Structure and assembly of turnip crinkle virus. II. Mechanism of reassembly in vitro

Inhibiting virus-capsid assembly by altering the polymerisation pathway

Virus capsids assemble through the repeated interaction of well-defined protein subunits in a highly specific process. Basic research into the mechanism of protein polymerisation and virus assembly suggest that inhibition of the protein-protein interactions necessary for assembly is a valid therapeutic strategy. Computer models of virus-capsid assembly have located vulnerable stages in assembly, and small-molecule inhibitors of virus assembly have been identified. The challenge will be identifying agents that block assembly with the required specificity.

Assembly of physalis mottle virus capsid protein in Escherichia coli and the role of amino and carboxy termini in the formation of the icosahedral particles

The coat protein gene of physalis mottle tymovirus (PhMV) was over expressed in Escherichia coli using pET-3d vector. The recombinant protein was found to self assemble into capsids in vivo. The purified recombinant capsids had an apparent s value of 56.5 S and a diameter of 29(+/-2) nm. In order to establish the role of amino and carboxy-terminal regions in capsid assembly, two amino-terminal deletions clones lacking the first 11 and 26 amino acid residues and two carboxy-terminal deletions lacking the last five and ten amino acid residues were constructed and overexpressed. The proteins lacking N-terminal 11 (PhCPN1) and 26 (PhCPN2) amino acid residues self assembled into T=3 capsids in vivo, as evident from electron microscopy, ultracentrifugation and agarose gel electrophoresis. The recombinant, PhCPN1 and PhCPN2 capsids were as stable as the empty capsids formed in vivo and encapsidated a small amount of mRNA. The monoclonal antibody PA3B2, which recognizes the epitope within region 22 to 36, failed to react with PhCPN2 capsids while it recognized the recombinant and PhCPN1 capsids. Disassembly of the capsids upon treatment with urea showed that PhCPN2 capsids were most stable. These results demonstrate that the N-terminal 26 amino acid residues are not essential for T=3 capsid assembly in PhMV. In contrast, both the proteins lacking the C-terminal five and ten amino acid residues were present only in the insoluble fraction and could not assemble into capsids, suggesting that these residues are crucial for folding and assembly of the particles.

Biochemical characterization of a smaller form of recombinant Norwalk virus capsids assembled in insect cells

The expression of the single capsid protein of Norwalk virus (NV) in Spodoptera frugiperda (Sf9) insect cells infected with recombinant baculovirus results in the assembly of virus-like particles (VLPs) of two sizes, the predominant 38-nm, or virion-size VLPs, and smaller, 23-nm VLPs. Here we describe the purification and biochemical characterization of the 23-nm VLPs. The 23-nm VLPs were purified to 95% homogeneity from the medium of Sf9 cultures by isopycnic CsCl gradient centrifugation followed by rate-zonal centrifugation in sucrose gradients. The compositions of the purified 23- and 38-nm VLPs were compared by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and protein immunoblots. VLPs of both sizes showed a doublet at 58 kDa, the size of the full-length capsid protein. Upon alkaline treatment, the 23-nm VLPs underwent dissociation into soluble intermediates that were able to reassemble into 23- and 38-nm VLPs upon dialysis, suggesting that the assembly of both types of structures has a common pathway. Antigenic and biochemical properties of the 38- and 23-nm VLPs were examined and found to be conserved. Immunoprecipitation assays using polyclonal and monoclonal antibodies indicated that immunodominant epitopes on the capsid protein as well as conformational epitopes are conserved in the two types of particles. The trypsin cleavage site at residue 227 was protected in the assembled particles of both sizes but exposed after alkaline dissociation. These results, and the conservation of the binding activity of both forms of recombinant NV VLPs to cultured cells (L. J. White, J. M. Ball, M. E. Hardy, T. N. Tanaka, N. Kitamoto, and M. K. Estes, J. Virol. 70:6589-6597, 1996), suggest that the tertiary folding of the capsid protein responsible for these properties is conserved in the two structures. We hypothesize that the 23-nm VLPs are formed when 60 units of the NV capsid protein assembles into a structure with T=1 symmetry.

Sindbis virus replicons and Sindbis virus: assembly of chimeras and of particles deficient in virus RNA

Alphaviruses are a well-characterized group of positive-strand RNA viruses. The identification of cis-acting elements in their genomes and their replication strategy have made them useful as vectors for the expression of heterologous genes. In infected cells, the nonstructural proteins, required for replication and transcription of the viral genes, are translated from the genomic RNA; the structural proteins, the capsid protein that interacts with the RNA to form the nucleocapsid and the proteins embedded in the lipid envelope, are translated from a subgenomic mRNA and can be replaced by heterologous genes. Such modified genomes are self-replicating (replicons); they can be introduced into the cells by transfection and can also be packaged into extracellular particles with defective helper (DH) RNAs. The particular DH RNA determines how well it is replicated and to what extent it is packaged. One potential complication of this system has been that recombination between the replicon genome and the DH RNA may occur. The studies described here were designed to prevent recombination by expressing the capsid protein from one DH RNA and the virus membrane proteins from a second helper RNA. Recombination to yield a nonsegmented infectious virus genome would then require several independent crossover events. There is a translational enhancer located downstream of the initiating AUG in the RNA of the capsid gene that had to be conserved in the second helper to achieve high-level expression of the viral glycoproteins. For this reason, we modified the capsid protein gene in two ways: the first was to use the capsid protein gene from a different alphavirus, Ross River virus, and the second was to make deletions in that gene to maintain the translational enhancer in the RNA but to eliminate the positively charged region in the protein that should be essential for the specific and nonspecific interactions with RNA. Transfections with replicon RNA and the deleted chimeric DH RNA as the only helper resulted in the high-level production of particles that were almost completely devoid of RNA. The inclusion of a helper expressing an intact Sindbis virus capsid protein gene led to the production of high levels of packaged replicons. Recombinants were not detected even after several undiluted passages.

Proline-dependent oligomerization with arm exchange

BACKGROUND

Oligomerization is often necessary for protein activity or regulation and its efficiency is fundamental for the cell. The quaternary structure of a large number of oligomers consists of protomers tightly anchored to each other by exchanged arms or swapped domains. However, nothing is known about how the arms can be kept in a favourable conformation before such an oligomerization.

