The Reovirus Mutant tsA279 L2 Gene Is Associated with Generation of a Spikeless Core Particle: Implications for Capsid Assembly

tsA279 selectively produces a novel subviral particle at the restrictive temperature.Previous experiments that used thin-section electron microscopy to examine group A mutants suggested there was no interference in assembly at 39°C in tsA201infections (33) and that both core-like and top-component particles were present in 39.5°C tsA279 cultures (43,44). Thin-section electron microscopy provides insufficient resolution to distinguish between structures with subtle differences. In addition to difficulties in identifying particle type, analyses of random cell sections may not include the population of particles released by cell lysis. An ideal way to determine the nature of various products is negative-stain electron microscopy because of the higher resolution of this technique (13, 42, 64). However, when the production of virally encoded products is reduced, as may occur in restrictive mutant infections, total populations of intermediates may be missed when determinations are made from gradient-purified fractions. To determine the exact nature and quantity of the different structures present, direct particle counting of negative-stained whole-cell cytoplasmic extracts was conducted after direct centrifugation onto an electron microscope grid (38, 63).

The relative proportions of structures produced by tsA279and T3D at both restrictive and permissive temperatures, as determined from direct particle counts of whole-cell lysates, are shown in Table1. Cells infected with tsA279and grown at a restrictive temperature produced approximately 50-fold fewer viral particles than those grown at permissive temperatures. The reduction in particle yield is consistent with previous reports of reduced viral protein production by this mutant under restrictive growth conditions (44). The particle/PFU ratio did not differ significantly between permissive and restrictive cultures of both T3D and tsA279 and ranged from 9 to 14 particles per PFU (data not shown). These values are similar to those in previous reports (80).

The predominant structural form produced at 32°C by bothtsA279 and T3D was the genome complete whole virion (diameter of approximately 80 nm) after both 36 and 72 h of incubation (between 77 and 86% of the total particles produced) (Table1). The next most common type of particle produced was the top component (6 to 19%, diameter of approximately 80 nm). Outer shell structures (diameter of approximately 90 nm) and core particles (approximately 50 nm in diameter) each comprised less than 6% of the particles produced. When infected cells were incubated at a restrictive temperature of 40°C, the distribution of particles produced was different. The proportion of whole virions was reduced to approximately 50 to 60% of all observed particles for both parental T3D and thets mutant. With T3D there was a trend toward increased proportions of top component and outer shells at the restrictive temperature at both 36 and 72 h postattachment. In addition, there was a trend toward an increased proportion of complete core particles, which contained both genome and the characteristic λ2 apical spike, at 72 h postinfection. The mutant produced similar proportions of top-component structures and about 10-fold fewer outer-shell structures than did T3D at a restrictive temperature. However, the proportion of core-like particles produced by the mutant clone increased from less than 2% at the permissive temperature to 18 and 31% at the nonpermissive temperature for the two different times tested (P < 0.0001 by χ² test with the Yates correction) (Table 1).

Core-like particles produced by tsA279 at a restrictive temperature (diameter of approximately 50 nm) appeared to be defective for the λ2 pentameric spike structure (Fig.1A), whereas core particles produced under permissive growth conditions for both T3D and tsA279and at a restrictive temperature for T3D clearly demonstrated the presence of the λ2 spike structure (data not shown). In addition, mutant core-like particles which were restrictively assembled (cultured and assembled under restrictive growth conditions) were penetrated by the negative stain, indicating that they were deficient for genomic dsRNA, while core particles produced under permissive and restrictive conditions by T3D and under permissive conditions by tsA279were not penetrated by the negative stain, indicating that these particles contained genomic dsRNA. The identity of the core-like particles produced by tsA279 under restrictive culture conditions was confirmed with polyclonal antireovirus IgG conjugated to a 12-nm colloidal gold probe (Fig. 1B, large arrow). The distributions of particles produced under the different temperature conditions were not significantly altered by input MOIs of either 5 or 50 PFU/cell (data not shown).

• Open in new tab
• Download powerpoint

Fig. 1.

