Identification of specific cucumber necrosis virus coat protein amino acids affecting fungus transmission and zoospore attachment.

Cucumber necrosis virus (CNV) is naturally transmitted in the soil by zoospores of the fungal vector Olpidium bornovanus. Successful transmission requires that virus particles attach to the surface of zoospores prior to zoospore encystment on host roots. Mechanically passaged CNV was screened for mutants deficient in fungus transmission. We found six such mutants, exhibiting transmission efficiencies ranging from approximately 14 to 76% of that of wild-type (WT) CNV. Results of in vitro virus-zoospore binding assays show that each mutant binds to zoospores less efficiently than WT CNV (21 to 68%), suggesting that defects in transmission for these mutants are at least partially due to inefficient zoospore binding. Analysis of the structure of the CNV coat protein subunit and trimer indicates that affected amino acids in all of the mutants are located in the shell or protruding domain and that five of six of them are potentially exposed on the surface of the virus particle. In addition, several of the mutated sites, along with a previously identified site in a region of subunit-subunit interaction in the coat protein shell domain (M. A. Robbins, R. D. Reade, and D. M. Rochon, Virology 234:138-146, 1997), are located on the particle quasi-threefold axis, suggesting that this region of the capsid may be important in recognition of a putative zoospore receptor. The individual sites may directly affect attachment to a receptor or could indirectly affect attachment via changes in virion conformation.

Expression of mutated poliovirus receptors in human neuroblastoma cells persistently infected with poliovirus.

Poliovirus (PV) is able to establish persistent infections in human neuroblastoma IMR-32 cells [Colbère-Garapin et al. (1989) Proc. Natl. Acad. Sci. USA 86, 7590]. During persistent infection, PV mutants are selected that display substitutions of residues in regions of the capsid known to interact with the PV receptor (PVR), a glycoprotein of the immunoglobulin superfamily. The mechanism of persistent infection in IMR-32 cells may therefore involve the selection of mutant PVRs. To test this hypothesis, the sequences of the PVR mRNAs in uninfected IMR-32 cells and in two independent IMR-32 cell cultures persistently infected with the Mahoney strain of PV type 1 (PV1/Mahoney) were determined. The PVR mRNA population of uninfected cells was homogeneous, and no mutation was repeatedly found, whereas that of persistently infected cells displayed missense mutations. Particular mutations were repeatedly detected, and all of them mapped to the N-terminal domain of PVR (domain 1), which interacts directly with PV. These mutations generated several types of PVR variants with the following substitutions: Ala67-->Thr alone, Ala67-->Thr associated with Gly39-->Ser, and Arg104-->Gln. Functional analysis of PVR in murine LM cells, stably expressing each of the PVR forms, showed that the PVR forms selected during persistent infection conferred on LM cells partial resistance to PV1/Mahoney-induced lysis. Although adsorption onto PVR seemed to be independent of the PVR form, an analysis of the conformational changes of the capsid during the early steps of the PV cycle provided evidence that the Ser39/Thr67 and Gln104 substitutions almost halved the conversion of 160S infectious particles into 135S A particles associated with the PV-PVR interaction. Altogether, these findings indicate that during persistent infection, specific mutations were selected in the domain 1 of PVR and that these mutations increased the resistance of cells to PV-induced lysis. These results are discussed in view of the position of the mutations on PVR.

Quasispecies development by high frequency RNA recombination during MHV persistence.

Recent studies suggest that infectious viruses and particularly persisting viral RNAs often exist as diverse populations or "quasispecies". We have developed an approach to characterize populations of the murine coronavirus mouse hepatitis virus (MHV) generated during persistent infection which has allowed us to begin to address the role of the viral quasispecies in MHV pathogenesis. We analyzed the population of persisting viral RNAs using reverse-transcription polymerase chain reaction amplification (RT-PCR) of the S1 "hypervariable" region of the spike gene followed by differential colony hybridization to identify spike deletion variants (SDVs) from acute and persistently infected mice. Sequence analysis revealed that mice with the most severe chronic paralysis harbored the most complex quasispecies. Mapping of the SDVs to the predicted RNA secondary structure of the spike RNA revealed that an isolated stem loop structure is frequently deleted. Overall, these results are consistent with high frequency recombination at sites of RNA secondary structure contributing to expansion of the viral quasispecies and persisting viral pathogenesis.

Molecular characterization of tomato spotted Wilt virus defective interfering RNAs and detection of truncated L proteins.

