Structural Glycoprotein E2 of Classical Swine Fever Virus Critically Interacts with Host Protein Torsin-1A during the Virus Infectious Cycle

ABSTRACT

The classical swine fever virus (CSFV) glycoprotein E2 is the major structural component of the virus particle. E2 is involved in several functions, such as virus adsorption to the cell, the elicitation of protective immune responses, and virus virulence in swine. Using a yeast two-hybrid system, we previously identified the swine host protein Torsin-1A, an ATPase protein residing in the endoplasmic reticulum and inner nucleus membrane of the cell, as a specific binding partner for E2. The interaction between Torsin-1A and E2 proteins was confirmed to occur in CSFV-infected swine cells using three independent methods: coimmunoprecipitation, confocal microscopy, and proximity ligation assay (PLA). Furthermore, the E2 residue critical to mediate the protein-protein interaction with Torsin-1A was identified by a reverse yeast two-hybrid assay using a randomly mutated E2 library. A recombinant CSFV E2 mutant protein with a Q316L substitution failed to bind swine Torsin-1A in the yeast two-hybrid model. In addition, a CSFV infectious clone harboring the E2 Q316L substitution, although expressing substantial levels of E2 protein, repetitively failed to produce virus progeny when the corresponding RNA was transfected into susceptible SK6 cells. Importantly, PLA analysis of the transfected cells demonstrated an abolishment of the interaction between E2 Q316L and Torsin-1A, indicating a critical role for that interaction during CSFV replication.

IMPORTANCE Structural glycoprotein E2 is an important structural component of the CSFV particle. E2 is involved in several virus functions, particularly virus-host interactions. Here, we characterized the interaction between CSFV E2 and swine protein Torsin-1A during virus infection. The critical amino acid residue in E2 mediating the interaction with Torsin-1A was identified and the effect of disrupting the E2–Torsin-1A protein-protein interaction was studied using reverse genetics. It is shown that the amino acid substitution abrogating E2–Torsin-1A interaction constitutes a lethal mutation, demonstrating that this virus-host protein-protein interaction is a critical factor during CSFV replication. This highlights the potential importance of the E2–Torsin-1A protein-protein interaction during CSFV replication and provides a potential pathway toward blocking virus replication, an important step toward the potential development of novel virus countermeasures.

INTRODUCTION

Classical swine fever virus (CSFV) causes a highly contagious disease of swine that can result in significant economic losses. CSFV is a small, enveloped virus possessing a positive-stranded RNA genome of 12.5 kb. The genome is transcribed as a single open reading frame encoding a 3,898-amino-acid polyprotein that, after its cleavage by viral and cellular proteases, produces 11 or 12 final proteins (NH2-Npro-C-Erns-E1-E2-p7-NS2-NS3-NS4A-NS4B-NS5A-NS5B-COOH) (1). The CSFV virus particle is composed of four structural proteins: the core protein, closely associated with the viral RNA, and three glycoproteins—E^rns, E1, and E2—structurally associated with the virus envelope. The functions of the structural proteins in virus replication and virulence in the natural host have been analyzed in many recent studies (2–11).

The identification of host proteins interacting with CSFV proteins and the functional characterization of those interactions is a relatively new area of study. Several host proteins have been shown to specifically interact with structural CSFV proteins. CSFV core protein has been demonstrated to interact with the proteins SUMO1 (small ubiquitin-related modifier 1), IQGAP1 (IQ motif containing GTPase-activating protein 1), UBC9 (ubiquitin-conjugating enzyme 9), and HB (hemoglobin subunit beta) (12–15), while E^rns has been shown to interact with the laminin receptor (16). Also, CSFV viroporin p7 has been shown to directly interact with the calcium signal-modulating cyclophilin ligand (CALMG), microtubule-associated protein RP/EB family member 1 (MAPRE1), and voltage-dependent anion channel 1 (VDAC1) (17, 18). In addition, E2 has been identified as an interacting partner of several different host proteins, including cellular actin (19), Anx2 (annexin 2) (20), Trx2 (thioredoxin) (21), MEK2 (mitogen-activated protein kinase 2) (22), PPP1CB (protein phosphatase 1 catalytic subunit beta) (23), SERTAD1 (SERTA domain-containing protein 1) (24), CCDC115 (coiled-coil domain containing 115) (25), and DCTN6 (dynactin subunit 6) (26). In most of these cases, these host-virus protein-protein interactions play a role in regulating the virus replication cycle, with a few also involved in virulence (12–14).