Negative-stain and immunogold-labeled λ2 spike-defective core particles produced at restrictive temperature. (A) Restrictive temperature cytoplasmic extracts oftsA279-infected cells were centrifuged directly onto electron microscope grids and negatively stained with PTA as described in Materials and Methods. (B) Same as in panel A except the grid samples were immunolabeled (12-nm gold) before negative staining. The core-like particles in both panels (large arrows) demonstrated apical depressions where the λ2 pentameric core spike structure would be expected (small arrows). (C) Permissive-temperature cytoplasmic extracts of tsA279-infected cells. (D) Restrictive-temperature cytoplasmic extracts of T3D-infected cells. Magnification, ×100,000. The bars represent 100 nm.

The production of λ2 deficient core particles maps to thetsA279 L2 gene.To determine which tsA279gene(s) was associated with the spike-defective phenotype, we analyzed the distribution of particles produced at the restrictive temperature by various reassortant clones which segregated the mutant L2 and M2 gene segments (Table 2). Each of the six clones that contain the mutant L2 gene produced significant amounts of spike-defective particles, whereas each clone with the T1L L2 gene produced insignificant amounts of spike-defective particles. All of the remaining gene segments were randomly associated with respect to the presence or absence of the spike.

tsA279 produces structures which do not contain the λ2 core spike protein nor the minor core protein μ2.The above results indicated that the tsA279 L2 gene was associated with the nonpermissive production of a λ2 defective core-like structure. To determine whether the core-like particles contained λ2, but in a conformation which was not resolved, or were devoid of λ2, we purified them and analyzed their protein content. Restrictively assembled tsA279 assembly intermediate structures labeled with [³⁵S]methionine-[³⁵S]cysteine were isolated by using cesium chloride buoyant density centrifugation as detailed in Materials and Methods. Scintillation counting of fraction aliquots identified a single peak of activity at a density of 1.33 to 1.35 gm/ml (Fig. 2A). Electron microscopic examination of this fraction revealed complete virions at this density (Fig. 2B1). No individual peaks of radioactivity could be distinguished in fractions of lower density. Core-like particles which were deficient for both genome and the λ2 spike structure were present at a density of 1.28 to 1.29 gm/ml (Fig. 2B2), and outer-capsid structures were isolated at a density range of 1.26 to 1.28 gm/ml (Fig. 2B3). No core-like particles were observed at higher buoyant densities. It was not possible to cleanly separate the λ2 protein-deficient core particles from the outer-shell structures by rate zonal centrifugation due to their similar sedimentation characteristics and the small amount of material present in the fractions after sucrose gradient centrifugation (data not shown).

• Open in new tab
• Download powerpoint

Fig. 2.

Purification of restrictively assembled core-like particles. L929 cells were infected with tsA279, viral proteins were labeled with [³⁵S]methionine-[³⁵S]cysteine during culture at 40°C, cell pellets were collected, the cytoplasmic membranes were solubilized with desoxycholate, and the samples were extracted two times with Freon 113. (A) Viral components from 6 × 10⁸ cells were separated on preformed CsCl gradients, fractionated, and the density of each fraction determined from the refractive  index; aliquots were then were counted for radioactivity as described in Materials and Methods. ●, Radioactivity in counts per minute/10 μl; +, fraction density in grams/cubic centimeter. (B) Electron micrographs of representative particles found in separate gradient fractions from the experiment shown in panel A. Panels: 1, whole virus (fraction 15); 2, λ2 spike-defective core particles (fraction 19); 3, outer-shell structures (gradient fraction 19); 4, separately prepared and purified T3D core particles prepared from permissively assembled whole T3D virions. A small amount of top component was present in each of fractions 17 through 20 (arrowhead, panel 3). Note the apical cavity present in the spike-defective core (small arrow, panel 2) (bar = 100 nm). (C) Fluorogram of a TGU-SDS-PAGE gel loaded with equivalent scintillation counts of permissively assembled T3D and tsA279 whole virions and from the same samples shown in panel B. Note the absence of λ2 in lanes 2 and 3, which are comprised primarily of core particles (lane 2) and outer shells (lane 3), with a small amount of top component in each fraction. T3, permissively assembled, purified T3D virions. Lanes 1 to 3 correspond to the samples represented in panel B and lane A shows permissively assembled, purified tsA279 virions. (D) Immunoblot of a TGU-SDS-PAGE gel run in parallel with and with the same samples as shown in panel C, with the addition of lane 4 which represents permissively assembled cores (same as panel B4). The gel was probed simultaneously by both anti-whole reovirus antiserum and anti-μ2 antiserum. Note the intense reaction of anti-μ2 in lane 4 and absence of reaction in the remaining lanes. The anti-whole reovirus antiserum shows weak reaction to λ2 in lane 4 at the concentration used but no reaction to λ2 in the remaining lanes. Similar results were obtained in each of four other experiments.