Junction sites of 25 different defective interfering (DI) RNAs of tomato spotted wilt virus (TSWV) were characterized. The DI RNAs varied in size from 2.0 to 5.2 kilobases (kb) and contained a single internal deletion. The absence of DI RNAs smaller than 2 kb suggested a size constraint for the survival of TSWV DI RNAs. This hypothesis was reinforced by the finding of a dimeric DI RNA formed by two 1.6-long monomers linked head to tail. Three types of junction sites were found, one type originating from a simple deletion; the second type contained a few extra nucleotides of unknown origin; and the third type contained a stretch of three to five nucleotides, originally occurring at both sides of the deletion and of which one was deleted. In 19 of the 25 DI RNAs studied, the original reading frame was maintained, suggesting a selective preference of DI RNAs with translational potency. Truncated proteins encoded by these DI RNAs were indeed detected in the nucleocapsid preparations. Folding studies of the complete L RNA revealed that the calculated minimal energy of folding was at 16 degreesC lower than at 23 degrees, indicating a higher stability of this molecule at low temperatures. The results suggest an involvement of locally folded secondary structures in the process of deletion, rather than the requirement of certain sequences around the deletion point. The DI RNA generation in TSWV is essentially, as discussed, similar to the process of RNA recombination described in many viruses.

IRIS explorer software for radial-depth cueing reovirus particles and other macromolecular structures determined by cryoelectron microscopy and image reconstruction.

Structures of biological macromolecules determined by transmission cryoelectron microscopy (cryo-TEM) and three-dimensional image reconstruction are often displayed as surface-shaded representations with depth cueing along the viewed direction (Z cueing). Depth cueing to indicate distance from the center of virus particles (radial-depth cueing, or R cueing) has also been used. We have found that a style of R cueing in which color is applied in smooth or discontinuous gradients using the IRIS Explorer software is an informative technique for displaying the structures of virus particles solved by cryo-TEM and image reconstruction. To develop and test these methods, we used existing cryo-TEM reconstructions of mammalian reovirus particles. The newly applied visualization techniques allowed us to discern several new structural features, including sites in the inner capsid through which the viral mRNAs may be extruded after they are synthesized by the reovirus transcriptase complexes. To demonstrate the broad utility of the methods, we also applied them to cryo-TEM reconstructions of human rhinovirus, native and swollen forms of cowpea chlorotic mottle virus, truncated core of pyruvate dehydrogenase complex from Saccharomyces cerevisiae, and flagellar filament of Salmonella typhimurium. We conclude that R cueing with color gradients is a useful tool for displaying virus particles and other macromolecules analyzed by cryo-TEM and image reconstruction.

Generation of coronavirus spike deletion variants by high-frequency recombination at regions of predicted RNA secondary structure.

Coronavirus RNA evolves in the central nervous systems (CNS) of mice during persistent infection. This evolution can be monitored by detection of a viral quasispecies of spike deletion variants (SDVs) (C. L. Rowe, S. C. Baker, M. J. Nathan, and J. O. Fleming, J. Virol. 71:2959-2969, 1997). We and others have found that the deletions cluster in the region from 1,200 to 1,800 nucleotides from the 5' end of the spike gene sequence, termed the "hypervariable" region. To address how SDVs might arise, we generated the predicted folding structures of the positive- and negative-strand senses of the entire 4,139-nt spike RNA sequence. We found that a prominent, isolated stem-loop structure is coincident with the hypervariable region in each structure. To determine if this predicted stem-loop is a "hot spot" for RNA recombination, we assessed whether this region of the spike is more frequently deleted than three other selected regions of the spike sequence in a population of viral sequences isolated from the CNS of acutely and persistently infected mice. Using differential colony hybridization of cloned spike reverse transcription-PCR products, we detected SDVs in which the hot spot was deleted but did not detect SDVs in which other regions of the spike sequence were exclusively deleted. Furthermore, sequence analysis and mapping of the crossover sites of 25 distinct patterns of SDVs showed that the majority of crossover sites clustered to two regions at the base of the isolated stem-loop, which we designated as high-frequency recombination sites 1 and 2. Interestingly, the majority of the left and right crossover sites of the SDVs were directly across from or proximal to one another, suggesting that these SDVs are likely generated by intramolecular recombination. Overall, our results are consistent with there being an important role for the spike RNA secondary structure as a contributing factor in the generation of SDVs during persistent infection.

Canyon rim residues, including antigenic determinants, modulate serotype-specific binding of polioviruses to mutants of the poliovirus receptor.

