Virus maturation involving large subunit rotations and local refolding.

Large-scale conformational changes transform viral precursors into infectious virions. The structure of bacteriophage HK97 capsid, Head-II, was recently solved by crystallography, revealing a catenated cross-linked topology. We have visualized its precursor, Prohead-II, by cryoelectron microscopy and modeled the conformational change by appropriately adapting Head-II. Rigid-body rotations ( approximately 40 degrees) cause switching to an entirely different set of interactions; in addition, two motifs undergo refolding. These changes stabilize the capsid by increasing the surface area buried at interfaces and bringing the cross-link-forming residues, initially approximately 40 angstroms apart, close together. The inner surface of Prohead-II is negatively charged, suggesting that the transition is triggered electrostatically by DNA packaging.

Bacteriophage T5 structure reveals similarities with HK97 and T4 suggesting evolutionary relationships.

Evolutionary relationships between viruses may be obscure by protein sequence but unmasked by structure. Analysis of bacteriophage T5 by cryo-electron microscopy and protein sequence analysis reveals analogies with HK97 and T4 that suggest a mosaic of such connections. The T5 capsid is consistent with the HK97 capsid protein fold but has a different geometry, incorporating three additional hexamers on each icosahedral facet. Similarly to HK97, the T5 major capsid protein has an N-terminal extension, or Delta-domain that is missing in the mature capsid, and by analogy with HK97, may function as an assembly or scaffold domain. This Delta-domain is predicted to be largely coiled-coil, as for that of HK97, but is approximately 70% longer correlating with the larger capsid. Thus, capsid architecture appears likely to be specified by the Delta-domain. Unlike HK97, the T5 capsid binds a decoration protein in the center of each hexamer similarly to the "hoc" protein of phage T4, suggesting a common role for these molecules. The tail-tube has unusual trimeric symmetry that may aid in the unique two-stage DNA-ejection process, and joins the tail-tip at a disk where tail fibers attach. This intriguing mix of characteristics embodied by phage T5 offers insights into virus assembly, subunit function, and the evolutionary connections between related viruses.

Human kinetochore-associated kinesin CENP-E visualized at 17 A resolution bound to microtubules.

The highly dynamic process of cell division is effected, in part, by molecular motors that generate the forces necessary for its enactment. Several members of the kinesin superfamily of motor proteins are implicated in mitosis, such as CENP-E, which plays essential roles in cell division, including association with the kinetochore to stabilize attachment of chromosomes to microtubules prior to and during their separation. Neither the functional assembly state of CENP-E nor its direction of motion along the polar microtubule are certain. To determine the mode of interaction between CENP-E and microtubules, we have used cryo-electron microscopy to visualize CENP-E motor domains complexed with microtubules and calculated a density map of the complex to 17 A resolution by combining helical and single-particle reconstruction methods. The interface between the motor domain and microtubules was modeled by docking atomic-resolution models of the subunits into the cryoEM density map. Our results support a plus end motion for CENP-E, consistent with features of the crystallographic structure. Despite considerable functional differences from the monomeric transporter kinesin KIF1A and the oppositely directed ncd kinesin, CENP-E appears to share many features of the intermolecular interactions, suggesting that differences in motor function are governed by small variations in the loops at the microtubule interface.

Structure of the dodecahedral penton particle from human adenovirus type 3.