We previously determined, using a yeast two-hybrid approach, that CSFV E2 specifically interacts with the swine host protein Torsin-1A (27). Torsin-1A is part of the AAA+ ATPase superfamily. AAA+ ATPases are a superfamily with a highly conserved ATP hydrolysis domain and are involved in various and diverse cellular processes (28). For example, they are known to be involved in cellular proteostasis, genome replication, and membrane fusion (29). Torsin-1A resides in the endoplasmic reticulum and the inner nuclear envelope. It is known to be involved in protein chaperoning and membrane displacements (30) and in cellular trafficking and cellular localization (31). Mutations in the Torsin-1A gene are commonly found to cause dystonia, a rare neurological disorder in humans (32).

Here, we confirm that the E2–Torsin-1A interaction, originally discovered using a yeast two-hybrid model, occurs in CSFV-infected swine cells, using three different methods: coimmunoprecipitation, confocal microscopy, and proximity ligation assay (PLA). In addition, the E2 residue critical for mediating the interaction with Torsin-1A was mapped using a reverse yeast two-hybrid approach. Reverse genetics using an infectious clone of CSFV was then used to create a recombinant CSFV mutant (CSFV.E2ΔTorsin-1A) harboring the specific substitutions disrupting the E2–Torsin-1A interaction, as assessed in a yeast two-hybrid assay. Cells transfected with mRNA encoded by CSFV.E2ΔTorsin-1A, although expressing substantial levels of E2 protein, failed to produce virus progeny. Importantly, PLA of the transfected cells demonstrated an abolishment of the interaction between E2 and E2–Torsin-1A, indicating a critical role for this interaction during CSFV replication.

RESULTS

CSFV structural glycoprotein E2 and swine host protein Torsin-1A interact in virus-infected cells.

Previously, we reported a group of swine host proteins that interact with the CSFV major structural glycoprotein E2 (27), which we identified using a yeast two-hybrid approach and employing a library constructed from swine macrophage mRNAs. Among proteins interacting with E2, we identified the host protein Torsin-1A (Fig. 1A). Torsin-1A belongs to the AAA+ ATPase superfamily and resides in the endoplasmic reticulum and the nuclear envelope of the cells. Torsin-1A is involved in protein chaperoning and membrane movement (30), ensuring protein stability, preventing protein misfolding, and promoting intracellular trafficking (31).

FIG 1.

(A) AH109 yeast cells were cotransformed with the indicated Gal4 binding domain (BD) or Gal4 activation domain (AD) plasmids. On the left, yeasts were spotted on selective medium lacking adenine, histidine, tryptophan, and leucine; on the right, yeasts were spotted on medium that lacks tryptophan and leucine for plasmid maintenance. PGADT7 is included as a negative control. (B) Coimmunoprecipitation of E2 and Torsin-1A in CSFV-infected or mock-infected SK6 cells. The input cell lysate was immunoprecipitated with antibodies for E2. Protein extracts were blotted for Torsin-1A as indicated in Materials and Methods, with bands observed at the expected molecular weights of 37 and 30 kDa for the two isoforms of Torsin-1A.

To examine if the interaction between E2 and Torsin-1A, identified by using the yeast two-hybrid method, in fact occurs in host cells infected with CSFV, three different and independent techniques were employed: coimmunoprecipitation, confocal microscopy, and proximity ligation assay (PLA).

