Computational protocol: Structural analysis of variant of Helicobacter pylori MotB in its activated form, engineered as chimera of MotB and leucine zipper

Similar protocols

Protocol publication

[…] Hp-chimMotB was thoroughly dialysed against buffer containing 100 mM sodium acetate pH 4.6 and 200 mM NaCl, with the dialysis buffer retained in order to determine its contribution to scattering. SAXS measurements were acquired at room temperature on the SAXS/WAXS beamline at the Australian Synchrotron, using a 1 M Pilatus detector (DECTRIS). Hp-chimMotB (0.25 mg mL−1, 0.11 mg mL−1 and 0.05 mg mL−1) and the respective matching dialysis buffer were exposed to X-rays (λ = 1.03 Å) for 1 sec, as the sample flowed through a 1.5 mm quartz capillary, and scattering data were collected over an s range of 0.015–0.5 (camera length = 1.6 meters, s is the magnitude of the scattering vector). The resulting 2-D scattering images were radially averaged and normalised to give absolute scattering intensities. After scaling, scattering intensities of the respective buffer and empty capillary were subtracted from the intensities of each Hp-chimMotB sample using PRIMUS. Data analysis was performed using the ATSAS suite of programs,. Guinier plots generated with PRIMUS were used to calculate values for forward scattering I(0) and radius of gyration (R g) using low resolution data (sR g < 1.3). An indirect Fourier transform of the scattering curve I(s) calculated by GNOM yielded intraparticle distance distribution function in real space P(r).The R g values were consistent throughout each different protein concentration tested. The molecular mass (MW) of chimMotB was calculated using the zero scattering angle value I(0) on the absolute scale and the known scattering of water as previously described; the partial specific volume of the protein was assumed to be 0.74 cm3g−1. An estimate of the MW (Da) of the protein samples, calculated by dividing the Porod volume Vp (Å3) (derived using DATPOROD) by 1.7 gave values close to those calculated from I(0) (Table ). The calculated values for the MW of Hp-chimMotB were consistent across each concentration, indicating that no aggregation was present. The highest protein concentration (0.25 mg mL−1) data in the s range between 0.026 and 0.238 nm−1 was used for shape reconstruction and modelling. [...] Particle shapes were restored from the experimental scattering profile of Hp-chimMotB using an ab initio simulated-annealing-based procedure implemented in DAMMIN. The algorithm fits models to the experimental data I(s) to minimise the discrepancy between the experimental and calculated scattering curves (χ2). The D max and R g values derived from the P(r) analysis were used. The starting model contained 6337 dummy atoms densely packed into a sphere of a diameter equal to D max (118 Å). Modelling was restrained to prolate shapes with a two-fold axis of symmetry along the direction of anisometry, as Hp-chimMotB exists as a homodimer in solution. Twenty simulations were performed, which generated very similar, but not identical, shapes. An averaged filtered structure was generated to determine common structural features using DAMAVER, and this structure was used as a fixed core in a final run of DAMMIN to refine the averaged model. [...] In order to assess the size distribution and interdomain flexibility within the ensemble of conformations adopted by Hp-chimMotB in solution, the SAXS data was analysed using the ensemble optimisation method (EOM),. Firstly, RanCh (random chain generator) was used to generate a pool of 10,000 possible Hp-chimMotB dimer models by treating the N-terminal leucine zipper/plug coiled coil region and the C-terminal domain as rigid bodies and connecting them by a non-clashing linker in random yet stereochemically reasonable conformations. The N-terminal leucine zipper/plug coiled coil region was modelled using the coordinates of a long homodimeric coiled coil observed in the crystal structure of the GCN4 zipper from S. cerevisiae fused with human vimentin coil 2B fragment (PDB ID: 1GK6, residues 355–393) as a three-dimensional template. Residues 355–379 of the template, matching in sequence with residues 1–25 in Hp-chimMotB, were left intact, whilst residues 380–393 were individually substituted to match the sequence of the MotB plug helix 41–54 (residue numbering 26–39 in Hp-chimMotB) by using the simple mutate function in COOT. The previously published crystal structure of Hp-MotB-C (PDB ID: 3CYP, residues 119–251) was used to model the C-terminal domain. All models in the pool had a two-fold symmetry axis going along the N-terminal coiled coil and C-terminal domain, consistent with the symmetry observed in the corresponding crystal structures. The theoretical scattering intensities of each model were calculated using CRYSOL. GAJOE (Genetic Algorithm Judging Optimisation of Ensembles), was then employed to select an ensemble of the pool structures that, when averaged, had similar theoretical scattering intensities to the experimental data. This algorithm was run 100 times, and the ensemble that showed least discrepancy with the experimental data was selected for the analysis of the D max distribution.The BUNCH algorithm was employed to generate individual models of Hp-chimMotB with the best fit to the experimental SAXS data. The models with P2 symmetry were assembled from the known crystal structures of the N-terminal and C-terminal fragments connected via a dummy-reside linker as described above, with the following modifications. Two reported conformations of the dimeric C-terminal domain, represented by the crystal structures of Hp-MotB-C/L90 (PDB ID: 3S0H) and Hp-MotB-C (PDB ID: 3CYP), were tested. The positions of both the N-terminal leucine zipper/plug coiled coil and the C-terminal domain were fixed while the simulated annealing procedure was used to find the probable conformations of the linker that would be in agreement with the experimental data. In separate BUNCH runs, the spacing between the leucine zipper/plug and the C-terminal domain was varied between 42 and 47 Å in 1 Å increments, such that the resulting D max values of the models (115–120 Å) were centred around the P(r)-derived D max value of 118 Å. For each of the six different spacing values, 20 BUNCH models were calculated and their quality was assessed by calculating the χ value between the scattering intensities of the model and the experimental data. […]

Pipeline specifications

Software tools ATSAS, DAMMIN, EOM, Coot, CRYSOL
Applications Small-angle scattering, Protein structure analysis
Organisms Helicobacter pylori, Dipturus trachyderma