The sub-viral dodecahedral particle of human adenovirus type 3, composed of the viral penton base and fiber proteins, shares an important characteristic of the entire virus: it can attach to cells and penetrate them. Structure determination of the fiberless dodecahedron by cryo-electron microscopy to 9 Angstroms resolution reveals tightly bound pentamer subunits, with only minimal interfaces between penton bases stabilizing the fragile dodecahedron. The internal cavity of the dodecahedron is approximately 80 Angstroms in diameter, and the interior surface is accessible to solvent through perforations of approximately 20 Angstroms diameter between the pentamer towers. We observe weak density beneath pentamers that we attribute to a penton base peptide including residues 38-48. The intact amino-terminal domain appears to interfere with pentamer-pentamer interactions and its absence by mutation or proteolysis is essential for dodecamer assembly. Differences between the 9 Angstroms dodecahedron structure and the adenovirus serotype 2 (Ad2) crystallographic model correlate closely with differences in sequence. The 3D structure of the dodecahedron including fibers at 16 Angstroms resolution reveals extra density on the top of the penton base that can be attributed to the fiber N terminus. The fiber itself exhibits striations that correlate with features of the atomic structure of the partial Ad2 fiber and that represent a repeat motif present in the amino acid sequence. These new observations offer important insights into particle assembly and stability, as well as the practicality of using the dodecahedron in targeted drug delivery. The structural work provides a sound basis for manipulating the properties of this particle and thereby enhancing its value for such therapeutic use.

Epitope diversity of hepatitis B virus capsids: quasi-equivalent variations in spike epitopes and binding of different antibodies to the same epitope.

To investigate the range of antigenic variation of HBV capsids, we have characterized the epitopes for two anti-capsid antibodies by cryo-electron microscopy and image reconstruction of Fab-labeled capsids to approximately 10A resolution followed by molecular modeling. Both antibodies engage residues on the protruding spikes but their epitopes and binding orientations differ. Steric interference effects limit maximum binding to approximately 50% average occupancy in each case. However, the occupancies of the two copies of a given epitope that are present on a single spike differ, reflecting subtle distinctions in structure and hence, binding affinity, arising from quasi-equivalence. The epitope for mAb88 is conformational but continuous, consisting of a loop-helix motif (residues 77-87) on one of the two polypeptide chains in the spike. In contrast, the epitope for mAb842, like most conformational epitopes, is discontinuous, consisting of a loop on one polypeptide chain (residues 74-78) combined with a loop-helix element (residues 78-83) on the other. The epitope of mAb842 is essentially identical with that previously mapped for mAb F11A4, although the binding orientations of the two monoclonal antibodies (mAbs) differ, as do their affinities measured by surface plasmon resonance. From the number of monoclonals (six) whose binding had to be characterized to give the first duplicate epitope, we estimate the total number of core antigen (cAg) epitopes to be of the order of 20. Given that different antibodies may share the same epitope, the potential number of distinct anti-cAg clones should be considerably higher. The observation that the large majority of cAg epitopes are conformational reflects the relative dimensions of a Fab (large) and the small size and close packing of the motifs that are exposed and accessible on the capsid surface.

Proteolytic and conformational control of virus capsid maturation: the bacteriophage HK97 system.

Bacteriophage capsid assembly pathways provide excellent model systems to study large-scale conformational changes and other mechanisms that regulate the formation of macromolecular complexes. These capsids are formed from proheads: relatively fragile precursor particles which mature by undergoing extensive remodeling. Phage HK97 employs novel features in its strategy for building capsids, including assembly without a scaffolding protein, and the formation of a network of covalent cross-links between neighboring subunits in the mature virion. In addition, proteolytic cleavage of the capsid protein from 42 kDa to 31 kDa is essential for maturation. To investigate the structural bases for proteolysis and cross-linking, we have used cryo-electron micrographs to reconstruct the three-dimensional structures of purified particles from four discrete stages in the assembly pathway: Prohead I, Prohead II, Head I and Head II. Prohead I has icosahedral T = 7 packing of blister-shaped pentamers and hexamers. The pentamers are 5-fold symmetric, but the hexamers exhibit an unusual departure from 6-fold symmetry, as if two trimers had undergone a shear dislocation of about 25 A. Proteolytic conversion to Prohead II leaves the outer surface largely unchanged, but a major loss of density from the inner surface is observed, which we infer to represent the excision of the amino-terminal domains of the capsid protein. Upon expansion to the Head I state, the capsid becomes markedly larger, thinner walled, and more polyhedral: moreover, the capsomer shapes change radically; especially notable is the disappearance of the large hexon dislocation. No differences between Head I and the covalently cross-linked Head II could be observed at the current resolution of about 25 A, from which we infer that it is the conformational rearrangements effected by expansion that create the micro-environments needed for the autocatalytic formation of the isodipeptide bonds found in the mature virions ("pseudo-active sites").

