A Single Mutation in the Carboxy Terminus of Reovirus Outer-Capsid Protein σ3 Confers Enhanced Kinetics of σ3 Proteolysis, Resistance to Inhibitors of Viral Disassembly, and Alterations in σ3 Structure

Abstract

Mammalian reoviruses undergo acid-dependent proteolytic disassembly within endosomes, resulting in formation of infectious subvirion particles (ISVPs). ISVPs are obligate intermediates in reovirus disassembly that mediate viral penetration into the cytoplasm. The initial biochemical event in the reovirus disassembly pathway is the proteolysis of viral outer-capsid protein σ3. Mutant reoviruses selected during persistent infection of murine L929 cells (PI viruses) demonstrate enhanced kinetics of viral disassembly and resistance to inhibitors of endocytic acidification and proteolysis. To identify sequences in σ3 that modulate acid-dependent and protease-dependent steps in reovirus disassembly, the σ3 proteins of wild-type strain type 3 Dearing; PI viruses L/C, PI 2A1, and PI 3-1; and four novel mutant σ3 proteins were expressed in insect cells and used to recoat ISVPs. Treatment of recoated ISVPs (rISVPs) with either of the endocytic proteases cathepsin L or cathepsin D demonstrated that an isolated tyrosine-to-histidine mutation at amino acid 354 (Y354H) enhanced σ3 proteolysis during viral disassembly. Yields of rISVPs containing Y354H in σ3 were substantially greater than those of rISVPs lacking this mutation after growth in cells treated with either acidification inhibitor ammonium chloride or cysteine protease inhibitor E64. Image reconstructions of electron micrographs of virus particles containing wild-type or mutant σ3 proteins revealed structural alterations in σ3 that correlate with the Y354H mutation. These results indicate that a single mutation in σ3 protein alters its susceptibility to proteolysis and provide a structural framework to understand mechanisms of σ3 cleavage during reovirus disassembly.

Many viruses use receptor-mediated endocytosis to enter host cells (reviewed in reference 32). For some viruses, exposure to the acidified environment in the endocytic compartment or proteolysis by proteases contained in these organelles facilitates structural alterations in viral capsid components required for delivery of the virus into the cytoplasm. Reoviruses are nonenveloped, double-shelled particles that contain a genome of 10 discontiguous segments of double-stranded RNA. Following attachment to cell-surface receptors, reoviruses are internalized into cells by receptor-mediated endocytosis (5, 6, 43, 46). In the endocytic pathway, the viral outer capsid is removed by acid-dependent proteases through an ordered series of proteolytic events resulting in the generation of infectious subvirion particles (ISVPs) (2, 6, 9, 44, 46). ISVPs are obligate intermediates in reovirus disassembly that mediate penetration of endosomal membranes, leading to delivery of the viral core into the cytoplasm (5, 21, 22, 31, 47). The earliest detectable proteolytic event in reovirus disassembly is the degradation of σ3 protein (6, 9, 44, 46), which likely triggers the cascade of disassembly steps that culminate in endosome penetration. Thus, the σ3 protein plays a central role in reovirus entry into cells.

Reoviruses are capable of establishing persistent infections of many types of cultured cells (reviewed in reference 12). These cultures are maintained by horizontal transmission of virus from cell to cell and therefore are more accurately termed “carrier cultures.” During maintenance of persistent reovirus infections of murine L929 (L) cells, mutations that affect viral disassembly are selected in cells and viruses (13). Mutant cells do not support acid-dependent proteolysis of the viral outer capsid (13) and do not express the enzymatically active form of the endocytic cysteine protease, cathepsin L (3). In contrast to wild-type viruses, viruses isolated from persistently infected cultures (PI viruses) demonstrate enhanced kinetics of viral disassembly (53) and can grow in cells treated with either ammonium chloride (13, 53), an inhibitor of vacuolar acidification (34, 41), or E64 (2), an inhibitor of cysteine proteases (4). These findings suggest that PI viruses have altered requirements for acidification and proteolysis to complete disassembly.

Analysis of wild-type × PI reassortant viruses indicates that growth of PI virus L/C in cells treated with ammonium chloride segregates with the S1 gene, whereas growth of PI 2A1 and PI 3-1 in ammonium chloride-treated cells segregates with the S4 gene (53). These results suggest that there are at least two acid-dependent disassembly events during conversion of virions to ISVPs: one involving viral attachment protein σ1, which is encoded by the S1 gene (30, 51), and another involving outer-capsid protein σ3, which is encoded by the S4 gene (35, 37). Growth of L/C, PI 2A1, and PI 3-1 in cells treated with E64 segregates exclusively with the S4 gene (2), suggesting that σ3 alone is the primary determinant of susceptibility of the viral outer capsid to proteolysis during viral disassembly. The deduced amino acid sequences of σ3 proteins of L/C, PI 2A1, and PI 3-1 contain from one to three mutations, including a common tyrosine-to-histidine mutation at amino acid 354 (Y354H) (53). In the case of PI 3-1 σ3 protein, Y354H is the only mutation observed, which provides strong evidence that Y354H influences protease susceptibility of σ3. However, it has not been possible to compare isogenic virus strains with the single σ3 polymorphism at amino acid 354 for cell entry and disassembly. Moreover, potential contributions of the other mutations in PI virus σ3 proteins in determining susceptibility to inhibitors of acidification and proteolysis have not been defined.