TGU-SDS-polyacrylamide gel electrophoresis (PAGE) protein analysis of the above fractions indicated that both the spike-defective cores and the outer-shell structures contained little, if any, protein λ2 (Fig.2C). Shorter exposures confirmed that these particles were also devoid of the ς1 protein (data not shown). Parallel TGU gels which were immunoprobed with the anti-λ2 monoclonal antibody 7F4 (93) confirmed the presence of λ2 in intact core samples and the absence of λ2 in the spike-less particle fractions (data not shown). No fractions which contained structures deficient for the λ2 protein were identified electron microscopically or by electrophoretic protein profile in the permissive cultures of either T3D or tsA279. To measure the copy numbers of various proteins present in the core-like particles, laser densitometric scans of fluorograms representing multiple samples were analyzed. The number of cysteines and methionines contributed by each structural protein to whole virions and core particles was determined by multiplying the copy number of the protein in the structure by the number of cysteines and methionines present in one copy of the relevant protein. The copy number of residues contributed by each protein was then standardized to the proportion of residues contributed by the major core protein λ1 (Table 3). The signal for λ1 was determined for each fraction and used to determine the predicted signal for the remaining structural proteins in that fraction. The actual signal for each protein was then expressed as a proportion of the predicted signal (Table 4).

When these densitometric calculations were applied to T3D andtsA279 virions which had been permissively assembled (cultured and assembled under permissive growth conditions), there was no significant difference between the wild type and the mutant. Restrictively assembled T3D and tsA279 virions also contained essentially the same number of copies of all structural proteins as permissively assembled virions (Table 4). However, there were significant differences in the relative amounts of core proteins λ2 and μ2 in the restrictive mutant core-like particle fractions; these structures contained approximately 2% of the expected amount of λ2, representing less than two copies per particle (Table 4). These fractions also contained outer-capsid proteins μ1/μ1c and ς3, which we assume to have been contributed by the contaminating outer-shell structures.

Gel analyses also indicated the minor protein μ2 produced by the mutant tsA279 at restrictive temperatures was either absent from the λ2-deficient core particles or migrated differently from μ2 produced at the permissive temperature (Fig. 2C, arrow). Laser densitometric scans indicated there were approximately normal amounts of the minor core protein λ3 but that there were about 5% the expected levels of μ2 (less than one copy per particle) in the core-like particles produced by the mutant at the nonpermissive temperature. Western blot evaluation of parallel TGU gels with polyclonal anti-μ2 antiserum confirmed that the core-like particles were deficient for this protein (Fig. 2D).

To determine whether the spike-deficient particles may represent normal assembly intermediates or “dead-end” products, we next determined the capacity of restrictively assembled spike-less core particles to accumulate spikes after temperature downshift experiments. After temperature downshift from 40 to 32°C in the presence of cycloheximide, the distribution of particles produced by the mutant was changed, with the proportion of core-like particles produced at a restrictive temperature being reduced from over 22 to below 2%, and the proportion of core particles which contained λ2 spikes increased from below 1 to approximately 20% (Table5). There was little difference between the proportion of other species of particles produced after temperature downshift. The proportion of various particles that accumulated in the presence of cycloheximide after 6 h in each of the cultures maintained at a constant temperature did not differ dramatically from the proportion of particles produced in comparable non-cycloheximide-treated cultures (Tables 1 and 5).

