Optimum Length and Flexibility of Reovirus Attachment Protein σ1 Are Required for Efficient Viral Infection

Abstract

Reovirus attachment protein σ1 is an elongated trimer with head-and-tail morphology that engages cell-surface carbohydrate and junctional adhesion molecule A (JAM-A). The σ1 protein is comprised of three domains partitioned by two flexible linkers termed interdomain regions (IDRs). To determine the importance of σ1 length and flexibility at different stages of reovirus infection, we generated viruses with mutant σ1 molecules of altered length and flexibility and tested these viruses for the capacity to bind the cell surface, internalize, uncoat, induce protein synthesis, assemble, and replicate. We reduced the length of the α-helical σ1 tail to engineer mutants L1 and L2 and deleted midpoint and head-proximal σ1 IDRs to generate ΔIDR1 and ΔIDR2 mutant viruses, respectively. Decreasing length or flexibility of σ1 resulted in delayed reovirus infection and reduced viral titers. L1, L2, and ΔIDR1 viruses but not ΔIDR2 virus displayed reduced cell attachment, but altering σ1 length or flexibility did not diminish the efficiency of virion internalization. Replication of ΔIDR2 virus was hindered at a postdisassembly step. Differences between wild-type and σ1 mutant viruses were not attributable to alterations in σ1 folding, as determined by experiments assessing engagement of cell-surface carbohydrate and JAM-A by the length and IDR mutant viruses. However, ΔIDR1 virus harbored substantially less σ1 on the outer capsid. Taken together, these data suggest that σ1 length is required for reovirus binding to cells. In contrast, IDR1 is required for stable σ1 encapsidation, and IDR2 is required for a postuncoating replication step. Thus, the structural architecture of σ1 is required for efficient reovirus infection of host cells.

INTRODUCTION

Attachment to cellular receptors is the first step in viral replication and serves an important role in viral tissue tropism and pathogenesis. This process may involve multistep adhesion accompanied by considerable conformational rearrangements of viral and host molecules (30) and stimulation of intracellular signaling (49). Enveloped viruses engage receptors using glycoproteins that stud the outside of their lipid bilayers, e.g., the glycoprotein complex of HIV (36, 38), gp350 of Epstein-Barr virus (47), and the hemagglutinin of influenza virus (22, 31). Nonenveloped viruses engage receptors by capsid protrusions, e.g., VP4 of rotavirus (40), or indentations, e.g., VP1 of rhinovirus (16, 56). Adenovirus and reovirus are exceptions among nonenveloped animal viruses. These viruses feature elongated attachment spikes that span the equivalent of a capsid radius in length (24, 53). Flexibility of the adenovirus fiber permits simultaneous engagement of multiple receptors (66). In turn, fiber length appears to influence adenovirus tropism (58). It is not understood how the conformation of the reovirus attachment molecule contributes to receptor engagement and subsequent replicative steps.

Mammalian orthoreoviruses (reoviruses) form nonenveloped icosahedral particles composed of two protein shells (19) that enclose 10 segments of double-stranded RNA (dsRNA) (28). The outer capsid contains four structural proteins: σ1, σ3, μ1, and λ2. The σ1 protein, which is anchored into pentameric λ2 turrets at the capsid vertices (19), functions as the reovirus attachment molecule (37, 64). This protein recognizes at least two cellular receptors: sialic acid (14, 55) and junctional adhesion molecule A (JAM-A) (4). JAM-A serves as a proteinaceous receptor for all reovirus serotypes (4, 9, 54), and sialic acid is a coreceptor for serotype 3 strains (14, 27, 50).

The σ1 protein is an important determinant of reovirus dissemination within the host and tropism for host cells and tissues (4, 5, 17). This long fiber-like molecule is comprised of three discernible structural regions: an N-terminal α-helical coiled coil, a central β spiral interrupted by a short stretch of α helix, and a C-terminal globular head (15, 48, 55). These domains are divided by two flexible segments termed interdomain region 1 (IDR1) and IDR2 (15, 24, 55). The σ1 protein engages its receptors using two distinct receptor-binding domains (RBDs) via adhesion strengthening (3). Sequences in the σ1 tail of type 3 reovirus bind sialic acid (14, 55), whereas sequences in the σ1 head engage JAM-A (4, 33).

It is possible that optimal interactions between σ1 and its receptors require that the reovirus attachment protein be long and flexible. Intramolecular mobility of σ1 at IDR1 and IDR2 (15, 24, 55) may allow movement of the spatially independent RBDs with respect to one another as well as to the rest of the virion and permit efficient, sequential engagement of sialic acid and JAM-A during adhesion strengthening. On the other hand, σ1 length may limit steric hindrance from the bulk of the virion and thus facilitate σ1-receptor interactions that result in productive infection. Considering that the utilization of molecular length and flexibility for receptor engagement is rare among nonenveloped animal viruses, the role of σ1 may extend beyond host cell binding. Some evidence suggests that σ1 is folded on the surface of virions and extends only upon proteolytic cleavage of the virus particle during viral disassembly (19, 25, 45). Hence, length and flexibility of σ1 might allow the attachment molecule to assume a conformation during viral particle assembly that primes it for subsequent disassembly events. The fate of σ1 upon reovirus entry into the endocytic compartment is not known. However, it is clear that σ1 must be released from the λ2 pentamers to allow exit of nascent mRNAs from transcribing core particles (10). Length and flexibility of σ1 may be required to facilitate this process.

In this study, we used reverse genetics (34, 35) to generate a panel of reovirus mutants with σ1 molecules of altered length and flexibility and evaluated the capacity of these mutants to bind the cell surface, internalize, uncoat, induce protein synthesis, assemble, and replicate. We reduced the length of the α-helical σ1 tail to engineer L1 and L2 reoviruses and deleted midpoint and head-proximal σ1 IDRs to generate ΔIDR1 and ΔIDR2 mutant viruses, respectively. We found that σ1 length and flexibility are required for efficient reovirus infectivity and replication. L1, L2, and ΔIDR1 viruses showed reduced capacity to attach to cells. In comparison with wild-type (WT) virus, none of the σ1 mutant viruses exhibited defects in internalization. Although not altered in attachment, ΔIDR2 virus was impeded at a postdisassembly replication step. None of these differences were attributable to changes in the folding of the mutant σ1 molecules. Surprisingly, ΔIDR1 virions harbored less σ1 on the outer capsid than WT virus. Thus, σ1 length is required for efficient receptor engagement, IDR1 is required for stable σ1 encapsidation, and IDR2 is required for efficient reovirus uncoating. This new information enhances an understanding of the functions mediated by the filamentous reovirus attachment protein.