Stoichiometry and domainal organization of the long tail-fiber of bacteriophage T4: a hinged viral adhesin.

The long-tail fibers (LTFs) form part of bacteriophage T4's apparatus for host cell recognition and infection, being responsible for its initial attachment to susceptible bacteria. The LTF has two parts, each approximately 70 to 75 nm long; gp34 (140 kDa) forms the proximal half-fiber, while the distal half-fiber is composed of gp37 (109 kDa), gp36(23 kDa) and gp35 (30 kDa). LTFs have long been thought to be dimers of gp34, gp37 and gp36, with one copy of gp35. We have used mass mapping by scanning transmission electron microscopy (STEM), quantitative SDS-PAGE, and computational sequence analysis to study the structures of purified LTFs and half-fibers of both kinds. These data establish that the LTF is, in fact, trimeric, with a stoichiometry of gp34: gp37: gp36: gp35 = 3:3:3:1. Averaged images of stained and unstained molecules resolve the LTF into a linear stack of 17 domains. At the proximal end is a globular domain of approximately 145 kDa that becomes incorporated into the baseplate. It is followed by a rod-like shaft (33 x 4 mm; 151 kDa) which correlates with a cluster of seven quasi repeats, each 34 to 39 residues long. The proximal half-fiber terminates in three globular domains. The distal half-fiber consists of ten globular domains of variable size and spacing, preceding a needle-like end domain (15 x 2.5 nm; 31 kDa). The LTF is rigid apart from hinges between the two most proximal domains, and between the proximal and distal half-fibers. The latter hinge occurs at a site of local non-equivalence (the "kneecap") at which density, correlated with the presence of gp35, bulges asymmetrically out on one side. Several observations indicate that gp34 participates in the sharing of conserved structural modules among coliphage tail-fiber genes to which gp37 was previously noted to subscribe. Two adjacent globular domains in the proximal half-fiber match a pair of domains in the distal half-fiber, and the rod domain in the proximal half-fiber resembles a similar domain in the T4 short tail-fiber (gp12). Finally, possible structures are considered; combining our data with earlier observations, the most likely conformation for most of the LTF is a three-stranded beta-helix.

Filamentous hemagglutinin of Bordetella pertussis. A bacterial adhesin formed as a 50-nm monomeric rigid rod based on a 19-residue repeat motif rich in beta strands and turns.

The filamentous hemagglutinin (FHA) of Bordetella pertussis is an adhesin that binds the bacteria to cells of the respiratory epithelium in whooping-cough infections. Mature FHA is a 220 kDa secretory protein that is highly immunogenic and has been included in acellular vaccines. We have investigated its structure by combining electron microscopy and circular dichroism spectroscopy (CD) with computational analysis of its amino acid sequence. The FHA molecule is 50 nm in length and has the shape of a horseshoe nail: it has a globular head that appears to consist of two domains; a 35 nm-long shaft that averages 4 nm in width, but tapers slightly from the head end; and a small, flexible, tail. Mass measurements by scanning transmission electron microscopy establish that FHA is a monomer. Its sequence contains two regions of tandem 19-residue pseudo-repeats: the first, of 38 cycles, starts at residue 344; the second, of 13 cycles, starts at residue 1440. The repeat motifs are predicted to consist of short beta-strands separated by beta-turns, and secondary structure measurements by CD support this prediction. We propose a hairpin model for FHA in which the head is composed of the terminal domains; the shaft consists mainly of the repeat regions conformed as amphipathic, hyper-elongated beta-sheets, with their hydrophobic faces apposed; and the tail is composed of the intervening sequence. Further support for the model was obtained by immuno-labeling electron microscopy. The 19-residue repeats of FHA have features in common with the leucine-rich repeats (LRRs) that are present in many eukaryotic proteins, including some adhesion factors. The model is also compared with the two other classes of filamentous proteins that are rich in beta-structure, i.e. viral adhesins and two beta-helical secretory proteins. Our proposed structure implies how the functionally important adhesion sites and epitopes of FHA are distributed: its tripeptide (RGD) integrin-binding site is assigned to the tail; the putative hemagglutination site forms part of the head; and two classes of immunodominant epitopes are assigned to opposite ends of the molecule. Possible mechanisms are discussed for two modes of FHA-mediated adhesion.

