'Hit and run' serial femtosecond crystallography of a membrane kinase in the lipid cubic phase

Abstract

The lipid-based bicontinuous cubic mesophase is a nanoporous membrane mimetic with applications in areas that include medicine, personal care products, foods and the basic sciences. An application of particular note concerns it use as a medium in which to grow crystals of membrane proteins for structure determination by X-ray crystallography. At least two variations of the mesophase exist. One is the highly viscous cubic phase, which has well developed long-range order. The other so-called sponge phase is considerably more fluid and lacks long-range order. The sponge phase has recently been shown to be a convenient vehicle for delivering microcrystals of membrane proteins to an X-ray free-electron laser beam for serial femtosecond crystallography (SFX). Unfortunately, the sponge phase approach calls for large amounts of protein that are not always available in the case of membrane proteins. The cubic phase offers the advantage of requiring significantly less protein for SFX but comes with its own challenges. Here, we describe the physico-chemical bases for these challenges, solutions to them and prospects for future uses of lipidic mesophases in the SFX arena.

Keywords: X-ray free-electron laser; crystal structure; enzyme; membrane protein; mesophase; monoacylglycerol.

Figure 1.

Schematic of the equilibrium temperature–composition phase diagram for the monoolein (9.9 MAG)–water system in the vicinity of 20°C. The different phases are shown as coloured zones and labelled accordingly. The cubic mesophase is extruded into the evacuated sample chamber for SFX under conditions indicated by the yellow star at 20°C and approximately 40% aqueous medium. Possible trajectories through the phase diagram taken upon dehydration, cooling and evaporative cooling are indicated by dashed arrows. The 20°C isotherm is identified by a horizontal dashed line. The liquid crystal-to-solid (L_c) transition is identified by the horizontal dashed line at 18°C. This schematic is based on the equilibrium phase diagram for 9.9 MAG reported in [15]. (Online version in colour.)

Figure 2.

Cartoon representation of the crystal-laden mesophase bolus as it is extruded through the nozzle (black triangles) of the LCP injector into the evacuated sample chamber at 20°C for serial femtosecond crystallographic measurements with an X-ray free-electron laser (XFEL). (a) Side view of the bolus where the gradient in colour from left to right corresponds to the gradient in temperature and composition along the length of the bolus induced by evaporative cooling. (b) End on view of the bolus where the gradient in colour corresponds to the gradient in temperature and composition (arrows) along the radius of the cylindrical bolus induced by evaporative cooling. Pristine, undamaged membrane protein crystals are coloured yellow and are shown dispersed in a blue cubic mesophase. Stars correspond to sites where the mesophase has transformed from the cubic to the solid L_c phase that may damage the crystals (red) and introduce defects (lightning bolt) in the bolus thereby affecting flow. The star in (b) with the enlarged grey background is drawn to suggest local heating due to the heat of fusion associated with the solidification reaction that may damage dispersed crystals nearby. (Online version in colour.)

Figure 3.

Microcrystals of DgkA grown in the cubic mesophase with 7.9 MAG as host lipid at 20°C in a 0.5 ml syringe. Details of sample preparation are described in [27]. (Online version in colour.)