Highly Divergent Strains of Porcine Reproductive and Respiratory Syndrome Virus Incorporate Multiple Isoforms of Nonstructural Protein 2 into Virions

Abstract

Viral structural proteins form the critical intermediary between viral infection cycles within and between hosts, function to initiate entry, participate in immediate early viral replication steps, and are major targets for the host adaptive immune response. We report the identification of nonstructural protein 2 (nsp2) as a novel structural component of the porcine reproductive and respiratory syndrome virus (PRRSV) particle. A set of custom α-nsp2 antibodies targeting conserved epitopes within four distinct regions of nsp2 (the PLP2 protease domain [OTU], the hypervariable domain [HV], the putative transmembrane domain [TM], and the C-terminal region [C]) were obtained commercially and validated in PRRSV-infected cells. Highly purified cell-free virions of several PRRSV strains were isolated through multiple rounds of differential density gradient centrifugation and analyzed by immunoelectron microscopy (IEM) and Western blot assays using the α-nsp2 antibodies. Purified viral preparations were found to contain pleomorphic, predominantly spherical virions of uniform size (57.9 nm ± 8.1 nm diameter; n = 50), consistent with the expected size of PRRSV particles. Analysis by IEM indicated the presence of nsp2 associated with the viral particle of diverse strains of PRRSV. Western blot analysis confirmed the presence of nsp2 in purified viral samples and revealed that multiple nsp2 isoforms were associated with the virion. Finally, a recombinant PRRSV genome containing a myc-tagged nsp2 was used to generate purified virus, and these particles were also shown to harbor myc-tagged nsp2 isoforms. Together, these data identify nsp2 as a virion-associated structural PRRSV protein and reveal that nsp2 exists in or on viral particles as multiple isoforms.

INTRODUCTION

Porcine reproductive and respiratory syndrome virus (PRRSV) is the cause of a complex systemic disease of swine, most notably affecting the respiratory and reproductive systems of infected hosts. PRRS is characterized by an acute viral infection of the porcine macrophage that leads to an immunologically altered state. In extreme cases, respiratory distress, metabolic dysregulation, and neuronal involvement result in significant mortality within days to weeks of experimental inoculation with highly pathogenic isolates (1, 2). Endemic disease from emerging and reemerging PRRSV results in estimated annual economic losses totaling greater than 560 million dollars, or $1.5 million per day, to the U.S. economy alone (3, 4).

A key hurdle to PRRSV investigation has been the difficulty in ascribing conserved functional characteristics to virally encoded nonstructural proteins. This hurdle is largely rooted in the significant genetic heterogeneity noted among strains, genotypes, and viral species of the family Arteriviridae and, more broadly, the order Nidovirales that also encompasses Coronaviridae and Roniviridae. Through the combined effects of a highly error-prone viral RNA-dependent RNA polymerase (RdRp) and a significant rate of genetic recombination, the evolution rate of PRRSV (4.71 × 10⁻² to 9.8 × 10⁻² synonymous substitution rate/site/year) is estimated to be nearly 10-fold higher than those of human immunodeficiency virus (HIV) and influenza virus (5–9). The narrative of the last 2 decades of PRRSV investigation has demonstrated the robust nature of PRRSV in undergoing extraordinary genetic change while maintaining pathogenicity.

