Date Published: June 01, 2018
Publisher: International Union of Crystallography
Author(s): Ieva Drulyte, Rachel M. Johnson, Emma L. Hesketh, Daniel L. Hurdiss, Charlotte A. Scarff, Sebastian A. Porav, Neil A. Ranson, Stephen P. Muench, Rebecca F. Thompson.
This paper describes different approaches that cryo-EM users can take to improve the quality of their sample distribution and ice for high-resolution single-particle cryo-EM.
The last ∼5 years have seen significant developments in the electron-microscopy (EM) field, with a rapid expansion in the use of single-particle approaches to determine high-resolution structures, including those of macromolecular complexes, membrane proteins and ribosomes (Nguyen et al., 2015 ▸; Fitzpatrick et al., 2017 ▸; Plaschka et al., 2017 ▸). From large viruses which may be over 80 nm in diameter and many tens of megadaltons to small soluble proteins of less than 150 kDa, single-particle analysis can offer insights into the structure and function of a diverse range of macromolecular complexes (Khoshouei et al., 2017 ▸; Hesketh et al., 2015 ▸; Vinothkumar et al., 2016 ▸; Rawson et al., 2018 ▸). Through developments in direct electron-detector technology, improved microscope hardware and more advanced image-processing algorithms, the expected resolution from EM has significantly changed, with 831 structures deposited in the EMDB with a resolution of <4 Å since 2013, compared with 18 in the preceding ten years (as of January 2018). If a macromolecular complex is unstable, intrinsically disordered or has buffer components that are incompatible with plunge freezing, it is futile to attempt to produce optimized cryo-EM grids with a good particle distribution. The best recipe for success in a cryo-EM experiment is proper characterization of the sample before attempting plunge freezing. Biochemical and biophysical tools such as SEC-MALLS, negative-stain EM, thermal melting temperature determined by circular-dichroism/thermal stability assays, and if available, binding/activity assays are all valuable ways of assessing sample suitability and stability for cryo-EM. Negative stain is a powerful tool for assessing sample heterogeneity, and where possible we would always recommend it as an initial characterization step on the pathway to a cryo-EM structure. Several different types of apparatus are available to aid plunge freezing, including the commercially available FEI Vitrobot, Leica EM GP (and GP2) and Gatan Cryoplunge 3, along with several home-built systems. All commercial systems allow the nominal control of humidity and temperature in the sample chamber to reduce unwanted evaporation from the blotted thin film, as well as to make the process more reproducible. After blotting, given the small volume that remains on the grid, even a small amount of evaporation will result in an unwanted concentration of the specimen and buffer components, which can lead to drastic changes in temperature, ionic strength and pH, and therefore macromolecular complex stability. This can also result in protein aggregation and increased partitioning of the molecules to the air–water interface (Passmore & Russo, 2016 ▸). Besides humidity and temperature control, commercially available systems offer a range of other controllable parameters, such as cryogen temperature control and blotting ‘force’, and have significant differences in their setup. For anyone looking to invest in or upgrade their current system, we have summarized the main characteristics of the blotting instruments that are currently available in Table 1 ▸. The blotting of grids using filter paper is a broadly applicable and straightforward approach to making grids that has been suitable for high-resolution structure determination by single-particle cryo-EM. However, there are a number of limitations including, as discussed above, the instability and irreproducibility of thin films and potential evaporation from the film leading to a concentration of buffer/sample components. In addition, the vast majority of the 3–4 µl of specimen applied to the grid is removed by blotting, leaving nanolitre volumes on the grid. Finally, the filter paper contains a number of divalent metals and contaminants. The blotting procedure can typically take 1–6 s, and during this time contaminants from the filter paper can reach levels within the blotted sample which are high enough to be problematic for metal-sensitive proteins such as myosin and when using polymers that are unstable in the presence of high levels of divalent ions (Parmar et al., 2017 ▸; Walker et al., 1994 ▸). Where there have been significant changes in the field of EM in microscope hardware, automation of data collection and methods of data processing in the last five years, for the vast majority of cryo-EM users plunge freezing has remained the only method of preparing macromolecular complexes for single-particle analysis. As detailed above, there are a large number of variables which can, in some cases dramatically, alter particle distribution and thin-film formation across a grid during plunge freezing. These, combined with issues of reproducibility of conditions both within and between grid-making sessions, mean that for some samples it can be a significant challenge to find the right conditions and obtain suitable grids for data collection. We see many cases where small changes to a single variable make the difference between unusable and high-quality cryo-EM grids; the ‘dark art’ of grid preparation. For now, researchers’ best hope of optimizing their sample for cryo-EM remains to pay careful consideration to the biochemistry and to robustly characterize the macromolecular complex, followed by, ideally, systematic optimization of cryo-EM grids by trial and error. As methods and technology develop, cryo-EM sample preparation will hopefully become more robust and reproducible. Source: http://doi.org/10.1107/S2059798318006496