Ab initiocrystal structure determination of spherical viruses that exhibit a centrosymmetric location in the unit cell

Abstract

Ab initio phasing by non-crystallographic symmetry averaging coupled with solvent flattening has previously been used to determine the structure of canine parvovirus. As CPV particles were located at general positions, initial phases were generated with a spherical shell deviating from the centre of symmetry. In many virus crystals, the viral particles are located at positions with a centre of symmetry in the unit cell. Thus, the initial phases calculated with a spherical shell model have a centre of symmetry. The inherent difficulty in structural determination was breaking the centre of symmetry in the initial phase. The centric nature of the initial phases of the rice dwarf virus crystal, however, was successfully broken at low resolution by iteration of the density-modification method described by Tsao et al. (1992). In this study, ab initio phasing was tested for seven viruses ranging from 82 to 344 A in radius. Although the crystal structures of the initial spherical shell models for each virus had a centre of symmetry, the centric natures of the initial phases were successfully broken by non-crystallographic symmetry averaging coupled with solvent flattening at a low resolution; these phases were then successfully extended to high-resolution measurements. A novel procedure of ab initio phasing for spherical virus crystals is proposed.

Ab initiophasing by molecular averaging in real space with new criteria: application to structure determination of a betanodavirus

“…In the second trial using a revised mask, which was made from the rough envelope, an interpretable map was obtained even with the mask cutting a small portion of the P-domain. This result is similar to that described in a previous report (Taka et al, 2005). In another simulation trial using the spherical mask with the addition of a hemispherical mask at the P-domain position that covered the entire P-domain region, the map was finally obtained as a Babinet inversion, which yielded a solution in which the phases were incremented by and in the production of negative electron density for the protein region (Tsao et al, 1992;Plevka et al, 2011).…”

“…As the starting model, a spherical-shell model with uniform density was used; the outer and inner radii of the model were 159 and 119 Å , respectively. These radius values were typical initial parameters for the T = 3 virus particles suggested in previous work (Taka et al, 2005) The initial mask created from the atomic structure of T = 3 FHV with a large mask radius of 13 Å around each atom. (a) The initial mask corresponding to the finally obtained icosahedral asymmetric unit (i-ASU) of T = 3 GNNV-LP.…”

“…The feasibility of ab initio phase determination has been demonstrated for spherical viruses (Tsao et al, 1992). Ab initio trials and ISSN 2059-7983 procedures have been reported for known icosahedral viruses (Taka et al, 2005). An assessment of the application of NCSA to icosahedral viruses, together with a detailed investigation of successful and unsuccessful cases, has also appeared (Plevka et al, 2011).…”

Crystal Structures of a Piscine Betanodavirus: Mechanisms of Capsid Assembly and Viral Infection

“…The diffraction data of GNNV-LP crystals contained 499,184 reflections and was 98% complete at a resolution range from 50 to 3.6 Å. To help initial phase determination by ab initio phasing [ 56 ], the very low-resolution data of the GNNV-LP crystals up to 266 Å were measured, and only a few reflections were not measured in the region of very low resolution (> 100 Å).…”

Iterative projection algorithms for ab initio phasing in virus crystallography

“…This approach has not developed into a general technique. Taka et al (2005) [24] conducted simulations of the reconstruction of icosahedral viruses starting with a spherical shell model and using conventional density modification. Although some success was achieved, the resolution was limited (maximum resolution between 4.3Å and 10Å), and some cases required incorporation of some calculated, rather than measured, diffraction amplitudes.…”