Computational protocol: Lipidic cubic phase serial millisecond crystallography using synchrotron radiation

Similar protocols

Protocol publication

[…] Serial crystallographic data were collected at the ESRF Microfocus Beamline (ID13; Grenoble, France) in the setup described in the Results section. Alignment of the X-ray beam onto the LCP stream was facilitated with a grid scan of the area around the tip of the injector. Diffraction patterns were collected from randomly oriented crystals with 10–50 ms exposure times at a rate of 10–17 Hz. Due to random failures in the Rayonix MX-170 CCD detector, the upper left quadrant was excluded during data analysis. The detector was set to 4 × 4 binning mode with a pixel size of 177 µm and a frame size of 960 × 960 pixels. Because of overheads due to saving data on the ESRF data server, the frame rate was limited to 17 frames s−1. Collected images (details in Fig. 1) were pre-processed using Cheetah (Barty et al., 2014) to exclude images without diffraction patterns. Higher hit rates and resolution were observed for crystals sized 20–40 µm. The CrystFEL program suite (White et al., 2012) was used for data processing.The conventional Cryo data set was collected at the PXI beamline of the Swiss Light Source (Paul Scherrer Institute, Villigen, Switzerland) in a cryostream at 100 K (details in Table. 1). The diffraction images were processed and scaled using XDS (Kabsch, 2010), followed by merging using Aimless (Evans & Murshudov, 2013). [...] A single molecular replacement solution was found using the SMX data set and the structure of sensory rhodopsin II [PDB code 1h68 (Royant et al., 2001), without ligands] as a search model in Phaser (McCoy et al., 2007). Initial phases were used to rebuild the bacteriorhodopsin model automatically with Phenix.autobuild (Adams et al., 2002), using simulated annealing refinement between iterative building steps. Manual building in Coot (Emsley & Cowtan, 2004) was used to complete the autobuild model, except for residues 1–4 and 234–249, and four side-chains for residues Q75, S158, K172 and R227. Further refinement was carried out using PDBREDO (Joosten et al., 2012), which suggested one TLS group for the whole protein chain. In a final round of model building and refinement in Refmac5 (Murshudov et al., 2011), the model was completed with ten water molecules, five lipid fragments and all-trans retinal. The retinal was refined with restrained geometry of the Schiff base at the covalent link to Lys216.The conventional Cryo data set was phased with Phaser, using bacteriorhodpsin [without ligands, PDB code 2ntu (Lanyi & Schobert, 2007)] as a search model. A single solution was found and the model was completed and refined using Refmac5. Retinal, 30 water molecules and eight lipid fragments were included in the last rounds of refinement. The unresolved region included the first four and last 17 residues, similar to the SMX structure. Moreover, a loop consisting of residues 157–163 was not included in the model obtained with the Cryo data, while for the SMX data it could be modelled into a weak electron density. Poorly resolved side-chains of K30, R164 and K172 were also not included in the model. The statistics are listed in Table. 1. The protein structures determined using the SMX and Cryo data sets have been deposited in the PDB with codes 4x31 and 4x32, respectively. […]

Pipeline specifications

Software tools XDS, CCP4, PHENIX, Coot, REFMAC5
Applications Small-angle scattering, Protein structure analysis
Organisms Dipturus trachyderma
Diseases Legg-Calve-Perthes Disease