RESULTS

Upon examination of such quaternary structures, we observe an extremely frequent occurrence of proline residues at the point where the arm leaves the protomer. Sequence alignment and site-directed mutagenesis confirm the importance of these prolines. The conservation of these residues at the hinge regions can be explained by the constraints that they impose on polypeptide conformation and dynamics: by rigidifying the mainchain, prolines favour extended conformations of arms thus favouring oligomerization, and may prevent interaction of the arms with the core of the protomer.

CONCLUSIONS

Hinge prolines can be considered as 'quaternary structure helpers'. The presence of a proline should be considered when searching for a determinant of oligomerization with arm exchange and could be used to engineer synthetic oligomers or to displace a monomers to oligomers equilibrium by mutation of this proline residue.

The structure of simian virus 40 refined at 3.1 A resolution

BACKGROUND

The structure of simian virus 40 (SV40), previously determined at 3.8 degree resolution, shows how its pentameric VP1 assembly units are tied together by extended C-terminal arms. In order to define more precisely the possible assembly mechanisms, we have refined the structure at 3.1 degree resolution.

RESULTS

New data from a high-intensity synchrotron source have been used for phase extension by electron-density averaging and refinement, exploiting only the strict 5-fold non-crystallographic symmetry for the real-space averaging steps. The accurate model enables us to study important structural features of the virus particle in detail. The remarkably invariant core of the VP1 pentamer bears the docking sites for the C-terminal arms from other pentamers. These contacts are the principal way in which pentameric assembly units are linked together in the capsid. Only at the interface between five-coordinated and six-coordinated pentamers do the pentamer cores appear to interact strongly. There are two cation-binding sites per VP1 monomer, seen in a soaking experiment with gadolinium nitrate. These sites are quite close to each other at the interfaces between pentamers.

CONCLUSION

We propose that the contact between five-coordinated and six-coordinated pentamers may help to generate a six-pentamer nucleus, with which further pentamers can assemble to generate the complete particle. Calcium ions probably stabilize the structure of the assembled particle, rather than direct its assembly.

Structure of sesbania mosaic virus at 3 A resolution

BACKGROUND

Sobemoviruses are a group of RNA plant viruses that have a narrow host range. They are characterized in vitro by their stability, high thermal inactivation point and longevity. The three-dimensional structure of only one virus belonging to this group, southern bean mosaic virus (SBMV), is known. Structural studies on sesbania mosaic virus (SMV), which is closely related to SBMV, will provide details of the molecular interactions that are likely to be important in the stability and assembly of sobemoviruses.

RESULTS

We have determined the three-dimensional structure of SMV at 3 A resolution. The polypeptide fold and quaternary organization are very similar to those of SBMV. The capsid consists of sixty icosahedral asymmetric units, each comprising three copies of a chemically identical coat protein subunit, which are designated as A, B and C and are in structurally different environments. Four cation-binding sites have been located in the icosahedral asymmetric unit. Of these, the site at the quasi-threefold axis is not found in SBMV. Structural differences are observed in loops and regions close to this cation-binding site. Preliminary studies on ethylene diamine tetra acetic acid (EDTA) treated crystals suggest asymmetry in removal of the quasi-equivalent cations at the AB, BC, and AC subunit interfaces.

CONCLUSIONS

Despite the overall similarity between SMV and SBMV in the nature of the polypeptide fold, these viruses show a number of differences in intermolecular interactions. The polar interactions at the quasi-threefold axis are substantially less in SMV and positively charged residues on the RNA-facing side of the protein and in the N-terminal arm are not particularly well conserved. This suggests that protein-RNA interactions are likely to be different between the two viruses.

Trinucleotide models for DNA bending propensity: comparison of models based on DNaseI digestion and nucleosome packaging data

DNaseI digestion studies (Brukner et al, EMBO J 14, 1812-1818 1995) and nucleosomebinding data (Satchwell et al, J. Mol. Biol. 191, 639-659 1986, Goodsell and Dickerson, Nucleic trinucleotides. A detailed comparison of the two models suggests that while both of them represent improvements with respect to dinucleotide based descriptions, the individual trinucleotide parameters are not highly correlated (linear correlation coefficient is 0.53), and a number of motifs such as TA-elements and CCA/TGG motifs are more realistically described in the DNaseI-based model. This may be due to the fact that the DNaseI-based model does not rely on a static geometry but rather captures a dynamic ability of ds DNA to bend towards the major grove. Future refinement of both models of both models on larger experimental data sets is expected to further improve the prediction of macroscopic DNA-curvature.

Stability of wild-type and temperature-sensitive protein subunits of the phage P22 capsid

Temperature-sensitive folding (tsf) mutants of the phage P22 coat protein prevent newly synthesized polypeptide chains from reaching the conformation competent for capsid assembly in cells, and can be rescued by the GroEL chaperone (Gordon, C., Sather, S., Casjens, S., and King, J. (1994) J. Biol. Chem. 269, 27941-27951). Here we investigate the stabilities of wild-type and four tsf mutant unpolymerized subunits. Wild-type coat protein subunits denatured at 40 degrees C, with a calorimetric enthalpy of approximately 600 kJ/mol. Comparison with coat protein denaturation within the shell lattice (Tm = 87 degrees C, delta H approximately 1700 kJ/mol) (Galisteo, M.L., and King, J. (1993) Biophys. J. 65, 227-235) indicates that protein-protein interactions within the capsid provide enormous stabilization. The melting temperatures of the subunits carrying tsf substitutions were similar to wild-type. At low temperatures, the tsf mutants, but not the wild-type, formed non-covalent dimers, which were dissociated at temperatures above 30 degrees C. Spectroscopic and calorimetric studies indicated that the mutant proteins have reduced amounts of ordered structure at low temperature, as compared to the wild-type protein. Although complex, the in vitro phenotypes are consistent with the in vivo finding that the mutants are defective in folding, rather than subunit stability. These results suggest a role for incompletely folded subunits as precursors in viral capsid assembly, providing a mechanism of reaching multiple conformations in the polymerized form.