MATERIALS AND METHODS

Cells.

Spinner-adapted L929 murine fibroblasts were grown in Joklik's spinner-modified minimum essential medium (SMEM; Lonza, Walkersville, MD) supplemented to contain 5% heat-inactivated fetal bovine serum (HI FBS; Invitrogen, Carlsbad, CA), 2 mM l-glutamine (Invitrogen), 100 U/ml penicillin (Invitrogen), 100 μg/ml streptomycin (Invitrogen), and 25 ng/ml amphotericin B (Sigma-Aldrich, St. Louis, MO). HeLa cells were grown in Dulbecco's modified Eagle medium (DMEM; Invitrogen) supplemented to contain 10% HI FBS, 2 mM l-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin, and 25 ng/ml amphotericin B. BHK-T7 cells were grown in DMEM supplemented to contain 5% HI FBS, 2 mM l-glutamine, 2% minimal essential medium amino acid solution (Invitrogen), and 1 mg/ml Geneticin (Invitrogen).

Viruses.

Recombinant reoviruses containing mutations in the σ1 protein were generated by plasmid-based reverse genetics using cloned type 3 Dearing (T3D) reovirus cDNAs (34, 35). The pT7-T3DS1.T249I template was used to engineer the pT7-T3DS1.T249I.R202W, pT7-T3DS1.T249I.L1, pT7-T3DS1.T249I.L2, pT7-T3DS1.T249I.ΔIDR1, and pT7-T3DS1.T249I.ΔIDR2 plasmids by QuikChange site-directed mutagenesis (Stratagene, La Jolla, CA). Monolayers of approximately 3 × 10⁶ BHK-T7 cells were cotransfected with 3.5 μg each of five plasmid constructs representing the T3D reovirus genome (pT7-T3DL1, pT7-T3DL2-M3, pT7-T3DL3-M1, and pT7-T3DM2-S2-S3-S4 in combination with pT7-T3DS1.T249I or one of the altered pT7-T3DS1.T249I plasmids) using 3 μl TransIT-LT1 transfection reagent (Mirus Bio LLC, Madison, WI) per μg of plasmid DNA. Following 48 h of incubation, recombinant viruses were isolated by plaque purification using L929 cells. Virus stocks were prepared as described previously (6). Virions were purified by Freon extraction from cell lysates and subsequent CsCl gradient centrifugation (25). Bands corresponding to reovirus particle density (1.36 mg/dl) were collected and dialyzed against virion storage buffer (150 mM NaCl, 15 mM MgCl₂, 10 mM Tris-HCl [pH 7.4]). The S1 gene sequences of mutant viruses were verified using PCR products generated from viral RNA subjected to OneStep reverse transcription-PCR (RT-PCR; Qiagen; Valencia, CA) using S1-specific primers (32). Sequences of the S1 gene-specific primers employed for mutagenesis and sequencing are available from the corresponding author upon request. The concentration of pure reovirus virions was determined by spectrophotometry using the following equivalence: 1 absorbance unit (AU) at 260 nm is equal to 2.1 × 10¹² particles/ml. Viral titer was determined by plaque assay using L929 cells (62).

Electrophoresis of reovirus genome.

Whole reovirus virions were incubated at 95°C for 10 min in sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) buffer (62.5 mM Tris HCl [pH 6.8], 2% sodium dodecyl sulfate, 10% glycerol, 0.004% bromophenol blue), loaded into wells of precast 4 to 20% gradient Tris-Tricine polyacrylamide gels (Bio-Rad Laboratories, Hercules, CA) at 5 × 10¹⁰ particles/well, and electrophoresed at a constant current of 15 mA for 16 h. Following electrophoresis, the gel was stained with ethidium bromide and visualized by UV transillumination.

Virus replication in L929 cells.

Monolayers of L929 cells seeded in 24-well plates (Costar; Corning Incorporated, Corning, NY) at 2 × 10⁵ cells/well were adsorbed with reovirus strains at a multiplicity of infection (MOI) of 0.01 PFU/cell at room temperature for 1 h. Inocula were aspirated. Cells were washed once with phosphate-buffered saline (PBS) and incubated in fresh medium at 37°C for various intervals. Virus titers in cell lysates were determined by plaque assay using L929 cells. Viral yields were determined using the following formula: log₁₀ yield_tx = log₁₀ (PFU/ml)_tx − log₁₀ (PFU/ml)_t0, where t_x is the time postadsorption and t₀ is time zero.

Assessment of reovirus infectivity by indirect immunofluorescence.

Monolayers of L929 cells or HeLa cells seeded in 24-well plates at 2 × 10⁵ cells/well were adsorbed with reovirus strains at an MOI of 2 PFU/cell at room temperature for 1 h. Inocula were aspirated. Cells were washed once with PBS and incubated in fresh medium at 37°C for various intervals. Cells were washed once with PBS and fixed with ice-cold methanol for at least 30 min. Cells were washed twice with PBS prior to blocking with 5% bovine serum albumin (BSA) in PBS at room temperature for 15 min. Cells were incubated with a polyclonal reovirus-specific serum (65) at a dilution of 1:500 in PBS containing 0.5% Triton X-100 (TX-100) at 37°C for 1 h. Primary antibody was removed, and cells were washed twice with PBS and incubated with 4′,6-diamidino-2-phenylindole (DAPI; Invitrogen) and Alexa Fluor 488-labeled goat anti-rabbit fluorescent secondary antibody (Invitrogen), both at a dilution of 1:1,000 in PBS containing 0.5% TX-100, at 37°C for 1 h. Cells were washed twice with PBS, and infected cells were visualized by fluorescence microscopy. Results are expressed as the percentage of infected cells present in a ×10 field of view.

