Limits...
Structure of CPV17 polyhedrin determined by the improved analysis of serial femtosecond crystallographic data.

Ginn HM, Messerschmidt M, Ji X, Zhang H, Axford D, Gildea RJ, Winter G, Brewster AS, Hattne J, Wagner A, Grimes JM, Evans G, Sauter NK, Sutton G, Stuart DI - Nat Commun (2015)

Bottom Line: Here we use an XFEL to determine the room temperature atomic structure for the smallest cytoplasmic polyhedrosis virus polyhedra yet characterized, which we failed to solve at a synchrotron.These protein microcrystals, roughly a micron across, accrue within infected cells.Our improved methods should open up more challenging biological samples to XFEL analysis.

View Article: PubMed Central - PubMed

Affiliation: Division of Structural Biology, The Wellcome Trust Centre for Human Genetics, University of Oxford, Roosevelt Drive, Oxford, Oxfordshire OX3 7BN, UK.

ABSTRACT
The X-ray free-electron laser (XFEL) allows the analysis of small weakly diffracting protein crystals, but has required very many crystals to obtain good data. Here we use an XFEL to determine the room temperature atomic structure for the smallest cytoplasmic polyhedrosis virus polyhedra yet characterized, which we failed to solve at a synchrotron. These protein microcrystals, roughly a micron across, accrue within infected cells. We use a new physical model for XFEL diffraction, which better estimates the experimental signal, delivering a high-resolution XFEL structure (1.75 Å), using fewer crystals than previously required for this resolution. The crystal lattice and protein core are conserved compared with a polyhedrin with less than 10% sequence identity. We explain how the conserved biological phenotype, the crystal lattice, is maintained in the face of extreme environmental challenge and massive evolutionary divergence. Our improved methods should open up more challenging biological samples to XFEL analysis.

No MeSH data available.


Related in: MedlinePlus

XFEL data analysis.(a) Distribution of Ewald sphere wavelengths for all reflections on a single image, before (grey, thick) and after (black, thin) orientation matrix refinement. Refinement resulted in a 0.117° rotational shift in the orientation model. (b) Cluster algorithm by a modified version of algorithm of Brehm and Diederichs24, showing positions of final artificially defined vectors corresponding to individual images (see Methods). Note the vectors fall into two sets, clearly separating the two indexing choices.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4403592&req=5

f1: XFEL data analysis.(a) Distribution of Ewald sphere wavelengths for all reflections on a single image, before (grey, thick) and after (black, thin) orientation matrix refinement. Refinement resulted in a 0.117° rotational shift in the orientation model. (b) Cluster algorithm by a modified version of algorithm of Brehm and Diederichs24, showing positions of final artificially defined vectors corresponding to individual images (see Methods). Note the vectors fall into two sets, clearly separating the two indexing choices.

Mentions: As an essential first step, we refined the orientation matrices of all crystals by an iterative process. Diffraction was modelled by describing the X-ray beam using an inflated energy bandwidth and infinitely small reciprocal lattice points (rlps), providing a ‘net’ to catch all possible reflections. These potential reflections were integrated, and strong reflections deemed ‘hits’. Refinement initially aimed to sharpen the histogram of frequency of hits vs wavelength by slightly adjusting the crystal orientation around the two axes orthogonal to the beam. This method worked well, a typical distribution before and after refinement is shown in Fig. 1a. The average combined angular correction was 0.075°. This refinement also optimized the estimate of the mean wavelength for each pulse. In contrast, the Kabsch paper reorients the crystal to minimize the rlps' distance from a nominally monochromatic Ewald sphere. An experimental model was also invoked, accounting for spot size (derived from crystal size), X-ray bandwidth and crystal mosaicity, each refined to best predict the observed selection of hits. For integration, spot positions were centred on the local maximal peak to allow for imperfect metrology. Images were metrology-corrected by cctbx.xfel16 and integrated using DIALS21. The bandwidth of the pulse was derived from the hit count histogram (Methods). The spot diameter refined to 1.8 × 10−4 Å−1, corresponding to a crystal size of ~0.5 μm, plausible based on electron microscopy analysis of the crystals. A mosaicity of 0.03° was used, which is consistent with synchrotron data but may also be an overestimate. Spots were generally well predicted (Supplementary Fig. 1). The limiting resolution for each image was chosen as the highest-resolution shell with I/σ(I) of 2.0, which almost always integrated to the edges of the image.