A significant fraction of attenuating and naturally occurring mutations have been observed within the region of replicase nonstructural protein 2 (nsp2) (10–12). nsp2 is encoded by 21% to 23% of the genome and consists of a papain-like protease domain (PLP2) near the amino terminus that has been shown to be a member of the ovarian tumor domain protease (OTU) family (13–16). The nsp2 PLP2/OTU region is followed by an extended central hypervariable (HV) region with strain-specific insertions or deletions, a putative transmembrane region (TM) that is predicted to encode up to four membrane-spanning regions, and a relatively conserved carboxyl (C-terminal) domain. Previous reports suggest a multifunctional role of nsp2 in critical replication and immune evasion functions (14–20). nsp2 is both the largest and most genetically diverse protein encoded within the PRRSV genome. The elevated frequency of mutation and recombination within the nsp2 coding region suggests there is significant selective pressure(s) against nsp2 genetic or protein content (17, 21, 22). Interestingly, whereas current data support only an intracellular role for nsp2 (19, 23–26), a strong antibody (Ab) response is generated against nsp2 during the course of experimental infection (27). nsp2 is recognized as an immunodominant target of the adaptive immune response resulting in significant anti (α)-nsp2 titers comparable to the response generated against the highly abundant/highly immunogenic nucleocapsid protein (N) (27, 28). Moreover, various B-cell and T-cell epitopes have been bioinformatically or experimentally defined within the nsp2 coding region (21, 29, 30). These attributes suggest that nsp2 might also maintain an extracellular localization or function during the viral replication cycle. Similarly, recent findings have described the incorporation of a PRRSV nsp2 orthologue, nsp3, into the virion of the related Coronaviridae, including severe acute respiratory syndrome virus (SARS) and transmissible gastroenteritis coronavirus (TGEV) (31, 32). In this study, we investigated the association of PRRSV nsp2 with the virion and provided evidence that nsp2 is a novel structural protein of the PRRSV particle. This finding was conserved across a range of five genetically diverse PRRSV strains, including the European (type 1) and North American (type 2) prototype strains Lelystad (33) and VR-2332 (34, 35), respectively, and a disparate type 2 strain, MN184 (12), as well as the Asian highly pathogenic PRRSV (HP-PRRSV) strains rJXwn06 (2) and rSRV07 (36) (Table 1). The presence of nsp2 in or on virions indicates a plausible immediate-early function of this protein within the viral life cycle and, importantly, that virion-associated extracellular localization may also contribute to the high selective pressure driving nsp2 genetic diversity.

Table 1.

Description of study strains

Strain	Description (yr of isolation)
Lelystad	European (type 1) prototype strain (1991)
VR-2332	North American (type 2) prototype strain (1989)
rV7-myc	Recombinant VR-2332 expressing a 3X c-myc epitope tag within nsp2 coding region
MN184C	Regional North American type 2 isolate (2001)
rJXwn06	Chinese highly pathogenic type 2 PRRSV isolate (2006)
rSRV07	Vietnamese highly pathogenic type 2 PRRSV isolate (2007)

MATERIALS AND METHODS

Antibodies.

Custom affinity-purified polyclonal antibodies (rabbit) to PRRSV nsp2 included α-nsp2-OTU (OTU domain; epitope SKFETTLPERVRPP), α-nsp2-HV (hypervariable region; epitope TRPKYSAQAIIDSG), α-nsp2-TM (transmembrane region; epitope SDPVGTACEFDSPE), and α-nsp2-C (C-terminal region; epitope NGLKIRQISKPSGG) (GenScript, Piscataway, NJ). Prior to Western blot probing, α-nsp2 antibodies were cross-absorbed with uninfected MARC-145 cell lysate for ≥1 h at 37°C to eliminate potential cross-reactivity with cellular products. Epitope sites were chosen based on predicted antigenicity (proprietary algorithm; Genscript, Piscataway, NJ), sequence conservation in the five study strains, and epitope location within the nsp2 coding region.

Immunofluorescence analysis.

MARC-145 cells were seeded at a density of ∼5 × 10⁴ cells/cm² and allowed to incubate for 2 to 3 days prior to infection in 1× minimum essential medium (MEM; SAFC Biosciences, Lenexa, KS) supplemented with 2.2% (wt/vol) sodium bicarbonate, 11% (wt/vol) sodium pyruvate (Life Technologies, Carlsbad, CA), and 10% fetal bovine serum (FBS; PAA Laboratories/GE Healthcare Bio-Sciences Corp., Piscataway, NJ) (bovine viral diarrhea virus-free). Monolayers were inoculated with passage 5 of VR-2332 at a multiplicity of infection (MOI) of 0.1, and the infections were allowed to progress for 48 h. At 48 h postinfection (h.p.i.), supernatants were removed and the monolayers were washed in phosphate-buffered saline (PBS; 10 mM Na₂HPO₄/KH₂PO₄, 137 mM NaCl, pH 7.4) prior to fixation with a 4% solution of paraformaldehyde (Electron Microscopy Sciences, Hatfield, PA)–PBS, followed by permeabilization with a 0.2% solution of Triton X-100 (Sigma-Aldrich, St. Louis, MO)–PBS. Fixed and permeabilized monolayers were analyzed by indirect immunofluorescence using the primary antibodies α-nsp2-OTU (1:100), α-nsp2-HV (1:100), α-nsp2-TM (1:50), and α-nsp2-C (1:50) or α-N SDOW-17-A (mouse) (1:50) (Rural Technologies Incorporated, Brookings, SD) followed by detection with the secondary Alexa Fluor 546 goat α-rabbit IgG (H+L) (Invitrogen, Carlsbad, CA) (at a final concentration of 80 μg/ml [1:25 dilution in PBS]) and fluorescein isothiocyanate (FITC)-conjugated α-mouse IgG (Sigma-Aldrich, St. Louis, MO) (1:12.5 dilution in PBS). Images were acquired with a Leica DM IRBE microscope in connection with a Leica DFC500 digital camera and the Leica Application Suite (v3.7.0) at magnifications between ×200 and ×400.