In this study, we used ISVPs recoated with mutant σ3 proteins containing mutations in PI virus σ3 proteins alone and in combination to precisely identify sequences in σ3 that modulate acid-dependent and protease-dependent steps in reovirus disassembly. This technique facilitates generation of particles differing by site-specific mutations for studies of steps in virus-cell interaction during a single cycle of replication (24). PI virions and recoated ISVPs (rISVPs) were tested for susceptibility to growth inhibition mediated by ammonium chloride and E64 and analyzed by cryo-electron microscopy (cryo-EM) and three-dimensional image analysis. The results demonstrate that Y354H enhances the kinetics of reovirus disassembly, confers resistance to inhibitors of acid-dependent proteolysis in cellular endosomes, and leads to alterations in σ3 structure.

MATERIALS AND METHODS

Cells and viruses.

L cells were grown in either suspension or monolayer cultures in Joklik's modified Eagle's minimal essential medium (Irvine Scientific, Santa Ana, Calif.) supplemented to contain 5% fetal bovine serum (Gibco BRL, Grand Island, N.Y.), 2 mM l-glutamine, and 100 U of penicillin per ml, 100 μg of streptomycin per ml, and 0.25 μg of amphotericin per ml (Sigma-Aldrich, St. Louis, Mo.). Spodoptera frugiperda (Sf21) cells (Clontech, Palo Alto, Calif.) were grown in Grace's insect cell medium (Gibco BRL) supplemented to contain 10% fetal bovine serum, 2 mM l-glutamine, 50 U of penicillin per ml, and 50 μg of streptomycin per ml.

Reovirus strains type 1 Lang (T1L) and type 3 Dearing (T3D) are laboratory stocks. PI viruses L/C (1), PI 2A1, and PI 3-1 (13) were isolated as previously described. Purified preparations of reovirus virions were made by using second-passage stocks as previously described (20). Baculovirus vector strains were derived from Autographa californica nuclear polyhedrosis virus (AcMNPV). Recombinant baculoviruses containing wild-type and mutant S4 gene cDNAs were generated by introduction of cDNAs into pBacPAK8 (Clontech), followed by lipofection-mediated cotransfer of plasmid recombinants and linearized BacPAK6 AcMNPV DNA (Clontech), respectively, into Sf21 cells according to the manufacturer's instructions. Recombinant viruses were isolated by plaque purification on monolayers of Sf21 cells and amplified by three passages in Sf21 cells.

Cloning and mutagenesis of S4 gene cDNAs.

The σ3-encoding S4 gene cDNAs of strains T3D, L/C, PI 2A1, and PI 3-1 were generated by reverse transcription-PCR (27) and introduced into the pCR2.1 vector (Invitrogen, San Diego, Calif.). S4 genes were amplified with primers specific for the noncoding regions of the T3D S4 gene (53).

Site-directed mutations of the T3D S4 gene cDNA in pCR2.1 were generated by PCR-based oligonucleotide mutagenesis (Stratagene, La Jolla, Calif.) according to the manufacturer's instructions. The following primer sets were used for mutagenesis (nucleotides different from those in the T3D S4 sequence are underlined): mutation A sense (5′-GGTCGTGTATCAATCTGCAGCGCGCAAGAGGG-3′) and mutation A antisense (5′-CCCTCTTGCGCGCTGCAGATTGATACACGACC-3′) and mutation B sense (5′-GAGGTTCACTAGTGAAGCTGAACCGGCTTCAG-3′) and mutation B antisense (5′-CTGAAGCCGGTTCAGCTTCACTAGTGAACCTC-3′). Mutant A-B was constructed from mutant B by insertion of mutation A with the mutation A primer set described above. Mutant A-C was constructed with XbaI and SpeI to excise the 5′ 437 nucleotides, including 24 nucleotides of polylinker sequence, of the L/C S4 gene cDNA in pCR2.1. These sequences were ligated into the BacPAK8 transfer vector containing the S4 gene cDNA of PI 3-1. Nucleotide sequences of the σ3-encoding regions of all S4 gene cDNAs in baculovirus transfer vectors were confirmed by using an ABI model 377 automated sequencer (PE-Applied Biosystems, Norwalk, Conn.). Error-free S4 gene cDNAs were used to construct recombinant σ3-expressing baculoviruses.

Expression of recombinant σ3 proteins.