Structure of CPV17 polyhedrin determined by the improved analysis of serial femtosecond crystallographic data.

Ginn HM, Messerschmidt M, Ji X, Zhang H, Axford D, Gildea RJ, Winter G, Brewster AS, Hattne J, Wagner A, Grimes JM, Evans G, Sauter NK, Sutton G, Stuart DI - Nat Commun (2015)

XFEL data analysis.(a) Distribution of Ewald sphere wavelengths for all reflections on a single image, before (grey, thick) and after (black, thin) orientation matrix refinement. Refinement resulted in a 0.117° rotational shift in the orientation model. (b) Cluster algorithm by a modified version of algorithm of Brehm and Diederichs24, showing positions of final artificially defined vectors corresponding to individual images (see Methods). Note the vectors fall into two sets, clearly separating the two indexing choices.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4403592&req=5

f1: XFEL data analysis.(a) Distribution of Ewald sphere wavelengths for all reflections on a single image, before (grey, thick) and after (black, thin) orientation matrix refinement. Refinement resulted in a 0.117° rotational shift in the orientation model. (b) Cluster algorithm by a modified version of algorithm of Brehm and Diederichs24, showing positions of final artificially defined vectors corresponding to individual images (see Methods). Note the vectors fall into two sets, clearly separating the two indexing choices.
Mentions: As an essential first step, we refined the orientation matrices of all crystals by an iterative process. Diffraction was modelled by describing the X-ray beam using an inflated energy bandwidth and infinitely small reciprocal lattice points (rlps), providing a ‘net’ to catch all possible reflections. These potential reflections were integrated, and strong reflections deemed ‘hits’. Refinement initially aimed to sharpen the histogram of frequency of hits vs wavelength by slightly adjusting the crystal orientation around the two axes orthogonal to the beam. This method worked well, a typical distribution before and after refinement is shown in Fig. 1a. The average combined angular correction was 0.075°. This refinement also optimized the estimate of the mean wavelength for each pulse. In contrast, the Kabsch paper reorients the crystal to minimize the rlps' distance from a nominally monochromatic Ewald sphere. An experimental model was also invoked, accounting for spot size (derived from crystal size), X-ray bandwidth and crystal mosaicity, each refined to best predict the observed selection of hits. For integration, spot positions were centred on the local maximal peak to allow for imperfect metrology. Images were metrology-corrected by cctbx.xfel16 and integrated using DIALS21. The bandwidth of the pulse was derived from the hit count histogram (Methods). The spot diameter refined to 1.8 × 10−4 Å−1, corresponding to a crystal size of ~0.5 μm, plausible based on electron microscopy analysis of the crystals. A mosaicity of 0.03° was used, which is consistent with synchrotron data but may also be an overestimate. Spots were generally well predicted (Supplementary Fig. 1). The limiting resolution for each image was chosen as the highest-resolution shell with I/σ(I) of 2.0, which almost always integrated to the edges of the image.

Bottom Line: Here we use an XFEL to determine the room temperature atomic structure for the smallest cytoplasmic polyhedrosis virus polyhedra yet characterized, which we failed to solve at a synchrotron.These protein microcrystals, roughly a micron across, accrue within infected cells.Our improved methods should open up more challenging biological samples to XFEL analysis.

View Article: PubMed Central - PubMed

Affiliation: Division of Structural Biology, The Wellcome Trust Centre for Human Genetics, University of Oxford, Roosevelt Drive, Oxford, Oxfordshire OX3 7BN, UK.

ABSTRACT
The X-ray free-electron laser (XFEL) allows the analysis of small weakly diffracting protein crystals, but has required very many crystals to obtain good data. Here we use an XFEL to determine the room temperature atomic structure for the smallest cytoplasmic polyhedrosis virus polyhedra yet characterized, which we failed to solve at a synchrotron. These protein microcrystals, roughly a micron across, accrue within infected cells. We use a new physical model for XFEL diffraction, which better estimates the experimental signal, delivering a high-resolution XFEL structure (1.75 Å), using fewer crystals than previously required for this resolution. The crystal lattice and protein core are conserved compared with a polyhedrin with less than 10% sequence identity. We explain how the conserved biological phenotype, the crystal lattice, is maintained in the face of extreme environmental challenge and massive evolutionary divergence. Our improved methods should open up more challenging biological samples to XFEL analysis.

No MeSH data available.


Related in: MedlinePlus