Probing sequence-specific RNA recognition by the bacteriophage MS2 coat protein

We present the results of in vitro binding studies aimed at defining the key recognition elements on the MS2 RNA translational operator (TR) essential for complex formation with coat protein. We have used chemically synthesized operators carrying modified functional groups at defined nucleotide positions, which are essential for recognition by the phage coat protein. These experiments have been complemented with modification-binding interference assays. The results confirm that the complexes which form between TR and RNA-free phage capsids, the X-ray structure of which has recently been reported at 3.0 A, are identical to those which form in solution between TR and a single coat protein dimer. There are also effects on operator affinity which cannot be explained simply by the alteration of direct RNA-protein contacts and may reflect changes in the conformational equilibrium of the unliganded operator. The results also provide support for the approach of using modified oligoribonucleotides to investigate the details of RNA-ligand interactions.

Virus structures and conformational rearrangements

Structural virology is a burgeoning subspecialty. Our understanding of the molecular organization of viruses has begun to contribute directly to the analysis of viral attachment and entry, assembly, antigenicity, and even viral pathogenesis, but there are still more puzzles than answers. Recent crystallographic results have helped us to understand the structural changes in viruses that affect their assembly and infectivity.

Structure-function relation of the NH2-terminal domain of the Semliki Forest virus capsid protein

The capsid (C) protein of alphaviruses consists of two protein domains: a serine protease at the COOH terminus and an NH2-terminal domain which is thought to interact with RNA in the virus nucleocapsid (NC). The latter domain is very rich in positively charged amino acid residues. In this work, we have introduced large deletions into the corresponding region of a full-length cDNA clone of Semliki Forest virus, expressed the transcribed RNA in BHK-21 cells, and monitored the autoprotease activity of C, the formation of intracellular NCs, and the release of infectious virus. Our results show that if the gene region encoding the whole NH2-terminal domain is removed, the expressed C protein fragment cannot assemble into NCs and virus particles but it is still able to function as an autoprotease. Thus, these results underline the general importance of the NH2-terminal domain in the virus assembly process and furthermore show that the serine protease domain can function independently of the NH2 terminus. Surprisingly, analysis of additional C protein deletion variants showed that not all of the NH2-terminal domain is required for virus assembly, but large deletions involving up to one-third of its positively charged residues are still compatible with NC and virus formation. The fact that so much flexibility is allowed in the structure of the NH2-terminal domain of C suggests that most of this region is involved in nonspecific interactions with the encapsidated RNA, probably through its positively charged amino acid residues.

Aura alphavirus subgenomic RNA is packaged into virions of two sizes

The alphavirus genome is 11.8 kb in size. During infection, a 4.2-kb subgenomic RNA is also produced. Most alphaviruses package only the genomic RNA into virions, which are enveloped particles with icosahedral symmetry, having a triangulation number (T) = 4. Aura virus, however, packages both the genomic RNA and the subgenomic RNA into virions. The genomic RNA is primarily packaged into a virion that has a diameter of 72 nm and which appears to be identical to the virions produced by other alphaviruses. The subgenomic RNA is packaged into two major, regular particles with diameters of 72 and 62 nm. The 72-nm-diameter particle appears to be identical in construction to virions containing genomic RNA. The 62-nm-diameter particle probably has T = 3. The large and small Aura virions can be partially separated in sucrose gradients. In addition to these two major classes of particles, there are other particles produced that appear to arise from abortive assembly. From these results and from previous studies of alphavirus assembly, we suggest that during assembly of alphavirus nucleocapsids in the infected cell there is a specific initiation event followed by recruitment of additional capsid subunits into the complex, that the triangulation number of the complex is not predetermined but depends upon the size of the RNA and interactions that occur during assembly, and that budding of assembled nucleocapsids results in the acquisition of an envelope containing glycoproteins arranged in a manner determined by the nucleocapsid.

Single-stranded DNA-protein interactions in canine parvovirus

BACKGROUND

Parvoviruses are small icosahedral single-stranded (ss) DNA viruses which replicate in rapidly proliferating cells, causing a variety of serious and often lethal diseases in mammals, including humans. The structure of canine parvovirus (CPV) showed an 11-nucleotide oligomeric fragment of its genome bound to 60 equivalent binding sites on the inside surface of the capsid. This provides an opportunity to study the conformation of ssDNA, its interactions with protein, and its role in viral assembly.

RESULTS

The icosahedrally ordered part of CPV ssDNA has an unusual loop conformation with the bases pointing outwards and the phosphates surrounding metal ions on the inside. The protein interacts with the bases, making 15 putative hydrogen bonds. The DNA electron density indicates preferences for particular base types in parts of the binding site. Statistical analysis of the genome yields approximately 30 regions with sequences similar to that observed in the structure, demonstrating a low level of sequence specificity for binding to capsid protein.

CONCLUSIONS

ssDNA can adopt unusual conformations upon association with protein by using phosphoribose backbone rotamers that are found in tRNA, but not in DNA duplexes. The CPV DNA-protein interactions differ from the non-specific backbone interactions seen in some plant and insect viruses. The sequence specificity, albeit low level, of the protein for CPV DNA may contribute both to distinguishing the viral DNA from other nucleic acids and to the DNA packaging process during viral assembly.

Efficient in vivo and in vitro assembly of retroviral capsids from Gag precursor proteins expressed in bacteria

The capsid precursor protein (Gag) of Mason-Pfizer monkey virus, the prototype type D retrovirus, has been expressed to high levels in bacteria under the control of the phage T7 promoter. Electron microscopic studies of induced cells revealed the assembly of capsid-like structures within inclusion bodies that formed at the poles of the cells 6 h after induction with isopropyl-beta-D-thiogalactopyranoside (IPTG). The inclusion bodies and enclosed capsid-like structures were solubilized completely in 8 M urea, but following renaturation, we observed assembly in vitro of capsid-like structures that demonstrated apparent icosahedral symmetry. These results demonstrate for the first time that retroviral capsid precursors have the propensity to self-assemble in vitro and point to new approaches for the analysis of retroviral assembly and structure.

Structures of the native and swollen forms of cowpea chlorotic mottle virus determined by X-ray crystallography and cryo-electron microscopy

BACKGROUND

RNA-protein interactions stabilize many viruses and also the nucleoprotein cores of enveloped animal viruses (e.g. retroviruses). The nucleoprotein particles are frequently pleomorphic and generally unstable due to the lack of strong protein-protein interactions in their capsids. Principles governing their structures are unknown because crystals of such nucleoprotein particles that diffract to high resolution have not previously been produced. Cowpea chlorotic mottle virions (CCMV) are typical of particles stabilized by RNA-protein interactions and it has been found that crystals that diffract beyond 4.5 A resolution are difficult to grow. However, we report here the purification of CCMV with an exceptionally mild procedure and the growth of crystals that diffract X-rays to 3.2 A resolution.