A mutant L2 gene (λ2 protein) is associated with the accumulation of spike-deficient core particles.The assembly pathways of simple helical plant viruses have been extrapolated from disassembly and reassembly studies of tobacco mosaic virus (15, 17, 34, 92). While it has been possible to perform disassembly studies with complete reovirus virions (68) and to assemble capsids of this (102) and related viruses from expressed proteins (45,91), sequential disassembly and reassembly has been marginally successful when applied to this (5) or other moderately complicated animal virus systems. Bacteriophage conditionally lethal mutants have been successfully used to derive assembly pathways of structurally complicated bacteriophage (29, 30). That success suggested that the Fields panel of reovirus conditionally lethal ts mutants (19, 32, 79) (reviewed in reference 21) could be used to determine the assembly pathways of a eucaryotic virus.

Direct particle counting of cytoplasmic extracts from restrictivetsA279-infected cultures confirmed the presence of and provided further insight into the nature of the core-like particles observed by thin-section electron microscopy (44). First, the proportion of core particles produced by the mutant under restrictive conditions was significantly greater than the proportion of cores produced under any of the permissive conditions (Table 1). Second, permissively assembled core particles observed in cultures of both T3D and tsA279 and restrictively assembled core particles from T3D demonstrated the λ2 apical spike and contained genomic material. Third, the core particles produced at restrictive temperature were deficient for both the λ2 and μ2 proteins and contained reduced levels of genomic material. Finally, when culture temperatures were shifted from 40 to 32°C during culture and maintained in the presence of cycloheximide to ensure the unavailability of permissibly transcribed viral protein, the proportion of spike-defective core-like particles declined, while the proportion of particles demonstrating spikes increased. This is the first report that core-like particles deficient for core proteins λ2 and μ2 (Table 4) may accumulate in the assembly pathway. Analysis of a panel of T1L×tsA279 intertypic reassortants indicated the production of the λ2 spike-deficient core particles was associated with the mutant L2 gene segment (Table 2).

The apparent inability to isolate comparable spike-less particles from wild-type infected cell extracts, coupled with the ability to readily disassemble virions and create spiked core particles (22, 28, 59,70), has led to the assumption that the core particles usually seen in infected cells by thin-section electron microscopy are complete core particles. This premise is also supported by the accumulation of core particles which contain the λ2 pentameric spike in restrictive cultures from representatives of recombination groups B (65) and G (23, 65, 88), which code for the core spike protein λ2 and major outer capsid protein ς3, respectively (21,67). Spike-less core particles have been prepared in laboratory settings from top component and whole virus (90, 96, 106). The core particles generated by White and Zweerink demonstrated apical cavities after the removal of the λ2 pentameric structure by high-pH treatment (96), as do those produced by treatment at high temperature (106). The core particles produced bytsA279 also demonstrated such cavities (Fig. 1A and 2B2, small arrow). Similarities in particle morphology of the spike-deficient core particle produced by tsA279 and the spike-deficient core particles produced in vitro (96, 106), coupled with the protein profiles of previously prepared spikeless cores (96, 106), suggested that all other core proteins were present. The unexpected observation that the μ2 protein was also not present in the core-like particles identified in this study provides further insight into the assembly pathway of reovirus.

Assembly pathway of reovirus.The accumulation of λ2 spike-deficient cores in the restrictively grown tsA279cultures implies that the assembly pathway is blocked as a result of the altered λ2 product. The identification of this structure in vivo suggests that initiation of capsid assembly may involve interactions between the core proteins λ1, λ3, and ς2 to form a “primary core particle” (Fig. 3B). This notion is supported by coexpression studies that show that λ1 and ς2 together, but not alone, are minimally required for core capsid assembly, and that coexpression of these two major proteins with any of the other core proteins resulted in incorporation of each protein into a genome-deficient core-like particle (102). The inclusion of λ3 in the primary core identified in the current study implies that this minor protein is incorporated into the structure at an early point in assembly, while the concomitant lack of both λ2 and μ2 in the primary core suggests that neither protein is necessary for the initiation of assembly.