Third-passage recombinant baculoviruses containing wild-type or mutant S4 gene cDNAs were used to infect Sf21 insect cells (5 × 10⁶) at a multiplicity of infection (MOI) of 2 PFU per cell. After incubation at 25°C for 72 h, cells were washed once with phosphate-buffered saline (PBS) and incubated on ice in 1 ml of cytoplasmic extraction buffer (10 mM HEPES [pH 7.9], 10 mM KCl, 1.5 mM MgCl₂, 0.5 mM dithiothreitol, 0.5 mM phenylmethysulfonyl fluoride) containing an EDTA-free protease inhibitor cocktail (Roche, Indianapolis, Ind.). After 30 min of incubation, 25 μl of 10% Igepal CA-630 (Sigma-Aldrich) was added, and the mixture was thoroughly vortexed. Cell nuclei and membranes were collected by centrifugation at 500 × g for 5 min. The supernatant was removed, and the cell pellet was incubated on ice in 1 ml of nuclear extraction buffer (20 mM HEPES [pH 7.9], 0.42 M NaCl, 1.5 mM MgCl₂, 25% glycerol, 0.2 mM EDTA) for 30 min. Cellular debris was collected by centrifugation at 22,000 × g for 10 min. The supernatant was removed, analyzed by electrophoresis, and used for recoating experiments.

SDS-PAGE of reovirus structural proteins.

Discontinuous sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was performed as previously described (28). Viral particles were solubilized by incubation in sample buffer (125 mM Tris, 2% 2-mercaptoethanol, 1% SDS, 0.01% bromophenol blue) at 65°C for 5 min. Samples were loaded into wells of 10% polyacrylamide gels and electrophoresed at a constant voltage of 200 V for 1 h. Following electrophoresis, gels were stained with Coomassie blue R-250 (Sigma-Aldrich) and dried between cellophane.

Recoating ISVPs with recombinant σ3 proteins.

ISVPs of strain T1L were prepared by treating 5 × 10¹² purified virions in virion storage buffer (150 mM NaCl, 10 mM MgCl₂, 10 mM Tris [pH 7.4]) with 200 μg of Nα-p-tosyl-l-lysine chloromethyl ketone-treated bovine α-chymotrypsin (Sigma-Aldrich) at 37°C for 60 min. T1L ISVPs were purified by cesium chloride (CsCl) density centrifugation as previously described (20) and dialyzed against virion storage buffer. Recombinant σ3 protein in nuclear extracts of baculovirus-infected Sf21 cells (5 × 10⁶) was mixed with T1L ISVPs (5 × 10¹² particles) in a volume of 1 ml and incubated at 25°C for 30 to 60 min. rISVPs were purified by CsCl density centrifugation and dialyzed against virion storage buffer. Infectious titers of rISVP preparations were determined by plaque assay with L-cell monolayers (49).

Treatment of reovirus virions and rISVPS with cathepsin L and cathepsin D.

Purified reovirus virions and rISVPS at a concentration of 4 × 10¹⁰ particles per ml in reaction buffer L (100 mM NaCl, 15 mM MgCl₂, 50 mM sodium acetate [pH 5.0]) were treated with 100 μg of purified, recombinant human cathepsin L (7) per ml in the presence of 5 mM dithiothreitol at 37°C for 0 to 16 h. Virions and rISVPS at a concentration of 4 × 10¹⁰ particles per ml in reaction buffer D (5 mM MgCl₂, 10 mM cysteine, 100 mM potassium acetate [pH 3.8]) were treated with 100 μg of purified, bovine cathepsin D (Sigma-Aldrich) per ml at 37°C for 0 to 6 h. Protease treatment was terminated by adding either 500 μM E64 (Sigma-Aldrich) to cathepsin L reactions or 100 mg of pepstatin A (Sigma-Aldrich) per ml to cathepsin D reactions and incubating on ice for 5 min. Reaction mixtures were analyzed by SDS-PAGE.

Densitometric analysis of reovirus structural proteins.

Dried, Coomassie-stained gels were scanned with Adobe Photoshop 5.0 (Adobe Systems, San Jose, Calif.). Band densities in the scanned gels were quantitated with the program Scion Image Beta 4 (Scion Corporation, Fredrick, Md.). σ3 stoichiometry on recoated particles was determined by comparing the densities of bands corresponding to σ3 to those of the δ protein of ISVPs, which is a virion-associated cleavage fragment of μ1C protein. Mean band densities corresponding to σ3 were divided by those corresponding to μ1C (virions) or δ (rISVPs) proteins to determine stoichiometry. In studies of virion and rISVP proteolysis, mean densities were determined for bands corresponding to the λ and σ3 proteins for each interval of protease treatment. Densities of bands corresponding to σ3 were divided by those corresponding to the λ proteins as a control for loading.

Growth of wild-type and PI virions, ISVPs, and rISVPs in the presence and absence of inhibitors of viral disassembly.

Monolayers of L cells (5 × 10⁵) in 24-well plates (Corning-Costar, Corning, N.Y.) were preincubated at 37°C for 4 h in untreated medium or for 4 h in medium supplemented to contain either 10 mM ammonium chloride or 100 μM E64. The medium was removed, and cells were adsorbed with virus particles at an MOI of 2 PFU per cell. After incubation at 25°C for 1 h, the inoculum was removed, cells were washed with PBS, and 1 ml of fresh, untreated medium or medium supplemented with either ammonium chloride or E64 was added. After incubation at 37°C for 24 h, cells were frozen and thawed twice, and viral titers in cell lysates were determined by plaque assay (49). Independent experiments were performed in duplicate.

Cryo-EM.