Coimmunoprecipitation experiments were performed using monoclonal antibodies specifically recognizing both interacting proteins. SK6 cells were infected with CSFV strain Brescia (BICv), a virus derived from an infectious clone-encoding Brescia strain (5), at a multiplicity of infection (MOI) of 10. Samples were harvested after 24 h in culture and evaluated as described in Materials and Methods. The mock-infected and infected cell lysates were collected, immunoprecipitated with the anti-E2 monoclonal antibody WH303, and subjected to Western blotting with a Torsin-1A-specific monoclonal antibody (Thermo Fisher catalog no. PA5-70624). A double band showing the molecular masses of the two isoforms (37 and 22 kDa, which correspond to the respective molecular weight predicted from their amino acid sequence) of Torsin-1A were clearly observed (Fig. 1). These results indicated that in CSFV-infected cells E2 can be coimmunoprecipitated along with Torsin-1A, confirming the previous discovery using yeast two-hybrid assays and suggesting that E2–Torsin-1A interaction occurs during viral infection of cell cultures.

Furthermore, the E2–Torsin-1A interaction in CSFV-infected cells was assessed by PLA (33). Recognition of both proteins was performed using the same reagents used in coimmunoprecipitation: monoclonal antibody WH303 for detection of E2 and a commercial polyclonal antibody for detection of Torsin-1A. Results of the proximity ligation assay, analyzed under experimental conditions similar to those used to perform the coimmunoprecipitation studies, confirmed that E2 and Torsin-1A interact in SK6 cell cultures infected with CSFV. This interaction appears as a distinct distribution throughout the cell cytoplasm (Fig. 2).

FIG 2.

Proximity ligation assay (PLA) for E2 and Torsin-1A in CSFV-infected SK6 cells. Interaction between E2 and Torsin-1A was determined by PLA in SK6 cells that were either mock infected or infected for 24 h with CSFV BICv (MOI = 10). Top and bottom panels show different images of the same treatment. Magnification, ×1,000. The figure shows representative results from one experiment, which was performed three times, with at least 100 cells analyzed in each experiment.

Last, to confirm that E2–Torsin-1A interaction occurs during CSFV infection of host cells, the localization of E2 and Torsin-1A during infection was assessed using double-labeled immunofluorescence and confocal microscopy in CSFV-infected cells. SK6 cells were infected or mock infected with BICv as described for coimmunoprecipitation and PLA. Cells were fixed on glass coverslips at 24 h postinfection and stained using monoclonal antibodies for E2 (WH303) and Torsin-1A. The results indicated a clear colocalization of CSFV E2 and Torsin-1A proteins; the appearance of distinct spots of Torsin-1A that colocalized with E2 during viral infection are distributed in the cell cytoplasm (Fig. 3), supporting the hypothesis that interaction between these two proteins occurs during viral infection.

FIG 3.

Analysis of the distribution of Torsin-1A and E2 proteins in CSFV-infected SK6 cells. Cells were infected or mock infected with CSFV BICv and processed by immunofluorescence staining at 24 hpi as described in Materials and Methods. CSFV E2 and Torsin-1A proteins were visualized with Alexa Fluor 488 (green) and Alexa Fluor 594 (red), respectively. Yellow indicates colocalization of Alexa Fluor 488 and Alexa Fluor 594 in the merged image. The figure shows representative results from one experiment, which was performed twice, with at least 100 cells analyzed in each experiment.

Therefore, three independent methods, coimmunoprecipitation, PLA, and confocal microscopy, confirmed the existence of the E2 and Torsin-1A interaction during CSFV infection of SK6 cell cultures.

Identification of CSFV E2 residues critical for Torsin-1A interaction.

Development of recombinant CSFV to study the role of interaction between virus and host proteins requires the identification of E2 residues mediating the interaction with Torsin-1A. These recombinant viruses will harbor residue substitutions shown to disrupt the virus-host interaction and can be used as a tool to study the role of the interaction in several virus functions. We have previously unsuccessfully attempted to map the E2 residues involved in the interaction with Torsin-1A using an alanine scanning approach involving a set of 76 CSFV E2 mutant proteins containing sequential stretches of amino acids where the native amino acid residues were replaced with alanine residues; however, this approach was not successful in identification of the E2 binding site for Torsin-1A (27).