Several mouse cell lines expressing hybrid human poliovirus receptors (hPVRs) bearing mutations in the first immunoglobulin-like domain were previously characterized for their defective binding and replication of poliovirus type 1 Mahoney (G. Bernhardt, J. Harber, A. Zibert, M. DeCrombrugghe, and E. Wimmer, Virology, 203, 344-356, 1994). Here we report that these mutant hPVRs were utilized to explore differences in the binding behavior of the three serotypes of poliovirus. Type 3 polioviruses (both Sabin and the neurovirulent Leon strain) clearly bound to the hPVR mutant Q130G/GD, but were incapable of initiating infection. Also, binding at 25 degrees of poliovirus types 2 and 3 to cell lines expressing the hPVR mutants P84SYS/HPGA, L99GAE/AAAA, and D117F was greater than type 1 poliovirus. Further study of the serotype-specific interaction with mutant hPVRs was accomplished with antigenic hybrid viruses. Improved binding by antigenic hybrid viruses demonstrated that serotype-specific binding to mutant hPVRs is, in part, determined by the amino acid sequence of neutralization antigenic sites (NAgs) and the probable conformational rearrangement of amino acids adjacent to the NAg sites. Finally, site-directed mutants of poliovirus were utilized to determine the relative contributions, to hPVR interactions, of individual amino acids with solvent accessible side chains in the viral canyon. Of the 18 viable virus mutants produced, 3 (D1226A, I1089A, and VPEK1166HPGA) expressed impaired replication phenotypes on the mutant hPVR cell lines P84SYS/HYSA and D117F. A location at the rim of the poliovirus canyon was implicated for the interaction of the amino terminal domain of the poliovirus receptor with conserved and serotype-specific viral surface amino acids. The possible involvement of elements of neutralization antigenic sites in receptor binding may explain, in part, why poliovirus exists in only three serotypes.

De novo generation of defective interfering-like RNAs in broad bean mottle bromovirus.

Broad been mottle virus (BBMV) is the only member of the bromoviruses that is known to accumulate defective-interfering (DI) RNAs (Romero et al., Virology 194, 576-584, 1993). De novo generation of DI-like RNAs was demonstrated during serial passages of BBMV in broad bean using either DI RNA-free virion RNA preparations or transcribed genomic RNA inocula. As for previously described DI RNAs, all but one of the characterized de novo generated DI-like RNAs were derived by a single in-frame deletion from the RNA2 component. The sole exception was derived by two shorter in-frame deletions from RNA2. The maintenance of an open reading frame by all DI-like RNAs suggests the importance of coding capacity and/or the shortened 2a protein in the accumulation of these RNAs during infection. The deletion junction sites were between nucleotides 1152 and 2366, suggesting that the retained regions are essential for the efficient accumulation of BBMV DI-like RNAs in planta. Short regions of sequence similarity and/or complementarity were revealed at the 5' and 3' junction borders. We speculate that these regions can facilitate DI (DI-like) RNA formation. In addition to DI-like RNAs, the full-length nucleotide sequences of RNA2 components of the Type and Morocco strains of BBMV are presented.

Analysis of the cell and erythrocyte binding activities of the dimple and canyon regions of the canine parvovirus capsid.

Canine parvovirus (CPV) binds to a number of cell and erythrocyte receptors, some of which are involved in cell infection, while others are used for other viral functions. Little is known about the regions of the virus capsid which bind to the cell receptors. CPV binds sialic acid through a region within or adjacent to the dimple on the surface of the capsid (Barbis, D. P., Chang, S-F., and Parrish, C. R., 1992, Virology 191, 301-308). In order to map the sialic acid binding site in more detail and to examine other regions of the capsid for cell receptor binding, a variety of mutant capsids were analyzed which had changes in two depressions within the surface of the capsid--the "canyon" and "dimple." In most cases recombinant VP1 and VP2 proteins were stably expressed together in canine A72 cells from a plasmid expression vector. The purified empty capsids were tested for their ability to bind sialic acid and thereby hemagglutinate (HA) erythrocytes and for binding to permissive host cells. In addition, the ability of neutralizing monoclonal antibodies to block cell attachment was also examined. Mutations of amino acids on a wall of the dimple eliminated or severely decreased HA. Changing various residues within the canyon had no effect on binding to either sialic acids or other receptors on feline lymphoblastoid cells, suggesting that the canyon is not the site of cell receptor attachment. Neutralizing monoclonal antibodies against both major antigenic determinants had variable effects on cell binding, but no consistent inhibition of binding was observed by antibodies directed against either of those two major antigenic determinants of the capsid.

WIN51711-resistant mutants of poliovirus type 3: capsid residues important in uncoating functions.

