Outrunning free radicals in room-temperature macromolecular crystallography

Abstract

A significant increase in the lifetime of room-temperature macromolecular crystals is reported through the use of a high-brilliance X-ray beam, reduced exposure times and a fast-readout detector. This is attributed to the ability to collect diffraction data before hydroxyl radicals can propagate through the crystal, fatally disrupting the lattice. Hydroxyl radicals are shown to be trapped in amorphous solutions at 100 K. The trend in crystal lifetime was observed in crystals of a soluble protein (immunoglobulin γ Fc receptor IIIa), a virus (bovine enterovirus serotype 2) and a membrane protein (human A(2A) adenosine G-protein coupled receptor). The observation of a similar effect in all three systems provides clear evidence for a common optimal strategy for room-temperature data collection and will inform the design of future synchrotron beamlines and detectors for macromolecular crystallography.

Figure 1

Comparison of FcγRIIIa crystal lifetimes in both stop–start and continuous data-collection modes. Data sets when the fast shutter was closed between images are termed ‘stop–start’ and are shown as blue open triangles, with the mean for each regime shown as a red circle. Data sets when the detector was read out continuously are shown as filled black triangles, with the mean shown as a red diamond (top). For all data, elapsed time is equal to the exposure time plus the shutter closed time between images (0 s in the case of continuous data collection). The bottom panel shows the lifetime of FcγRIIIa crystals as a function of frame rate. All data were collected in continuous data-collection mode, i.e. with the X-ray shutter left open for the duration of the experiment. Data for individual crystals are shown as black triangles and the means for each dose-rate regime are shown as red diamonds.

Figure 2

Plot showing variation in the lifetime of BEV crystals as a function of frame rate. At frame rates above 15 Hz an increase in the mean lifetime of the crystals is apparent. The dose per frame is kept constant for all crystals and frame rates.

Figure 3

Plot showing variation in the lifetime of A_2AAR crystals as a function of frame rate. A linear relationship between frame rate and lifetime can be observed. The dose per frame is kept constant for all crystals and frame rates.

Figure 4

Change in the UV–Vis absorption spectra of FcγRIIIa crystallization solution at 100 K upon exposure to X-rays. (a) shows a contour plot showing changes at all wavelengths between 200 and 750 nm, while (b) shows the change in absorbance at 240 and 593 nm. The X-ray shutter is opened at α (t = 3.40 s), the 593 nm peak reaches a maximum at β (6.50 s), the X-ray shutter is closed at γ (12.40 s) and the 100 K nitrogen stream is blocked at δ (83 s).

Figure 5

Change in UV–Vis absorbance of FcγRIIIa crystallization solution at 240 and 593 nm irradiated at 100 K and then warmed to 180 K. The x axis reflects the readback temperature from the cryojet controller; the temperature at the sample will differ from this. Note that this axis is nonlinear and approximate. At approximately 150 K absorbance at low wavelengths decreases significantly, indicating increased mobility of radicals. The absorbance at 240 K does not return to zero owing to discolouration and icing of the sample upon slow warming.

Figure 6

Lifetime of FcγRIIIa, BEV and A_2AAR crystals as a function of dose. Note that the FcγRIIIa data have been scaled by a factor of two to facilitate comparison with the BEV 2 and A_2A data. All crystal types show an increase in lifetime at higher frame rates.