Over the last century, X-ray crystallography has been critical in elucidating and understanding the structure-function paradigm of biological macromolecules. To date, it continues to be one of the most successful methods in elucidating atomic resolution structures of many uniquely different proteins that are crucial to the fundamental understanding of cell biochemistry, medicine, and early drug discovery1,2. However, protein crystallization remains a bottleneck in studying many protein targets, particularly membrane proteins and large protein complexes3. Consequently, protein crystallization is almost always considered an art due to the labor-intensive trial-and-error approaches employed4,5,6.
A precipitating agent is usually added to a protein solution at high concentration to form a well-ordered, regular, and repeating lattice arrangement of protein molecules, known as crystals. Under favorable conditions, such as temperature, pH, concentration, and precipitant agent, a supersaturated solution eventually forms, followed by crystal nucleation and growth7,8. Although there have been many advances in crystallization trial setups, predominantly with the development of high-throughput robotic systems and the availability of ready-made "sparse matrix" screens, the general approaches to protein crystallization have largely remained unchanged over the years. Common experimental protein crystallization techniques include vapor diffusion (hanging drop and sitting drop)9, microbatch (under oil)10,11, free-interface diffusion (microfluidic devices)12, and dialysis (using buttons and other techniques)13,14,15. However, other more specialized setups also exist, such as mesophase approaches for crystallizing membrane proteins16,17. While the majority of X-ray protein structures deposited in the Protein Data Bank have so far been solved through crystallization by vapor diffusion methods6,18,other approaches, such as crystallization by dialysis, seem to be underutilized, likely due to the practical aspects related to their experimental setup.
Crystallization by dialysis simply relies on the slow diffusion of solutes (precipitants, ions, additives, and buffers) through a semi-permeable membrane that simultaneously prevents protein molecules from circulating. This way, the protein solution is slowly brought into equilibrium, with the precipitant reaching the necessary concentration to crystallize. The system's kinetics depends on temperature, precipitant concentration, and the cellulose membrane molecular weight cut-off (MWCO)19. To date, the most popular crystallization setup by dialysis has been using microdialysis buttons made of transparent acrylic sheets. These are usually immersed in reservoirs (mostly using vapor diffusion hanging drop plates) containing the crystallization precipitant solutions. However, this lower-throughput method also requires specific assembly to seal the protein solution within the dialysis membrane placed over the button chamber, as illustrated in Figure 1. Moreover, air bubbles trapped between the dialysis membrane and the protein solution are a frequent problem that impair crystal growth. Another constraint of the method is the sample requirements, whereby much higher concentrations and volumes are necessary compared to vapor diffusion methods, to accommodate the dialysis buttons. Therefore, crystallization using microdialysis buttons has been perceived as an unappealing method, especially for difficult targets such as membrane proteins, whose purification yields are frustratingly low. Recently, microfluidic devices have been developed to facilitate protein crystallization by dialysis15. These chips have also been designed to have high X-ray transparency with low background, allowing the chips to be used for in situ data collection at room temperature, thus eliminating the inconvenience of harvesting and cryocooling crystals. Despite these advances, the approach is still very low-throughput and expensive.
Figure 1: Schematic representation of crystallization by dialysis using dialysis buttons. (A)Schematic representation of a crystallization dialysis button. (B) The protein solution is added to the microdialysis button chamber. (C) The dialysis membrane is held to the microdialysis button with the help of a rubber ring (O-ring) applied via an applicator. (D) The dialysis button is ready to be immersed in the reservoir containing the crystallization solution (dialysis solution), as shown in (E). The vial containing the immersed dialysis button must be sealed to avoid evaporation. Please click here to view a larger version of this figure.
Here, a straightforward protocol is presented for screening protein crystallization conditions and crystal growth using the 96-well high-throughput dialysis plate. These disposable plates are designed to be used similarly to the vapor diffusion crystallization plates (pipette then seal), as shown in Figure 2. The plates can accommodate up to 3.2 µL of protein and 350 µL of dialysis solution. Each well features a separate regenerated cellulose membrane to prevent cross-contamination between the wells. The setup takes around 10 min to complete and does not require any specialized equipment besides what can be found in all standard crystallization laboratories. Four different proteins, including two membrane proteins, are used to demonstrate and validate this approach as an effective method for high-throughput (HTP) protein crystallography.
Figure 2: Crystallization workflow using the microdialysis plate. (A) Removal of the red adhesive "cover film". (B) Dispensing the protein droplets into each of the drop wells. (C) The wells are covered with the UV "cover film". (D) The plate is inverted to add the dialysis solutions (or crystallization screen). (E) The plate is sealed and incubated. (F,G) Microscope inspection of the drops. Please click here to view a larger version of this figure.
The use of this crystallization by dialysis protocol was demonstrated using the 0.5 mL dialyzer tube (Figure 3) for the large-scale (hundreds to thousands) production of microcrystals, suitable for state-of-the-art data collection methods such as serial crystallography at both XFEL facilities20,21,22,23,24 and synchrotrons25,26,27, as well as for MicroED28,29,30 approaches.
Figure 3: Large-scale microdialysis crystallization using the dialyzer tube. (A) Schematic representation of the 0.5 mL dialyzer tube. (B) Side view of a beaker containing the crystallization solution and the floating tube rack holding a dialyzer tube. Please click here to view a larger version of this figure.