Turnover Rate of NS3 Proteins Modulates Bluetongue Virus Replication Kinetics in a Host-Specific Manner

ABSTRACT

Bluetongue virus (BTV) is an arbovirus transmitted to livestock by midges of the Culicoides family and is the etiological agent of a hemorrhagic disease in sheep and other ruminants. In mammalian cells, BTV particles are released primarily by virus-induced cell lysis, while in insect cells they bud from the plasma membrane and establish a persistent infection. BTV possesses a ten-segmented double-stranded RNA genome, and NS3 proteins are encoded by segment 10 (Seg-10). The viral nonstructural protein 3 (NS3) plays a key role in mediating BTV egress as well as in impeding the in vitro synthesis of type I interferon in mammalian cells. In this study, we asked whether genetically distant NS3 proteins can alter BTV-host interactions. Using a reverse genetics approach, we showed that, depending on the NS3 considered, BTV replication kinetics varied in mammals but not in insects. In particular, one of the NS3 proteins analyzed harbored a proline at position 24 that leads to its rapid intracellular decay in ovine but not in Culicoides cells and to the attenuation of BTV virulence in a mouse model of disease. Overall, our data reveal that the genetic variability of Seg-10/NS3 differentially modulates BTV replication kinetics in a host-specific manner and highlight the role of the host-specific variation in NS3 protein turnover rate.

IMPORTANCE BTV is the causative agent of a severe disease transmitted between ruminants by biting midges of Culicoides species. NS3, encoded by Seg-10 of the BTV genome, fulfills key roles in BTV infection. As Seg-10 sequences from various BTV strains display genetic variability, we assessed the impact of different Seg-10 and NS3 proteins on BTV infection and host interactions. In this study, we revealed that various Seg-10/NS3 proteins alter BTV replication kinetics in mammals but not in insects. Notably, we found that NS3 protein turnover may vary in ovine but not in Culicoides cells due to a single amino acid residue that, most likely, leads to rapid and host-dependent protein degradation. Overall, this study highlights that genetically distant BTV Seg-10/NS3 influence BTV biological properties in a host-specific manner and increases our understanding of how NS3 proteins contribute to the outcome of BTV infection.

INTRODUCTION

Bluetongue virus (BTV) is an arthropod-borne virus responsible for a hemorrhagic disease of domestic and wild ruminants (1, 2) and is transmitted between mammalian hosts by biting midges of the Culicoides family (3, 4). C. imicola represents the major vector species in Africa (3) and in southern Europe (5, 6), whereas C. obsoletus and C. pulicaris species have been involved in the transmission of BTV in central and northern Europe (7–9). The impact of bluetongue disease in Europe was relatively low before the 1990s (6, 10, 11). However, since 1998, several seasonal incursions of distinct BTV serotypes/strains have occurred virtually every year, causing massive economic losses to animal husbandry (7, 12–16). BTV is a nonenveloped virus that belongs to the Orbivirus genus within the family Reoviridae (17, 18). It possesses a ten-segmented double-stranded RNA genome encoding seven structural proteins (VP1 to VP7) and four nonstructural proteins (NS1 to NS4) (19–22). The virion consists of an outer capsid formed by VP2 and VP5 and an inner capsid composed of VP7 and VP3 (18, 23, 24). The viral transcription complex is composed of VP1 (polymerase), VP4 (capping enzyme), VP6 (helicase), and the viral genome (23, 25). NS1 has been associated with cytopathogenicity (26) and the transactivation of BTV protein translation (27). NS2 plays a key role in the formation of viral inclusion bodies (VIBs) where the assembly of new virions takes place (28, 29), while the recently discovered NS4 enhances BTV replication in cells pretreated with interferon (20, 22). NS3 proteins are encoded by BTV segment 10 (Seg-10) as two isoforms, NS3 and NS3a, the latter being translated from a second in-frame initiation codon and lacking the first N-terminal 13 amino acid residues (30). NS3 proteins are glycoproteins that localize in the Golgi apparatus and at the plasma membrane of infected cells (31–33), where they facilitate viral particle release either by increasing plasma membrane permeability through a viroporin activity (34) or by viral budding (35). In insect cells, BTV particles egress from the plasma membrane, whereas in mammalian cells they are released by budding only for a short time period before the virus induces very rapidly cell lysis. Viral trafficking at the plasma membrane and egress by budding are mediated through the interactions of NS3 with BTV outer capsid proteins (VP2 and VP5) (36, 37) and cellular proteins, the calpactin light chain (P11) (36, 38), which interacts with annexin-2, a lipid raft-associated protein (39), and TSG101 and NEDD4-like ubiquitin ligases, involved in the pathway of endosomal sorting complexes required for transport (ESCRT) (40, 41). Interestingly, both NS3 and NS3a proteins are essential for BTV egress in Culicoides-derived cells, whereas only the full-length NS3 isoform seems to significantly influence viral particle release and cytopathic effect (CPE) in BSR (baby hamster kidney) cells, favoring BTV propagation (38, 42). In addition, the identification and characterization of a BTV NS3 mutant as a vaccine candidate revealed that NS3 proteins are crucial for BTV virulence and the induction of viremia in sheep (43). Unlike in insects, the type I interferon production and signaling pathway represents one of the first lines of defense against viral infection in mammals (44), and BTV NS3 proteins have been shown to inhibit the activation of the IFN-β promoter in reporter-based cell assays (45, 46). Overall, NS3 are multifunctional proteins that play critical and different roles in the BTV replication cycle depending on host cells. Thus far, BTV exists as at least 27 immunologically distinct serotypes based on differences of the VP2 outer core protein (47–50). In addition, most BTV segment sequences, including those of Seg-10, have been divided into three main topotypes, Western groups 1 and 2 and Eastern group 1, based on their geographical origins (49, 51–54). Interestingly, Seg-10 sequences are highly variable in BTV and also in other midge-borne orbiviruses, such as African horse sickness virus (AHSV), enzootic hemorrhagic disease virus (EHDV), and equine encephalosis virus (EEV), and may be involved in determining both vector competence and virulence (55, 56). In BTV, Seg-10 displays the highest degree of variability among the segments encoding nonstructural proteins (53, 57). Previous studies showed frequent reassortment events between several BTV strains in the field (57, 58). Indeed, during coinfection of the same host cell, any BTV segment, including Seg-10, can be exchanged, resulting in reassortant viruses that may possess different features from those of the original viruses. Along these lines, recent studies showed that a monoreassortant virus with Seg-10 of BTV-8 in the backbone of BTV-1 (Seg-1 to Seg-9) induces the formation of plaques smaller than those of wild-type BTV-1, and that NS3 is one of the primary determinants of BTV-8 virulence in IFN-α/β receptor knockout (IFNAR^−/−) mice (59, 60). Together, these data raise the question of whether genetically distant Seg-10 proteins alter BTV properties such as viral replication, cytopathogenicity, virulence, and/or vector competence. In this study, we investigated the role of BTV Seg-10 and encoded NS3 proteins in modulating BTV-host interactions. To this end, BTV reassortant viruses were generated by reverse genetics, harboring different Seg-10 derived from the three genetically distant BTV topotypes. The results of this study show that NS3 proteins from different BTV can significantly influence its replication kinetics in ovine but not in Culicoides cells. In addition, our data reveal that a single amino acid residue leads to a rapid and host-specific turnover of NS3 proteins in ovine cells and to the attenuation of BTV virulence in mice.

MATERIALS AND METHODS

Aortic endothelial cell isolation and culture.