To identify E2 amino acids involved in the interaction with Torsin-1A, we used an alternative approach utilizing the yeast two-hybrid system. This method allows the identification of E2 residues mediating Torsin-1A interaction by using a random mutagenesis approach. This approach evaluates the ability of Torsin-1A to interact with a library of randomly mutated E2 where an average of 5 amino acid residues were randomly substituted throughout E2. In this procedure, 1 × 10⁶ different mutations in E2 were tested using the library screening approach described in Materials and Methods. E2 mutants lacking reactivity with Torsin-1A were also tested for their ability to bind protein HPRT1 (hypoxanthine phosphoribosyl transferase 1), another host protein that was also found to be a specific host binding partner for E2. This was performed to rule out the possibility that the random mutants introduced had gross conformational changes, frameshifts, or truncated forms of E2 that could mediate a nonspecific loss of protein binding. One E2 mutant showed a decreased ability of E2 protein to bind the Torsin-1A protein. This mutant of E2 (E2ΔTorsin-1A) harbors a residue substitution at position Q316L and was used to further study the role of Torsin-1A binding to E2 in CSFV replication (Fig. 4).

FIG 4.

Yeast two-hybrid protein-protein interaction of CSFV protein E2 or E2ΔTorsin-1A (Q316L) coupled to the Gal4 binding domain with Torsin-1A or HPRT coupled to the Gal4 activation domain. Portions (10 μl) of strains expressing the indicated constructs were spotted on (A) the selective medium SD−ALTH or (B) the nonselective medium SD−TL.

Q316L substitution interferes with the E2–Torsin-1A interaction in SK6 cells.

The effect of the E2 Q316L substitution on the interaction between E2 and Torsin-1A was further studied in SK6 cells transfected with E2ΔTorsin-1A RNA and evaluated by PLA. Recognition of both proteins, CSFV E2 and host Torsin-1A, was performed using reagents that were the same as and experimental conditions that were similar to those used in the coimmunoprecipitation, PLA, and confocal microscopy procedures described earlier: monoclonal antibody WH303 for detection of E2 and a commercial polyclonal antibody for detection of Torsin-1A (Thermo Fisher catalog no. PA5-70624). The PLA confirmed the presence of interaction between E2 and Torsin-1A in cells transfected with BICv RNA while showing a complete absence of interaction in SK6 cell cultures transfected with E2ΔTorsin-1A RNA (Fig. 5). These results indicate that the Q316L substitution in glycoprotein E2 is lethal and disrupts the E2–Torsin-1A interaction, suggesting that the presence of a functional interaction between virus protein E2 and cell protein Torsin-1A is necessary for CSFV replication.

FIG 5.

PLA for E2 and Torsin-1A in transfected SK6 cells. Interaction between E2 and Torsin-1A was determined by PLA in SK6 cells that were transfected with RNA from either BICv or E2ΔTorsin-1A or mock transfected. The top and bottom panels show two different images of the same treatment. The figure shows representative results from one experiment, which was performed three times, with at least 100 cells analyzed in each experiment.

Attempt to develop the CSFV E2ΔTorsin-1A mutant.

To further evaluate the interaction of CSFV E2 with host Torsin-1A, a recombinant mutant infectious clone (IC) based on virulent strain Brescia (BICv) was constructed using reverse genetics. The E2 amino acid residue identified as mediating the E2–Torsin-1A interaction (Q316L) was introduced into the IC encoding a full-length cDNA of CSFV strain Brescia (pBIC) (2) to produce the pE2ΔTorsin-1A construct. Infectious RNA was in vitro transcribed from full-length infectious clones of pBIC or pE2ΔTorsin-1A and further used to transfect susceptible SK6 cells. Attempts to rescue de novo-synthesized virus were performed at day 4 posttransfection, when SK6 cell monolayers either transfected with RNA or mock transfected were stained by immunocytochemistry using monoclonal antibody WH303 to detect the presence of CSFV E2 expression. E2 expression was readily detected in cells transfected with either BICv or E2ΔTorsin-1A RNA, although the number of E2 positively stained cells was clearly lower in cells transfected with E2ΔTorsin-1A RNA (Fig. 6).