Capsid-binding drugs that inhibit the first stage of picornaviral uncoating were used to select drug-resistant mutants of the Sabin strain of poliovirus type 3. Such mutants provide information about parts of the capsid that are important for functions blocked by the drugs, and also about pathways to drug resistance. Amino-acid substitutions allowing virus to produce progeny in the presence of drug were mapped to 13 different residues occupying three distinct locations: (I) the canyon base; (II) the lining of the drug-binding pocket; and (III) the base of the protomer. These loci might be thought of as action points for transmitting the uncoating signal from receptor, through the pocket, and to the base of the protomer. All of the mutations in a special class of drug-dependent mutants were clustered at site (III) and all were hyperlabile, i.e., uncoated spontaneously (without receptor) at growth temperature unless prevented from doing so by the presence of drug in the pocket. Thus, site (III) seems to represent a kind of thermostat which regulates the temperature at which the uncoating transition (release of VP4 to form A particles) is triggered.

Distribution of drug resistance mutations in type 3 poliovirus identifies three regions involved in uncoating functions.

We have previously described the use of an uncoating inhibitor, WIN 51711, to select drug-resistant mutants of the Sabin strain of poliovirus type 3. Two-thirds of the mutants proved to be dependent on the drug for plaque formation because of extreme thermolability (A. G. Mosser and R. R. Rueckert, J. Virol. 67:1246-1254, 1993). Here we report the responsible mutations; all were traced to single amino acid substitutions. Mutations conferring dependence and thermolability occurred in all four capsid proteins (VP1 to VP4), but all were clustered near residue 53 of VP4 at the inner capsid surface. Amino acid substitutions of the remaining non-drug-dependent mutants were mapped to three distinct loci: (i) on or near the inner capsid surface, at VP4 residue 46 or VP1 residue 129, in the vicinity of the drug dependence substitutions; (ii) at residues 192, 194, and 260 in the lining of the VP1 beta barrel, which is the drug-binding site; and (iii) at VP1 residue 105 on the edge of the canyon surrounding the fivefold axis of symmetry, the putative receptor-binding site. All of the mutations increased the eclipse rate of cell-attached virus. Such mutants help identify parts of the capsid that play a role in viral uncoating functions.

Two dominant neutralizing antigenic determinants of canine parvovirus are found on the threefold spike of the virus capsid.

The 25-nm diameter parvovirus capsid is assembled from 60 copies of a sequence common to the overlapping VP1 and VP2 proteins. Here we examine the epitope specificity's of 28 monoclonal antibodies (MAb) prepared against canine parvovirus (CPV), feline panleukopenia virus (FPV), and raccoon-dog parvovirus or blue (Arctic) fox parvovirus. Comparing the reactivity of those MAb with various MAb-selected escape mutants, or with natural variants of CPV or mink enteritis virus (MEV) which differ at known sequences, showed that the binding of 20 of those MAb was strongly affected by variations of two regions on the threefold spike of the CPV capsid. One region was adjacent to the tip of the threefold spike, and the second was around VP2 residue 300, on the shoulder of that structure. MAb recognizing both antigenic sites efficiently neutralized the virus infectivity and inhibited hemagglutination. Mutations leading to natural antigenic variation have also been observed in both those sites in naturally variant strains of CPV or MEV, suggesting that they are important antigenic structures on these parvoviruses. The bindings of several MAb were not affected by the mutations at those antigenic sites, indicating that they recognized other, and perhaps conserved, structures.

Multiple amino acids in the capsid structure of canine parvovirus coordinately determine the canine host range and specific antigenic and hemagglutination properties.

Canine parvovirus (CPV) and feline panleukopenia virus (FPV) are over 98% similar in DNA sequence but have specific host range, antigenic, and hemagglutination (HA) properties which were located within the capsid protein gene. In vitro mutagenesis and recombination were used to prepare 16 different recombinant genomic clones, and viruses derived from those clones were analyzed for their in vitro host range, antigenic, and HA properties. The region of CPV from 59 to 91 map units determined the ability to replicate in canine cells. A complex series of interactions was observed among the individual sequence differences between 59 and 73 map units. The canine host range required that VP2 amino acids (aa) 93 and 323 both be the CPV sequence, and those two CPV sequences introduced alone into FPV greatly increased viral replication in canine cells. Changing any one of aa 93, 103, or 323 of CPV to the FPV sequence either greatly decreased replication in canine cells or resulted in an inviable plasmid. The Asn-Lys difference of aa 93 alone was responsible for the CPV-specific epitope recognized by monoclonal antibodies. An FPV-specific epitope was affected by aa 323. Amino acids 323 and 375 together determined the pH dependence of HA. Amino acids involved in the various specific properties were all around the threefold spikes of the viral particle.