RESULTS

The 3.2 A X-ray structure of native CCMV, an icosahedral (T = 3) RNA plant virus, shows novel quaternary structure interactions based on interwoven carboxyterminal polypeptides that extend from canonical capsid beta-barrel subunits. Additional particle stability is provided by intercapsomere contacts between metal ion mediated carboxyl cages and by protein interactions with regions of ordered RNA. The structure of a metal-free, swollen form of the virus was determined by cryo-electron microscopy and image reconstruction. Modeling of this structure with the X-ray coordinates of the native subunits shows that the 29 A radial expansion is due to electrostatic repulsion at the carboxyl cages and is stopped short of complete disassembly by preservation of interwoven carboxyl termini and protein-RNA contacts.

CONCLUSIONS

The CCMV capsid displays quaternary structural interactions that are unique compared with previously determined RNA virus structures. The loosely coupled hexamer and pentamer morphological units readily explain their versatile reassembly properties and the pH and metal ion dependent polymorphism observed in the virions. Association of capsomeres through inter-penetrating carboxy-terminal portions of the subunit polypeptides has been previously described only for the DNA tumor viruses, SV40 and polyoma.

The three-dimensional distribution of RNA and protein in the interior of tomato bushy stunt virus: a neutron low-resolution single-crystal diffraction study

BACKGROUND

The published high-resolution model of the isometric T = 3 plant virus tomato bushy stunt virus (TBSV) shows the packing in three different environments (A, B, C) of the 180 coat protein subunits of the capsid. It does not, however, account for the localization of either the viral RNA or approximately 25% of the amino acids of the protein subunits, although at least the RNA is rigidly linked to the viral capsid. Solution studies have shown that most of the missing protein is located in an inner shell, and that most of the RNA is sandwiched between the two protein shells.

RESULTS

We have determined the organization of TBSV at 16 A resolution, using neutron single-crystal diffraction. Connections between the two protein shells are confined to the 20 three-fold axes of the virion, where three C-type subunits meet. Much more RNA density is located under the 30 C-C dimers than under the 60 A-B dimers, where we could even identify lagoons of solvent.

CONCLUSIONS

Our results emphasize the importance of the amino termini of the 60 C-type protein subunits not only in the RNA-protein interactions but also in the organization of the coat protein, and, probably, in the assembly of the virion. The lack of equivalence between subunits of classes A or B and subunits of class C is even more pronounced in the interior of the virion than in the outer shell, which possesses icosahedral symmetry.

Crystal structure of an RNA bacteriophage coat protein-operator complex

The RNA bacteriophage MS2 is a convenient model system for the study of protein-RNA interactions. The MS2 coat protein achieves control of two distinct processes--sequence-specific RNA encapsidation and repression of replicase translation--by binding to an RNA stem-loop structure of 19 nucleotides containing the initiation codon of the replicase gene. The binding of a coat protein dimer to this hairpin shuts off synthesis of the viral replicase, switching the viral replication cycle to virion assembly rather than continued replication. The operator fragment alone can trigger self-assembly of the phage capsid at low protein concentrations and a complex of about 90 RNA operator fragments per protein capsid has been described. We report here the crystal structure at 3.0 A resolution of a complex between recombinant MS2 capsids and the 19-nucleotide RNA fragment. It is the first example of a structure at this resolution for a sequence-specific protein-RNA complex apart from the transfer RNA synthetase complexes. The structure shows sequence-specific interactions between conserved residues on the protein and RNA bases essential for binding.

Molecular mechanism of RNA phage morphogenesis

Recent progress on the molecular mechanism of RNA phage morphogenesis is described. Functional studies, both in vivo and in vitro, are correlated with the latest structural studies on phages, their capsids and the assembly initiation RNA stem-loop.

Local rule-based theory of virus shell assembly

A local rule-based theory is developed which shows that the self-assembly of icosahedral virus shells may depend on only the lower-level interactions of a protein subunit with its neighbors--i.e., on local rules rather than on larger structural building blocks. The local rule theory provides a framework for understanding the assembly of icosahedral viruses. These include both viruses that fall in the quasiequivalence theory of Caspar and Klug and the polyoma virus structure, which violates quasi-equivalence and has puzzled researchers since it was first observed. Local rules are essentially templates for energetically favorable arrangements. The tolerance margins for these rules are investigated through computer simulations. When these tolerance margins are exceeded in a particular way, the result is a "spiraling" malformation that has been observed in nature.

Three-dimensional structure of baculovirus-expressed Norwalk virus capsids

The three-dimensional structure of the baculovirus-expressed Norwalk virus capsid has been determined to a resolution of 2.2 nm using electron cryomicroscopy and computer image processing techniques. The empty capsid, 38.0 nm in diameter, exhibits T = 3 icosahedral symmetry and is composed of 90 dimers of the capsid protein. The striking features of the capsid structure are arch-like capsomeres, at the local and strict 2-fold axes, formed by dimers of the capsid protein and large hollows at the icosahedral 5- and 3-fold axes. Despite its distinctive architecture, the Norwalk virus capsid has several similarities with the structures of T = 3 single-stranded RNA (ssRNA) viruses. The structure of the protein subunit appears to be modular with three distinct domains: the distal globular domain (P2) that appears bilobed, a central stem domain (P1), and a lower shell domain (S). The distal domains of the 2-fold related subunits interact with each other to form the top of the arch. The lower domains of the adjacent subunits associate tightly to form a continuous shell between the radii of 11.0 and 15.0 nm. No significant mass density is observed below the radius of 11.0 mm. It is suspected that the hinge peptide in the adjoining region between the central domain and the shell domain may facilitate the subunits adapting to various quasi-equivalent environments. Architectural similarities between the Norwalk virus capsid and the other ssRNA viruses have suggested a possible domain organization along the primary sequence of the Norwalk virus capsid protein. It is suggested that the N-terminal 250 residues constitute the lower shell domain (S) with an eight-strand beta-barrel structure and that the C-terminal residues beyond 250 constitute the protruding (P1+P2) domains. A lack of an N-terminal basic region and the ability of the Norwalk virus capsid protein to form empty T = 3 shells suggest that the assembly pathway and the RNA packing mechanisms may be different from those proposed for tomato bushy stunt virus and southern bean mosaic virus but similar to that in tymoviruses and comoviruses.