FIG 6.

In vitro transfection of SK6 cells with either BICv or E2ΔTorsin-1A RNA. Cells were transfected with the corresponding in vitro-transcribed RNA from each of the corresponding infectious clones or mock transfected. At day 4 posttransfection, cell monolayers were fixed with 50% (vol/vol) ethanol/acetone and stained by immunohistochemistry with anti-CSFV E2 monoclonal antibody WH303, as described in Materials and Methods.

The presence of infectious CSFV in the cell extract and cell supernatants obtained from cell cultures transfected with either BICv or E2ΔTorsin-1A RNA were assessed by virus isolation using SK6 cells. High titers (6.95 50% tissue culture infectious doses [TCID₅₀]/ml; standard deviation = 0.49) of virus were detected in cell cultures transfected with BICv RNA, while no virus titers were detected (sensitivity of detection, ≥1.8 TCID₅₀/ml) in three consecutive, independently performed transfection experiments with E2ΔTorsin-1A RNA. In addition, material from the three E2ΔTorsin-1A transfections was blind passaged three times in SK6 cell cultures in an attempt to detect the potential presence of infectious virus. All blind passages were stained at 4 days postinfection with monoclonal antibody WH303 to detect CSFV E2 expression. All blind passages performed with the material obtained from the three E2ΔTorsin-1A RNA transfections were negative for the presence of CSFV. Therefore, the Q316L substitution in glycoprotein E2 constitutes a lethal mutation for CSFV replication under the experimental conditions tested here.

DISCUSSION

The mechanisms involved in replication of CSFV in the host cell still are not completely understood. It is clear that direct interaction between host and viral proteins have profound implications for the modulation of cell mechanisms allowing virus replication to proceed during infection. Studying the interaction between CSFV and host proteins has been an active field of research in the last few years. Our preceding work as well as that from other laboratories has been focused on the identification of cellular proteins interacting with CSFV proteins to understand their role in virus replication, immunomodulation and virulence during virus infection both in cell cultures and the natural host. We demonstrated that structural CSFV core protein interacts with host SUMO1, IQGAP1, and UBC9 (12–14), and more recently, we described the interaction of major structural glycoprotein E2 with the following host proteins: protein phosphatase 1 catalytic subunit beta (PPP1CB) (23), protein coiled-coil domain-containing 115 (CCDC115) (25), dynactin subunit 6 (DCTN6) (26), and SERTA domain containing protein 1 (SERTAD1) (24). In addition, others have shown the interaction of E2 with several host proteins, including cellular actin (19), thioredoxin (21), annexin 2 (20), and mitogen-activated protein kinase 2 (22). In addition to modulating the virus replication cycle, these host-virus protein interactions have also been shown to influence virulence in swine (12–14). Here, we report the characterization of the interaction between CSFV E2 protein and host protein Torsin-1A.

The host protein Torsin-1A is part of the AAA+ ATPase superfamily, whose members are associated with a variety of cellular activities (34). Torsin-1A is the only member of this family of proteins residing in both the endoplasmic reticulum and the inner nuclear envelope (35, 36). One striking deviation of Torsin-1A from canonical AAA+ proteins is that it lacks inherent ATPase activity (37). Instead, Torsin-1A requires interactions with one of two transmembrane ATPase-stimulating proteins located within the endoplasmic reticulum and the nuclear envelope membranes (38). These functional protein cofactors are the lamina-associated polypeptide 1 (LAP1), which localizes to the inner nuclear membrane, and the luminal domain-like LAP1 (LULL1), which remains within the peripheral endoplasmic reticulum (35, 39). Through an active-site complementation mechanism, LAP1 and LULL1 are individually able to stimulate Torsin-1A ATPase activity by donating a catalytic arginine residue to the Torsin-1A active site that is necessary for transition state stabilization (34).