Hepatitis B virus capsid: localization of the putative immunodominant loop (residues 78 to 83) on the capsid surface, and implications for the distinction between c and e-antigens.

Hepatitis B virus capsid protein comprises a 149 residue "assembly" domain that polymerizes into icosahedral particles, and a 34 residue RNA-binding "protamine" domain. Recently, the capsid structure has been studied to resolutions below 10 A by cryo-electron microscopy, revealing much of its alpha-helical substructure and that it appears to have a novel fold for a capsid protein; however, the resolution is still too low for chain-tracing by conventional criteria. Aiming to establish a fiducial marker to aid in the process of chain-tracing, we have used cryo-microscopy to pinpoint the binding site of a monoclonal antibody that recognizes the peptide from residues 78 to 83. This epitope resides on the outer rim of the 30 A long spikes that protrude from the capsid shell. These spikes are four-helix bundles formed by the pairing of helix-turn-helix motifs from two subunits; by means of a tilting experiment, we have determined that this bundle is right-handed. Variants of the same protein present two clinically important and non-crossreactive antigens: core antigen (HBcAg), which appears early in infection as assembled capsids; and the sentinel e-antigen (HBeAg), a non-particulate form. Knowledge of the binding site of our anti-HBcAg antibody bears on the molecular basis of the distinction between the two antigens, which appears to reflect conformational differences between the assembled and unassembled states of the capsid protein dimer, in addition to epitope masking in capsids.

Shared motifs of the capsid proteins of hepadnaviruses and retroviruses suggest a common evolutionary origin.

The structure of the dimeric C-terminal domain of the HIV-1 capsid protein (CA), recently determined by X-ray crystallography (Gamble et al. (1997)), has a notable resemblance to the structure of the hepatitis B virus (HBV) capsid protein (Cp) dimer, previously determined by cryo-electron microscopy (Conway et al. (1997), Böttcher et al. (1997)). In both proteins, dimerization is effected by formation of a four-helix bundle, whereby each subunit contributes a helix-loop-helix and most of the interaction between subunits is mediated by one pair of helices. These are the first two observations of a motif that is common to the capsid proteins of two enveloped viruses and quite distinct from the eight-stranded anti-parallel beta-barrel found in most other virus capsid proteins solved to date (Harrison et al. (1996)). Motivated by the structural resemblance, we have examined retroviral and HBV capsid protein sequences and found weak but significant similarities between them. These similarities further support an evolutionary relationship between these two virus families of great medical importance -- the hepadnaviruses (e.g. HBV) and retroviruses (e.g. HIV).

Structural features in the heptad substructure and longer range repeats of two-stranded alpha-fibrous proteins.

Considerable sequence data have been collected from the intermediate filament proteins and other alpha-fibrous proteins including myosin, tropomyosin, paramyosin, desmoplakin and M-protein. The data show that there is a clear preference for some amino acids to occur in specific positions within the heptad substructure that characterizes the sequences which form the coiled-coil rod domain in this class of proteins. The results also indicate that although there are major similarities between the various proteins there are also key differences. In all cases, however, significant regularities in the linear disposition of the acidic and the basic residues in the coiled-coil segments can be related to modes of chain and molecular aggregation. In particular a clear trend has been observed which relates the mode of molecular aggregation to the number of interchain ionic interactions per heptad pair.