Reconstitution of Flock House provirions: a model system for studying structure and assembly

Assembly of Flock House virus in infected Drosophila cells proceeds through an intermediate, the provirion, which lacks infectivity until the coat precursor protein, alpha, undergoes a spontaneous "maturation" cleavage (A. Schneemann, W. Zhong, T. M. Gallagher, and R. R. Rueckert, J. Virol 6:6728, 1992). We describe here methods for purifying provirions in a state which permitted dissociation and reassembly. Dissociation, to monomeric alpha protein and free RNA, was accomplished by freezing at pH 9.0 in the presence of 0.5 M salt and 0.1 M urea. When dialyzed at low ionic strength and pH 6.5, the dissociation products reassembled spontaneously to form homogeneous provirions with a normal complement of RNA as judged by cosedimentation with authentic virions and by ability to undergo maturation cleavage with acquisition of substantial, though subnormal, infectivity. Reconstitution experiments, i.e., remixing components after separating RNA from capsid protein, generated abnormal particles, suggesting the presence in the unfractionated dissociation products of an unidentified "nucleating" component.

The atomic structure of Carnation Mottle Virus capsid protein

The structure of the Carnation Mottle Virus (CMtV) capsid protein has been determined at 3.2 A resolution by the method of molecular replacement. Three-dimensional data were collected from a small number of crystals (sp.g. I23, a = 382.6 A) using the synchrotron radiation with an image plate as detector. The coordinates of Tomato Bushy Stunt Virus (TBSV) were used as a searching model. Refinement of the coordinates of 7,479 non-hydrogen atoms performed by the program XPLOR, has led to an R-factor of 18.3%. It was found that the amino acid chain fold of capsid protein is very similar to that in other icosahedral viruses. However, there are some differences in the contact regions between protein subunits and also the lack of the beta-annulus around the 3-fold icosahedral axes. The structural and biochemical results lead us to consider an alternative assembly pathway.

Interaction of human papillomavirus (HPV) type 16 capsid proteins with HPV DNA requires an intact L2 N-terminal sequence

Encapsidation of papillomavirus DNA involves DNA-protein and protein-protein interactions. We sought to define the role of each human papillomavirus (HPV) capsid protein in HPV DNA encapsidation. HPV16 major (L1) and minor (L2) capsid proteins purified from recombinant vaccinia virus-infected cells were compared for their ability to bind nucleic acids. L2 protein, but not L1 protein, could bind HPV DNA. To map the DNA-binding region of L2, a series of truncated or point-mutated L2 protein open reading frames were used to show that only the N terminal of L2 was required for L2-DNA binding. This interaction depends critically on charged amino acids (Lys or Arg) in the first 12 amino acids of the N terminal of the protein. Several techniques were used to show that L2 interaction with DNA did not require specific DNA sequences. We propose that HPV L2 protein may play a major role in papillomavirus capsid assembly by introducing HPV DNA to the virus particles formed by the self assembly of the L1 major structural protein.

Identification of domains in brome mosaic virus RNA-1 and coat protein necessary for specific interaction and encapsidation

Even though many single-stranded RNAs are present in the cytoplasm of infected cells, encapsidation by brome mosaic virus (BMV) coat protein is specific for BMV RNA. Although the highly conserved 3' region of each of the three BMV genomic RNAs is an attractive candidate for the site of recognition by the coat protein, band shift and UV cross-linking assays in the presence of specific and nonspecific competitors revealed only nonspecific interactions. However, BMV RNA-1 formed a retarded complex (complex I) with the coat protein in the absence of competitors, and two domains of RNA-1 that specifically bound coat protein in a small complex (complex II), presumably early in the encapsidation process, were identified. Strong nonspecific, cooperative binding was observed in the presence of high concentrations of coat protein, suggesting that this provides the mechanism leading to rapid encapsidation seen in vivo. In contrast, no binding to a coat protein mutant lacking the N-terminal 25 amino acids that has been shown to be incapable of encapsidation in vivo (R. Sacher and P. Ahlquist, J. Virol. 63:4545-4552, 1989) was detected in vitro. The use of deletion mutants of RNA-1 revealed the presence of domains within the coding region of protein 1a that formed complexes with purified coat protein. One deletion mutant (B1SX) lacking these domains was only slightly more effective in dissociating RNA-1-coat protein complexes than were nonspecific competitors, further suggesting that regions other than the 3' end can participate in the selective encapsidation of BMV RNAs.

Nucleation and growth phases in the polymerization of coat and scaffolding subunits into icosahedral procapsid shells

The polymerization of protein subunits into precursor shells empty of DNA is a critical process in the assembly of double-stranded DNA viruses. For the well-characterized icosahedral procapsid of phage P22, coat and scaffolding protein subunits do not assemble separately but, upon mixing, copolymerize into double-shelled procapsids in vitro. The polymerization reaction displays the characteristics of a nucleation limited reaction: a paucity of intermediate assembly states, a critical concentration, and kinetics displaying a lag phase. Partially formed shell intermediates were directly visualized during the growth phase by electron microscopy of the reaction mixture. The morphology of these intermediates suggests that assembly is a highly directed process. The initial rate of this reaction depends on the fifth power of the coat subunit concentration and the second or third power of the scaffolding concentration, suggesting that pentamer of coat protein and dimers or trimers of scaffolding protein, respectively, participate in the rate-limiting step.

Deletion analysis of the capsid protein of Sindbis virus: identification of the RNA binding region

The capsid protein of Sindbis virus has multiple functions in the life cycle of the virus. One essential function is to interact with the genomic RNA of the virus to form the nucleocapsid. The experiments described in this article define a region of the protein that is required for binding to Sindbis virus RNA. The assay we used measured the binding of in vitro-translated proteins to RNA on the basis of their migration with the RNA during electrophoresis in an agarose gel. Binding to RNA showed specificity; more protein bound to an RNA containing the previously defined packaging signal in Sindbis virus RNAs than to a similar RNA lacking this sequence. We were able to produce a variety of deleted forms of the capsid protein by constructing cDNAs with in-frame deletions throughout the coding region of the capsid protein gene. These cDNAs were then transcribed into mRNAs and translated in vitro. C-terminal deletions in the capsid protein were obtained by preparing transcripts from cDNAs linearized at sites within the coding region. Our studies identified a 32-amino-acid region that is essential for the specificity in RNA binding, and they defined a 68-amino-acid minimal sequence which displays almost the complete specific RNA binding activity of the intact Sindbis virus capsid protein containing 264 amino acids.