Torsin-1A is implicated in broad and critical functions, including lipid synthesis (40), regulation of membrane morphology (41), protein quality control (42, 43), and endoplasmic reticulum redox sensing (37, 38, 43). In addition, Torsin-1A performs distinct functions at the nuclear membrane. Torsin-1A and its cofactor LAP1 are essential for proper assembly of fibroblast nuclear envelope-anchored transmembrane actin-associated nuclear (TAN) lines (44), which are composed of linker arrays of nucleoskeleton and cytoskeleton complexes associated with retrograde-flowing actin. Torsin-1A modulates the rearward motion of nuclei during centrosome positioning and is implicated in maintaining cell polarity in migrating cells (45). LULL1 is localized throughout the ER (35), while the nuclear domain of LAP1 binds to the nuclear lamina and therefore resides in the inner nuclear membrane (46).

No interaction between Torsin-1A and any virus protein has been reported so far. Knockout of genes for Torsin-1A and its ligands LAP1 and LULL1 was studied in HeLa cells to demonstrate the potential role of these molecules in herpes simplex virus 1 (HSV-1) replication. While Torsin-1A knockout reduces herpesvirus production only marginally, knockout of LULL1 reduces HSV-1 production by more than 10-fold. The assembly and packaging of viral capsids in the nucleus are severely delayed in LULL1 knockout cells, and the viral genome copy number per cell is decreased by 10-fold. These findings established that LULL1 is required for herpesvirus maturation at a step prior to nuclear egress, suggesting that LULL1 has additional functions independent of the previously established role of Torsin-1A ATPases in nuclear membrane budding (47).

The results reported here characterize, for the first time, cellular protein Torsin-1A as a protein interaction partner for CSFV protein E2 in the infected cell. At his moment, we cannot speculate whether disruption of the interaction of E2 with Torsin-1A can affect CSFV replication. The involvement of Torsin-1A in multiple cell functions opens multiple options to hypothesize the role of its interaction with E2 as critical for virus viability. Several Torsin-1A functions, such as protein chaperoning and intracellular trafficking, along with its location in the endoplasmic reticulum may facilitate the posttranscriptional modifications of E2, its correct folding, or its proper location in the process of production of the virus particles. Regardless of the exact molecular mechanism involved, our results indicate that interaction between host protein Torsin-1A and structural CSFV glycoprotein E2 is critical in CSFV replication in vitro since, as shown here, the disruption of the E2–Torsin-1A interaction completely abrogated virus replication during CSFV infection in susceptible cell cultures. This presents new possibilities to explore CSFV requirements of virus-host interaction necessary to produce virus replication and the development of novel alternative countermeasures.

MATERIALS AND METHODS

Viruses and cells.

Swine kidney (SK6) cells (2), free of bovine viral diarrhea virus (BVDV), were cultured in Dulbecco's minimal essential medium (DMEM) (Gibco, Grand Island, NY) with 10% fetal calf serum (FCS) (Atlas Biologicals, Fort Collins, CO). Titration of CSFV from samples of inoculated animals’ samples was performed using SK6 cells in 96-well plates (Costar, Cambridge, MA). Viral infectivity was detected after 4 days in culture by an immunoperoxidase assay using the CSFV monoclonal antibody WH303 (48) and the Vectastain ABC kit (Vector Laboratories, Burlingame, CA). Titers were calculated and expressed as TCID₅₀ per milliliter as previously described (49). As performed, virus detection sensitivity was ≥1.8 TCID₅₀/ml.

Immunoblotting and antibodies.