The effect of σ1 blockade on reovirus infectivity was tested by incubating reovirus strains with PBS, PBS containing mouse IgG2α, or PBS containing σ1-specific monoclonal antibody (MAb) 9BG5 (8) at room temperature for 1 h prior to adsorption onto L929 cells seeded in 24-well plates at 2 × 10⁵ cells/well at an MOI of 2 PFU/cell at room temperature for 1 h. The dependence of reovirus infection on sialic acid and JAM-A was tested by treatment of HeLa cells seeded in 24-well plates at 2 × 10⁵ cells/well with PBS, 40 mU Arthrobacter ureafaciens neuraminidase, or 10 μg/ml human JAM-A (hJAM-A)-specific MAb J10.4 (39) at room temperature for 1 h prior to adsorption with reovirus strains at an MOI of 500 PFU/cell at room temperature for 1 h. Cells were washed and incubated in fresh medium at 37°C for 20 h. Infectivity was assessed by indirect immunofluorescence.

HA assays.

Purified virions were distributed into 96-well U-bottom microtiter plates (Costar) at an initial concentration of 10¹¹ particles/well and serially diluted 1:2 in 50 μl PBS. Calf erythrocytes (Colorado Serum Company, Denver, CO) were washed twice with PBS, resuspended at 1% (vol/vol) in PBS, delivered to virus-containing wells at 50 μl/well, and incubated at 4°C for 3.5 h. The hemagglutination (HA) titer is defined as 10¹¹ particles divided by the number of particles per HA unit. One HA unit is the particle number sufficient to produce HA (a partial or complete shield of erythrocytes on the well bottom).

Fluorescence-linked immunosorbent assay (FLISA).

Purified virions were adsorbed onto 96-well Immulon 2HB flat-bottom microtiter plates (Fisher Scientific, Suwanee, GA) at a concentration of 4 × 10¹⁰ particles/well in 0.05 M carbonate-bicarbonate buffer (pH 9.6; Sigma-Aldrich). Following incubation at 4°C overnight, plates were blocked with 2% BSA in 0.05 M carbonate-bicarbonate buffer (pH 9.6) at 37°C for 2 h. Wells were washed four times with PBS containing 0.05% Tween 20 prior to incubation with either σ1-specific MAb 9BG5 (8) or σ3-specific MAb 4F2 (63) at 52 and 7 μg/ml, respectively, in PBS containing 1% BSA. Plates were incubated at 37°C for 1 h, washed five times with 0.05% Tween 20 in PBS, incubated with IRDye 800 CW goat anti-mouse secondary antibody (LI-COR, Lincoln, NE) at a dilution of 1:1,000 in PBS containing 1% BSA, washed four times with 0.05% Tween 20 in PBS, rinsed with PBS, and visualized using an Odyssey infrared imaging system (LI-COR). Well fluorescence intensity was determined using Odyssey application software, version 3.0 (LI-COR). The integrated well fluorescence signal for σ1 (9BG5) was divided by that obtained for σ3 (4F2) to quantify the results.

Generation of σ1 head-specific antiserum.

The C-terminal head domain of T3D σ1 (amino acids 293 to 455) was expressed and purified as described previously (57). Rabbit polyclonal serum specific for the T3D σ1 head was generated by Cocalico Biologicals. A single New Zealand White rabbit was immunized with the T3D σ1 C-terminal head domain and boosted at 2, 3, and 7 weeks postimmunization.

Immunoblot of reovirus σ1 and σ3 proteins.

Purified reovirus virions were incubated at 95°C for 10 min in SDS-PAGE buffer, placed into wells of precast 4 to 20% gradient Tris-Tricine polyacrylamide gels at 10¹⁰ particles/well, and electrophoresed at 100 V for 1.5 h. Viral proteins were transferred onto nitrocellulose membranes at 12 V for 45 min and blocked in Odyssey blocking buffer (LI-COR) at room temperature for 1 h. Membranes were incubated with rabbit polyclonal serum specific for the σ1 head and mouse MAb 4F2 specific for σ3 (63) at a 1:1,000 dilution and 7 μg/ml, respectively, in Odyssey blocking buffer supplemented to contain 0.2% Tween 20 at room temperature for 1 h. Membranes were incubated with Alexa Fluor 647-labeled goat anti-rabbit (Invitrogen) and IRDye 800CW goat anti-mouse secondary antibodies at 1:5,000 and 1:10,000 dilutions, respectively, in Odyssey blocking buffer supplemented to contain 0.2% Tween 20 at room temperature for 1 h and scanned using an Odyssey infrared imaging system. Protein band intensity was quantified using Odyssey application software, version 3.0. Integrated protein band fluorescence for σ1 was divided by that for σ3 to normalize the results.

Assay of ammonium chloride bypass kinetics.

Monolayers of L929 cells seeded in 24-well plates at 2 × 10⁵ cells/well were adsorbed with reovirus strains at an MOI of 25 PFU/cell at 4°C for 1 h. Following adsorption, virus inocula were removed, cells were washed with cold PBS, and 1 ml of prewarmed medium was added to each well. Ammonium chloride was delivered to wells at various times postadsorption to provide a final concentration of 25 mM. At 20 h postadsorption, cells were washed once with PBS, fixed with ice-cold methanol, and processed for immunofluorescence staining. The percent infected cells was determined in a ×10 field of view and normalized to the untreated condition to quantify the results.

Evaluation of reovirus internalization by confocal microscopy.