Virus purification.

For the generation of purified viral stocks used in this study, low-passage-number (≤15 passages) MARC-145 cells were seeded at a density of ∼5.0 × 10⁴ cell/cm² and allowed to incubate for 3 days in 1× MEM prior to inoculation. Low-passage-number viral stocks (passage 5 or below) of field isolates VR-2332 (U87392), recombinant JXwn06 (rJXwn06) (EF641008), rSRV07 (JX512910), MN184C (EF488739), Lelystad virus (M96262), and a rVR-2332 strain (rV7) encoding a myc tag in the hypervariable region of nsp2 (rV7-myc) (FJ524377) (37) were used to infect confluent MARC-145 monolayers (2,500 cm² per isolate) at an MOI of 0.1 and maintained at 37°C and 5% CO₂ until a 60% to 80% cytopathic effect (CPE) was reached (roughly 2 days). (Please refer to previous publications [36, 38, 39] for analysis of the growth kinetics for each study strain). Cell supernatants were collected and clarified at 5,000 × g and 4°C for 1 h to remove nonadherent cells and large cellular debris prior to downstream purification of cell-free virions. The resulting clarified viral supernatants were pelleted through the use of sterile 0.5 M sucrose in a TNE buffer (10 mM Tris-HCl [pH 7.0], 0.1 M NaCl, 1 mM EDTA) cushion at 104,000 × g for 3 to 4 h at 4°C to selectively exclude the pelleting of viral and cellular components of <1 g/ml density. Viral pellets were gently reconstituted in low volumes of TNE buffer prior to overlaying on top of 1.29 g/ml density cesium chloride (CsCl) (optical grade; Life Technologies, Carlsbad, CA) in TNE buffer and subjected to banding at 151,000 × g for 12 to 24 h at 4°C. Viral bands were collected by side tapping with a sterile needle followed by a second round of purification through CsCl gradient banding (2× CsCl). The virus (2× CsCl purified) was diluted 10-fold in TNE buffer and pelleted at 151,000 × g for 2 h at 4°C. Final purified viral pellets were eluted in low volumes of TNE buffer–10% protease inhibitor cocktail (P8340; Sigma-Aldrich, St. Louis, MO) and subjected to downstream analysis by immunoelectron microscopy (IEM) and Western blotting. Alternatively, after the initial pelleting step through the 0.5 M sucrose cushion, viral and mock isolates were purified by semidiscontinuous sucrose gradients generated as previously described (40). Since a viral band was not visible by sucrose purification, final samples were collected by side-tapping the sucrose gradients and collecting density ranges between 20% and 30% sucrose plus TNE, estimated to contain the density range of the PRRSV virion (1.18 to 1.19 g/ml [35]).

Immunoelectron microscopy.