Reovirus particles were prepared for cryo-EM by previously described techniques (17, 38). A 4-μl aliquot of each specimen (virions or rISVPs) was applied to one side of a holey carbon grid. The grid was then blotted and plunged into a bath of liquid ethane (−180°C). The frozen-hydrated sample was transferred to a precooled GATAN cryoholder (GATAN, Inc., Pleasanton, Calif.) and imaged with a JEOL 1200 transmission electron microscope (JEOL, Inc., Peabody, Mass.) operated at 100 kV and maintained at a specimen temperature of −163°C. Regions of interest were imaged at a magnification of ×30,000 with an electron dose of 5 electrons/Ų. From each region, a focal pair was recorded with intended defocus values of 1 and 2 μm. The electron images were recorded with a 1-s exposure on Kodak SO-163 film (Kodak, Rochester, N.Y.). Film was developed in Kodak D-19 developer at 21°C for 12 min and fixed in Kodak fixer at 21°C for 10 min.

Three-dimensional image reconstructions.

Micrographs were selected based on particle concentration, quality of ice, and appropriate defocus conditions. Images were digitized on a Zeiss SCAI microdensitometer (Carl Zeiss, Inc., Englewood, Colo.) with a 7-μm step size. Pixels were averaged to give a 14-μm step size that corresponded to 4.67 Å per pixel in the object. Particles were boxed with an area of 256 by 256 pixels. Determination of orientational parameters, refinement of these parameters, and three-dimensional reconstructions were performed by using the ICOS Toolkit software suite (29). The orientations of the particles were determined by the common lines approach (11) and refined by the cross-common lines method (19). Three-dimensional image reconstructions from a set of particles that adequately represented the icosahedral asymmetric unit were computed by cylindrical expansion methods (11). The further-from-focus micrograph in each focal pair was processed first to obtain a low-resolution reconstruction. This reconstruction then was used to determine the correct orientations of particles imaged in the corresponding closer-to-focus micrograph.

Image reconstructions were computed to a resolution within the first zero of the contrast transfer function (CTF) of the corresponding micrograph. Defocus values were determined from CTF ring positions in the sum-of-particle Fourier transforms. Defocus values of various specimens in the closer-to-focus micrographs ranged from 1.2 to 1.4 μm. Image reconstructions were corrected for the effects of the CTF by previously described procedures (55). The final resolution for each reconstruction was determined by Fourier ring correlation analysis (48). Image reconstructions were computed to an ∼24-Å resolution, which was the lowest resolution among the various specimens, for comparative analysis. Contour levels in each reconstruction were chosen to represent equal volumes between radii at ∼305 and ∼425 Å. Reconstructions of various specimens were radially scaled to match the radial extension of the innermost layer (λ1 layer) with an assumption that mutations in σ3 are unlikely to alter the radial disposition of this capsid layer. Reconstructions were viewed on a Silicon Graphics Workstation (SGI, Mountain View, Calif.) with IRIS Explorer v3.5 software (Numerical Algorithms Group, Inc., Oxford, United Kingdom).

Fitting of X-ray structure into the cryo-EM reconstruction.

The X-ray coordinates of σ3 (42) were initially rigid-body fitted into its portion of the cryo-EM density map by visual inspection with the graphical program O (25). Subsequent refinements were performed with the Situs program package, a rigid-body correlation fitting procedure that allows the fitting of X-ray structures into low-resolution cryo-EM density maps (54).

RESULTS

Expression of recombinant σ3 protein.

The deduced amino acid sequences of PI virus σ3 proteins contain from one to three mutations (Table 1), only one of which, Y354H, has been consistently linked to viral growth in the presence of inhibitors of reovirus disassembly (2, 53). To determine whether Y354H is sufficient to confer resistance to ammonium chloride and E64 and to assess possible contributions of other mutations selected in σ3 during persistent infection, we engineered recombinant baculoviruses to express wild-type T3D σ3 protein and σ3 proteins containing mutations observed in PIviruses. We also generated four additional recombinant baculoviruses by using either site-directed mutagenesis or exchange of PI virus S4 gene cDNA sequences to express novel σ3 mutants (Table 1). Recombinant baculoviruses were used to infect insect cells, and nuclear extracts were prepared. Asdemonstrated by SDS-PAGE (Fig. 1A), the major protein band in nuclear extracts migrated with the electrophoretic mobility of native σ3. Thus, insect cells infected with recombinant baculoviruses produce σ3 protein, as reported previously (24).

TABLE 1.

Mutations in the deduced amino acid sequences of σ3 proteins used to define sequences that modulate susceptibility of σ3 to proteolysis

σ3 protein	Location and nature of mutation(s)
Mutant A	26, Y→C
Mutant B	130, E→K
PI 3-1	354, Y→H
Mutant A-B	26, Y→C; 130, E→K
Mutant A-C	26, Y→C; 354, Y→H
PI 2A1	130, E→K; 354, Y→H
L/C	26, Y→C; 130, E→K; 354, Y→H

FIG. 1.

