The structures of the Cabb-B and CM1841 strains of cauliflower mosaic virus (CaMV) have been solved to about 3 nm resolution from unstained, frozen-hydrated samples that were examined with low-irradiation cryo-electron microscopy and three-dimensional image reconstruction procedures. CaMV is highly susceptible to distortions. Spherical particles, with a maximum diameter of 53.8 nm, are composed of three concentric layers (I-III) of solvent-excluded density that surround a large, solvent-filled cavity (approximately 27 nm dia). The outermost layer (I) contains 72 capsomeric morphological units, with 12 pentavalent pentamers and 60 hexavalent hexamers for a total of 420 subunits (37-42 kDa each) arranged with T = 7 icosahedral symmetry. CaMV is the first example of a T = 7 virus that obeys the rules of stoichiometry proposed for isometric viruses by Caspar and Klug (1962, Cold Spring Harb. Symp. Quant. Biol. 27, 1-24), although the hexameric capsomers exhibit marked departure from the regular sixfold symmetry expected for a structure in which the capsid protein subunits are quasi-equivalently related. The double-stranded DNA genome is distributed in layers II and III along with a portion of the viral protein. The CaMV reconstructions are consistent with the model based on neutron diffraction studies (Kruse et al., 1987, Virology 159, 166-168) and, together, these structural models are discussed in relation to a replication-assembly model (Hull et al., 1987, J. Cell Sci. (Suppl.) 7, 213-229). Remarkable agreement between the reconstructions of CaMV Cabb-B and CM1841 suggests that other strains of CaMV adopt the same basic structure.

The structure of rhesus rotavirus was examined by cryoelectron microscopy and image analysis. Three-dimensional reconstructions of infectious virions were computed at 26- and 37-A resolution from electron micrographs recorded at two different levels of defocus. The major features revealed by the reconstructions are (a) both outer and inner capsids are constructed with T = 13l icosahedral lattice symmetry; (b) 60 spikelike projections, attributed to VP4, extend at least 100 A from the outer capsid surface; (c) the outer capsid, attributed primarily to VP7, has a smoothly rippled surface at a mean radius of 377 A and is perforated by 132 aqueous holes ranging from 40-65 A in diameter; (d) the inner capsid has a "bristled" outer surface composed of 260 trimeric-shaped columns of density, attributed to VP6, which merge with a smooth, spherical shell of density at a lower, mean radius of 299 A, and which is perforated by holes in register with those in the outer capsid; (e) a "core" region contains a third, nonspherical shell of density at a mean radius of 225 A that encapsidates the double-stranded RNA genome; and (f) the space between the outer and inner capsids forms an open aqueous network that may provide pathways for the diffusion of ions and small regulatory molecules as well as the extrusion of RNA. The assignment of different viral structural proteins to specific features of the reconstruction has been tentatively made on the basis of excluded volume estimates and previous biochemical characterizations of rotavirus.

The three-dimensional structure of Nudaurelia capensis beta virus (N beta V) was reconstructed to 3.2-nm resolution from images of frozen-hydrated virions. The distinctly icosahedral capsid (approximately 40-nm diameter) contains 240 copies of a single 61-kDa protein subunit arranged with T = 4 lattice symmetry. The outer surface of unstained virions compares remarkably well with that previously observed in negatively stained specimens. Inspection of the density map, volume estimates, and model building experiments indicate that each subunit consists of two distinct domains. The large domain (approximately 40 kDa) has a cylindrical shape, approximately 4-nm diameter by approximately 4-nm high, and associates with two large domains of neighboring subunits to form a Y-shaped trimeric aggregate in the outer capsid surface. Four trimers make up each of the 20 planar faces of the capsid. Small domains (approximately 21 kDa) presumably associate at lower radii (approximately 13-16.5 nm) to form a contiguous, non-spherical shell. A T = 4 model, constructed from 80 trimers of the common beta-barrel core motif (approximately 20 kDa) found in many of the smaller T = 3 and pseudo T = 3 viruses, fits the dimensions and features seen in the N beta V reconstruction, suggesting that the contiguous shell of N beta V may be formed by intersubunit contacts between small domains having that motif. The small (approximately 1800 kDa), ssRNA genome is loosely packed inside the capsid with a low average density.

The structure of a neutralizing immunoglobulin (monoclonal antibody mAb17-IA), bound to human rhinovirus 14 (HRV14), has been determined by cryo-electron microscopy and image reconstruction. The antibody bound bivalently across icosahedral twofold axes of the virus, and there were no detectable conformational changes in the capsid. Thus, bivalently bound IgGs do not appear to cause gross deformations in the capsid. Differences between the electron density of the constant domains of the bound Fab fragment and IgG structures suggested that conformational changes occur about elbow axes upon bivalent attachment as was previously predicted. No significant density was observed for the Fc fragment, which adds further evidence for a high degree of mobility about the hinge region.

Alphaviruses are a group of icosahedral, positive-strand RNA, enveloped viruses. The membrane bilayer, which surrounds the approximately 400 A diameter nucleocapsid, is penetrated by 80 spikes arranged in a T = 4 lattice. Each spike is a trimer of heterodimers consisting of glycoproteins E1 and E2. Cryoelectron microscopy and image reconstruction of Ross River virus showed that the T = 4 quaternary structure of the nucleocapsid consists of pentamer and hexamer clusters of the capsid protein, but not dimers, as have been observed in several crystallographic studies. The E1-E2 heterodimers form one-to-one associations with the nucleocapsid monomers across the lipid bilayer. Knowledge of the atomic structure of the capsid protein and our reconstruction allows us to identify capsid-protein residues that interact with the RNA, the glycoproteins, and adjacent capsid-proteins.