Glass coverslips (1.5 mm) were placed in 24-well plates and treated with 1.6 mg/ml BD Matrigel basement membrane matrix (BD Life Sciences, Franklin Lakes, NJ) diluted in incomplete medium at room temperature for 1 h with continuous rocking. Coverslips were rinsed with incomplete medium, seeded with 6.5 × 10⁴ L929 cells each, and incubated at 37°C overnight. The medium was removed, and cells were adsorbed with WT, length (L), and IDR reoviruses labeled with Alexa Fluor 488 as described previously (23) at 5 × 10⁴ virus particles per cell at room temperature for 30 min with continuous rocking. Virus inocula were aspirated, and cells were either washed twice with PBS and fixed immediately with 10% formalin or incubated in complete medium at 37°C for 1 h, rinsed once with PBS, and fixed. Formalin was quenched with an equivalent volume of 0.1 M glycine. Cells were washed twice with PBS, blocked with PBS-BG (0.5% BSA, 0.1% glycine) at room temperature for 10 min, and incubated with polyclonal reovirus-specific serum at a 1:1,000 dilution in PBS-BG at room temperature for 1 h. Cells were rinsed three times with PBS-BG at room temperature for 15 min with continuous rocking and incubated with an Alexa Fluor 647-labeled goat anti-rabbit secondary antibody (Invitrogen) and Alexa Fluor 546-labeled phalloidin (Invitrogen) at 1:1,000 and 1:100 dilutions in PBS-BG, respectively, at room temperature for 1 h. Cells were rinsed three times with PBS-BG at room temperature for 15 min with continuous rocking. Coverslips were mounted onto glass slides using Aqua-Poly/Mount (Polysciences, Warrington, PA) and viewed using a Zeiss LSM 510 Meta inverted confocal microscope. Cells were selected for imaging in the phalloidin channel to limit observer bias.

Assessment of reovirus attachment to cells by flow cytometry.

L929 cells (10⁶) in 1.5-ml tubes were adsorbed with reovirus strains at 5 × 10⁴ particles/cell at 4°C for either 30 or 60 min with continuous rotation. Following adsorption, cells were washed with ice-cold PBS supplemented to contain 5% BSA (PBS-BSA) and incubated with reovirus-specific rabbit polyclonal antiserum at a 1:2,500 dilution in PBS-BSA at 4°C for 30 min with continuous rotation. Cells were washed twice with ice-cold PBS-BSA and incubated with Alexa Fluor 647-labeled goat anti-rabbit secondary antibody (Invitrogen) at a 1:1,000 dilution in PBS-BSA at 4°C for 30 min with continuous rotation. Following two washes with ice-cold PBS-BSA, cells were fixed in 1% paraformaldehyde in PBS and scored for virus binding by flow cytometry using a 3-laser BD LSRII flow cytometer (BD Biosciences, Franklin Lakes, NJ). Results were quantified by normalizing the mean fluorescence intensities (MFIs) of cell populations adsorbed with the L and IDR reovirus mutants for either 30 or 60 min to the MFI of the cell population adsorbed with WT virus for 60 min.

Statistical analysis.

Mean values were compared using two-tailed Student's t tests. P values of <0.05 were considered statistically significant. Standard error of the mean (SEM) was used as a measure of variability for means calculated for independent experiments containing multiple replicates. Standard deviation (SD) was used as a measure of variability among individual data points.

RESULTS

Construction and characterization of reovirus σ1 length and flexibility mutants.

To investigate the function of σ1 length and flexibility in the reovirus life cycle, we engineered a panel of recombinant reoviruses bearing σ1 molecules of varied length and flexibility using reverse genetics (34, 35) (Fig. 1A). Mutations altering the length and flexibility of σ1 were introduced into the S1 gene-containing plasmid pT7-T3DS1.T249I by site-directed mutagenesis. We used a parental plasmid encoding a T249I substitution in the short α-helical region of the T3D σ1 body domain (Fig. 1A) to prevent proteolytic degradation of σ1 (13). As the sialic acid-binding pocket in σ1 resides in the second and third β-spiral repeats of the σ1 body domain (55), we truncated sequences in the α-helical region of the σ1 tail (20, 48) to alter σ1 length. Amino acid residues 51 to 100 were removed from mutant L1, and residues 83 to 155 were removed from mutant L2 (Table 1; Fig. 1A). IDR1 (residues 155 to 164) and IDR2 (residues 291 to 294) were deleted from σ1 to engineer ΔIDR1 and ΔIDR2 reoviruses, respectively (Table 1; Fig. 1A). Despite three attempts, a reovirus mutant lacking both IDR1 and IDR2 (ΔIDR1/2 virus) could not be recovered (Table 1). L and IDR mutant reovirus S1 gene sequences were verified using cDNA obtained by RT-PCR from purified viral dsRNA. Electrophoresis of the dsRNA gene segments of mutant reoviruses revealed the anticipated genotypes (Fig. 1B). Thus, reovirus mutants with alterations in σ1 length and truncations of each of the IDRs can be recovered by reverse genetics.

Fig 1.

L and IDR reovirus mutants. (A) Schematic of WT, L1, L2, ΔIDR1, and ΔIDR2 reovirus σ1 proteins. The α helices, β-spiral repeats, and β barrel are shown in blue, yellow, and red, respectively. Receptor-binding domains are indicated by underlines. The position of the T249I mutation is shown with black dots. Deletions of σ1 sequence are indicated with inverted triangles. SA, sialic acid; JAM-A, junctional adhesion molecule A; IDR1, interdomain region 1; IDR2, interdomain region 2. (B) Segmented dsRNA genomes of L and IDR mutant viruses. Viral gene segments from virions of WT, R202W, L1, L2, ΔIDR1, and ΔIDR2 reovirus strains were resolved by SDS-PAGE, stained with ethidium bromide, and visualized by UV transillumination. Large (L), medium (M), and small (S) class gene segments are designated. S1 gene segments are indicated by white arrowheads.

Table 1.

Characterization of L and IDR mutant reoviruses

Virus	Mutation in σ1	Particle/PFU ratioa
WTb		149.1 (8)
R202Wc	R202W	162.1 (3)
L1	Δ51Q-100S	245.3 (3)
L2	Δ83R-155Q	213.4 (6)
ΔIDR1	Δ155Q-164T	543.0 (5)d
ΔIDR2	Δ291S-294P	285.3 (3)
ΔIDR1/2	Δ155Q-164T/Δ291S-294P	NRe

^aVirus particle concentration was determined by spectrophotometry using the equivalence of 1 AU at 260 nm is equal to 2.1 × 10¹² particles/ml. Titers (in PFU/ml) were determined by plaque assay. The number of independent viral purification stocks tested is shown in parentheses.

^bT3D.σ1.T249I (WT).

^cT3D.σ1.R202W/T249I (R202W).

^dP < 0.05 in comparison with WT virus.

^eNR, not recoverable.