(A) Expression of σ3 protein in insect cells with recombinant baculoviruses. The expressed σ3 proteins of wild-type T3D, three PI viruses, and four site-directed σ3 mutants are shown. A band corresponding to expressed σ3 proteins is labeled, and T3D virions are included as a control. An empty vector control was loaded in the lane labeled V, and molecular mass markers (in kilodaltons) were loaded in the lane labeled M. (B) ISVPs recoated with wild-type and mutant σ3 proteins. ISVPs of reovirus strain T1L were incubated with insect cell nuclear extracts containing the σ3 proteins shown to generate rISVPs. Equal numbers of rISVPs were analyzed by SDS-PAGE with T1L virions included as controls. Viral proteins are labeled. The ratio of σ3 band density to that of μ1C protein (virions) or δ protein (rISVPs) is indicated at the bottom of the gel.

Recoating ISVPs with wild-type and mutant σ3 proteins.

To generate rISVPS with recombinant σ3 proteins, T1L ISVPs were incubated with the expressed σ3 protein isolated from Sf21 cells. Although T3D would have been an appropriate strain for σ3 recoating studies, it was not selected due to the reduced infectivity of T3D ISVPs, resulting from cleavage of the T3D σ1 protein during virion-to-ISVP conversion in vitro with chymotrypsin (10, 39). After incubation, rISVPs were isolated by CsCl gradient centrifugation and analyzed by SDS-PAGE (Fig. 1B). Each of the recombinant σ3 proteins was capable of recoating T1L ISVPs, resulting in the generation of rT3D, r3-1, rL/C, r2A1, rA, rB, rA-B, and rA-C. To assess the stoichiometry of σ3 protein on these recoated particles, the densities of bands corresponding to σ3 were compared to those of the δ protein of ISVPs, which is a virion-associated cleavage fragment of μ1C protein. For each rISVP species, the σ3/δ ratio approximated the σ3/μ1C ratio determined for virions. These results demonstrate that expressed σ3 proteins recoat ISVPs from a different reovirus strain with native stoichiometry.

Treatment of virions and rISVPs with endocytic proteases.

To identify mutations in σ3 protein that influence the susceptibility of σ3 to proteolytic disassembly in vitro, virions of wild-type and PI viruses and rISVPs containing expressed wild-type and mutant σ3 proteins were treated with either of the endocytic proteases cathepsin L or cathepsin D. Cathepsin L is a cysteine protease that is capable of producing ISVPs when used to treat reovirus virions (3, 18, 18a). Cathepsin D is an aspartic protease that is incapable of producing ISVPs (18, 26). Virions and rISVPs were treated with purified, recombinant, human cathepsin L (7) for 0 to 16 h, and proteolysis of viral outer-capsid proteins was monitored by SDS-PAGE (Fig. 2). Treatment of wild-type T3D and PI 3-1 virions with cathepsin L resulted in proteolysis of outer-capsid proteins indicative of ISVP formation with degradation of σ3 and cleavage of μ1C to form δ. In comparison to T3D, proteolysis of PI 3-1 during cathepsin L treatment occurred with substantially faster kinetics, consistent with previous results obtained with the intestinal protease chymotrypsin (53). Treatment of rISVPS with cathepsin L resulted in degradation of σ3 alone, since conversion of μ1C to δ had occurred during ISVP preparation prior to recoating. In comparison to rT3D, proteolysis of rL/C, r2A1, r3-1, and rA-C with cathepsin L occurred with faster kinetics and was similar to that of PI 3-1 virions. For virions and rISVPS, the densities of bands corresponding to σ3 were compared to those of the λ proteins as a control for loading. Analysis of σ3/λ band intensities for PI 3-1 virions and rL/C, r2A1, r3-1, and rA-C particles demonstrated that >90% of σ3 was removed from the viral outer capsid by 2 h (Fig. 2). In sharp contrast, proteolysis of wild-type T3D virions and rT3D, rA, rB, and rA-B particles was incomplete after 16 h (digestion of rA and rB was equivalent to that of rA-B [data not shown]). The σ3 proteins of virions and rISVPS demonstrating faster kinetics of σ3 proteolysis after treatment with cathepsin L contain a common mutation, Y354H.

FIG. 2.

Electrophoretic analysis of viral proteins of wild-type (wt) and PI virions and rISVPs following treatment with cathepsin L. (A to D) Purified virions of the particles shown were treated with human cathepsin L (pH 5.0). Equal numbers of viral particles were analyzed by SDS-PAGE. Molecular mass markers (in kilodaltons) were loaded in lanes labeled M. Positions of viral proteins are indicated. (E and F) For each interval of protease treatment, the densities of bands corresponding to σ3 were divided by those corresponding to the λ proteins as a control for loading. Shown are the mean σ3:λ ratios for two or three independent experiments. Error bars indicate the standard errors of the means.