Three-stranded alpha-fibrous proteins: the heptad repeat and its implications for structure.

Amino acid sequence data have been collected for the coiled-coil rod domains of three-stranded alpha-fibrous proteins--fibrinogen, laminin, tenascin, macrophage scavenger receptor protein and the leg fibre protein from bacteriophage. Such domains are characterized by a heptad substructure in which apolar residues occur alternately three and four residues apart. The distribution of residues in each position of the heptad has been analysed, and the results compared with those obtained for the two-stranded alpha-fibrous proteins, which include the intermediate filament and myosin families. Distinctions can be drawn between the sequences in two- and three-stranded coiled-coil structures and these provide criteria that will prove useful in predicting secondary and tertiary structure purely from sequence data.

Structure of the pseudorabies virus capsid: comparison with herpes simplex virus type 1 and differential binding of essential minor proteins.

The structure of pseudorabies virus (PRV) capsids isolated from the nucleus of infected cells and from PRV virions was determined by cryo-electron microscopy (cryo-EM) and compared to herpes simplex virus type 1 (HSV-1) capsids. PRV capsid structures closely resemble those of HSV-1, including distribution of the capsid vertex specific component (CVSC) of HSV-1, which is a heterodimer of the pUL17 and pUL25 proteins. Occupancy of CVSC on all PRV capsids is near 100%, compared to ~50% reported for HSV-1 C-capsids and 25% or less that we measure for HSV-1 A- and B-capsids. A PRV mutant lacking pUL25 does not produce C-capsids and lacks visible CVSC density in the cryo-EM-based reconstruction. A reconstruction of PRV capsids in which green fluorescent protein was fused within the N-terminus of pUL25 confirmed previous studies with a similar HSV-1 capsid mutant localizing pUL25 to the CVSC density region that is distal to the penton. However, comparison of the CVSC density in a 9-Å-resolution PRV C-capsid map with the available crystal structure of HSV-1 pUL25 failed to find a satisfactory fit, suggesting either a different fold for PRV pUL25 or a capsid-bound conformation for pUL25 that does not match the X-ray model determined from protein crystallized in solution. The PRV capsid imaged within virions closely resembles C-capsids with the addition of weak but significant density shrouding the pentons that we attribute to tegument proteins. Our results demonstrate significant structure conservation between the PRV and HSV capsids.

Development of the dodecahedral penton particle from adenovirus 3 for therapeutic application.

The subviral dodecahedral particle of adenovirus 3, which assembles spontaneously in insect cells expressing the viral penton base protein, shows promise as a vector for drug delivery. Its ability to gain cell entry has been demonstrated and recent structural analysis has outlined details of the interfaces between penton bases and the importance of proteolysis of the penton base N terminus for assembly, providing a basis for understanding particle assembly and stability. Here, work in manipulating the assembly status of the dodecahedron by changing buffer conditions and subsequent success in passively encapsidating a marker molecule is described. This represents an important stage towards development of the dodecahedral particle for use as a delivery vehicle capable of targeting therapeutic molecules to specific cell types.

Spectral signal-to-noise ratio and resolution assessment of 3D reconstructions.

Measuring the quality of three-dimensional (3D) reconstructed biological macromolecules by transmission electron microscopy is still an open problem. In this article, we extend the applicability of the spectral signal-to-noise ratio (SSNR) to the evaluation of 3D volumes reconstructed with any reconstruction algorithm. The basis of the method is to measure the consistency between the data and a corresponding set of reprojections computed for the reconstructed 3D map. The idiosyncrasies of the reconstruction algorithm are taken explicitly into account by performing a noise-only reconstruction. This results in the definition of a 3D SSNR which provides an objective indicator of the quality of the 3D reconstruction. Furthermore, the information to build the SSNR can be used to produce a volumetric SSNR (VSSNR). Our method overcomes the need to divide the data set in two. It also provides a direct measure of the performance of the reconstruction algorithm itself; this latter information is typically not available with the standard resolution methods which are primarily focused on reproducibility alone.