The crystal structure of Fab17-IA, an antigen-binding fragment from a murine immunoglobulin that neutralizes human rhinovirus 14 (HRV14), has been solved to 2.7 A resolution. Fab17-IA crystallized into three different space groups depending upon the method used to purify the intact antibody. The structure was determined by use of molecular and isomorphous replacement methods. The current model has a crystallographic R-factor of approximately 19% for 10,192 independent reflections between 8 and 2.7 A. Correlation coefficient calculations showed that the Fab17-IA structure can be fit into the Fab17-IA/HRV14 image reconstruction density to within 5 A positional accuracy and to within a few degrees of rotation. The resulting interface of the docked antibody was examined and showed extensive charge and shape complementarity with the virus surface that was supported by site-directed mutagenesis experiments. The success of this approach validates the utility of combining X-ray crystallography with cryo-electron microscopy of complex macromolecular assemblies.

The structures of Sindbis virus and Ross River virus complexed with Fab fragments from monoclonal antibodies have been determined from cryoelectron micrographs. Both antibodies chosen for this study bind to regions of the virions that have been implicated in cell-receptor recognition and recognize epitopes on the E2 glycoprotein. The two structures show that the Fab fragments bind to the outermost tip of the trimeric envelope spike protein. Hence, the same region of both the Sindbis virus and Ross River virus envelope spike is composed of E2 and is involved in recognition of the cellular receptor.

Cryoelectron microscopy has been used to determine the first structure of a virus when complexed with its glycoprotein cellular receptor. Human rhinovirus 16 (HRV16) complexed with the two amino-terminal, immunoglobulin-like domains of the intercellular adhesion molecule-1 (ICAM-1) shows that ICAM-1 binds into the 12 A deep "canyon" on the surface of the virus. This is consistent with the prediction that the viral receptor attachment site lies in a cavity inaccessible to the host's antibodies. The atomic structures of HRV14 and CD4, homologous to HRV16 and ICAM-1, showed excellent correspondence with observed density, thus establishing the virus-receptor interactions.

The structures of bovine papillomavirus type 1 (BPV-1) and human papillomavirus type 1 (HPV-1) were determined at 2.5 nm resolution by cryoelectron microscopy and three dimensional image reconstruction techniques. As expected, the reconstructions showed that both viruses consist of a T = 7 icosahedral capsid (approximately 60 nm in diameter) which surrounds a nucleohistone core. The capsid morphologies of the two viruses are nearly indistinguishable. Each capsid consists of a shell layer (approximately 2 nm thick) of nearly continuous density from which capsomers project radially to a maximum height of approximately 5.8 nm. The five-coordinate (pentavalent) and six-coordinate (hexavalent) capsomers both exhibit distinct five-fold axial symmetry as was observed for SV40 and polyoma viruses. Thus, both genera (papilloma and polyoma) of the papovavirus family have now been shown to have the characteristic "all-pentamer" capsid construction. BPV-1 and HPV-1 capsomers consist of a thick (8.6 nm diameter) trunk that broadens distally to form a regular five-pointed, star-shaped head, and proximally to create the shell layer where capsomers associate. A cylindrical channel (approximately 2.8 nm diameter) extends along the axis of each capsomer from the interior of the virus to a point approximately half way to the capsomer surface. Computationally sectioned views of individual capsomers displayed at decreasing radii show that each of the five capsomer subunits (in both pentavalent and hexavalent capsomers) makes a pronounced (30 degrees) left-handed twist just above the outer surface of the capsid shell. Similar views of the reconstructions also clarify the morphology of intercapsomer contacts. For example, they show how hexavalent capsomers coordinate six neighboring capsomers despite the fact that they contain only five subunits. The system of intercapsomer contacts is indistinguishable in BPV-1 and HPV-1, but quite different from that reported for polyoma virus capsids assembled in vitro from the major capsid protein, VP1 (D. M. Salunke, D. L. D. Caspar, and R. L. Garcea. 1989. Biophys. J. 56:887-900). Thus, because both polyoma and papilloma viruses have all-pentamer capsids, it appears that intracapsomer subunit-subunit interactions which stabilize pentameric capsomers are better preserved evolutionarily than those involved in capsomer-capsomer contacts.

Background

Studies of simple RNA animal viruses show that cell attachment, particle destabilization and cell entry are complex processes requiring a level of capsid sophistication that is difficult to achieve with a shell containing only a single gene product. Nodaviruses [such as Flock House virus (FHV)] are an exception. We have previously determined the structure of FHV at 3 A resolution, and now combine this information with data from cryo-electron microscopy in an attempt to clarify the process by which nodaviruses infect animal cells.

Results

A difference map was computed in which electron density at 22 A resolution, derived from the 3.0 A resolution X-ray model of the FHV capsid protein, was subtracted from the electron density derived from the cryo-electron microscopy reconstruction of FHV at 22 A resolution. Comparisons of this density with the X-ray model showed that quasi-equivalent regions of identical polypeptide sequences have markedly different interactions with the bulk RNA density. Previously reported biphasic kinetics of particle maturation and the requirement of subunit cleavage for particle infectivity are consistent with these results.

Conclusions

On the basis of this study we propose a model for nodavirus infection that is conceptually similar to that proposed for poliovirus but differs from it in detail. The constraints of a single protein type in the capsid lead to a noteworthy use of quasi-symmetry not only to control the binding of a 'pocket factor' but also to modulate maturation cleavage and to release a pentameric helical bundle (with genomic RNA attached) that may further interact with the cell membrane.