To determine whether mutations in σ3 protein correlate with altered susceptibility to proteolysis by an endocytic protease incapable of generating ISVPs, virions and rISVPS were treated with purified bovine cathepsin D for 0 to 6 h, and proteolysis of viral outer-capsid proteins was monitored by SDS-PAGE (Fig. 3). Cathepsin D treatment of PI 3-1 virions and rL/C, r2A1, r3-1, and rA-C particles resulted in degradation of σ3 with little proteolysis of other viral proteins, except at the 6-h time point. However, degradation of σ3 during cathepsin D treatment of rT3D, rA, rB, and rA-B particles (digestion of rA and rB was equivalent to that of rA-B [data not shown]) was minimal. Digestion of wild-type T3D virions by cathepsin D was intermediate to that of rT3D and r3-1 particles. These findings were confirmed by densitometric analysis of σ3/λ band intensities, which demonstrated that >90% of σ3 was removed from PI 3-1 virions and rL/C, r2A1, r3-1, and rA-C particles by 1 h (Fig. 3). Similar to our findings with cathepsin L, the σ3 proteins of virions and rISVPS demonstrating specific cleavage of σ3 with cathepsin D contain Y354H. These results indicate that σ3 proteins containing an isolated mutation, Y354H, are more susceptible to cleavage by either cathepsin L, which is capable of mediating reovirus disassembly (3, 18, 18a), or cathepsin D, which is not (18, 26).

FIG. 3.

Electrophoretic analysis of viral proteins of wild-type (wt) and PI virions and rISVPs following treatment with cathepsin D. (A to D) Purified virions of the particles shown were treated with bovine cathepsin D (pH 3.8). Equal numbers of viral particles were analyzed by SDS-PAGE. Molecular mass markers (in kilodaltons) were loaded in lanes labeled M. Positions of viral proteins are indicated. (E and F) For each interval of protease treatment, densities of bands corresponding to σ3 were divided by those corresponding to the λ proteins as a control for loading. Shown are the mean σ3:λ ratios for two or three independent experiments. Error bars indicate the standard errors of the means.

Growth of rISVPS in the presence and absence of inhibitors of viral disassembly.

To identify mutations in σ3 protein that confer viral growth in the presence of inhibitors of viral disassembly, virions and rISVPs were tested for the capacity to grow in the presence of either the acidification inhibitor ammonium chloride or the protease inhibitor E64. L cells were infected with wild-type viruses, T1L or T3D, T1L ISVPs, PI 3-1, or rISVPs generated by using the mutant σ3 proteins shown in Table 1 in the presence or absence of ammonium chloride or E64. After 24 h of growth under each condition, viral titers in cell lysates were determined by plaque assay (Fig. 4). Each of the particles tested produced approximately equivalent yields after 24 h of growth in the absence of inhibitors of disassembly. However, in the presence of ammonium chloride, yields of PI 3-1 virions and r3-1, rL/C, r2A1, and rA-C particles were approximately 10- to 20-fold greater than those of wild-type T1L and T3D virions and rA, rB, and rA-B particles. Virions and rISVPs with greater yields in the presence of ammonium chloride each contain σ3 proteins with Y354H. Similarly, in the presence of E64, yields of PI 3-1 virions and r3-1, rL/C, r2A1, and rA-C particles were approximately 10- to 30-fold greater than those of wild-type T1L and T3D virions and rA, rB, and rA-B particles. These results indicate that σ3 proteins containing Y354H are resistant to inhibitors of acid-dependent and protease-dependent steps that occur during viral disassembly.

FIG. 4.

Growth of reovirus virions, ISVPs, and rISVPs in the presence and absence of inhibitors of viral disassembly. Monolayers of L cells (5 × 10⁵) were preincubated for 4 h in untreated medium or for 4 h in medium supplemented with either 10 mM ammonium chloride or 100 μM E64. The medium was removed, and cells were infected with the particles shown at an MOI of 2 PFU per cell. After 1 h of adsorption, the inoculum was removed, fresh medium with or without either ammonium chloride or E64 was added, and cells were incubated for 24 h. Viral titers in cell lysates were determined by plaque assay. The results are presented as mean viral yields, calculated by dividing the titer at 24 h by the titer at 0 h for each condition, for two to three independent experiments. Error bars indicate the standard error of the means.

Structural analysis of virions of wild-type and PI reoviruses and rISVPs.

To gain insight into the structural basis for the alterations in kinetics of viral disassembly and resistance to disassembly inhibitors associated with σ3-Y354H, we used cryo-EM and three-dimensional image analysis to compare the structures of wild-type and PI virions and rISVPs. Cryo-images of unstained, frozen-hydrated wild-type virions, PI virions, and rISVPs are shown in Fig. 5. These images indicate that PI virions and rISVPs have morphological characteristics that are similar to those of wild-type virions. Reconstructions of native virions and their corresponding rISVPs were computed from their respective cryo-images. These reconstructions revealed that both wild-type and PI virions have a similar structural organization. The structures of wild-type T3D (computed to a 24-Å resolution with 183 particles) and PI virus L/C (computed to a 24-Å resolution with 496 particles) are shown in Fig. 6. In all reconstructions, the fingerlike projections of σ3 protein are organized on an incomplete T=13 icosahedral lattice with six σ3 subunits at the local sixfold axis and an incomplete ring of four σ3 subunits adjacent to the fivefold axis surrounding the λ2 turret. Three-dimensional image reconstructions of rT3D, rL/C, and r3-1 (each computed to a 24-Å resolution with 113, 90, and 121 particle images, respectively) revealed identical structural organizations, both with respect to the σ3 protein and the remainder of the structure (Fig. 6 and data not shown). These results demonstrate that the overall arrangements of σ3 proteins in the outer capsid of wild-type and PI virions and rISVPs are similar and that interactions of σ3 with other structural proteins on rISVPs recapitulate those of native reovirus particles.

