Computational protocol: Crystal structures of a double-barrelled fluoride ion channel

Similar protocols

Protocol publication

[…] Diffraction data for Bpe-S7 were processed by the Xia2 pipeline to XDS and scaled using aimless. The space group was determined to be P21212 with two Bpe dimers and four S7 monobodies in the asymmetric unit (). A phasing strategy was devised that used pre-derivatisation of Bpe mutated with a single cysteine residue (E94C) with Hg (II) Acetate prior to crystallisation (see above). None of the native crystals were isomorphous with the Hg derivatised crystals (Riso > 40 %), despite having similar cell dimensions. Indeed, Hg derivatised crystals were observed to diffract X-rays to slightly higher resolution than native crystals, therefore effort was directed at these samples for phasing and refinement. The four Hg sites were located using the single wavelength anomalous dispersion (SAD) method as implemented in SHELX with the positions further refined and initial phases calculated using SHARP with solvent flattening in SOLOMON. To improve the phases a second and third dataset were also collected at the Se edge using both Hg and seleno-L-methionine derivatized protein and another Hg derivatised dataset respectively (). All 16 Se plus 4 Hg sites were located using SHELX and this dataset was combined with the initial 3.6 Angstrom Hg derivatized data. Phases were substantially improved using SIRAS combining the three datasets in SHARP. We did not observe higher resolution diffraction in the Native crystals, which typically gave diffraction between 3.6 - 3.8 Angstroms. Our highest resolution dataset with optimal scaling statistics was one of the Hg-derivatised crystals, we therefore used this dataset for subsequent refinement of the model built into the experimental electron density maps calculated from SHARP (see below). For the Bpe-L2 crystals, data were similarly processed and scaled in spacegroup P1. Phases were calculated using molecular replacement as implemented in PHASER using the experimentally determined Bpe model and a homology model of the L2 monobody using a previously determined structure of a loop-library monobody (PDB: 3RZW). The unambiguous solution showed two Bpe homodimers and four L2 monobodies. The electron density maps clearly showed major differences in the selected, variable regions of the monobody. For the Ec2-S9 crystals the data were processed as above with space group P41. Phases were calculated using Se-SAD with 8 Se sites, and processed as above. The experimental electron density maps were of exceptional quality following phase extension to the highest resolution shell of 2.58 Å. Data were collected at Advanced Light Source beamlines 8.2.1 and 8.2.2, and Diamond Light Source, beamlines I24 and I04. [...] For the Bpe-S7 complex structure, a model for the channel was built into the experimental electron density maps calculated from SHARP using O with sigma-A-weighted 2Fo-Fc and mFo-DFc electron density maps. The S7 monobodies were initially built using a homology model based a previously determined structure of a side library monobody (PDB: 4JEG). These models were placed into the experimental electron density maps using MolRep. The partial models were further cycled back into phase calculation in SHARP to improve the initial solvent envelope used for the solvent flipping procedure. The amino acid side chains were then built using the Se and Hg sites to determine the correct register. Refinement of the Bpe-S7 model was carried out in Refmac5, against the highest resolution dataset for these crystals, 3.6 Å, which came from one of the Hg-derivatised crystals used for phasing (). No prior phase information was used during the refinement, however refinement was improved following anisotropic truncation of the structure factors. To avoid biasing the model, NCS was not employed except at the final round of refinement to improve model geometry. Model validation was carried out using the Molprobity server. The Ec2-S9 model was built directly into the experimental maps, using Se sites to ensure the correct register, and then monobodies were placed by molecular replacement using PHASER with a homology model based on S7. The Bpe-L2 model was built into the electron density maps calculated from PHASER following iterative rounds of structure refinement in PHENIX and Refmac5. The structural model was revised in real space with the program COOT. […]

Pipeline specifications

Software tools xia2, XDS, CCP4, SHELX, Molrep, REFMAC5, MolProbity, PHENIX, Coot
Applications Small-angle scattering, Protein structure analysis
Chemicals Fluorides, Sodium