The structural biology community is abuzz with excitement over X-ray free-electron lasers (XFELs). This technology has been available for only a few years, but rapid methodological developments are already having a strong impact on the protein crystallography field. These lasers are a billion times brighter than the best synchrotron radiation sources and produce ultrashort, ∼50-femtosecond X-ray pulses, ideal for collecting high-quality single diffraction patterns from single, micrometer-sized protein crystals before radiation damage destroys the crystal. Single-crystal diffraction snapshots are then computationally merged to solve a protein structure. This powerful approach is known as serial femtosecond crystallography (SFX).

Most SFX studies to date have utilized a liquid injector to stream a steady supply of millions of microcrystals into the path of the XFEL beam. “The problem with these injectors is that the flow is too fast for the repetition rate of the [XFEL], which is 120 hertz,” explains Vadim Cherezov of The Scripps Research Institute. “Only one crystal out of tens of thousands of crystals gets hit by the beam, and all others are wasted between the pulses. It's not possible to slow it down because the flow becomes unstable.”

An LCP injector enables serial femtosecond crystallography of membrane proteins crystallized in gel-like LCP medium. Figure from Weierstall et al., Nature Publishing Group.

Recent work from a large international team of researchers to develop a new injector may greatly improve the efficiency of SFX. Cherezov's group has been using an established method of crystallizing G protein–coupled receptors (GPCRs) in lipidic cubic phase (LCP) medium. This approach often produces high-quality but small crystals, perfectly sized for SFX. The problem is that LCP is a gel and therefore incompatible with a liquid-based injector. The solution, developed by Uwe Weierstall and coworkers at Arizona State University, is a new device called an LCP microextrusion injector. With this tool, a gel-like crystal stream can be injected into the XFEL beam at lower speeds than the liquid injector allows, thereby minimizing waste of precious sample (Weierstall et al., 2014)2.

Using the LCP-based approach, Cherezov's team recently solved two GPCR structures at high resolution: the human serotonin 5-HT2B receptor, bound to its agonist ergotamine at 2.8-angstrom resolution (Liu et al., 2013), and the human smoothened receptor, bound to its antagonist cyclopamine at 3.2-angstrom resolution (Weierstall et al., 2014)1. Their success is particularly notable given that GPCRs are notoriously difficult either to produce in quantity for SFX using liquid injectors or to generate as large crystals for traditional synchrotron diffraction.

Unlike synchrotron diffraction, which requires cooling crystals to cryogenic temperatures to minimize radiation damage, with SFX, diffraction snapshots can be taken at room temperature before radiation damage has a chance to set in. “Cryo-cooling introduces sample artifacts into the structure,” explains Cherezov. “The room-temperature data give us a much better idea of the structure in the native environment.” Because the data processing steps for traditional crystallography and SFX approaches are completely different, his team was heartened to see that a traditional cryo-cooled structure and a room-temperature structure for the serotonin receptor were largely similar, though loop regions in the room-temperature structure showed greater dynamic flexibility, as one might expect.

For crystallographers wanting to try their own hand at LCP crystallization, the Arizona State engineering team is happy to provide construction and use details.