Characterization of an internal element in turnip crinkle virus RNA involved in both coat protein binding and replication

The major coat-protein-binding element of turnip crinkle virus RNA was previously mapped in the region of the UAG termination codon in the viral polymerase gene. This region encompasses two of the high-affinity coat-protein-binding sites (Fa and Ff) that we suggested were physically associated in a stem-loop in a ribonucleoprotein complex involved in assembly initiation (Wei, Heaton, Morris, and Harrison, J. Mol. Biol. 214, 85-95, 1990). We have also demonstrated that this RNA element was capable of specific coat protein binding in vitro (Wei and Morris, J. Mol. Biol. 222, 437-443, 1991). We now provide physical evidence, by in vitro chemical and enzymatic probing of the viral RNA, that support the suggestion that the two coat-protein-binding sites base pair to form a stem structure (A/F stem) surrounding the UAG terminator in wild-type RNA. We have shown here that a mutant with seven conservative nucleotide substitutions in Fa does not accumulate to detectable levels in plants or protoplasts and that the A/F stem structure is drastically altered in this mutant. We suggest that the primary effect of this mutation is on replication rather than on a reduction in RNA stability resulting from a defect in encapsidation of the virion RNA because previous results have shown that encapsidation-deficient mutants have little or no effect on viral RNA replication (Hacker, Petty, Wei, and Morris, Virology 186, 1-8, 1992). The analysis of the A/F stem was extended by construction and characterization of a series of mutants and revertants that displayed variable levels of replication deficiency but minimal concomitant defect in encapsidation efficiency. The extent of the replication defect correlated with the predicted destabilization of the A/F stem structure. We conclude from these results that this RNA element is involved in viral replication, and we tentatively suggest that the A/F stem structure may be functionally involved in the readthrough translation of the viral polymerase.

Topological analysis of the hepatitis B virus core particle by cysteine-cysteine cross-linking

The nucleocapsid, or core particle, of hepatitis B virus is formed by 180 subunits of the core protein, which contains Cys at positions 48, 61, 107 and 183, the latter constituting the C terminus. Upon adventitious oxidation, some or all of these cysteine residues participate in the formation of disulphide bridges, leading to polymerization of the subunits within the particle. To utilize the cysteine residues as topological probes, we reduced the number of possible intersubunit crosslinks by replacing these residues individually, or in all combinations, by serine. A corresponding set of variants was constructed within the context of an assembly-competent core protein variant that lacks the highly basic C-terminal region. Analysis, by polyacrylamide gel electrophoresis under non-reducing conditions, of the oxidative crosslinking products formed by the wild-type and mutant proteins expressed in Escherichia coli, revealed a clear distinction between the three N-proximal, and the C-terminal Cys: N-proximal Cys formed intermolecular disulphide bonds only with other N-proximal cysteine residues, leading to dimerization. Cys48 and Cys61, in contrast to Cys107, could be crosslinked to the homologous cysteine residues in a second subunit, and are therefore located at the dimer interface. Cys 183 predominantly formed disulphide bonds with Cys183 in subunits other than those crosslinked by the N-proximal cysteine residues. Hence, the polymers generated by oxidation of the wild-type protein are S-S-linked dimeric N-terminal domains interconnected via Cys183/Cys183 disulphide bonds. The intermolecular crosslinks between the N-proximal cysteine residues were apparently the same in the C-terminally truncated and in the full-length proteins, corroborating the model in which the N-terminal domain and the C terminus of the HBV core protein form two distinct and structurally independent entities. The strong tendency of the N-terminal domain for dimeric interactions suggests that core protein dimers are the major intermediates in hepatitis B virus nucleocapsid assembly.

Calculation of the free energy of association for protein complexes

We have developed a method for calculating the association energy of quaternary complexes starting from their atomic coordinates. The association energy is described as the sum of two solvation terms and an energy term to account for the loss of translational and rotational entropy. The calculated solvation energy, using atomic solvation parameters and the solvent accessible surface areas, has a correlation of 96% with experimentally determined values. We have applied this methodology to examine intermediates in viral assembly and to assess the contribution isomerization makes to the association energy of molecular complexes. In addition, we have shown that the calculated association can be used as a predictive tool for analyzing modeled molecular complexes.

The role of proteolytic processing in the morphogenesis of virus particles

Proteinases are encoded by many RNA viruses, all retroviruses and several DNA viruses. They play essential roles at various stages in viral replication, including the coordinated assembly and maturation of virions. Most of these enzymes belong to one of three (Ser, Cys or Asp) of the four major classes of proteinases, and have highly substrate-selective and cleavage specific activities. They can be thought of as playing one of two general roles in viral morphogenesis. Structural proteins are encoded by retroviruses and many RNA viruses as part of large polyproteins. Their proteolytic release is a prerequisite to particle assembly; consequent structural rearrangement of the capsid domains serves to regulate and direct association and assembly of capsid subunits. The second general role of proteolysis is in assembly-dependent maturation of virus particles, which is accompanied by the acquisition of infectivity.

Turnip crinkle virus genes required for RNA replication and virus movement

We have used infectious in vitro transcripts from mutagenized turnip crinkle virus (TCV) cDNA clones to identify the gene products required for viral RNA replication, virion assembly, and intercellular movement. Previous sequence analysis of the TCV genome revealed the presence of five open reading frames which had the potential to encode gene products of 88, 38, 28, 9, and 8 kDa. Inoculation of protoplasts with infectious RNA revealed that only the p28 and p88 gene products are required for viral RNA synthesis. Although the p8 and p9 gene products were dispensable for RNA replication and virion assembly in protoplasts, mutations in the p8 and p9 genes prevented the production of systemic infections in plants. No viral RNA or protein was observed in the inoculated or systemic leaves of plants inoculated with transcripts synthesized from p8 or p9 mutant cDNAs. In contrast to these results, viral RNA was recovered from the inoculated, but not the systemic leaves, of plants inoculated with an RNA lacking the coat protein (CP) gene. With the CP mutant, no symptoms were observed on normally systemic hosts, but small local lesions were induced on Chenopodium amaranticolor. These results indicate that p8, p9, and CP are required for viral movement.