To prepare purified samples for immunoelectron microscopy, a small volume (5 to 10 μl) of each purified PRRSV preparation was chemically cross-linked using electron-microscopy-grade paraformaldehyde (PFA) (Electron Microscopy Sciences, Hatfield, PA) at a final concentration of 4% for 5 min at room temperature to maintain the integrity of the virion during processing. Fixed samples were loaded onto nickel Formvar support grids (Electron Microscopy Sciences, Hatfield, PA), and remaining reactive aldehydes were inactivated by treatment of the grids with 0.5 M glycine–PBS followed by blocking overnight with 1% bovine serum albumin (BSA) (CAS 9048-46-8; Sigma-Aldrich, St. Louis, MO)–PBS at 4°C. Immunodetection of virions was accomplished by probing with primary α-nsp2 antibody diluted in 1% BSA–PBS to a final dilution of 1:100 (2 h at 4°C), and the secondary antibody goat α-rabbit was conjugated to 1- to 6-nm-diameter colloidal gold (CG) particles, used at 1:10 dilution (Electron Microscopy Sciences, Hatfield, PA) (4°C for 30 min). After each antibody incubation step, grids were washed a minimum of 4 times for 5 min each time with 1% BSA–PBS at room temperature. Probed samples (grids) that were labeled with 1-nm-diameter CG were further processed with an Aurion R-gent SE-EM high-efficiency silver enhancement system per the manufacturer's instruction (Electron Microscopy Sciences, Hatfield, PA) to increase visibility, resulting in large, slightly pleomorphic punctate labels of various sizes greater than the original 1 nm diameter (at times larger than 20 nm). Samples labeled with 6-nm-diameter CG were not silver enhanced. Images were acquired using an FEI Tecnai G² transmission electron microscope operated at 80 kV with a Hamamatsu Photonics ORCA-HR digital charge-coupled-device (CCD) camera in conjunction with the FEI Tecnai G² software platform.

Western blot analysis.

Purified virus samples were resolved under reducing and denaturing conditions using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and 4% to 12% Bis-Tris Novex NuPage gels in conjunction with an XCell Surelock mini-cell electrophoretic module with 2-(N-morpholino)ethanesulfonic acid (MES) running buffer per the manufacturer's instructions (Life Technologies, Carlsbad, CA). Resolved proteins were transferred to nitrocellulose membranes for Western blot detection by the use of an iBlot dry transfer system (Life Technologies, Carlsbad, CA) and blocked at room temperature for ≥1 h in 5% (wt/vol) dehydrated milk–PBS–0.05% Tween 20 (PBST) with shaking. nsp2 detection proceeded by probing with the primary α-nsp2 antibodies (OTU and HV) (rabbit) or with the α-myc monoclonal antibody (MAb) 9E10 (mouse) (Developmental Studies Hybridoma Bank, University of Iowa, IA), diluted in filtered 5% milk–PBST at a ratio of 1:1,000 (α-nsp2 antibodies) or 1:1,500 (9E10) overnight at 4°C with rocking, followed by secondary goat α-rabbit conjugated horseradish peroxidase (HRP) (IgG HinL chain; SouthernBiotech, Birmingham, AL) or goat α-mouse-conjugated horseradish peroxidase (Santa Cruz Biotechnology, Dallas, TX) antibody diluted in filtered 5% milk–PBST at a ratio of 1:2,500 and incubated overnight at 4°C with rocking. To detect nsp2 isoforms, Western blots were treated with chemiluminescent/chemifluorescent ECL Plus substrate (Pierce, Rockford, IL) and imaged using chemiluminescent film (Kodak, Rochester, New York) or digital capture (G:Box Chemi XT4; Syngene, Frederick, MD).

RESULTS

Characterization of intracellular nsp2 during PRRSV infection.

Polyclonal affinity-purified anti-peptide antibodies were produced in rabbits against four different regions of the nsp2 protein denoted α-nsp2-OTU (VR-2332 amino acids [aa] 69 to 82), α-nsp2-HV (VR-2332 aa 633 to 646), α-nsp2-TM (VR-2332 aa 630 to 643), and α-nsp2-C (VR-2332 aa 1183 to 1196) (Fig. 1). The selected epitopes represented conserved regions of the PRRSV strains examined, except for type 1 strain Lelystad in some cases. Three antibodies, α-nsp2-OTU, α-nsp2-HV, and α-nsp2-C, were found to specifically stain PRRSV-infected MARC-145 cells (200×) in primarily perinuclear regions and did not cross-react with any uninfected cell component (Fig. 2A). However, in a subset of infected cells, a diffuse punctate pattern, characteristic of endoplasmic reticulum (ER)/Golgi compartments, was evident throughout much of the cell by α-nsp2-OTU staining at higher magnification (×400) (Fig. 2B). A similar staining pattern was noted with the α-nsp2-HV antibody, although to a lesser extent, and the bulk of staining was localized to the perinuclear region (Fig. 2). nsp2 was not detected by the α-nsp2-TM antibody by immunofluorescence assay (IFA) (Fig. 2), perhaps because this region is predicted to be embedded in the membrane, rendering the epitope inaccessible under the experimental conditions used, or because the antibody itself was ineffective. Finally, α-nsp2-C detection localized primarily to regions directly surrounding the cell nuclei (Fig. 2). Surprisingly, in some cases, nsp2 densely localized to long cellular extensions as detected by each of the three antibodies (Fig. 3A). These extensions extend the length of two or more cells and are found to correspond to moderate to high-intensity nsp2 detection along the length of the filament. PRRSV nucleoprotein (N) colocalized with nsp2 at the greatest intensity at the perinuclear regions of virally infected cells. N was additionally identified to localize to a lesser extent near the distal and proximal regions of the nsp2 (+) cellular projections (Fig. 3B).