• Open in new tab
• Download powerpoint

Fig. 3.

The proposed model for the assembly of reovirus. (A) Assembly of the dodecahedral base unit. Five dimers of the core protein λ1, ten units of ς2, and one copy of λ3 associate to form the dodecahedral base structure. (B) Assembly of the primary core particle. Twelve dodecahedral base units assemble to form the primary capsid structure. (C) Assembly of the intermediate particle. Two copies of μ2 associate with each apex of the primary core particles immediately prior to or collateral with assembly of the twelve λ2 pentameric spikes. The trimeric ς1 attachment protein is most likely coassembled with the λ2 spike, with the amino termini of the ς1 fibers interacting with the carboxy termini of the λ2 spike proteins. (D) The outer capsid proteins ς3 and μ1 associate with each other and then condense around the λ2 pentameric spike to form the complete virion.

Previous reports have identified l3, m3, s3, and s4 as the first viral mRNAs present in reovirus-infected cultures (52). The encoded proteins for three of these—μNS, ςNS, and ς3—are the first viral proteins to associate with viral single-stranded RNA (ssRNA) to form ssRNA-containing complexes (ssRCCs) in restrictive cultures of T3D and the dsRNA-negative tsmutants tsC447, tsD357, and tsE320, which are defective for ς2, λ3, and ςNS, respectively (3). Small quantities of λ1 (encoded by L3) and λ2 were the next constituents identified in the ssRCC complexes. The major core protein ς2 was not identified in these complexes from T3D or in the different ts mutants examined. These complexes were not examined for the presence of λ3 nor μ2 (3). The next structures identified in that study were dsRNA-containing complexes (dsRCCs), which were comprised of the proteins μNS, ςNS, ς3, and λ2. Coprecipitation studies with antibodies directed against the major core proteins ς2 and λ1 also precipitated dsRCCs from restrictive infections with T3D (3). The results from that study are consistent with previous reports of an ssRNA transcriptase particle comprised of the proteins λ1, ς2, μNS, and reduced quantities of λ2 (66, 107, 108). Review of the data in those reports does not preclude the presence of μ2 or λ3, which comigrate with other proteins in the gel system used. While the primary core particle identified in the present study differs from the dsRCCs reported with other ts mutants (3) in that it lacks the nonstructural proteins, it bears some similarity to the previously identified transcriptase particle (66). This 40-nm particle has been reported to contain all core structural proteins plus the nonstructural protein μ0 (now called μNS) (66).

These studies are consistent with a dodecahedral assembly model based upon a five-sided apical complex which contains 1 central copy of the minor core protein λ3, 5 dimers of the major core protein λ1 and 10 copies of ς2, in which λ3 interacts with the λ1 amino termini (27, 68) to form the dodecahedral base unit (Fig. 3A) from which the primary core particle is subsequently formed. Nonstructural proteins would be excluded as base units assemble, resulting in the primary core particle identified in the current studies (Fig. 3B). Dodecahedral structures constructed from penton base units have been identified in adenovirus type 3 infections (73). Coexpression of the adenovirus 3 penton base and fiber proteins also produce dodecahedral particles comprised of 12 penton and fiber base units (31, 83). Since each multimeric penton base unit is comprised of 5 penton protein monomers, these particles are made up of 60 penton monomers, plus 12 trimeric apical fibers, making them analogous to miniature capsids (83), with a triangulation lattice T=1 (16, 50). Dodecahedral models have also been proposed for the assembly of small capsids with triangulation lattice numbers equal to 1, such as the picornaviruses (reviewed in reference2), bacteriophage φ6 (14), and the yeast virus L-A (18). Bacteriophage φ6 and yeast virus L-A are other dsRNA viruses constructed from 120 copies of a major capsid protein (14, 18). The core particle of reovirus shares similar copy numbers of major core proteins and an apparent triangulation lattice value of 1 (20, 68).