Ovine aortic endothelial cells (OvEC) were isolated and characterized as described previously (61). Briefly, aortas from freshly killed animals were cleaned and placed, intima layer down, into collagenase (2 mg/ml in Dulbecco's modified Eagles medium [DMEM]) (Sigma-Aldrich) for 1 h at 37°C. After incubation, the endothelial cells were removed by scraping and seeded in 6-well plates. The cells were maintained at 37°C, 5% CO₂, and 3% O₂ in DMEM supplemented with 10% fetal bovine serum (FBS), 1:100 dilution of human large vessel endothelial cell growth supplement (no. A14608-01; Life Technologies), 25 μg/ml penicillin-streptomycin solution (p/s), and 50 ng/ml amphotericin B. Cells were confirmed as endothelial cells by light microscopy and only passaged once before experimental infection.

Cell cultures.

BSR cells (cloned from baby hamster kidney cells [BHK-21] kindly provided by Karl Conzelmann) were grown in DMEM. CPT-Tert is an immortalized cell line of sheep choroid plexus cells (62), and they were grown in Iscove's modified Dulbecco's medium (IMDM). Mammalian cell lines were cultivated at 37°C in a 5% CO₂ humidified atmosphere. KC cells (established from Culicoides sonorensis embryos) were grown in Schneider's insect medium and incubated at 28°C (63). All cell lines were supplemented with 10% FBS and 25 μg/ml p/s.

Plasmids.

The plasmids used to rescue the BTV-1 strain have been described previously (20). Segment 10 of BTV-2RSA (GenBank accession number JN255931) and BTV-16RSA (GenBank accession number FJ713328), here referred to as Seg-10/-2 and Seg-10/-16, respectively, were synthesized and cloned into the pUC57 plasmid by GenScript. Seg-10/-2 and Seg-10/-16, encoding the same NS3 amino acid proteins as Seg-10/-1 and being referred to as Seg-10/-2AA and Seg-10/-16AA, respectively, also were synthesized by GenScript. Seg-10/-2PL was derived from Seg-10/-2 plasmid by site-directed mutagenesis using the QuikChange kit (Stratagene) as suggested by the manufacturer to exchange the proline (P24) with a leucine (L24) within the NS3 protein of Seg-10/-2. Each of these plasmids contains a single BTV segment flanked at the 5′ end by a T7 promoter and at the 3′ end by a restriction site (SapI or BsaI) to allow linearization and the in vitro transcription of BTV-like capped RNA using the mMESSAGE mMACHINE T7 Ultra transcription kit (Ambion) by following the manufacturer's instructions.

Reverse genetics.

Segment 10 reassortant BTVs were rescued by reverse genetics as previously described (20, 64–66). Briefly, 2 × 10⁵ BSR cells grown in 12-wells plate were transfected twice with BTV RNAs using Lipofectamine 2000 (Invitrogen). A total of 1 × 10¹¹ RNA copies of each segment encoding VP1, VP3, VP4, NS1, NS2, and VP6 proteins was used for the first transfection. After 18 h, the cells were further transfected with all 10 BTV segments. Four hours after the second transfection, the medium of cells was removed and cells were overlaid with 2 ml of 2× minimal essential media (MEM) containing 1% agarose type VII (Sigma-Aldrich), 2% FBS, and 25 μg/ml p/s. The plate was incubated at 37°C until plaques appeared.

Virus stocks and titrations.

After the reverse genetics, individual BTV rescued clones were picked through the agarose overlay and resuspended in 500 μl of DMEM. Two hundred microliters of each rescued virus was added to F75 flasks of KC cells in a final volume of 20 ml. Five days postinfection (p.i.), the supernatants were collected and the virus titers were determined by standard plaque assays using serial 10-fold dilutions in CPT-Tert cells. Briefly, monolayers of CPT-Tert cells in 12-wells plates were infected for 2 h at 37°C using 10-fold dilutions of the virus stocks, the medium containing virus then was removed, and the cells were washed once with IMDM. Two milliliters of a semisolid overlay then was added (1.2% Avicel in 2× MEM), supplemented with 2% FBS and 25 μg/ml p/s, and the plate was incubated at 37°C for 3 days. Finally, the overlay was discarded and the cells were washed with phosphate-buffered saline (PBS) and stained with 0.2% crystal violet–3.7% formaldehyde–20% ethanol solution. Viral titers are expressed as PFU/ml. Three independent measurements of the diameters (in millimeters) of at least 70 plaques of each BTV Seg-10 reassortant virus (at 10⁻⁵- and 10⁻⁶-fold dilutions) were performed using Inkscape software and, as a reference, the known diameter of a 12-well plate's well (22.09 mm). The median values (Mdn) and interquartile ranges (IQR) were calculated, and statistical analysis was performed using a Wilcoxon nonparametric test and NCSS9 software. Viral stocks also were titrated using limiting-dilution assays in CPT-Tert cells as previously described (20, 67), and the viral titers are expressed as 50% tissue culture infective doses (TCID₅₀)/milliliter. Finally, to assess viral physical titers, BTV RNAs were extracted in triplicate from 100 μl of each viral stock with the QIAamp viral RNA minikit (Qiagen) and then pooled. Subsequently, RNA samples were 10-fold diluted to measure the quantification cycle (C_q) values in the exponential phase of the standard curve prior to use of 4.5 μl to perform quantitative reverse transcription-PCR (RT-qPCR) targeting Seg-2 of BTV with an Adiavet BTV type 1 kit (Adiagene) according to the manufacturer's instructions. Three independent RT-qPCR experiments were performed. The C_q values were determined as a cycle number at which fluorescence has increased above background, as defined by the manufacturer. In parallel, Seg-2 RT-qPCR efficiency was systematically determined by standard curves using 10-fold dilutions of the corresponding in vitro-transcribed RNA. Note that for each virus, several viral stocks from different reverse genetics were established, titrated, and used in this study.

Virus growth curves.

A total of 5 × 10⁴ OvEC, 7.5 × 10⁴ CPT-Tert, or 5 × 10⁵ KC cells were plated in 24-well plates 2 days prior to infection. Cells then were infected by BTV Seg-10 reassortant viruses at a multiplicity of infection (MOI) of 0.001 (CPT-Tert and KC) and 0.1 (OvEC) and incubated at 37°C or 28°C for 2 h. The viral inocula were kept for titration; the cells were washed three times with free media and incubated with 1 ml of the appropriate fresh growth medium (referred to as 0 h p.i.). One hundred microliters of supernatant was collected and replaced with fresh growth medium at different times from 0 h to 72 h p.i. for OvEC and CPT-Tert cells and 0 h to 96 h for KC cells. Inocula and supernatants were titrated by limiting-dilution assays (TCID₅₀/ml) in CPT-Tert as previously described (20, 67). Each experiment was performed at least twice in triplicate using two viral stocks produced from independent reverse genetics. Statistical analyses were performed at time points p.i. with viral titers equal to or above 2 log₁₀ (TCID₅₀/ml) with Kruskal-Wallis test and using the NCSS9 software.

Cytopathic effect assays.

A total of 5 × 10⁴ OvEC and 7.5 × 10⁴ CPT-Tert cells were plated in 24-well plates 2 days prior to infection. Cells were infected with the viruses at different MOIs (from 0.1 to 0.0001), incubated at 37°C for 2 h, and washed three times with fresh medium. Fresh growth medium was added, and 72 h p.i., the supernatants were removed and cells were stained by a crystal violet-formaldehyde solution. The Image-Pro Plus software (MediaCybernetics) was used to quantify in each well the percentage of the cell monolayer that was disrupted after BTV replication. Each experiment was performed at least twice in triplicate using two viral stocks produced from independent reverse genetics.

Intracellular and extracellular BTV viral titers and RNA levels.