As a first step to determine whether any properties attributable to σ1 are altered in the L and IDR mutants, we tested the infectivity of the L and IDR reoviruses by calculating particle-to-PFU ratios for several independent purifications of each virus (Table 1). The majority of virus stocks had particle/PFU ratios between 150 and 650, which is consistent with previously reported values for reovirus (18, 29). Despite some sample-to-sample variation, only the ΔIDR1 virus had a significantly higher particle/PFU ratio than WT virus (Student's t test, P < 0.01). This result suggests that removal of IDR1 from the σ1 molecule diminishes the efficiency of reovirus infection.

σ1 length and flexibility are required for reovirus replication in cell culture.

To directly test whether σ1 length and flexibility influence reovirus replication, we quantified viral yields following infection of L929 cells over a 24-h time course. Cells were infected with WT, L, or IDR reoviruses at an MOI of 0.01 PFU/cell, and viral titers were determined by plaque assay at various intervals postadsorption (Fig. 2A and B). The L and IDR mutants exhibited longer eclipse periods than WT virus. L2 and ΔIDR2 viruses produced yields similar to those of WT following completion of the first replicative cycle (24 h) (6), but the replication kinetics of these viruses were significantly delayed (Student's t test, P < 0.05). On the other hand, L1 and ΔIDR1 viruses replicated more slowly than WT virus throughout the course of the first replicative cycle (Student's t test, P < 0.05). These data suggest that length and flexibility of σ1 are required for reovirus replication in cultured cells.

Fig 2.

Mutant reovirus replication and infectivity in L929 cells. (A, B) Cells were adsorbed with WT, L (A), and IDR (B) mutant reoviruses at an MOI of 0.01 PFU/cell. The inoculum was removed, fresh medium was added, and cells were incubated at 37°C for the times shown. Titers of virus in cell lysates were determined by plaque assay using L929 cells. Results are expressed as mean viral yield, which is defined as log₁₀(titer)_tx − log₁₀(titer)_t0, where t_x is the time postadsorption and t₀ is time zero, for three independent experiments. Error bars represent SEM. (C, D) Cells were adsorbed with WT, L (C), and IDR (D) mutant reoviruses at an MOI of 2 PFU/cell and processed for indirect immunofluorescence at the times shown postadsorption. Results are expressed as mean percentage of infected cells in a ×10 field of view for three independent experiments. Error bars represent SEM.

σ1 length and flexibility are required for reovirus infection in cell culture.

Given the key role of σ1 in reovirus attachment (3, 4, 14, 33, 55), we next quantified L and IDR mutant infectivity in a viral protein production assay to evaluate the efficiency of the initial stages of infection. Monolayers of L929 cells were adsorbed with WT, L, and IDR reoviruses at an MOI of 2 PFU/cell, and viral infectivity was scored at various times postadsorption by indirect immunofluorescence (Fig. 2C and D). The L and IDR reoviruses showed delayed kinetics of viral protein synthesis, indicative of decreased infection compared with that by WT virus (Student's t test, P < 0.05). The percentage of cells infected by L1, L2, and ΔIDR1 viruses was only 25% of that produced by WT virus at 24 h postadsorption. In comparison with the other mutant viruses, ΔIDR2 virus appeared to be less disabled. The percentage of cells infected by this virus was 60% of that observed for WT virus at 24 h postadsorption. These findings suggest that an impediment to L and IDR reovirus mutant replication occurs at a step in the viral life cycle prior to or during translation of viral proteins.

IDR1 is required for maintenance of σ1 within the outer capsid.

We next evaluated the conformation and functional profile of mutant L and IDR σ1 molecules. To assess whether the mutant σ1 proteins are folded properly, we determined whether pretreatment of L and IDR mutant virions with MAb 9BG5, which binds to a conformational epitope in the σ1 head (52), inhibits infection (Fig. 3). Cells were adsorbed with either vehicle- or antibody-treated virus stocks at an MOI of 2 PFU/cell, incubated in fresh medium at 37°C for 22 h, and stained with a polyclonal reovirus-specific serum. Following adsorption with a control IgG2α antibody, each of the viruses tested retained a full capacity to infect L929 cells. In contrast, the infectivity of WT, L, and IDR reoviruses was substantially diminished by incubation with MAb 9BG5. These results suggest that the L and IDR mutant reoviruses harbor σ1 molecules with head regions that are properly folded.

Fig 3.

Infection by L and IDR mutant reoviruses is dependent on σ1. WT, L, and IDR reoviruses were incubated in PBS, PBS containing mouse IgG2α, or PBS containing σ1-specific MAb 9BG5 at room temperature for 1 h prior to adsorption at an MOI of 2 PFU/cell onto L929 cells at room temperature for 1 h. Cells were washed and incubated in fresh medium at 37°C for 20 h. Infectivity was assessed by indirect immunofluorescence. Results are expressed as the mean percentage of infected cells in a ×10 field of view for three independent experiments. Error bars represent SEM. *, P < 0.01 in comparison to the mock-treated condition (Vehicle).

To assess the capacity of σ1 to engage sialic acid, we performed HA assays using bovine erythrocytes. The capacity of reovirus to produce HA is determined by binding to sialylated glycans on the erythrocyte surface (26, 51). Serial dilutions of virus were incubated with red blood cells at 4°C, and the HA titer was determined after 3.5 h of incubation (Fig. 4A and B). The HA capacity of L1, L2, and ΔIDR2 viruses was comparable to that of WT virus. The T3D σ1-R202W mutant cannot bind sialic acid (55) and, consequently, produced no HA in this assay. In contrast, the HA titer of the ΔIDR1 virus was approximately 80% less than that of WT. From these data, we concluded that the reduced capacity of the ΔIDR1 mutant to cross-link erythrocytes is attributable to either altered folding of the σ1 sialic acid-binding region or decreased encapsidation of σ1 onto the ΔIDR1 virion.

Fig 4.

HA assay of L and IDR mutant reoviruses. (A) Purified virions were serially diluted in PBS and incubated with bovine erythrocytes (1% [vol/vol] in PBS) at 4°C for 3.5 h. Erythrocyte shields indicate HA, and erythrocyte buttons indicate absence of erythrocyte cross-linking. (B) Results are expressed as mean log₂ (HA titer) for four independent experiments. HA titer is defined as 10¹¹ particles divided by the number of particles per HA unit. One HA unit equals the particle number sufficient to produce HA. Error bars represent SEM. *, P < 0.001 in comparison to WT virus.