Interactions between viral coat protein and a specific binding region on turnip crinkle virus RNA

The turnip crinkle virus coat protein binding sites in the ribonucleoprotein complex resulting from virion dissociation have been identified previously. In this study, RNA binding characteristics of viral coat protein to a region encompassing the protected RNA fragments Fa, Ff, and Fc (Fafc) have been investigated further using an RNA transcript (the Fafc fragment). These experiments have shown that coat protein requires no additional viral RNA elements to bind to this region. Such binding was shown to be specific for turnip crinkle virus coat protein using an ultra-violet light cross-linking assay. Gel mobility shift analyses demonstrated that the protein-RNA interactions produced two complexes: a homogeneous small ribonucleoprotein complex, and larger complexes which failed to migrate into gels. High salt and limiting protein concentrations favored the formation of the small ribonucleoprotein complex, whereas low salt and excess protein concentrations favored the larger complexes. RNA competition experiments demonstrated that small ribonucleoprotein complex formation coincided with specific RNA binding of the coat protein to the Fafc fragment. In addition, the coat protein possessed a poly(U)-binding site(s), which enabled it to interact with single-stranded RNA in a sequence non-specific manner to form large complexes. The results suggest that the coat protein contains both specific and non-specific RNA binding activities located at physically distinct sites. These results are consistent with the proposed assembly model for turnip crinkle virus.

Structure of simian virus 40 at 3.8-A resolution

The crystallographically determined structure of simian virus 40 shows that the 72 pentamers of viral protein VP1, which form the outer shell, have identical conformations except for the C-terminal arms of their subunits. Five arms emerge from each pentamer and insert into neighbouring pentamers. This tying together of standard building blocks allows for the required variability in packing geometry without sacrificing specificity.

Hyperreactivity of adenines and conformational flexibility of a translational repression site

We have used a diethylpyrocarbonate (DEPC) modification [(1976) Prog. Nucl. Acids Res. 16, 189-262] to probe the accessibility of adenines essential for coat protein binding in the MS2 translational operator [(1983) Biochemistry 22, 2601-2610, 2610-2615, 4723-4730; (1987) Biochemistry 26, 1563-1568]. The essential adenines are apparently hyperreactive with this reagent relative to other sites within the same molecule. Variation of ionic strength, pH and divalent cation concentrations reveal the existence of two distinct conformers of the RNA operator as judged by DEPC reactivity. We propose that the hyperreactivity observed is due to the participation of neighbouring bases in the DEPC modification reaction and can be used as a novel structural probe.

Structure and assembly of turnip crinkle virus. VI. Identification of coat protein binding sites on the RNA

Structural studies of turnip crinkle virus have been extended to include the identification of high-affinity coat protein binding sites on the RNA genome. Virus was dissociated at elevated pH and ionic strength, and a ribonucleoprotein complex (rp-complex) was isolated by chromatography on Sephacryl S-200. Genomic RNA fragments in the rp-complex, resistant to RNase A and RNase T1 digestion and associated with tightly bound coat protein subunits, were isolated using coat-protein-specific antibodies. The identity of the protected fragments was determined by direct RNA sequencing. These approaches allowed us to study the specific RNA-protein interactions in the rp-complex obtained from dissociated virus particles. The location of one protected fragment downstream from the amber terminator codon in the first and largest of the three viral open reading frames suggests that the coat protein may play a role in the regulation of the expression of the polymerase gene. We have also identified an additional cluster of T1-protected fragments in the region of the coat protein gene that may represent further high-affinity sites involved in assembly recognition.

Hepatitis B virus nucleocapsid assembly: primary structure requirements in the core protein

As a step toward understanding the assembly of the hepatitis B virus (HBV) nucleocapsid at a molecular level, we sought to define the primary sequence requirements for assembly of the HBV core protein. This protein can self assemble upon expression in Escherichia coli. Applying this system to a series of C-terminally truncated core protein variants, we mapped the C-terminal limit for assembly to the region between amino acid residues 139 and 144. The size of this domain agrees well with the minimum length of RNA virus capsid proteins that fold into an eight-stranded beta-barrel structure. The entire Arg-rich C-terminal domain of the HBV core protein is not necessary for assembly. However, the nucleic acid content of particles formed by assembly-competent core protein variants correlates with the presence or absence of this region, as does particle stability. The nucleic acid found in the particles is RNA, between about 100 to some 3,000 nucleotides in length. In particles formed by the full-length protein, the core protein mRNA appears to be enriched over other, cellular RNAs. These data indicate that protein-protein interactions provided by the core protein domain from the N terminus to the region around amino acid 144 are the major factor in HBV capsid assembly, which proceeds without the need for substantial amounts of nucleic acid. The presence of the basic C terminus, however, greatly enhances encapsidation of nucleic acid and appears to make an important contribution to capsid stability via protein-nucleic acid interactions. The observation of low but detectable levels of nucleic acid in particles formed by core protein variants lacking the Arg-rich C terminus suggests the presence of a second nucleic acid-binding motif in the first 144 amino acids of the core protein. Based on these findings, the potential importance of the C-terminal core protein region during assembly in vivo into authentic, replication-competent nucleocapsids is discussed.

Structural studies of virus particles

X-ray crystallographic studies of complete virus particles have resulted in models with resolution at atomic level. They have begun to give some insight into the mechanisms behind processes in the viral life cycle such as assembly, disassembly and cell attachment.

The three-dimensional structure of the bacterial virus MS2

The structure of the icosahedral bacteriophage MS2 has been determined to 3.3 A resolution by X-ray crystallography. The phase determination involved both molecular replacement at low resolution using a known structure and heavy-atom substitution. The coat protein has no structural similarity to that of any other known RNA virus.