A total of 7.5 × 10⁴ CPT-Tert cells were plated in 24-well plates 2 days before infection. The cells then were infected with BTV Seg-10 reassortant viruses at an MOI of 0.001 for 2 h. At 18 h p.i., before the appearance of CPE, the supernatants were harvested and stored at 4°C. The cells were disrupted by freeze-thawing and resuspended in 0.5 ml of IMDM. The supernatant and cellular fractions were clarified by low-speed centrifugation and the supernatants titrated by limiting-dilution assay. Each experiment was repeated three times. The ratios between intracellular and extracellular average viral titers then were calculated for each virus. In addition, to assess the intracellular and extracellular Seg-2 and Seg-10 RNA levels, the same experiments were performed, except that the supernatants and the cells were kept to extract BTV RNAs with the QIAamp viral RNA minikit (Qiagen) by following the manufacturer's instructions. A volume of 4.5 μl of each RNA sample then was used to perform Seg-2 and Seg-10 RT-qPCR in triplicate with an Adiavet BTV group + type 1 kit and by following the manufacturer's instructions (Adiagene). In parallel, RT-qPCR efficiencies were systematically determined for BTV-1 Seg-2 and each Seg-10 by standard curves using serial 10-fold dilutions of the corresponding in vitro-transcribed RNAs. The intracellular Seg-2/Seg-10 average C_q value ratios were calculated for each virus.

In vitro infection assays.

A total of 7.5 × 10⁴ CPT-Tert and 5 × 10⁵ KC cells were plated in 24-well plates 2 days prior to infection. CPT-Tert and KC cells were infected as described previously at an MOI of 0.01 and 0.001, respectively. After 18 h p.i. for CPT-Tert and 5 days p.i. for KC cells, the supernatants were removed and cells were collected in 100 μl of 1× Laemmli buffer for Western blot analysis. Each experiment was performed in triplicate. To analyze the NS3/NS3a turnover rates, CPT-Tert and KC cells were infected at an MOI of 0.1, and 18 or 48 h p.i., respectively, they were treated with cycloheximide (50 μg/ml) and harvested in 100 μl of 1× Laemmli buffer at 0 h (T0), 2 h (T2), and 4 h (T4) posttreatment for Western blot analysis. In parallel, the same experiments were performed by treating the cells with cycloheximide in combination with MG132 (10 μM) (which is a potent proteasome inhibitor that may also impede some lysosomal proteases [68, 69]) prior to collecting the cells 4 h posttreatment and treating them as mentioned above. Western blot signals for the expression of NS3a and NS1 were quantified using Image Studio Lite software. For each BTV Seg-10 reassortant virus, values of NS3a and NS1 obtained at T0 were arbitrarily assigned a value of 100 and compared to those yielded at the other time points posttreatment with cycloheximide. In addition, signals of NS3a obtained at T4 (with or without MG132 inhibitor) were normalized to those of NS1, and fold changes of NS3 levels upon MG132 treatment were calculated for each BTV Seg-10 reassortant virus. Each experiment was performed at least three times independently. Statistical analyses were performed with Kruskal-Wallis test and using NCSS9 software.

In vitro translation assays.

In vitro translation assays were performed using a one-step human coupled IVT kit according to the manufacturer's instructions (Thermo Scientific). Reactions were performed for 2.5 h at 30°C either without viral RNA (mock) or with 1 μg of Seg-10/-1 or Seg-10/-2 and 750 ng of BTV-1 Seg-7 in vitro-transcribed RNAs. Finally, 10 μl of 2× Laemmli buffer was added to 10 μl of each reaction mix for Western blot analysis of NS3, VP7, and α-tubulin proteins. Each experiment was performed independently three times with at least two different sets of in vitro-transcribed RNAs.

Transfection assays.

A total of 1 × 10⁵ CPT-Tert or BSR cells were plated in 24-well plates. The following day, 500 ng of different in vitro-transcribed Seg-10 RNAs (NS3/NS3a) and Seg-5 (NS1) or Seg-7 (VP7) of BTV-1 were cotransfected using Lipofectamine 2000. Cells were harvested at 24 h posttransfection in 100 μl of 1× Laemmli buffer for Western blot analysis. Each experiment was performed independently three times with at least two different sets of in vitro-transcribed RNAs.

Western blotting.

Proteins were separated by SDS-PAGE, transferred to nitrocellulose membranes (Bio-Rad), and incubated overnight at 4°C with polyclonal rabbit antibodies against BTV NS3 (1/8,000), NS1 (1/5,000), or VP7 (1/8,000) protein or monoclonal mouse antibody against α-tubulin (1/500) (Sigma-Aldrich). Membranes were exposed to the appropriate peroxidase-conjugated secondary antibodies and then developed by chemiluminescence using SuperSignal West Pico substrate (Thermo Scientific) followed by exposure to X-ray film (Blue Devil Interleaved; Genesee Scientific).

Immunofluorescence and confocal microscopy.

The day before infection, 3 × 10⁴ CPT-Tert cells were plated in 8-well glass chamber slides (LabTek) in 200 μl of growth medium. Cells then were infected with rBTV-1 or rBTV-2 at an MOI of 0.1 and incubated at 37°C. After 22 h p.i., cells were fixed with ice-cold methanol for 10 min and essentially processed as already described (70, 71). Briefly, cells were immunolabeled with polyclonal rabbit anti-NS3, and a goat anti-rabbit IgG conjugated to Alexa Fluor 488 (Molecular Probes) was used as a secondary antibody. Slides were mounted with medium containing DAPI (4′,6-diamidino-2-phenylindole; Vectashield, Vector Laboratories) and analyzed with a Leica TCS SP2 confocal microscope. Each set of experiments was repeated twice independently, and the NS3 protein distribution was assessed from at least 25 individual cells per experiment. Statistical analysis was performed with a χ² test to compare the intracellular NS3 distribution patterns of rBTV/-1 and rBTV/-2.

Animal experiments.

Animal experiments were carried out at the Virology Laboratory for Animal Health (Maisons-Alfort, France), which maintains a colony of transgenic mice deficient for the type I interferon receptor (IFNAR^−/−) used to determine the virulence of BTV Seg-10 reassortant viruses. Groups of adult mice (n ≥ 5 per group) were mock infected or inoculated intraperitoneally with 10 PFU of each virus (adjusted to 100 μl with PBS). Clinical signs and survival rates were examined daily until the end of the experiment at 14 days p.i. After 4 days, EDTA blood samples were collected and tested by BTV Seg-2 RT-qPCR using an Adiavet BTV type 1 kit (Adiagene). The presence of BTV-specific antibodies was detected by competition enzyme-linked immunosorbent assay (ELISA) 14 days p.i. according to the manufacturer's instructions (ID screen bluetongue competition; IDvet). Each experiment was performed at least twice independently for each virus. Kaplan-Meier survival plots of RT-qPCR C_q values and statistical analyses with a one-way analysis of variance (ANOVA) test were performed using NCSS9 software.

Culicoides infection assays.

Culicoides nubeculosus females were obtained from the colony maintained at the Pirbright laboratory (72). Three-day-old females (n = 20) were intrathoracically injected with 0.345 μl of BTV Seg-10 reassortant viruses (1 × 10⁶ PFU/ml) as previously described (73). After injection, Culicoides were incubated at 28°C for 5 days, and the surviving females were individually homogenized in a volume of 1 ml IMDM containing 10% FBS, 2.5 μg/ml amphotericin B, 100 U/ml nystatin, 50 μg/ml gentamicin, and 25 μg/ml p/s as previously described (74). Each homogenate then was titrated by limiting-dilution assays in CPT-Tert cells. Each experiment was performed at least twice using two different viral stocks. Statistical analyses were performed with a Wilcoxon nonparametric test using NCSS9 software.

Ethics statement.

All animal work was conducted according to the recommendations in the directive on the protection of animals used for experimental and other scientific purposes of the EU. The protocols were approved by the animal experiments Ethics Committees at ANSES in France (license number 12/04/11-5). All efforts were made to minimize suffering of animals.