To distinguish between these two possibilities, we quantified the relative amounts of capsid-associated σ1 and σ3 in WT, L, and IDR virions using σ1-specific MAb 9BG5 (8) and σ3-specific MAb 4F2 (63) by FLISA (Fig. 5A and B). After normalizing the σ1 signal intensity to the σ3 signal intensity to control for viral particle number, the σ1 signal for ΔIDR1 virus but not the other mutants was decreased compared with that for WT virus. To confirm these results, we quantified the relative amounts of σ1 on L and IDR virus capsids by immunoblotting virion proteins using a polyclonal serum specific for the σ1 head and σ3-specific MAb 4F2 (63) as a loading control (Fig. 6A to D). Concordant with the FLISA results, the relative amount of encapsidated σ1 was diminished exclusively for ΔIDR1 virus. Taken together, these results suggest that the σ1 molecules of the L and IDR reoviruses are folded properly, but the amount of σ1 expressed on the ΔIDR1 capsid is diminished. It is possible that the smaller amount of σ1 displayed by ΔIDR1 virus occurs as a consequence of either reduced stability of σ1 at capsid vertices or decreased encapsidation of σ1 during viral assembly.

Fig 5.

FLISA of L and IDR mutant virus particles using σ1-specific MAb 9BG5 and σ3-specific MAb 4F2. (A) Purified virions (4 × 10¹⁰) were adsorbed onto 96-well high-binding enzyme-linked immunosorbent assay plates, incubated with primary antibodies, exposed to fluorescent secondary antibodies, and visualized using an infrared imaging system. This panel shows representative signals obtained for σ1 and σ3. (B) Quantification of the σ1/σ3 FLISA signal intensity ratio for L and IDR mutant reoviruses. Results are expressed as mean well fluorescence intensity ratio for two independent experiments. Error bars represent SEM. *, P < 0.05 in comparison to WT virus.

Fig 6.

(A, B) Immunoblot analyses of L (A) and IDR (B) mutant reovirus σ1 and σ3 proteins. WT, L1, L2, ΔIDR1, and ΔIDR2 virion proteins were resolved by SDS-PAGE and transferred onto nitrocellulose membranes. Membranes were probed with a polyclonal σ1 head-specific serum and σ3-specific MAb 4F2. After incubation with fluorescent secondary antibodies, protein bands were visualized using an infrared imaging system. The σ3 band serves as an internal loading control. (C, D) Quantification of the σ1/σ3 immunoblot band intensity ratio for L (C) and IDR (D) mutant reoviruses. Results are expressed as mean protein band fluorescence intensity ratio for three independent experiments. Error bars represent SEM. *, P < 0.05 in comparison to WT virus.

L and IDR reovirus infection is dependent on sialic acid and JAM-A.

To test whether the L and IDR reovirus mutants engage the known reovirus receptors, we tested the infectivity of WT and mutant reoviruses using HeLa cells after treatment with vehicle, A. ureafaciens neuraminidase, or JAM-A-specific MAb J10.4 (39). Pretreated cells were adsorbed with reovirus strains at an MOI of 500 PFU/cell (calculated on the basis of titers determined using L929 cells), incubated in fresh medium at 37°C for 22 h, and stained with a polyclonal reovirus-specific antiserum (Fig. 7). Following treatment with neuraminidase, the infectivity of the L and IDR viruses was decreased by 50 to 70% in comparison to the vehicle-treated condition. Similarly, incubation of cells with MAb J10.4 diminished infectivity of the L and IDR viruses by 90% in comparison to that for cells treated with vehicle. These data suggest that the L and IDR reoviruses engage sialic acid and JAM-A in a mechanism that leads to infection.

Fig 7.

Infectivity of L and IDR mutant reoviruses is dependent on sialic acid and JAM-A. HeLa cells were treated with PBS (−), 40 mU/ml A. ureafaciens neuraminidase (N), or 10 μg/ml hJAM-specific MAb J10.4 (J) prior to adsorption with WT, L, and IDR mutant reoviruses at an MOI of 500 PFU/cell at room temperature for 1 h. Cells were washed and incubated in fresh medium at 37°C for 20 h. Infectivity was assessed by indirect immunofluorescence. Results are expressed as mean percentage of infected cells in a ×10 field of view for three independent experiments. Error bars represent SEM. *, P < 0.01 in comparison to the untreated condition (PBS).

The L and IDR replication defect occurs prior to endosomal escape.

Following attachment to host cells, reovirus is internalized into the endocytic pathway, where the viral particle undergoes acid-dependent proteolytic disassembly (2, 7, 12, 21, 59, 61), which allows the virus to gain access to the cytoplasm (1, 11). To determine whether σ1 length and flexibility are required for steps in the reovirus life cycle preceding endosomal escape, monolayers of L929 cells were adsorbed with reovirus strains at an MOI of 25 PFU/cell at 4°C to synchronize attachment, incubated at 37°C in fresh medium, and exposed to ammonium chloride to prevent endosomal acidification at defined intervals postadsorption. After an overnight incubation, cells were scored for infection by indirect immunofluorescence (Fig. 8A and B). L1, L2, and ΔIDR1 mutant viruses bypassed the ammonium chloride blockade more slowly than WT reovirus (Student's t test, P < 0.05). These findings indicate that σ1 length and IDR1 are required for steps in reovirus replication that precede endosomal escape. The ΔIDR2 mutant bypassed the pH blockade with kinetics similar to those observed for WT virus (Student's t test, P > 0.05), suggesting that IDR2 is required for efficient completion of replication steps that temporally fall between endosomal escape and viral protein synthesis.

Fig 8.

Kinetics of ammonium chloride blockade bypass by L (A) and IDR (B) mutant reoviruses in L929 cells. Cells were adsorbed with the reovirus strains shown at an MOI of 25 PFU/cell at 4°C for 1 h. Following adsorption, cells were washed with cold PBS and incubated in fresh medium at 37°C for 20 h. Ammonium chloride was added to the medium at the indicated times postadsorption to achieve a final concentration of 25 mM. Infectivity was assessed by indirect immunofluorescence. Results are expressed as mean percentage of infected cells in a ×10 field of view normalized to the untreated condition (no ammonium chloride) for three independent experiments. Error bars represent SEM.