FIG. 5.

Electron cryo-micrographs of wild-type and PI virions and rISVPs embedded in a thin layer of vitreous ice. (A) Wild-type T3D, (B) L/C, (C) PI 2A1, (D) PI 3-1, (E) rT3D, (F) rL/C, (G) r2A1, and (H) r3-1. Scale bar in panel A, 1,000 Å.

FIG. 6.

Three-dimensional image reconstructions of wild-type and PI virions and rISVPs. Reconstructions viewed along the icosahedral threefold axis of wild-type T3D (A), L/C (B), rT3D (C), and rL/C (D) are shown. Closeup views of the hexameric rings of σ3 proteins from the reconstructions of recoated particles (E) rT3D and (F) rL/C are shown to indicate that the assembly of σ3 in rISVPs is native.

Analysis of conformational changes in σ3 protein.

To identify structural differences in σ3 proteins of wild-type and PI viruses, we compared three-dimensional image reconstructions of wild-type and PI virions (Fig. 7). In this analysis, we found two significant differences. First, the radial extension of the σ3 protein of L/C virions was less than that of the other PI viruses and wild-type T3D (Fig. 7C), which were equivalent. Second, T3D virions contained a cleft in the σ3 density (Fig. 7A and C) that was filled in reconstructions of PI virions and rISVPs containing σ3-Y354H (Fig. 7B and C and data not shown). To ensure that these features were reproducible, a minimum of two reconstructions with independent sets of particles were generated. Altered σ3 density in the region of the T3D cleft was consistently observed for all PI viruses. These results demonstrate that PI viruses have an alteration in σ3 protein density, which likely indicates a conformational change associated with Y354H.

FIG. 7.

Structural alterations in σ3 protein. (A) A thin slice of density representing two of the hexameric σ3 proteins, represented as a mesh, from two independent wild-type T3D reconstructions (shown in yellow and red) superimposed. An arrow indicates the position of the cleft in σ3 density. The atomic structure of σ3 (in white) is docked into one of the σ3 densities (right). The ribbon diagram of σ3 in the same orientation is shown (inset) with an arrow indicating the location of the cleft. Also shown (inset) are mutations found in PI virus σ3 protein: Y26C (purple), E130K (red), and Y354H (green). (B) Density maps from two independent reconstructions of PI virus L/C are superimposed in cyan and green. (C) Density maps for wild-type T3D (red) and PI virus L/C (cyan) are shown superimposed. The cleft seen in the wild-type T3D reconstruction is absent in L/C.

DISCUSSION

In this study, we show that a single mutation in the reovirus σ3 protein enhances the kinetics of viral disassembly, mediates resistance to inhibitors of endocytic acidification and proteolysis, and produces structural alterations in σ3. Y354H is present in the σ3 proteins of PI viruses L/C, PI 2A1, and PI 3-1, which were isolated from independent persistently infected L-cell cultures (1, 13). A Y354H mutation also is found in E64-adapted variant viruses selected by serial passage of wild-type T3D in L cells treated with E64 (18). By using rISVPs containing individual or combinations of σ3 mutations found in PI viruses, we found that Y26C and E130K do not contribute to altered disassembly phenotypes. Although it is possible that these mutations play other roles during persistent reovirus infection, these effects are not apparent from the experiments reported here. Three-dimensional image analysis of PI viruses indicates that σ3 proteins with Y354H have a common structural alteration. We propose that this mutation alters protease access to a σ3 cleavage site and enhances the susceptibility of σ3 to proteolysis.

Proteolytic removal of σ3 is the initial step in the reovirus disassembly cascade that results in formation of ISVPs. Three lines of evidence support a central role for σ3 cleavage in reovirus disassembly. First, pharmacologic inhibitors of either endocytic acidification (33, 46) or proteolysis (2, 18) block σ3 cleavage, which leads to an arrest of reovirus disassembly and inhibition of the virus replication cycle. Second, PI viruses grow in the presence of these inhibitors (2, 13, 53) and exhibit accelerated proteolysis of σ3 according to in vitro assays of viral disassembly (53). Third, reovirus variants selected specifically for resistance to protease inhibitor E64 have mutations in σ3 protein (18). The findings reported here extend those reported previously and demonstrate that Y354H in an otherwise isogenic background accelerates σ3 cleavage and confers resistance to viral disassembly inhibitors. These results support the conclusion that σ3 plays an essential role in the regulation of reovirus disassembly.