RESULTS

Production and titration of Seg-10 BTV reassortant viruses by reverse genetics.

This study aimed at assessing whether the variability of Seg-10 sequences (nucleic acids and/or corresponding NS3 proteins) could alter BTV-host interactions in terms of viral replication, CPE, and/or virulence. As shown in Fig. 1A and previously demonstrated by others (53, 57), the majority of Seg-10 sequences cluster into three main topotypes: two Western (group 1 and group 2) and one Eastern (group 1) group. By reverse genetics, we generated BTV Seg-10 reassortant viruses harboring a BTV-1 backbone (i.e., BTV-1 RNA Seg-1 to Seg-9) and different Seg-10 from the three main topotypes: (i) BTV-1RSA (Western group 2), (ii) BTV-2RSA (Western group 1), and (iii) BTV-16RSA (Eastern group 1), here referred to as rBTV/-1, rBTV/-2, and rBTV/-16, respectively (Fig. 1B). Viral stocks (supernatants of infected KC cells) then were titrated both by quantitative RT-PCR (RT-qPCR), to assess BTV-1 Seg-2 RNA levels, and limiting-dilution or plaque assays in ovine choroid plexus cells (CPT-Tert). As shown in Fig. 1C to E, the three viral stocks rBTV/-1, rBTV/-2, and rBTV/-16 displayed similar amounts of viral particles as assessed by RT-qPCR and limiting-dilution and plaque assays. However, rBTV/-2 induced smaller plaques than rBTV/-1 and rBTV/-16 in CPT-Tert cells (Fig. 1E).

FIG 1.

Production and titration of Seg-10 BTV reassortant viruses by reverse genetics. (A) Seg-10 phylogenetic tree of different serotypes and strains of BTV. Sequences cluster into three major topotype groups, Western 1 and 2 and Eastern 1. The different Seg-10 selected for this study and their sequence references are indicated as BTV-1 (JX680466.1), BTV-2 (JN255931.1), and BTV-16 (FJ713328.1). The scale bar indicates the number of substitutions per site. (B) Strategy to produce, by reverse genetics, BTV reassortant viruses (rBTV/-1, rBTV/-2, and rBTV/-16) with the BTV-1 backbone (Seg-1 to -9) and a different Seg-10 (Seg-10/-1, Seg-10/-2, and Seg-10/-16). (C) Titers of viral stocks by RT-qPCR targeting the Seg-2 double-stranded RNA (dsRNA) of BTV-1. (D) Infectious titers of viral stocks by limiting-dilution assays on CPT-Tert cells. (E) Plaque assays of viral stocks on CPT-Tert cells. The diameters of the plaques (n > 70/virus) were measured in millimeters. The median values (Mdn), interquartile ranges (IQR), and significances (P < 10⁻⁶ by Wilcoxon rank-sum test) are indicated on the right side of the panel. Note that despite a similar infectious titer (≈5 × 10⁶ PFU/ml), rBTV/-2 induced smaller plaques than the two other Seg-10 reassortant viruses. Each of these experiments was repeated three times independently. Bars in panels C and D indicate standard errors.

Seg-10/NS3 modulates BTV replication kinetics in ovine cells but not in Culicoides.

Endothelial cells are primary sites of BTV replication in vivo (75). Thus, the replication kinetics and cytopathogenicity of the three BTV Seg-10 reassortants were investigated in primary ovine aortic endothelial cells (OvEC). rBTV/-1 reached significantly higher titers than rBTV/-16 in the supernatants of infected cells at 48 h and 72 h postinfection (p.i.), whereas rBTV/-2 displayed the lowest replication kinetics (Fig. 2A). Accordingly, rBTV/-2 induced less CPE in OvEC than rBTV/-1 and rBTV/-16 (Fig. 2B). As mentioned before, NS3 proteins mediate BTV egress but also interfere with the type I IFN synthesis in mammalian cells (45). Hence, replication kinetics of the three Seg-10 BTV reassortants next were assessed in ovine CPT-Tert cells, which are known to be deficient for type I IFN synthesis (61, 62). Similar to what was previously observed in OvEC cells, rBTV/-2 replicated less efficiently than rBTV/-1 and rBTV/-16; however, unlike in OvEC cells, the replication kinetics of the latter was higher than that of rBTV/-1 (Fig. 2C). Accordingly, rBTV/-2 and rBTV/-16 induced less or more severe CPE, respectively, than rBTV/-1 in CPT-Tert cells (Fig. 2D). Together, these data indicate that Seg-10/NS3 from different BTV strains differentially modulates replication kinetics and CPE in ovine cells. Indeed, rBTV/-2 replicates less efficiently than rBTV/-1 and rBTV/-16 in the two ovine cell lines used in this study, whereas the replication kinetics of rBTV/-16 is cell type dependent (i.e., lower or higher than that of rBTV/-1 in OvEC or CPT-Tert cells, respectively). As BTV is a Culicoides-borne virus, we questioned whether the different Seg-10/NS3 also could modulate BTV replication kinetics in insects belonging to this genus. To this end, growth curve experiments of the three Seg-10 reassortant BTV viruses were performed in KC Culicoides cells, and as shown in Fig. 2E, all of them replicated to a similar level. These results also were confirmed in vivo by intrathoracical injection of rBTV/-1, rBTV/-2, and rBTV/-16 in C. nubeculosus females. Indeed, no statistically significant difference was observed in the viral titers of the three Seg-10 reassortant viruses from infected Culicoides at day 5 p.i. (P > 0.05 by Wilcoxon rank-sum test) (Fig. 2F). Overall, these data show that the three Seg-10 analyzed in this study alter BTV replication kinetics in a host-specific manner.

FIG 2.

Seg-10/NS3 modulates BTV replication kinetics in ovine cells but not in Culicoides. Growth curve experiments were performed in OvEC (A), CPT-Tert (C), and KC Culicoides cells (E) with the three BTV Seg-10 reassortant viruses at the indicated MOI. Inocula and cell supernatants were collected from 0 h to 72 h p.i., and the viral titers were obtained by limiting-dilution assays in CPT-Tert cells. Each experiment was performed in triplicate, and bars indicate the standard errors. Dashed lines indicate the threshold of virus detection [1.5 log₁₀ (TCID₅₀/ml)]. Kruskal-Wallis statistical analyses were performed, and significance (compared to rBTV/-1 data) is presented: P < 0.05 (*), P < 0.01 (**), and P < 0.001 (***). Cytopathic effect assays in OvEC (B) and CPT-Tert (D) cells were performed with the three viruses at different MOI as indicated on the side of each panel. CPE was quantified at the most relevant MOI (indicated in boldface; e.g., 0.01 for OvEC and 0.001 for CPT-Tert) and expressed as the percentage of cell layer destroyed by each virus (% CPE). Each experiment was performed in triplicate. (F) C. nubeculosus females were intrathoracically injected with the different BTV Seg-10 reassortant viruses. Two sets of experiments (rBTV-1 versus rBTV-2 and rBTV-1 versus rBTV-16) were performed at least twice independently. At 5 days p.i., midges were individually homogenized and titrated by limiting-dilution assays in CPT-Tert cells. No significant difference was observed between the viral titers of the three Seg-10 reassortant viruses (P > 0.05 by Wilcoxon tests).

NS3 proteins per se modulate BTV replication kinetics and cytopathogenicity.