Rapid antigenic-type replacement and DNA sequence evolution of canine parvovirus.

Analysis of canine parvovirus (CPV) isolates with a panel of monoclonal antibodies showed that after 1986, most viruses isolated from dogs in many parts of the United States differed antigenically from the viruses isolated prior to that date. The new antigenic type (designated CPV type 2b) has largely replaced the previous antigenic type (CPV type 2a) among virus isolates from the United States. This represents the second occurrence of a new antigenic type of this DNA virus since its emergence in 1978, as the original CPV type (CPV type 2) had previously been replaced between 1979 and 1981 by the CPV type 2a strain. DNA sequence comparisons showed that CPV types 2b and 2a differed by as few as two nonsynonymous (amino acid-changing) nucleotide substitutions in the VP-1 and VP-2 capsid protein genes. One mutation, resulting in an Asn-Asp difference at residue 426 in the VP-2 sequence, was shown by comparison with a neutralization-escape mutant selected with a non-CPV type 2b-reactive monoclonal antibody to determine the antigenic change. The mutation selected by that monoclonal antibody, a His-Tyr difference in VP-2 amino acid 222, was immediately adjacent to residue 426 in the three-dimensional structure of the CPV capsid. The CPV type 2b isolates are phylogenetically closely related to the CPV type 2a isolates and are probably derived from a common ancestor. Phylogenetic analysis showed a progressive evolution away from the original CPV type. This pattern of viral evolution appears most similar to that seen in some influenza A viruses.

Structural homology among four nodaviruses as deduced by sequencing and X-ray crystallography.

The genomic RNA2s of nodaviruses encode a single gene, that of protein alpha, the precursor of virion proteins beta and gamma. We compared the sequences of the RNA2s of the nodaviruses, black beetle virus (BBV), flock house virus, boolarra virus and nodamura virus, with the objective of identifying homologies in the primary and secondary structure of these RNAs and in the structure of their encoded protein. The sequences of the four RNAs were found to be similar, so that homologous regions relating to translation and RNA replication were readily identified. However, the overall, secondary structures in solution, deduced from calculations of optimal Watson-Crick base-pairing configurations, were very different for the four RNAs. We conclude that a particular, overall, secondary structure in solution within host cells is not required for virus viability. The partially refined X-ray structure of BBV (R = 26.4% for the current model) was used as a framework for comparing the structure of the encoded proteins of the four viruses. Mapping of the four protein sequences onto the BBV capsid showed many amino acid differences on the outer surface, indicating that the exteriors of the four virions are substantially different. Mapping in the beta-barrel region showed an intermediate level of differences, indicating that some freedom in choice of amino acid residues is possible there although the basic framework of the capsids is evidently conserved. Mapping onto the interior surface of the BBV capsid showed a high degree of conservation of amino acid residues, particularly near the protein cleavage site, implying that that region is nearly identical in all four virions and has an essential role in virion maturation, and also suggests that all four capsid interior surfaces have similar surfaces exposed to the viral RNA. Apart from a small portion of the C promoter, the amino terminus of the BBV protein (residues 1 to 60) is crystallographically disordered and the amino acid residues in that region are not well conserved. The disordered portion of the BBV protein clearly projects from the capsid inner surface into the interior of the virion, the region occupied by the viral RNA. In all four viruses, residues 1 to 60 had a high proportion of basic residues, suggesting a virus-specific interaction of the amino terminus with the virion RNA.

Identification of regions of brome mosaic virus coat protein chemically cross-linked in situ to viral RNA.

RNA-protein cross-links were introduced into brome mosaic virus in situ by using the heterobifunctional agent p-azidophenylglyoxal. An improved RNA isolation method, without phenol extraction, was used to isolate RNA cross-linked with protein. RNA of the covalently linked complex was acid-digested and the oligonucleotides still attached to protein were 5'-end-labelled with 32P. The complexes were digested with trypsin and the tryptic peptides were purified by reversed-phase high-performance liquid chromatography. Amino acid analyses of cross-linked tryptic peptides revealed that out of the total 188 amino acids of brome mosaic virus coat protein only the 80 N-terminal amino acids are involved in the interaction with viral RNA. These results are discussed in connection with a predicted secondary structure of the coat protein. Both alpha helix (for amino acids 11-19) and other structures (between amino acids 20 and 80) are implicated in the coat protein-viral RNA interactions.