Growth of PI viruses in the presence of ammonium chloride segregates with either the σ1-encoding S1 gene (L/C) or the σ3-encoding S4 gene (PI 2A1 and PI 3-1) (53), whereas growth of PI viruses in the presence of E64 segregates exclusively with the S4 gene (2). Using rISVPs containing wild-type and mutant σ3 proteins, we found that ISVPs recoated with σ3 proteins containing Y354H were capable of growth in the presence of either ammonium chloride or E64. No other mutation in σ3 contributed to these phenotypes. Interestingly, rISVPs containing the σ3 protein of L/C were capable of growth in the presence of ammonium chloride, although this phenotype segregates with the L/C S1 gene (53). It is possible that interactions between σ1 and σ3 in rISVPs differ from those in native virions, which might obviate a requirement for σ1 mutations in conferring resistance to ammonium chloride of rISVPs containing the L/C σ3 protein. In support of this possibility, the σ1 protein undergoes a dramatic conformational change from a retracted to an extended form during ISVP generation (16, 20). Mutations in the L/C σ1 protein that segregate with ammonium chloride resistance might influence this conformational change and confer acid-independent viral disassembly in the absence of mutations in σ3. Such effects might not be operant for rISVPs, in which any conformational changes in σ1 presumably would have occurred prior to recoating. Nonetheless, our results clearly demonstrate that mutations in σ3 alone can confer growth in the presence of ammonium chloride. A precise delineation of the role of σ1 and σ3 mutations in conferring resistance to ammonium chloride will require analysis of core particles recoated with various combinations of wild-type and PI virus outer-capsid proteins (8).

How might a Y354H mutation enhance the susceptibility of σ3 to proteolysis? The crystallographic structure of T3D σ3 reveals a bilobed molecule consisting of a virion-proximal small lobe and a virion-distal large lobe (42). Independent image reconstructions of L/C, PI 2A1, and PI 3-1 demonstrate a bulge in σ3 density in the same region of wild-type T3D that contains a cleft between the two lobes. These findings indicate that Y354H alters the structure of σ3. By virtue of a histidine at residue 354, L/C, PI 2A1, and PI 3-1 σ3 are anticipated to be more sensitive than wild-type σ3 to decreases in pH within endosomes that would ultimately result in alterations in interactions between amino acid side chains in the vicinity of residue 354. We predict that such alterations modulate access to potential σ3 cleavage sites.

Tyr 354 in T3D σ3 lies at the end of an amino acid sequence connecting an α-helix in the virion-proximal small lobe to two C-terminal β-strands in the virion-distal large lobe of the molecule (Fig. 8). Contact amino acids in this network include the carbonyl oxygen of Tyr 354 with the ɛ-amino nitrogen of Lys 234 (42). In addition, the Tyr 354 side chain faces a hydrophobic pocket (Tyr 223, Tyr 225, Leu 228, Tyr 238, Leu 242, Val 243, and Phe 349) in the vicinity of two acidic residues (Asp 224 and Glu 227), the carboxyl side chains of which are directed away from the hydrophobic pocket (Fig. 8B). In virions containing Tyr 354 in σ3, the predominate interactions between the side chains would be hydrophobic, and access to this region by solvent and other molecules would be hindered. However, His 354 in σ3 is likely repulsed from the hydrophobic pocket and possibly forms salt bridges with either Asp 224 or Glu 227 located nearby. Such interactions might alter σ3 structure, resulting in the density changes observed by cryo-EM associated with His 354 in σ3. Since Tyr 354 is among the residues joining the small and large lobes of the protein, conformational alterations in σ3 associated with His 354 might be transmitted to a lower region of the protein. Furthermore, structural changes in σ3 proteins containing His 354 also would allow enhanced access to residues 238 to 250 within the hydrophobic pocket, which form a protease-sensitive region identified in T1L and T3D σ3 (18a, 23). Therefore, we propose that C-terminal residues in σ3 act as a latch to protect σ3 cleavage sites and that His 354 may act as a hinge, which becomes activated for optimal movement as pH decreases from neutral to acidic within endosomes. Thus, His 354 in σ3 may mimic a lower-pH σ3 disassembly intermediate in which the density shift seen in PI viruses corresponds to an open C-terminal latch. To test these predictions, high-resolution structural studies of PI 3-1 σ3 are in progress.

FIG. 8.

Atomic structure of T3D σ3 with mutations. (A) Ribbon diagram of the carbon tracing of the 1.8-Å crystal structure of the T3D σ3 protein (42). Mutations found in PI virus σ3 protein are Y26C (purple), E130K (red), and Y354H (green). The area of increased density found in PI virus σ3 containing His 354 visualized by cryo-EM is marked by an asterisk. The C terminus and N terminus are marked with arrows. (B) The C-terminal inset in panel A (dashed lines) is rotated approximately 90^o to highlight residues that potentially interact with Y354H. In panel B1, a hydrophobic pocket is created by residues Y223, Y225, L228, Y238, L242, V243, and F349. In panel B2, the charged residues are D224, E227, and K234. The C terminus is indicated by the letter “C.”

Nonenveloped viruses must maintain a balance between extracellular stability and the capacity to disassemble following internalization into host cells. Reovirus virions are remarkably resistant to various inactivating agents (14, 15, 50, 52), and σ3 is hypothesized to provide enhanced environmental stability of virions compared to that of ISVPs (40, 45). While mutations in σ3 selected during persistent infection enhance viral disassembly, these mutations actually attenuate viral virulence (36). Therefore, enhanced susceptibility to proteolytic disassembly comes at a cost in the virus-host interaction. The findings reported here identify the molecular basis for susceptibility of σ3 to proteolytic cleavage and lead to an improved understanding of mechanisms by which σ3 regulates reovirus stability and disassembly.