As mentioned before, NS3 proteins play a major role in BTV particle release (18). Thus, we wondered whether the variations observed in the replication kinetics of rBTV/-1, rBTV/-2, and rBTV/-16 in ovine cells were due to differences in the efficiency of viral egress. As shown in Fig. 3A, the intracellular and extracellular titers of rBTV/-2 were lower than those of rBTV/-1 and rBTV/-16 in infected CPT-Tert cells at 18 h p.i. (before the appearance of CPE). Moreover, no significant differences were observed in the intracellular/extracellular ratio of the titers in cells infected with the three Seg-10 BTV reassortants (P > 0.05 by Kruskal-Wallis test) (Fig. 3B), suggesting that these viruses did not display major intrinsic differences in viral particle release efficiency. At the same time, RT-qPCR experiments were carried out to assess the intracellular and extracellular levels of Seg-10 and Seg-2 RNAs for each virus. In accordance with data described before, the extracellular quantitation cycle (C_q) values of rBTV/-2 were significantly higher than those of the other viruses (Fig. 3C); however, the intracellular Seg-2/Seg-10 average C_q value ratios were comparable (P > 0.05 by Kruskal-Wallis test) (Fig. 3D). Therefore, the Seg-10 RNA levels of each BTV reassortant virus were proportionally similar to their RNA Seg-2 levels, suggesting that the intracellular stability and replication efficiency of each RNA Seg-10 are comparable.

FIG 3.

BTV Seg-10 reassortant viruses do not display differences of viral egress efficiency or Seg-10 RNA stability. (A) CPT-Tert cells were infected with rBTV/-1, rBTV/-2, and rBTV/-16 at an MOI of 0.001. At 18 h p.i., supernatants were collected and cells were disrupted by freeze-thawing. Intracellular and extracellular infectious titers were obtained by limiting-dilution assays in CPT-Tert cells [log₁₀ (TCID₅₀/ml)]. One-way ANOVA was performed, and significance (compared to rBTV/-1 data) is presented as P < 0.05 (*), P < 0.01 (**), and P < 0.001 (***). (B) Ratios of intracellular versus extracellular average viral titers were calculated for each BTV Seg-10 reassortant virus. Statistical analyses were performed with a Kruskal-Wallis test. No significant difference was observed in the intracellular/extracellular ratio of the titers in cells infected with the three Seg-10 BTV reassortants (P > 0.05). (C) CPT-Tert cells were infected at an MOI of 0.001. At 18 h p.i., BTV RNAs were extracted from the cells and the supernatants, and the intracellular and extracellular Seg-2 and Seg-10 RNA levels were estimated by RT-qPCR. One-way ANOVA was performed, and significance (compared to rBTV/-1 data) is presented as P < 0.05 (*) and P < 0.001 (***). (D) The intracellular Seg-2/Seg-10 average C_q value ratios were calculated for each BTV Seg-10 reassortant virus. Each experiment was performed in triplicate, and bars indicate standard errors. Statistical analyses were performed with a Kruskal-Wallis test. No significant difference was observed in the intracellular Seg-2/Seg-10 average C_q value ratios in cells infected with the three Seg-10 BTV reassortants (P > 0.05).

The NS3 proteins encoded by the different Seg-10 used in this study are 94.3 to 98.3% identical but display some differences in their amino acid sequences (Fig. 4A). Thus, to assess the effect of these NS3 amino acid variations, Seg-10/-2 and Seg-10/-16, encoding NS3 proteins with strictly identical amino acid residues of Seg-10/-1 (named Seg-10/-2AA and Seg-10/-16AA, respectively), were synthesized by GenScript, and the corresponding viruses, rBTV/-2AA and rBTV/-16AA, were produced by reverse genetics. As shown in Fig. 4B and C, rBTV/-2AA developed larger plaques and induced more severe CPE in CPT-Tert cells than rBTV/-2. In addition, rBTV/-2AA reached higher titers than rBTV/-2 and displayed a phenotype similar to that of rBTV/-1 in CPT-Tert and OvEC cells (Fig. 4D and data not shown). On the other hand, rBTV/-16AA induced smaller plaques and less CPE in CPT-Tert cells than rBTV/-16 (Fig. 4B and C). The replication kinetics of rBTV/-16AA was slightly lower than that of rBTV/-16 during the first 24 h p.i. but caught up with it at later time points (Fig. 4D).

FIG 4.

Role of NS3 amino acids in modulating BTV in vitro characteristics. (A) Alignment of NS3 protein sequences encoded by Seg-10 of BTV-1, -2, and -16. Interaction domains with calpactin (P11-BD) and BTV VP2 (VP2-BD) and the two late domains (L-Domains) and the viroporin-associated region are represented in boxes. The two transmembrane (TM) and the extracellular (EC) domains are also indicated. G, predicted glycosylation sites. The boldfaced and underlined amino acids in red are unique to one sequence, those in blue are common to two sequences, and those in green are the residues different among the three of them. (B) Infectious titers of viral stocks by limiting-dilution assays in CPT-Tert cells. Seg-10/-2 and Seg-10/-16 mutated to encode NS3 proteins with identical amino acid residues of Seg-10/-1 were used to produce by reverse genetics rBTV/-2AA and rBTV/-16AA, respectively. Representative pictures of plaque assays in CPT-Tert cells with the five BTV Seg-10 reassortant viruses are presented. The diameters of the plaques (n > 70/virus) were measured in millimeters. The median values (Mdn), interquartile ranges (IQR), and significance (P < 10⁻⁶ or P < 10⁻³ by Wilcoxon sum-of-rank test) are indicated on the right side of the panel. (C) Cytopathic effect assays in CPT-Tert cells were performed with the five viruses at the different MOI indicated on the side of each panel. CPE was quantified at the most relevant MOI (indicated in boldface; e.g., 0.01) and expressed as the percentage of cell layer destroyed by each virus (% CPE). Each experiment was performed in triplicate. (D and E) Growth curve experiments were performed in CPT-Tert cells (D) and KC Culicoides cells (E) with the five BTV Seg-10 reassortant viruses at the indicated MOI. Inocula and cell supernatants were collected from 0 h to 72 h (CPT-Tert) or 96 h (KC) p.i., and the viral titers were obtained by limiting-dilution assays in CPT-Tert cells. Each experiment was performed in triplicate, and bars indicate the standard errors. Dashed lines indicate the threshold of virus detection [1.5 log₁₀ (TCID₅₀/ml)]. Kruskal-Wallis statistical analyses were performed, and significance (compared to rBTV/-1 data) is presented as follows: P < 0.05 (*), P < 0.01 (**), and P < 0.001 (***).

Finally, and as previously observed, despite some weakly significant differences at 48 h p.i., all of the viruses exhibited less variable replication kinetics in KC than in ovine cells, suggesting once again that the influence of the different NS3 proteins on BTV replication occurs mainly in ovine cells and, thus, in a host-specific manner (Fig. 4D and E). Together, these results indicate that the amino acid sequences of the different NS3 proteins are key determinants in modulating BTV replication kinetics and cytopathogenicity in ovine cells.

Seg-10/-2 yields a low steady-state level of NS3 proteins.

