Date Published: March 01, 2017
Publisher: International Union of Crystallography
Author(s): Patrick M. Collins, Jia Tsing Ng, Romain Talon, Karolina Nekrosiute, Tobias Krojer, Alice Douangamath, Jose Brandao-Neto, Nathan Wright, Nicholas M. Pearce, Frank von Delft.
A high-throughput method is described for crystal soaking using acoustic droplet ejection, and its effectiveness is demonstrated.
Obtaining protein–ligand complexes, the workhorse experiment in structure-based ligand design (SBLD), relies on two methods for achieving the prerequisite crystals: crystal soaking and co-crystallization (Hassell et al., 2007 ▸). Co-crystallization is achieved by adding the small molecule of interest to the protein prior to setting up a crystallization experiment, or by simply including it as a component in the crystallization condition. The potential ligand is free to bind to the protein in solution prior to the formation of a crystal lattice, allowing the freedom for potential structural changes. Crystal soaking is the process of taking pre-grown crystals and soaking them with the small molecule of interest. The potential ligand can access the binding sites by diffusing through solvent channels within the crystal lattice, as long as the sites are not involved in crystal packing or otherwise obscured (Danley, 2006 ▸).
We have developed a method for the soaking of protein crystals that is both gentle and rapid. By using the precision of acoustic dispensing to target the transfer of solvent and compounds away from sensitive protein crystals, an increase in solvent tolerance and X-ray diffraction reproducibility can be achieved, even at a rate of 100 crystals per minute. The use of control experiments to empirically determine the optimal experimental conditions for each crystal system has enabled fragment screening on multiple diverse protein targets in the XChem facility at Diamond. We typically observe fragment hit rates of between 1 and 10%, which is in good agreement with what other groups have observed (Bauman et al., 2013 ▸; Schiebel et al., 2016 ▸; Spurlino, 2011 ▸); hence, we conclude that the described protocol works efficiently. While the method may not be ideal for all crystal systems, its great advantage is that it is generic, since it can be applied to any crystal system with minimal preparatory work. The solvent-tolerance screen provides a means to detect problems in the soaking protocol in a systematic manner and avoids time being lost on soaking, collecting and analysing fragment-soaked crystals with little chance of success. A fast and efficient soaking protocol is instrumental for the XChem user program at Diamond Light Source, and the more than 30 000 crystals that have been soaked and data collected from them and the 800 fragment hits that have been seen so far are testament to its robustness.
The X-ray structures presented here are part of a larger fragment screen, with data available from the Zenodo data repository, https://doi.org/10.5281/zenodo.345833.