For immunoblotting analysis, infected and mock-infected (control) cell monolayers were washed in ice-cold phosphate-buffered saline (PBS) and lysed in radioimmunoprecipitation assay (RIPA) buffer (Teknova, Hollister, CA) in the presence of a protease inhibitor cocktail (Roche, Basel, Switzerland). Proteins were resolved on NuPAGE 4 to 12% (wt/vol) bis-Tris gels (Invitrogen, Carlsbad, CA) and transferred to polyvinylidene difluoride (PVDF) membranes following manufacturer instructions. Immunodetection was performed using the following antibodies: polyclonal anti-Torsin-1A (catalog no. PA5-70624; Thermo Fisher Scientific, Waltham, MA) and anti-CSFV E2 protein monoclonal antibody WH303 (22), and Pierce goat anti-mouse and anti-rabbit IgG peroxidase-conjugated secondary antibody reagent (catalog no. 31430 and 31460; Thermo Fisher Scientific, Waltham, MA), respectively. Western blots were imaged using an Azure C400 system and analyzed with cSeries capture software (Azure Biosystems, Dublin, CA).

Immunoprecipitation.

Immunoprecipitation procedures were performed in triplicate using the Pierce coimmunoprecipitation kit following the manufacturer’s instructions (Thermo Fisher Scientific). In brief, cell suspensions were washed in ice-cold PBS and lysed with the Pierce lysis buffer plus protease inhibitor cocktail (Roche, Basel, Switzerland). Anti-E2 WH303 was conjugated to the Pierce beads and incubated with precleared lysate overnight. The beads were then washed with Pierce lysis buffer plus 0.1% Triton X-100 (Sigma-Aldrich, St. Louis, MO) and protease inhibitor cocktail (Roche) and eluted in Pierce elution buffer.

Proximity ligation assay.

The proximity ligation assay (PLA) was performed in triplicate using guidelines from the Duolink-PLA kit (Sigma-Aldrich). SK6 cells were plated onto 12-mm round coverslips (Thomas Scientific, Swedesboro, NJ, USA) in a 24-well plate (Corning, Corning, NY, USA) at a density of 25,000 cells/well. Next day, cells were infected (MOI = 10) or mock infected (uninfected control cells), and 24 h later, cells were fixed with 4% (wt/vol) formaldehyde in PBS at room temperature for 20 min, followed by permeabilization buffer (0.3% Triton X-100 in PBS) for 10 min. Fixed cells were then blocked with Duolink blocking buffer for 30 min at 37°C, followed by incubation with primary antibodies anti-E2 WH303 (31) and anti-Torsin-1A (Thermo Fisher catalog no. PA5-70624) at 4°C for 1 h. Cells were then washed twice with Duolink (Sigma-Aldrich) wash buffer A and incubated with the plus and minus PLA probes for 1 h at 37°C, followed by 2 washes with Duolink wash buffer A and a 30-min incubation at 37°C with Duolink ligase in ligation buffer. Fixed cells were then washed twice with Duolink wash buffer A followed by incubation with Duolink polymerase in amplification buffer at 37°C for 100 min. The fixed cells were then washed twice with Duolink wash buffer B and mounted with Duolink PLA mounting medium with DAPI (4′,6-diamidino-2-phenylindole).

Colocalization studies.

SK6 cells were plated in 24-well tissue culture dishes containing 12-mm round coverslips at a density of 2.5 × 10⁴ cells per/well and infected with CSFV Brescia virus (BICv) at an MOI of 0.1 for 24 h or mock infected. Twenty-four hours later, the cells were fixed on the coverslips with 4% paraformaldehyde (EMS, Hatfield, PA) for 15 min. The paraformaldehyde was removed, the tissue culture dishes were washed three times with PBS, and then the cells were permeabilized with 0.5% Triton X-100 for 5 min at room temperature, followed by incubation in blocking buffer (PBS, 5% normal goat serum, 2% bovine serum albumin, 10 mM glycine, 0.01% thimerosal) for 1 h at room temperature. The fixed cells were then incubated with the primary antibodies for E2 and Torsin-1A overnight at 4°C. After being washed three times with PBS, the cells were incubated with the appropriate secondary antibodies, goat anti-rabbit immunoglobulin G (IgG) (1/400; Alexa Fluor 594; Molecular Probes) and goat anti-mouse IgG (Alexa Fluor 488; Molecular Probes), for 1 h at room temperature. After this incubation, the coverslips were washed three times with PBS, mounted with ProLong Gold antifade mountant (Life Technologies) containing DAPI, and then left for 24 h to cure. The samples were examined using a Zeiss LSM 710 confocal microscope. The captured images were adjusted for contrast and brightness using Zeiss Zen Blue software.