We next analyzed by Western blotting the levels of NS3 proteins produced by the different Seg-10 BTV reassortant viruses in cell lysates of both CPT-Tert- and KC-infected cells. In mammal cells, NS3 proteins typically are detected by Western blotting as several bands that correspond to (i) the native NS3 and NS3a proteins (26 kDa and 24 kDa, respectively), which are partially glycosylated into two main forms, NS3-G and NS3a-G (29 kDa and 27 kDa, respectively), and (ii) a smear of high-molecular-mass forms (from 30 kDa to 39 kDa) that represents different carbohydrate groups that are covalently attached to NS3/NS3a glycoproteins (33, 40). Notably, unlike in mammal cells, the glycosylated forms of NS3/NS3a are not present in insect cells due to differences in the glycosylation pathways between these two animal phyla (76). As shown in Fig. 5A, rBTV/-2 yielded a smaller amount of native and glycosylated forms of NS3 and NS3a proteins than rBTV-1 and rBTV-16 in CPT-Tert cells, whereas, interestingly, such differences were not observed in KC cells (Fig. 5B). However, in the latter cells, rBTV/-1 appeared to produce fewer NS3 proteins than rBTV/-2, most probably due to its amino acid sequences, as rBTV/-2AA displayed levels of NS3 proteins similar to those of rBTV-1 (Fig. 5B). The small amount of NS3 proteins detected in rBTV/-2-infected CPT-Tert cells could be explained by the low replication kinetics of this virus; however, among the Seg-10 reassortant viruses, this variation was less evident for the NS1 proteins. Thus, in order to assess whether the difference in NS3/NS3a levels was due to Seg-10/-2 per se (and not solely to the low replication kinetics of this virus), CPT-Tert cells were cotransfected with the different Seg-10 RNAs (i.e., Seg-10/-1, Seg-10/-2, or Seg-10/-2AA) and BTV-1 Seg-5, encoding NS1, which is known to transactivate viral protein translation (27). Importantly, by Western blot analyses, Seg-10/-2 yielded a lower level of native and glycosylated forms of NS3 and NS3a proteins than Seg-10/-1 and Seg-10/-2AA, whereas the amount of NS1 proteins was comparable under each condition (Fig. 5C). A previous study showed that the wild-type BTV-2RSA strain (BTV-2RSA), even if was more virulent in type I IFN receptor knockout mice, replicated to a lower level than its corresponding vaccine strain (BTV-2RSAVacc) in BSR cells (67). Interestingly, the NS3 protein sequences of BTV-2RSA and BTV-2RSAVacc differ by only one amino acid residue at position 24, a proline (P24) in BTV-2RSA and a leucine (L24) in BTV-2RSAVacc. The NS3 proteins encoded by Seg-10/-1 and Seg-10/-16 also possess a leucine residue at position 24 (Fig. 4A). Therefore, we reasoned that the P24 residue could be responsible for the small amount of NS3 proteins yielded from Seg-10/-2. Thus, we generated a Seg-10/-2 with a P24L mutation (here referred to as Seg-10/-2PL). As shown in Fig. 5C, Seg-10/-2PL produced amounts of NS3 and NS3a proteins similar to those of Seg-10/-1 or Seg-10/-2AA, demonstrating that a single amino acid residue (P24) is responsible for the low steady-state level of NS3/NS3a yielded by Seg-10/-2, a defect that is fully rescued by changing P24 to L24.

FIG 5.

Western blotting of NS3 proteins produced from the different BTV Seg-10 reassortant viruses. (A and B) Infectious assays with rBTV/-1, -2, -2AA, -16, and -16AA were performed in CPT-Tert (A) and KC (B) cells at the indicated MOI. Cells were collected at 18 h (CPT-Tert) and 5 days (KC) p.i. and analyzed by Western blotting with antisera raised against NS3, NS1, and α-tubulin as specified. NS3 and NS3a glycosylated forms are indicated as NS3-G and NS3a-G, respectively. An asterisk indicates a nonspecific band. Each experiment was performed three times independently, and representative blots are presented. (C) The proline residue at position 24 of NS3/NS3a encoded by Seg-10/-2 was changed by mutagenesis to a leucine to generate Seg-10/-2PL. CPT-Tert cells were transfected with 500 ng of the indicated Seg-10 (NS3) and BTV-1 Seg-5 (NS1) in vitro-transcribed RNAs. Twenty-four hours posttransfection, cells were collected and Western blot analyses were performed with the appropriate antisera as indicated in each panel. Note that this experiment was performed independently three times with two different sets of RNAs, and a representative blot is shown.

NS3 proteins of rBTV/-2 display a rapid turnover rate and are more sensitive to protein degradation in ovine but not in Culicoides cells.

We next investigated the reasons which may explain the smaller amount of NS3 proteins produced in mammalian cells by Seg-10 of BTV-2RSA compared to that of BTV-1RSA. Initially, we assessed whether NS1 could differentially transactivate the translation efficiency of Seg-10/-1 and Seg-10/-2. To this end, BSR cells were transfected with Seg-10/-1 or Seg-10/-2 RNAs in the presence or absence of BTV-1 Seg-5 (encoding NS1). These experiments were performed in BSR cells, as no NS3 signal was detected in CPT-Tert cells in the absence of Seg-5/NS1. The translation of Seg-10/-2 was transactivated by NS1 (Fig. 6A, i) and yielded less NS3 and NS3a proteins than Seg-10/-1, both in the presence or absence of NS1 (Fig. 6A, ii), ruling out the hypothesis that NS1 transactivation is responsible for the small amount of NS3 proteins. Thus, we next checked whether the translation efficiency of Seg-10/-1 and Seg-10/-2 could differ. As shown in Fig. 6B, in vitro HeLa cell-free translation assays failed to show any difference between the amounts of NS3/NS3a translated from the Seg-10/-1 and Seg-10/-2 RNAs. The intracellular distribution of NS3 proteins next was assessed by immunofluorescent assays. Although NS3 proteins were occasionally detected at the plasma membrane of BTV-infected CPT-Tert cells, two main different patterns of NS3 staining were observed, as revealed by confocal microscopy: dispersed in the cytoplasm and concentrated in a perinuclear region (Fig. 6C, i). However, no significant difference of NS3 distribution was observed in rBTV-1- and rBTV-2-infected CPT-Tert cells (P = 0.7140 by χ² test) (Fig. 6C, ii).

FIG 6.

Low level of NS3 yielded by Seg-10/-2 is not related to the intracellular distribution of NS3, the translation efficiency, or the transactivation by NS1. (A) BSR cells were transfected with 500 ng of Seg-10 (NS3/NS3a) and the Seg-7 (VP7) or Seg-5 (NS1) in vitro-transcribed RNAs. Twenty-four hours posttransfection, cells were collected and Western blot analyses were performed with the appropriate antisera as indicated in each panel. Samples were loaded on separate blots with different exposure times to better appreciate NS3 protein levels under each condition. (i) NS1 protein was able to transactivate the NS3/NS3a translation of Seg-10/-1 and Seg-10/-2. (ii) Seg-10/-2 yielded less NS3/NS3a than Seg-10/-1 in the absence (Seg-7) or presence (Seg-5) of NS1 protein expression. An asterisk indicates nonspecific bands. NS3 and NS3a glycosylated forms are indicated as NS3-G and NS3a-G, respectively. Each experiment was performed independently in triplicate with two different sets of RNAs. (B) In vitro translation assays (HeLa lysates) were performed with 1 μg of Seg-10/-1 or Seg-10/-2 and 750 ng of BTV-1 Seg-7 in vitro-transcribed RNAs. Blots were incubated with the appropriate antisera as indicated in each panel. NS3 and NS3a glycosylated forms are indicated as NS3-G and NS3a-G, respectively. (C) CPT-Tert cells infected with rBTV/-1 and rBTV/-2 were analyzed by immunofluorescent assays at 22 h p.i. (i) The intracellular distribution of NS3 was scored as dispersed (Disp.) or concentrated (Conc.) using confocal microscopy. The scale bar represents 10 μm. (ii) The graph represents the number (percent) of cells in which the intracellular distribution of NS3 proteins displays a dispersed or concentrated staining pattern. At least 25 cells in random fields from two independent experiments were counted.

