphenea | N/A | 10 mm x 10 mm, pack of 4 |
n-dodecyl β-D-maltoside (DDM) | GLYCON Biochemicals GmbH | D97002-C | |
Quantifoil R1.2/1.3 300 mesh copper grids | Enzo Life Sciences | JBS-X-101-Cu300 | |
Quantifoil R2/1 300 mesh copper grids | Enzo Life Sciences | JBS-X-102-Cu300 | |
Quantifoil R2/1 300 mesh gold grids | Electron Microscopy Sciences | Q350AR1 | |
Scissors | Agar Scientific | AGT577 | |
Uranyl Acetate | TAAB Laboratories Equipment | U001 | |
Vitrobot Mark IV | FEI | N/A | |
Whatman filter paper 55 mm | GE Healthcare Life Sciences | 1441-055 | |
Whatman filter paper 70 mm | GE Healthcare Life Sciences | 1441-070 |