Internalization of L and IDR mutants is not altered.

To directly assess the role of σ1 length and flexibility in reovirus internalization, we adsorbed L929 cells with Alexa Fluor 488-labeled L and IDR mutant reoviruses at 50,000 particles/cell and fixed cells at 0 and 60 min postadsorption. We then stained cells without prior permeabilization with phalloidin to visualize actin and polyclonal reovirus-specific serum (65) to visualize extracellular reovirus particles. We quantified internalization of reovirus virions by confocal microscopy. The inoculum dose used in this experiment allowed us to detect a signal for the most disabled L and IDR mutant viruses and has been used previously for evaluation of reovirus internalization (41–44). Representative confocal micrographs of reovirus-infected L929 cells are shown in Fig. 9A. Cell surface-associated and intracellular particles are depicted in aquamarine and green, respectively. The average number of total particles per cell (extracellular and intracellular) was significantly diminished for L1, L2, and ΔIDR1 viruses at 0 min (Fig. 9A; data not shown) and 60 min (Fig. 9A and B) postadsorption, suggesting a defect in attachment for these viruses. However, at 0 and 60 min postadsorption, 30 and 60% of the virus particles were internalized, respectively, for all strains tested (Fig. 9C). Therefore, native σ1 length and flexibility are not required for reovirus internalization, but these features of σ1 appear to augment attachment to host cells.

Fig 9.

Internalization of L and IDR mutant reoviruses. L929 cells were adsorbed with 50,000 particles/cell of Alexa Fluor 488-labeled WT, L, and IDR reoviruses (green) at room temperature for 30 min. The inoculum was removed, and cells were incubated in fresh medium for the intervals shown, stained without cellular permeabilization for actin (red) and extracellular reovirus (blue), and imaged by confocal microscopy. (A) Representative digital fluorescence images of cells infected with L and IDR viruses at 0 and 60 min postadsorption. Actin, extracellular reovirus, and intracellular reovirus are depicted in red, aquamarine, and green, respectively. Scale bars, 10 μm. (B) Quantification of the total number of reovirus particles at 60 min postadsorption in single planes of view for 15 to 20 cells per virus strain for three independent experiments. Error bars indicate SEM. *, P < 0.05 in comparison to WT virus. (C) Quantification of the percent internalized reovirus particles at 0 and 60 min postadsorption in single planes of view for 15 to 20 cells per virus strain for three independent experiments. Error bars indicate SEM. *, P < 0.05 in comparison to WT virus.

Length and IDR1 of σ1 are required for reovirus attachment.

To directly investigate the role of σ1 length and flexibility in reovirus attachment, we adsorbed L929 cells with the L and IDR mutants at 4°C for 30 or 60 min and quantified virus binding by flow cytometry (Fig. 10). Compared with WT virus, we found that cell binding was reduced by 65 to 78% for L1, L2, and ΔIDR1 viruses at both 30 and 60 min postadsorption. The binding of ΔIDR2 virus to cells was reduced by about 10 to 17% compared with that of WT virus, but this difference was not statistically significant. These results show that σ1 length and IDR1 are required for efficient reovirus attachment to L929 cells.

Fig 10.

Attachment of L and IDR mutant reoviruses. L929 cells were adsorbed with 50,000 particles/cell of WT, L, and IDR viruses at 4°C for either 30 or 60 min. Following adsorption, cells were incubated with reovirus-specific polyclonal antiserum, and virus attachment was assessed by flow cytometry. Results are presented as the percentage of WT virus binding after 60 min of incubation for two independent replicates. Error bars represent SD. *, P < 0.02 in comparison with WT virus.

DISCUSSION

Adenovirus and reovirus display elongated filamentous attachment spikes that extend from their capsids (24, 53). This architectural feature distinguishes these viruses from other nonenveloped animal viruses that engage receptors via capsid protrusions (40) or indentations (56). Thus, the structures of adenovirus fiber and reovirus σ1 raise the possibility that these molecules serve functions in viral replication in addition to cell attachment. In fact, flexibility of the adenovirus fiber is required for efficient internalization of the virus (66). It is not known how the conformation of the reovirus attachment molecule contributes to receptor engagement and subsequent replicative steps. In this study, we used plasmid-based reverse genetics (34, 35) to engineer mutant viruses bearing σ1 molecules of altered length and flexibility (Fig. 1A and B; Table 1) and defined the capacity of these mutants to bind to cells, internalize, uncoat, induce viral protein synthesis, replicate, and assemble. We found that σ1 length and flexibility are required for efficient reovirus attachment, uncoating, and stable σ1 encapsidation during viral particle assembly.

We shortened the reovirus attachment protein by deleting parts of the α-helical tail domain (48) to avoid altering the sialic acid RBD located in the β-spiral body domain of the molecule (14, 55). Seven and 10 heptad repeats were removed from σ1 to generate the L1 and L2 reoviruses, respectively (Fig. 1A; Table 1). We also deleted the midpoint and head-proximal IDRs of σ1 to engineer ΔIDR1 and ΔIDR2 viruses, respectively (Fig. 1A; Table 1). We were unable to recover a virus lacking both IDR1 and IDR2 (Table 1). Since our results indicate that both of these sequences are required for efficient reovirus replication, we think that a double-IDR-deletion virus is not viable.

The mutations altering the length and flexibility of σ1 characterized in this study do not appear to affect σ1 folding. L and IDR virus infection of L929 cells was neutralized by MAb 9BG5, which recognizes a conformational epitope in the σ1 head domain (8, 52) (Fig. 3). This finding suggests that there are no gross alterations in the σ1 head. Moreover, following adsorption of HeLa cells with JAM-A-specific MAb J10.4 (39), infection by the L and IDR mutants was abolished (Fig. 7), suggesting that the JAM-A RBD of these reoviruses is not misfolded. Surprisingly, FLISAs using 9BG5 revealed that the relative amount of encapsidated σ1 is reduced in ΔIDR1 virus but not in L1, L2, or ΔIDR2 virus (Fig. 5A and B). Immunoblot analysis of L and IDR capsid proteins using a polyclonal antiserum specific for the σ1 head also showed that only the ΔIDR1 virus of the mutants tested harbors less σ1 than WT virus (Fig. 6A to D). Finally, the particle/PFU ratio for ΔIDR1 virus was increased (Table 1), suggesting that a higher proportion of ΔIDR1 virions than WT virions is functionally impaired. Hence, the decreased attachment to L929 cells observed for ΔIDR1 virus (Fig. 10) is most likely attributable to stoichiometrically less σ1 on the ΔIDR1 capsid.