We then tested if NS3/NS3a turnover rate variations could be involved in such differences. First, Seg-10/-2PL was used to generate the corresponding rBTV/-2 mutant virus in a BTV-1 backbone (here referred to as rBTV/-2PL). CPT-Tert or KC cells then were infected with rBTV/-1, rBTV/-2, and rBTV/-2PL for 18 h or 48 h, respectively, and treated with cycloheximide to arrest the global translation and assess over time the quantity of NS3 proteins (from 0 to 4 h posttreatment). The effect of cellular protein degradation machinery on NS3/NS3a amounts also was evaluated using MG132, a potent proteasomal degradation inhibitor that also may impede some lysosomal proteases (68, 69). As shown in Fig. 7A, in infected CPT-Tert cells, the rBTV/-2 NS3 proteins displayed a higher relative turnover rate than those of rBTV/-1, which is stabilized in rBTV/-2PL (P < 0.01 at T2 by Kruskal-Wallis test). Surprisingly, the speed of decay of rBTV/-2PL NS3 proteins appeared even slower than those of rBTV/-1 (Fig. 7A, i and iii) (P < 0.01 at T2 by Kruskal-Wallis test). Most interestingly, by inhibiting the protein degradation machinery of infected CPT-Tert cells with MG132 inhibitor, rBTV/-2 NS3 proteins level significantly increased 8.7-fold, whereas those of Seg-10/-1 and Seg-10/-2PL only displayed 2.3 and 1.4-fold increases, respectively (Fig. 7A, i to iv) (P < 0.05 by Kruskal-Wallis test). Previously, we showed that rBTV/-2 yields fewer NS3 proteins in infected CPT-Tert cells (Fig. 5A); however, in this experiment, the amounts of NS3 produced by the three viruses appear similar at T0, because they were analyzed on separate Western blots with different exposure times (Fig. 7A, i to iii). The same experiments were performed in KC cells, and while the NS3 proteins of all of these viruses also were subjected to intracellular proteolysis, no significant differences in their decay rates or sensitivity to protein degradation inhibition were observed between the different viruses (P > 0.05 by Kruskal-Wallis test) (Fig. 7B). Notably, in the presence of the MG132 inhibitor, a third and unknown NS3-related protein appeared at about 18 kDa (below NS3a) in KC cells (Fig. 7B). Together, these data show that NS3 proteins are targeted for degradation both in CPT-Tert and KC cells, whereas their relative turnover rates vary in ovine but not in Culicoides cells. Moreover, NS3 proteins of rBTV/-2 appear more sensitive to MG132 inhibitor than those of rBTV/-1 and rBTV/-2PL in CPT-Tert but not in KC cells, suggesting that P24 leads to a more efficient and host-specific degradation of BTV NS3 proteins.

FIG 7.

Effect of P24 on NS3/NS3a turnover rate and cellular protein degradation in ovine and Culicoides cells. (A and B) The CPT-Tert (A) and KC (B) cells were infected at an MOI of 0.1 with the indicated BTV Seg-10 reassortant viruses. Eighteen hours (CPT-Tert) or 48 h (KC) p.i., the cells were treated with cycloheximide (Cyclo) alone or in combination with MG132 (a proteasome and lysosome inhibitor). Cycloheximide-treated cells then were collected for Western blot analyses at 0 h (T0), 2 h (T2), and 4 h (T4) posttreatment, whereas cycloheximide- and MG132-treated cells were collected only at 4 h (T4) posttreatment. Western blotting of CPT-Tert and KC cell lysates infected with rBTV/-1 (i), -2 (ii), and -2PL (iii) using antisera raised against NS3, NS1, and α-tubulin, as indicated, are shown. NS3 and NS3a glycosylated forms are indicated as NS3-G and NS3a-G, respectively. An asterisk indicates nonspecific bands. A question mark indicates an unknown NS3-related protein that appeared at about 18 kDa in KC-infected cells treated with MG132. Signals of NS3a and NS1 proteins were quantified from three experiments using Image Lite Studio software, and the average values obtained are presented below each panel (% NS3a and % NS1). Values represent arbitrary units relative to the values of NS3a and NS1 signals at T0 (which were assigned a value of 100%). Representative blots from three independent experiments are shown. (iv) Dot plots represent the relative fold changes of NS3a levels upon MG132 treatment of CPT-Tert and KC cells infected with rBTV/-1, rBTV/-2, or rBTV/-2PL. Horizontal bars represent the mean values of the data obtained. Statistical analyses were performed by Kruskal-Wallis test. P values indicate a significant difference (P < 0.05) between NS3a fold change of rBTV/-2 and those of rBTV/-1 and rBTV/-2PL in CPT-Tert but not in KC cells.

P24 of NS3 proteins reduces the replication kinetics and attenuates the virulence of rBTV/-2.

Finally, the influence of P24 on BTV characteristics was investigated. rBTV/-2PL formed larger plaques than rBTV/-2 (Fig. 8A), induced CPE, and reached titers comparable to those of rBTV/-1 in CPT-Tert cells (Fig. 8B and C, respectively). As previously shown, rBTV/-2 displayed reduced replication kinetics compared to those of the two other BTV reassortant viruses, both in IFN-competent and -deficient ovine cells. Hence, IFNAR^−/− mice represent a relevant experimental model to assess the effect of P24 on in vivo properties of BTV (67, 77, 78). IFNAR^−/− mice were intraperitoneally inoculated with the three Seg-10 reassortant BTV viruses, rBTV/-1, rBTV/-2, and rBTV/-2PL, and their clinical signs and survival rates was observed daily up to 14 days p.i. As shown in Fig. 8D, 100% of mice infected with rBTV/-1 and rBTV/-2PL died between 6 and 7 days p.i.; on the other hand, all of the mice (n = 14) inoculated with rBTV/-2 survived throughout the duration of the experiment (Fig. 8D). At 4 days p.i., RT-qPCR assays were performed on viral RNA extracted from the blood of infected mice. As shown in Fig. 8E, the average C_q values observed for rBTV/-2-infected mice were significantly lower than those of rBTV/-1 and rBTV/-2PL (C_q of 32.5, 27.1, and 27.6, respectively; P < 0.01 by one-way ANOVA test). As an additional control, VP7 competitive ELISAs were performed at 14 days p.i. from the blood of mock- and rBTV/-2-inoculated mice, showing that these animals developed an immune response against BTV (Fig. 8F). Overall, these data demonstrate that P24 of the NS3 proteins leads to low replication kinetics of rBTV/-2 and its virulence attenuation in IFNAR^−/− mice.

FIG 8.

Impact of P24 on BTV in vitro characteristics and virulence in IFNAR^−/− mice. (A) Infectious titers of viral stocks by limiting-dilution assays in CPT-Tert cells. Seg-10/-2PL was used to produce rBTV/-2PL by reverse genetics. Representative pictures of plaque assays in CPT-Tert cells with the rBTV/-1, rBTV/-2, and rBTV/-2PL viruses are shown. The diameters of the plaques (n > 70/virus) were measured in millimeters. The median values (Mdn), interquartile ranges (IQR), and significance (P < 10⁻⁶ by Wilcoxon rank-sum test) are indicated on the right side of the panel. Note that rBTV/-2PL induced larger plaques than rBTV/-2. (B) Cytopathic effect assays in CPT-Tert cells were performed with the three Seg-10 reassortant viruses at different MOI as indicated on the side of each panel. CPE was quantified at the most relevant MOI (indicated in boldface; e.g., 0.01) and expressed as the percentage of cell layer destroyed by each virus (% CPE). Each experiment was performed in triplicate. (C) CPE assay supernatants (MOI, 0.001) were collected at 3 days p.i. and titrated by limiting dilution in CPT-Tert cells. The graph represents viral titers of each virus as indicated. Bars indicate the standard errors. One-way ANOVA was performed: *, P < 0.05; ***, P < 0.001. (D) Kaplan-Meier survival plots of IFNAR^−/− mice (n ≥ 10) intraperitoneally inoculated with the indicated BTV Seg-10 reassortant viruses (10 PFU). Note that, unlike rBTV/-2, rBTV/-2PL killed 100% of the inoculated animals at 7 days p.i. (E) Graph representing the average C_q values of Seg-2 RNAs at 4 days p.i. with the three BTV Seg-10 reassortant viruses. Bars indicate the standard errors. One-way ANOVA was performed: ***, P < 0.001. (F) The presence of BTV VP7-specific antibodies in mock- and rBTV/-2-infected mice (n = 10 per condition) was detected by competition ELISAs 14 days p.i. The graph represents the frequency distribution of ELISA results (expressed as percentages of negativity) from serum samples of mock- or rBTV/-2-infected mice. Positive, doubtful, or negative were assigned according to the cutoff values recommended by the manufacturer (≤35%, positive; 35% to ≤45%, doubtful; >45%, negative).