Yeast two-hybrid screening for disruption of Torsin-1A binding to E2.

Plasmids containing E2 fused to the Gal4 binding domain E2-BD and Torsin-1A fused to the Gal4 activation domain (Torsin-1A-AD), PGADT7 and HPRT1-AD, were used as negative controls and were previously identified or constructed (27). E2-BD was randomly mutated using a mutagenic PCR approach to give an average of 5 nucleotide substitutions across the E2 open reading frame (this random mutant library was constructed by Epoch Bioscience [Bothell, WA]). The random E2 mutant library was then cotransformed into Saccharomyces cerevisiae strain AH109 along with Torsin-1A-AD with a transformation rate of at least 1 × 10⁶ individual colonies representing full coverage of the E2 mutagenic library. Transformed yeast colonies were initially selected on synthetic defined (SD) agar medium lacking leucine and tryptophan (SD−TL), the selection markers for Torsin-1A-AD and E2-BD, respectively. The yeasts were then replica plated onto agar plates lacking histidine and adenine (SD−ALTH) and incubated at 30°C. Colonies that grew on SD−ALTH plates contained a mutant E2-BD plasmid that maintained the ability of E2 to interact with Torsin-1A. However, individual yeast colonies that were no longer able to grow on SD−ALTH plates and grew only on SD−TL agar plates were tested further. Individual colonies that met this criterion were further tested by growing in liquid SD−TL medium overnight at 30°C, spot plated as previously described (12) on both SD−TL and SD−ALTH agar plates at 30°C, and observed daily for growth in comparison with native E2-BD and Torsin-1A. Individual colonies that were negative for growth on SD−ALTH plates were then subjected to previously described plasmid recovery procedures (12), where the plasmid could be isolated using a Qiagen miniprep kit (Hilden, Germany) and Sanger sequenced. Sanger sequencing often revealed that the mutated E2 protein contained stop codons or out-of-frame mutations, thus explaining why the loss of E2–Torsin-1A interaction occurred. Plasmids that contained stop codons or out-of-frame mutations were discarded and not studied further. In the case where only individual amino acids were mutated, the E2 mutant plasmids were tested individually by cotransformation with Torsin-1A-AD, HPRT1-AD (hypoxanthine phosphoribosyl transferase 1), a previously identified (27) positive E2 protein interactor, or PGADT7 (negative control) and selected on SD−TL plates. Then, individual colonies were grown overnight in SD−TL liquid medium at 30°C and spot plated on both SD−TL and SD−ALTH to assess the ability of individual E2 mutants to bind both positive and negative controls (HPRT1-AD and PGADT7) and to confirm reduction of binding to Torsin-1A. This second selection was performed to discard any mutant E2 proteins that lost the ability to bind Torsin-1A because of a gross structural change in E2.

Construction of the CSFV E2ΔTorsin-1A mutant.

A full-length IC of the virulent CSFV Brescia strain (pBIC) (2) was used as a template in which the E2 amino acid substitutions disrupting the E2–Torsin-1A interaction, as mapped by the reverse yeast two-hybrid method, were included. Therefore, the residue substitution Q316L was introduced into the native E2 sequence in the designed E2ΔTorsin-1A construct. The E2ΔTorsin-1A plasmid was obtained by DNA synthesis (Epoch Life Sciences, Sugar Land, TX).

The CSFV E2ΔTorsin-1A full-length genomic clone was linearized with SrfI and in vitro transcribed using the T7 MEGAscript system (Ambion, Austin, TX). RNA was precipitated with LiCl and transfected into SK6 cells by electroporation at 500 V, 720 Ω, and 100 W with a BTX 630 electroporator (BTX, San Diego, CA). Cells were seeded in 12-well plates and incubated for 4 days at 37°C and 5% CO₂. E2 expression was detected by immunoperoxidase staining as described above.