Evidence for specificity in the encapsidation of Sindbis virus RNAs

We investigated the interaction of the capsid protein of Sindbis virus with Sindbis viral RNAs and defined a region of the genome that is required for binding in vitro and for packaging in vivo. The binding studies were performed with purified capsid protein immobilized on nitrocellulose and 32P-labeled RNAs transcribed in vitro from viral and nonspecific cDNAs. Genomic and defective interfering (DI) RNAs bound capsid protein significantly better than either the subgenomic (26S) RNA or nonspecific RNAs. Transcripts prepared from either truncated or deleted cDNAs were used to define the segment required for binding. This segment, which is represented twice in DI RNA, lies between nucleotides 746 and 1226 of the genomic RNA and is within the coding region of the nonstructural protein nsP1. Insertion of a domain covering these sequences into a nonviral RNA was able to convert it from a background level of binding to an activity that was 80% that of the Sindbis virus DI RNA. We analyzed DI RNA transcripts in detail because they could be studied not only for the ability to bind capsid protein in vitro but also for the ability to be replicated and packaged in vivo in the presence of helper virion RNA. The results obtained with three DI RNAs are reported. One (CTS14), which has one copy of the binding domain, bound efficiently to capsid protein in vitro and was packaged in vivo as measured by amplification on passaging. In contrast, a DI RNA (CTS1) which lacked this region did not bind to capsid protein and was not detected on passaging. By using lipofectin (P. L. Felgner, T. R. Gadek, M. Holm, R. Roman, H. W. Chan, M. Wenz, J.P. Northrop, G. M. Ringold, and M. Danielson, Proc. Natl. Acad. Sci. USA 84:7413-7417, 1987) to enhance RNA uptake, we were able to demonstrate that CTS1 RNA was replicated in the transfected cells. It was replicated to the same level as another DI RNA (CTS253) which has only the 3' 279 nucleotides of the binding domain and these are located near the 3' terminus of the RNA. CTS253 bound capsid protein to an intermediate level but was amplified on passaging. The binding studies and the in vivo packaging data, taken together, provide strong support for the conclusion that there is a specific capsid recognition domain in Sindbis virus RNA that plays a role in nucleocapsid assembly.

Effects of deletions in the N-terminal basic arm of brome mosaic virus coat protein on RNA packaging and systemic infection

The first 25 amino acids of brome mosaic virus (BMV) coat protein include 8 basic and no acidic residues and are implicated in binding the encapsidated RNA. Using infectious transcripts from BMV RNA3 cDNA clones, we modified this region of the coat gene. A coat protein mutant with the first 25 amino acids deleted failed to direct either packaging of viral RNA in protoplasts or systemic infection of whole barley plants. Neither symptoms, virions, nor viral RNA was detectable in plants inoculated with this mutant or a mutant with a frameshift mutation in the coat gene. Mutants with the normal start codon changed to AAG or with the first eight codons deleted allowed translation to start at a downstream AUG, resulting in a deletion of the first 7 amino acids of the mature wild-type coat protein. These mutants not only packaged viral RNA in protoplasts but directed symptomatic, systemic infections that developed with normal speed and degree of spread within the host. The AUG-to-AAG point substitution did not revert to the wild type after long-term culture in planta. Wild-type BMV virions were also found to contain small amounts of a protein that coelectrophoresed with the truncated coat protein produced by the viable AAG and eight-codon-deletion mutants. This minor coat protein species presumably arose by infrequent translation initiation at the second AUG in the wild-type coat protein gene. Absence of encapsidation-competent coat protein appeared to stimulate production of nonstructural proteins in protoplast infections.

Structure-function relationships of icosahedral plant viruses

X-ray diffraction studies on single crystals of a few viruses have led to the elucidation of their three dimensional structure at near atomic resolution. Both the tertiary structure of the coat protein subunit and the quaternary organization of the icosahedral capsid in these viruses are remarkably similar. These studies have led to a critical re-examination of the structural principles in the architecture of isometric viruses and suggestions of alternative mechanisms of assembly. Apart from their role in the assembly of the virus particle, the coat proteins of certian viruses have been shown to inhibit the replication of the cognate RNA leading to cross-protection. The coat protein amino acid sequence and the genomic sequence of several spherical plant RNA viruses have been determined in the last decade. Experimental data on the mechanisms of uncoating, gene expression and replication of several classes of viruses have also become available. The function of the non-structural proteins of some viruses have been determined. This rapid progress has provided a wealth of information on several key steps in the life cycle of RNA viruses. The function of the viral coat protein, capsid architecture, assembly and disassembly and replication of isometric RNA plant viruses are discussed in the light of this accumulated knowledge.

The genome structure of turnip crinkle virus

The nucleotide sequence of turnip crinkle virus (TCV) genomic RNA has been determined from cDNA clones representing most of the genome. Segments were confirmed using dideoxynucleotide sequencing directly from viral RNA, and the 3' terminal sequence was confirmed by chemical sequencing of end-labeled genomic RNA. Three open reading frames (ORFs) have been identified by examination of the deduced amino acid sequences and by comparison with the ORFs found in the genome of carnation mottle virus. ORF 1 initiates near the 5' terminus of the genome and is punctuated by an amber termination codon. Translation of ORF 1 would yield a 28-kDa protein and an 88-kDa read-through product. The read-through domain possesses amino acid sequence similarities with putative viral RNA polymerases. ORFs 2 and 3 encode products of 38 (coat protein) and 8 kDa, respectively, which are expressed from subgenomic mRNAs. The organization of the TCV genome suggests that TCV is closely related to carnation mottle virus and distinct from members classified in other small RNA virus groups, such as the tombus- and sobemoviruses.

Turnip crinkle virus infection from RNA synthesized in vitro

Genome-length cDNA clones of turnip crinkle virus (TCV) were constructed with SmaI and XbaI restriction sites engineered at the 5' and 3' termini, respectively. The genome-length cDNAs were positioned downstream of modified lambda and T7 phage promoters such that in vitro transcription resulted in RNAs with 5 extra nucleotides at the 3' end, and 1, 2, or 14 extra nucleotides at the 5' end depending on the construction. Transcripts with 14 extraviral 5' nucleotides were not infectious, while transcripts with 1 or 2 additional 5' nucleotides, with or without 5'-cap analog included in transcription reactions, were biologically active. These were approximately an order of magnitude less infectious than RNA extracted from TCV virions. The additional 5' nucleotides were not maintained in progeny viral RNAs isolated from plants.