Structure of the Month: October 2009 [see all]

de novo Phasing at Home Using the Anomalous Signal of Iodide: The Structure Solution of Cytoplasmic Domain of the Inner Membrane Protein EpsF

Jan Abendroth¹, Daniel D. Mitchell², Konstantin V. Korotkov², Tonya L. Johnson³, 
Allison C. Kreger², Maria Sandkvist³, Wim G. Hol²

Pathogens use sophisticated protein machineries for the transport of virulence factors through their outer membrane. In Vibrio cholerae the Type II Secretion System (T2SS) is responsible for the extracellular transport of e.g. the major virulence factor cholera toxin. The T2SS consists of up to 15 Eps proteins (Extracellular Protein Secretion), most of them in several copies, and spans from the cytoplasm, via inner membrane and periplasm, to the actual pore in the outer membrane. As the entire complex is a very long shot to crystallize, we focused on crystallizing sub-complexes or individual proteins.

Among several inner membrane proteins of the T2SS, EpsF is the only one that crosses the membrane multiple times. Even though the full-length protein (406 residues) from different Vibrio species could be expressed and purified in a number of detergents in milligram amounts, we were not able to obtain crystals. Constructs of the soluble cytoplasmic N-terminal domain (residues 1-171) did not yield crystals either. Sequence analysis revealed a stretch of residues with very little sequence homology within the broader sequence family. A truncated version of the N-terminal domain comprising residues 56-171 (cyto-EpsF) could readily crystallized under a wide range of conditions: As the precipitant a spectrum of PEGs (400 to 8,000) could be used; the pH could be varied between 6.0 and 8.0; crystallization was dependent on divalent cations, Ca²⁺ yielded crystals of much better quality than Mg²⁺, while even small amounts of Sr²⁺ or Ba²⁺ led to precipitation of the protein. The crystallization conditions optimized to: 12.5% PEG 400, 200 mM CaOAc₂, 100 mM MES pH 7.0.

For cryoprotection, the crystals were sequentially transferred in buffers containing 200 mM CaOAc₂, 100 mM MES pH 7.0, plus increasing concentrations of PEG 400 and NaCl in the following ratios: 20%/300 mM, 20%/600 mM, 25%/600 mM, 30%/600 mM. The sequential transfer could keep the diffraction spots less mosaic. A data set up to 1.7 Å resolution could be collected at our home-source (see Data collection details below): MicroMax-007 HF rotating anode (Rigaku) equipped with VariMax HF (Osmic) monochromator and a Saturn 944 (Rigaku) CCD detector. Without a useable search model at hand the structure could not be solved by Molecular Replacement.

We therefore attempted to obtain experimental phase information from selenated samples at the home source by taking advantage of the NaCl in the purification buffer of cyto-EpsF, which comprises half of the crystallization drop: 20 mM Tris-HCl, pH 8, 300 mM NaCl, 1 mM TCEP. During the sequence of cryo buffers outlined in the previous paragraph, NaCl was gradually replaced with NaI. From a crystal soaked in NaI buffer, 720 images with 0.5° phi slicing were collected on our in-house system, yielding a data set up to 1.9 Å resolution with a 12-fold overall redundancy. Data were processed with both the d*TREK and XDS packages. A strong anomalous signal could be detected with SHELXC. Sixteen sites for anomalous diffracters were located with SHELXD. The sites were refined, and initial phases were calculated with SHARP. After density modification with SOLOMON/DM, the maps were readily interpretable for automatic model building by ARP/wARP which could build an almost complete model.

Several different atom types contributed to the anomalous scattering: 12x I^- (f" = 6.75e-), 2x Ca²⁺ (f" = 1.28e-), 10x Se (f" = 1.14e-), and 2x S (f" = 0.56e-). The 16 sites that SHELXD could locate were later identified as twelve iodide sites, two Ca²⁺ sites and two false sites. The selenium and sulphur sites were not strong enough, however, could be detected later in anomalous Fourier maps. Iodide could be found in two chemical environments, either in hydrophobic or in basic patches on the protein surface (Figure 1-3).

In a separate experiment, a highly redundant data set was collected at the SSRL beamline 9-2 at longest wavelength accessible at this tunable beamline (λ =2.066 Å), see Data collection details below. Two strong Ca²⁺ sites (f"=2.13 e-) and 13 weak sulphur sites (0.95 e-) were used for phasing. Using the same set of programs as outlined above, a complete model could be built into the experimental electron densities.

The first example shows how the strong anomalous signal of iodide at the CuKα wavelength can be used to obtain useful experimental phase information at the home source. The property of iodide to bind to both hydrophobic and basic pockets certainly is of advantage. The second example shows how the much weaker anomalous signal of calcium and sulphur can be used to phase a protein at long wavelength.

Citation: The Three-Dimensional Structure of the Cytoplasmic Domains of EpsF from the Type 2 Secretion System of Vibrio Cholerae. Jan Abendroth, Daniel D. Mitchell, Konstantin V. Korotkov, Tonya L. Johnson, Allison C. Kreger, Maria Sandkvist, Wim G. Hol (2009), J. Structural Biology 166(3), 303-15.

Data collection details