A simple pressure-assisted method for MicroED specimen preparation

Micro-crystal electron diffraction (MicroED) has emerged as a powerful technique for structure determination of macromolecules using nano- and micron-sized crystals, overcoming limitations of X-ray diffraction for small crystals. Despite its potential, MicroED's application has been hindered by the challenges of specimen preparation, a process typically delicate and time-consuming. MicroED necessitates sub-micrometer-thick crystals to allow electron beam penetration and minimize multiple scattering, requiring thin surrounding ice for optimal signal-to-noise ratio while protecting crystals from dehydration. Existing methods, like adjusting crystallization conditions, mechanical crystal segmentation, or cryo-focused ion beam milling (cryo-FIB), have limitations. Methods for handling crystals grown in viscous buffers (e.g., those containing high molecular weight polymers like polyethylene glycols or lipid cubic phase) are particularly challenging, often yielding insufficient crystal density on the TEM grid and thick ice layers. Pipetting-blotting-plunging routines, manual back-side blotting, and direct crystallization on EM grids have been attempted but lack simplicity and universality. Therefore, there is a critical need for a simple and widely applicable MicroED specimen preparation method. This paper introduces Preassis, a novel pressure-assisted method designed to address these challenges.

Several methods have been explored to address the challenges of MicroED specimen preparation. Adjusting crystallization conditions to produce smaller crystals is one approach, but control over crystal size remains difficult. Mechanical methods, such as vigorous pipetting or sonication, can fragment larger crystals, but these techniques can damage delicate crystals and lead to inconsistent results. Cryo-FIB milling offers precise control over crystal thickness but requires specialized equipment and expertise. The pipetting-blotting-plunging method, adapted from cryo-EM, often removes many microcrystals during blotting and struggles with viscous solutions. Other methods, like manual back-side blotting and direct crystallization on EM grids, have also been explored but lack the simplicity and universality needed for widespread adoption. High molecular weight polymers, commonly used in protein crystallization, increase solution viscosity, posing a significant challenge for existing preparation methods. Similarly, lipid cubic phase (LCP) crystallization, often employed for membrane proteins, results in extremely viscous samples that are difficult to handle with current methods. The need for a robust and straightforward method capable of handling a wide range of samples, particularly those with high viscosity, remains unmet.

Preassis is a pressure-assisted method where excess liquid is removed from a sample applied to an EM grid through the use of suction/pressure. A simple setup involves placing the grid on a filter paper above a Buchner flask connected to a vacuum pump. A droplet of the crystal suspension is placed onto the grid; the vacuum pulls excess liquid through the grid, leaving crystals behind. The grid is then plunged into liquid ethane for vitrification. The vitrified ice thickness can be controlled by adjusting the pressure (achieved by varying pumping speed), and the carbon hole sizes of the EM grids used. The methodology section of the paper details the setup, step-by-step procedures, and parameters used. The influence of pressure, grid type (hole size), and humidity on ice thickness is also investigated. The researchers compared Preassis to the standard double-sided blotting Vitrobot method, evaluating crystal density, MicroED data quality, and ice thickness. Different protein crystal samples with varying viscosities were tested using both methods, including tetragonal lysozyme (high concentration), R2lox (high viscosity), and GTPase (high viscosity). The study included analyzing TEM images to assess ice thickness and using MicroED data to evaluate data quality, comparing resolutions and signal-to-noise ratios between Preassis and Vitrobot prepared samples. Systematic comparisons were conducted to determine the optimal pressure and grid combinations for various sample viscosities, including samples with different concentrations of PEG 6000. The study also examines the ideal characteristics of grid squares and crystals for high-quality MicroED data collection.

Preassis demonstrates significant advantages over existing methods, particularly for samples with low crystal concentrations and high viscosity. Compared to the Vitrobot method, Preassis retains significantly more crystals (up to two orders of magnitude more) on the TEM grid, enabling MicroED analysis even with highly diluted samples. The ice thickness is effectively controlled by manipulating pressure and grid properties (hole size). Preassis successfully processes samples with a wide range of viscosities, from non-viscous samples requiring no additional pressure to highly viscous samples (e.g., those with high concentrations of PEG) requiring higher pressures. High-resolution MicroED data were obtained for crystals grown in viscous media (44% PEG 400 and 30% PEG 4000), significantly improving data quality compared to Vitrobot results. For samples with high viscosity, such as those containing high concentrations of polyethylene glycol (PEG), larger hole sizes and high pressure are needed for optimal results. Analysis showed that humidity has a far smaller effect on ice thickness using Preassis, when compared to Vitrobot. Preassis consistently produces usable grids under both ambient and high humidity conditions, unlike the Vitrobot method. The study details how to optimize pressure and grid selection for new protein samples. The paper identifies optimal characteristics for grid squares and crystals, emphasizing the importance of thin ice and appropriate hydration of the crystals, even in the presence of viscous buffers. Notably, the improved sample preparation using Preassis was essential for the successful structure determination of the R2lox protein, previously unsuccessful using the Vitrobot method. In summary, Preassis was shown to be a superior method for specimen preparation for MicroED experiments, particularly for samples with low crystal concentration or high viscosity.

The development of Preassis addresses a significant bottleneck in MicroED, namely the preparation of high-quality specimens. The method's simplicity, universality, and effectiveness across a wide range of sample conditions significantly expands the applicability of MicroED. The ability to handle highly viscous samples, previously a major limitation, opens new possibilities for structural studies of proteins typically difficult to crystallize using other methods, including membrane proteins. The ability to control ice thickness allows for optimization of data quality and signal-to-noise ratio. The high retention of crystals, even at low concentrations, makes Preassis advantageous for samples with limited material availability. The improved data quality demonstrated in the study reinforces the significant impact of the Preassis method. This is particularly evident in the case of R2lox, which shows significant improvement in data resolution compared to previous attempts using Vitrobot.

Preassis offers a simple, versatile, and high-performing method for preparing MicroED specimens. Its ability to control ice thickness and handle high-viscosity samples expands MicroED applications significantly. Future directions include automating the process and adding an environmental chamber to further enhance throughput and reproducibility. The authors suggest additional research to focus on techniques for growing smaller crystals directly and on developing screening methods suited for these small crystals.

The current Preassis setup results in non-uniform ice thickness across the EM grid. While this might offer an advantage by increasing the chance of finding suitable crystals, the non-uniformity necessitates manual searching for suitable regions. For extremely viscous samples, like those crystallized in LCP, additional methods might be needed to reduce viscosity before using Preassis. The optimal pressure for successful crystal preparation may need to be adjusted depending on the specific properties of the crystals, such as size and shape, and the viscosity of the solution. Further optimization may be needed for different types of crystals.