Since an MAb specific for the σ1 sialic acid RBD is not available, we tested the capacity of the L and IDR mutants to engage sialic acid using HA assays (Fig. 4A and B). The HA capacities of L1, L2, and ΔIDR2 viruses were comparable to the HA capacity of WT virus, suggesting that the sialic acid RBD of these viruses is properly folded. However, the HA titer of ΔIDR1 virus was decreased by approximately 80%. This reduction is not likely caused by misfolding of the ΔIDR1 virus sialic acid RBD. Following treatment of HeLa cells with neuraminidase to remove cell-surface sialic acid, infection of WT, L, and IDR reoviruses was reduced to the same extent (Fig. 7), suggesting that all of the viruses tested rely in part on sialic acid binding for infection and, therefore, that their σ1 sialic acid RBD is properly folded. These results provide further evidence that the conformations of the L and IDR mutant σ1 molecules are not grossly altered and that the amount of σ1 on the ΔIDR1 virus capsid is reduced.

The hydrophobic N terminus of σ1 directly interacts with pentameric turrets formed by λ2 at the capsid vertices and thus tethers the reovirus attachment molecule to the virion surface (19). Other outer-capsid proteins have not been reported to directly interact with σ1, but such contacts cannot be excluded based on the available evidence. The reovirus attachment molecule is likely collapsed in some way on the surface of virions and extends outward only upon proteolytic removal of σ3 during the formation of infectious subvirion particles (ISVPs) during viral disassembly (19, 25, 45). Interestingly, ISVPs display decreased infectivity following exposure to heat (46), which might be attributable to decreased stability of the extended conformer of σ1 at the capsid vertices. Therefore, IDR1 may stabilize σ1 by allowing it to adopt a folded conformation on the virion surface. IDR1 also could permit σ1 to assume a conformation amenable to capsid assembly. Alternatively, IDR1 may interact with other outer-capsid proteins (λ2, μ1, or σ3) to allow stable σ1 encapsidation in the virion.

Our findings suggest that flexibility of σ1 at IDR2 is required for a replicative step that takes place after protease-dependent uncoating of reovirus in the endocytic pathway but before initiation of viral protein synthesis. Following an initial delay, ΔIDR2 virus produced WT-level yields of viral progeny at the end of a single infectious cycle (24 h) (Fig. 2B). In addition, ΔIDR2 virus infected a significantly lower percentage of cells than WT virus (Fig. 2D), again consistent with a prolonged eclipse phase. However, the interval required by ΔIDR2 virus to bypass a block to replication imposed by ammonium chloride did not differ from that for WT (Fig. 8B). Concordantly, ΔIDR2 virus did not exhibit defects in attachment (Fig. 10) or internalization (Fig. 9).

The underlying mechanism for the defect in ΔIDR2 virus replication is not clear from our study. However, σ1 must be shed from the λ2 turrets to allow export of nascent transcripts through unobstructed λ2 channels at the icosahedral vertices of transcribing reovirus cores (10, 19). Since σ1 dissociates from λ2 at some point after endosomal proteolysis of the reovirus outer capsid (10, 11), we think that IDR2 of σ1 is required to facilitate this process. An alternative possibility is that IDR2 is involved in viral RNA transcription or protein synthesis in some way, but this explanation does not seem likely. Little σ1 would be available following disassembly and endosomal escape to serve a function in viral transcription or translation.

It is clear from our results that optimum length of the σ1 molecule is required for efficient reovirus binding to cells. L1 and L2 both showed delayed replication kinetics (Fig. 2A) and infected approximately 75% fewer cells than WT virus (Fig. 2C). Both mutants exhibited delayed bypass of an ammonium chloride replication block (Fig. 8), which we think is entirely due to inefficient attachment (Fig. 10), as internalization of these mutants is not altered (Fig. 9). Moreover, the reduction in binding to cells detected for L1 and L2 (Fig. 10) corresponds to the decrease in infectivity observed for these mutants (Fig. 2C). Therefore, σ1 length could limit steric hindrance from the bulk of the virion to allow appropriate cell-surface receptor engagement. Although the α-helical fragment of L1 σ1 is longer than that of L2 σ1 (Fig. 1A; Table 1), we did not note any obvious phenotypic differences between L1 and L2 in the assays employed in this study. It is possible that mutants with even shorter σ1 proteins would be more severely compromised. However, we did not attempt to generate such viruses in our study.

The length and flexibility of adenovirus fiber are important determinants of adenovirus tropism (58) and internalization efficiency (66), respectively. In this study, we found that despite extensive structural similarities between the adenovirus and reovirus attachment proteins (60), the function of σ1 length and flexibility in reovirus replication differs from the role of fiber length and flexibility in the adenovirus life cycle. We do not think that the length of the reovirus attachment protein contributes significantly to reovirus tropism in the infected host. Despite dramatic serotype-dependent differences in tropism (17, 64), σ1 varies only between 455 and 470 amino acids for strains of each of the three reovirus serotypes (48). However, we found that σ1 length is required for the efficient binding of reovirus to cells, perhaps because this feature of σ1 reduces steric hindrance from the much larger virion during receptor engagement. On the other hand, flexibility of σ1 at IDR2 is required for a replicative event that follows proteolytic disassembly of reovirus particles in the endocytic pathway and may permit efficient σ1 dissociation from λ2 turrets prior to viral endosomal escape. Finally, flexibility of σ1 at IDR1 is required for stable encapsidation of the σ1 protein during particle assembly. This work reveals new functions for reovirus σ1 and provides insights into molecular events at the virus-cell interface that lead to productive infection.