Subsequent assembly of the complete core particle is proposed to involve association of the minor core protein μ2 with the apices of the primary core particle either immediately prior to or collaterally with λ2 (Fig. 3C). Electron cryomicroscope reconstructions of T1L top-component cores further suggest that λ2 and μ2 interact with λ3 at the apices of the primary core particle (27). While the precise location of the two minor proteins in the apical complex have not been established, all three proteins have been associated with functions of the RNA-dependent RNA polymerase either separately or in different combinations (72, 87, 106). The finding that expressed λ2 exists as a monomer (61) suggests that λ2 condenses into pentamers during association with the primary core particle. Mutation(s) in the λ2 protein could affect the appearance of the primary core particle in two ways: by either preventing pentamerization of the protein or blocking the interactions of the protein and/or its pentamer with the core to produce the complete core particle. In addition, the accumulation of outer-shell structures comprised of proteins μ1 and ς3, but clearly lacking proteins λ2 and ς1 (Fig. 2), suggests that the mutant λ2 protein or its pentamer is not able to interact with the outer capsid proteins when assembled at restrictive temperatures, in contrast to wild-type λ2 protein, which can interact with the outer capsid proteins (28,62). The observation that temperature shiftdown experiments (Table 5) lead to condensation of spikes onto the previously spike-less restrictive particles is consistent with the proposed condensation of these proteins to form the complete core particle. In the proposed dodecahedral assembly path for the reovirus primary core particle, the replication of the minus sense viral RNA to complete the genome at this stage could also result in a maturational event similar to that reported for φ6 (14), where the dodecahedral structure of the base unit would be displaced radially from the pentameric faces during assembly to provide a triangulation lattice T=1 icosahedral structure.

The precise nature of the interactions of ς1 with the λ2 spike is not clear (27, 68). This protein was not examined during previous coexpression and assembly studies. However, incorporation of the trimeric ς1 attachment fiber with the λ2 pentamer most likely occurs at the time the λ2 spike condenses and associates with the primary core, with the amino terminus of the ς1 fibers interacting with the carboxy terminus of the λ2 proteins (56, 60), thereby forming an intermediate particle.

The final stage of reovirus capsid assembly involves condensation of the outer capsid around the core lattice (Fig. 3D). Previous reports had identified the presence of outer-shell structures (L top component) in restrictive cultures of tsC447 that contained λ2, μ1, ς1, and ς3 (62). Electron cryomicroscope examination of whole virions and ISVPs have demonstrated interactions between λ2 and both μ1 and ς3. The same study showed little contact between μ1 and ς3 and the core shell (28). Those reports suggested that the λ2 complex may serve as a nucleation site for condensation of μ1 into the outer capsid lattice (28). However, the presence of μ1/ς3 outer-shell complexes that clearly lack both λ2 and ς1 in tsA279 restrictive cultures indicates that these proteins are not required for condensation of an outer-capsid-like structure. Coprecipitation studies indicated that the outer capsid proteins μ1 and ς3 form complexes in permissive cultures of T3D (55) and tsG453 (88). The inability to coprecipitate these outer capsid proteins in restrictive cultures oftsG453, combined with the observation that the restrictive product of tsG453 infections is a core-like particle (23, 88) rather than an intermediate subviral particle, indicate that the outer capsid proteins μ1 and ς3 must interact with each other before they can condense to form the outer capsid (88).

In conclusion, investigations into the intermediate assembly structures that accumulate in restrictive cultures of tsA279, a member of the Fields panel of reovirus ts mutants, have identified a novel primary core capsid comprised of λ1, λ3, and ς2. This structure may represent the transitional stage between the RNA-containing complexes described by Antczak and Joklik (3) and the λ2 spike-containing core structure identified in restrictive cultures of tsB352 (65) and tsG453(23, 65, 88). We propose that these complexes combine through a dodecahedral mechanism of assembly to form the primary core capsid as the first capsid structure in the reovirus assembly pathway. The remaining core proteins λ2 and μ2 then combine with the primary core capsid, along with the ς1 trimeric attachment complex, to form the more complete intermediate particle. Finally, the outer capsid proteins μ1 and ς3 interact and coprecipitate around this intermediate particle to form the complete reovirus capsid.