Arrangement of protein subunits and the distribution of nucleic acid in turnip yellow mosaic virus. II. Electron microscopic studies

Abstract

Earlier electron microscope investigations by the method of negative staining with sodium phosphotungstate showed that the surface of the particle of turnip yellow mosaic virus is composed of 32 morphological units arranged at the vertices of a polyhedron approximating in shape to a rhombic triacontahedron. The substructure of the morphological units was not resolved, although it was suggested on the basis of the X-ray diffraction studies, which showed the presence of strict 532 symmetry within the protein shell, and the observed equality of size of the morphological units, that the twenty 6-co-ordinated morphological units were each composed of six equivalent subunits and the twelve 5-co-ordinated morphological units of five equivalent subunits. We have confirmed the suggested rotational symmetry of the morphological units by electron microscopy, using uranyl acetate as a negative stain. In some cases the morphological units appear polygonal—pentagonal in the 5-co-ordinated positions and hexagonal in the 6-co-ordinated positions—and in the most favourable cases they can be seen to consist of discrete units about 33 A apart. We identify these parts with the centres of contrast (by the negative stain) of the outward protruding tips of the ultimate structure units of which the protein shell is built. The arrangement of structure units arrived at by inspection of single-particle images is consistent with the symmetry of the T = 3 icosahedral surface lattice, so that there are 180 structure units. This structure has been confirmed by the analysis of the patterns discernible when the particles are lined up in 2-dimensional crystals over holes in the carbon substrate. These patterns can be followed across the array of particles, and so are less sensitive to the imperfections in the quality of individual particle images, and provide a more accurate means of measurement of the co-ordinates of the structure units. The arrangement of structure units is fairly uniform at the surface of mean contrast of the protein shell, with no clumping into hexamers and pentamers. In the images of the virus particles there is, however, additional density linking the centres of the units into 32 large morphological units, which as a result are much more dominant in appearance than in the images of the top-component particles. We attribute this difference in appearance to the presence of RNA in the virus particles, so that the RNA must be distributed in such a way that local concentrations occur in the regions of the 32 surface lattice points. This picture of the virus structure is consistent with the X-ray diffraction studies reported in the preceding paper. The particle images and the modes of packing of the virus particles have also enabled deductions to be made regarding the morphology of the particles. The structure units do not point radially outwards from the main body of the particle, which is inaccessible to negative stain, but, at an outer radius, are splayed towards the centres of hexamers and pentamers. The surface of the particle is highly grooved, the outer tips of the 32 morphological units lying at a radius of about 150 A, while the minimum radius in the valleys between the morphological units is about 130 A. The proportion of particle images on which the structure units can be resolved is quite small. In the Appendix, it is shown by tilting experiments that the majority of particle images, especially over holes in the carbon substrate, arise from superposition of detail from two sides of the particle; hence a very high proportion of the particles is completely enveloped in the negative stain. In these films of stain over holes, the particle structure is also often well preserved. The particle images on the carbon substrate tend, however, to be more one-sided and to originate from a flattened surface of contrast. The majority of two-side images are quite complex; only in views of the particle seen exactly down a 2-fold axis do the centres of the large morphological units and, in fact, the positions of the structure units for the symmetrical arrangement found on the two sides of the particle, superpose exactly. We have, however, been able to account for about half of the two-side particle images that occur over holes in the carbon substrate with a limited gallery of analogues (shadowgraphs) of two-side images of the structure.