DISCUSSION

BTV Seg-10 encodes two isoforms of a nonstructural protein, NS3 and NS3a, which play critical roles in BTV infection (18, 46). Seg-10 sequences display some genetic variability between BTV isolates, and coinfection of a single host with two different BTV strains may lead, by reassortment, to viral progeny with different biological properties than those of the original BTV. Thus, this study was aimed at assessing whether genetically distant Seg-10 and their corresponding NS3 proteins could affect BTV host interactions in terms of replication, cytopathogenicity, or virulence. To this end, a reverse genetics system was used to swap BTV-1RSA Seg-10 with that of two other serotypes, BTV-2RSA or BTV-16RSA. Remarkably, the NS3 proteins encoded by BTV-2RSA significantly reduced BTV replication kinetics in two ovine cell lines but not in Culicoides. This host-dependent effect of rBTV/-2 NS3 was due to a single amino acid residue (P24), which leads to a low steady-state level of NS3 and NS3a isoforms compared to those of rBTV/-1 and rBTV/-2PL in ovine but not in Culicoides cells. Our results further show that the P24 residue of BTV-2RSA NS3 in the BTV-1 backbone significantly reduced BTV titers 4 days p.i. and abrogated its lethality in IFNAR^−/− mice. Previously, Caporale and colleagues demonstrated that the BTV-2RSA strain (harboring P24 in its NS3 proteins) replicated to lower titers in BSR cells than its corresponding vaccine strain, BTV-2RSAVacc, bearing L24 (67). However, BTV-2RSA is fully virulent in IFNAR^−/− mice, whereas BTV-2RSAVacc is not. Altogether, these data and those obtained in this study suggest that the amino acid at position 24 (P24 or L24) also influences BTV replication kinetics in its natural backbone, but the difference of virulence between BTV-2RSA and BTV-2RSAVacc may be due to additional viral factors that could compensate for the replication kinetics defect induced by P24 in NS3 proteins. Accordingly, BTV pathogenicity most likely is multifactorial both in sheep and experimental mouse models (59, 67, 79). A recent study revealed that NS3 is one of the primary determinants of BTV-8 virulence in IFNAR^−/− mice, along with VP2, and that several other BTV viral proteins also contribute to its pathogenicity (59). Hence, the results obtained by placing the Seg-10 RNAs in their original backbone may have different outcomes due to a preferred and more adapted BTV genome constellation. Nonetheless, our study unveils at least one way adopted by NS3 proteins to indirectly contribute to BTV pathogenicity by modulating its replication kinetics. Notably, a leucine residue is found at position 24 of the majority of BTV NS3 sequences available in the database, but among those belonging to the Western group 1 topotype, P24 is not unique to BTV-2RSA. Indeed, P24 also is present in the NS3 protein sequences of at least three BTV-2 (isolated from India or Italy) and two viruses of other serotypes (BTV-11 from the United States and BTV-12 from Jamaica) (51, 80–82). Interestingly, NS3/NS3a bearing P24 instead of L24 appears very unstable in ovine but not in Culicoides cells, most likely due to a more efficient and host-specific protein degradation. Typically, ubiquitination on certain lysine residue(s) is a posttranslational modification that is important to regulate protein turnover through the cell's protein degradation machinery (83). Interestingly, in silico ubiquitin prediction analyses (www.ubpred.org/) revealed that BTV NS3 proteins display putative ubiquitination on two lysine residues at positions 13 and 15 (K13 and K15). Thus, although other viral factors come into play to determine BTV virulence (59, 67), mono- and/or polyubiquitination of NS3 proteins may have a key role in differentially modulating their intracellular fate and stability in a host-specific manner and, consequently, BTV replication kinetics and its related pathogenicity in mammals. Notably, a recent study showed that NS3 proteins are not essential to intracellular BTV RNA replication in BSR cells (42). Hence, according to this study and our data, the small amount of NS3 proteins yielded by Seg-10 of BTV-2RSA is unlikely to affect viral RNA replication per se but rather would decrease viral egress and, ultimately, BTV propagation and virulence. Future experiments to assess the ubiquitin status of these different BTV NS3 proteins and, more generally, the role of ubiquitination and protein degradation pathways during BTV infection will certainly help to understand the molecular bases of their intracellular degradation and stability and the difference observed between mammals and insects.

In this study, depending on the ovine cell types that were used, rBTV/-16 replication kinetics varied compared to that of rBTV/-1. Indeed, rBTV/-16 reached titers slightly lower than those of rBTV/-1 in IFN-competent primary OvEC cells while it was higher in CPT-Tert cells (deficient for type I IFN synthesis) (61, 62). Interestingly, NS3 proteins recently have been shown to interfere in vitro with the type I IFN synthesis (45); thus, we reason that this activity could account for the different replication kinetics exhibited by rBTV/-1 and rBTV/-16 in OvEC cells. On the contrary, in CPT-Tert cells, rBTV/-16 replicated at higher titers than rBTV/-1, and the substitution of the NS3 amino acids of BTV-16RSA for those of BTV-1RSA (i.e., Seg-10/-16AA) reduced only slightly BTV replication kinetics and solely at early time points p.i. (from 12 h to 24 h p.i.), suggesting an effect of the RNA Seg-10 sequences per se. However, a more drastic effect was observed on CPE, with rBTV/16AA inducing less CPE than rBTV/-16 in CPT-Tert cells. NS3 proteins display a viroporin activity that may facilitate viral egress at early time points and contribute to the CPE in mammalian cells (34). Therefore, it also is possible that the NS3 proteins encoded by BTV-16RSA display stronger viroporin activity than those of BTV-1RSA. Additional experiments to unravel the mechanisms adopted by the different NS3 to inhibit type I IFN synthesis and better characterize their viroporin activity will surely shed some light on the reasons why rBTV/-16 replication kinetics is ovine cell type dependent. Notably, as for rBTV-2 and unlike in ovine cells, no difference was detected both in vitro (KC cells) and in vivo (injected midges) between the replication of rBTV/-1 and that of rBTV/-16. However, by Western blot analyses, rBTV/-1 produced slightly less NS3 (but not NS3a) than rBTV/-2 and rBTV/-16 in infected Culicoides cells. Thus, even though these three BTV Seg-10 do not have a major effect on BTV replication kinetics in KC cells or in midges infected intrathoracically, we cannot completely exclude that they might influence viral replication and/or dissemination in orally infected Culicoides. On the other hand, inhibition of type I IFN synthesis and viroporin activity exerted by NS3 proteins surely is more specific to mammals, where they seem to play multiple roles, unlike in insects, where their sole function described so far is their ability to mediate viral particle release by budding at the plasma membrane (35, 41). Hence, the amino acid sequence variability of NS3 proteins is most likely to influence one or more of their functions and, consequently, BTV biological properties in mammals rather than in insects. Overall, our study reveals that, unlike that in Culicoides, the sequence variability within NS3/NS3a alters BTV replication kinetics and, therefore, its virulence in mammals, with a host-specific variation of NS3 protein turnover rate being the phenotype. Given the tendency of this segmented virus to reassort during coinfection, our results may help us to predict, in a more accurate manner, how dangerous cocirculating BTV serotypes or strains could be if we endure novel outbreaks.