Fig 1.

Peptide sequence alignment of nsp2 epitopes. The nsp2 peptide sequences from genetically diverse study strains of PRRSV, including the European type 1 prototype Lelystad virus and the North American type 2 prototype VR-2332 strain, the regional North American MN184 strain, and the Asian highly pathogenic PRRSV rJXwn06 (China) and rSRV07 (Vietnam) isolates, were aligned using the ClustalW algorithm (55) in the parent program Geneious 6.1.5 (Biomatters Limited) and transferred to Jalview 2.8 (56) for percent identity coloring. The intensity of shading depicts the level of sequence similarity between strains; the darkest shaded areas represent a fully conserved residue or region, and absence of shading represents a low level of sequence conservation at that position. Alignment regions are shown for each epitope sequence targeted by custom α-nsp2 antibodies. The linear epitope sequence range is denoted above the alignment in reference to the VR-2332 nsp2 protein sequence (AA = amino acid). The diagram depicts nsp2 in reference to VR-2332 nsp2, PLP2 (papain-like protease 2/OTU enzymatic domain), HV (large central hypervariable region), and TM1 to TM4 (hydrophobic region; predicted transmembrane regions 1 to 4).

Fig 2.

Immunofluorescence detection of nsp2 within virally infected cells. (A) MARC-145 cell monolayers inoculated with VR-2332 at an MOI = 0.1 (48 h.p.i.) were fixed and permeabilized prior to detection with the primary antibodies α-nsp2-OTU (1:50), α-nsp2-HV (1:50), α-nsp2-C (1:100), and α-nsp2-C (1:50) followed by indirect fluorescence detection with secondary goat α-rabbit Alexa Fluor 488 antibody (1:25) (Invitrogen, Carlsbad, CA), imaged at ×200 to ×400 magnification. Images were collected using a consistent exposure and saturation setting for all images within each channel. For each antibody, an uninfected cell control (negative) is shown, labeled with identical primary and secondary antibody concentrations. DAPI, 4′,6-diamidino-2-phenylindole. (B) Selected high-magnification images of virally infected monolayers. Representative images of α-nsp2-OTU, α-nsp2-HV, α-nsp2-TM, and α-nsp2-C are shown.

Fig 3.

nsp2 localizes to a high concentration within cell-to-cell protrusions. (A) nsp2 was confirmed to localize to these projections by the use of a variety of α-nsp2 antibodies, including α-nsp2-OTU (I and II), α-nsp2-HV (III and IV), and α-nsp2-C (V). (B) Nucleocapsid colocalizes with nsp2 at the proximal and distal regions of the cell-to-cell projections. VR-2332-infected MARC-145 cell monolayers dually stained for nsp2 (α-nsp2-OTU) and N (α-nucleocapsid; SDOW17-A) revealed punctate foci of N along the axis of some nsp2(+) cell-to-cell projections (Merge, right cellular projection).

Procedural validation of PRRSV particle purification.

Viral supernatants were generated from infections at low multiplicities and were harvested prior to any overt destruction of the cellular monolayer or cellular necrosis (60% to 80% CPE, 0 to 5% monolayer detachment [by surface area]), thereby limiting release of any cellular nsp2 into the supernatant. Cell-free virions were pelleted through a 0.5 M sucrose exclusion cushion to limit collection of cellular and viral components of <1 g/ml density. Sucrose cushion pellets were resuspended and subsequently purified by two sequential continuous CsCl self-forming gradients, where banded viral particles were harvested at each step by side-tapping with a syringe. Alternatively, virions were initially also purified by semidiscontinuous (10% to 60%) sucrose gradients (see Fig. 6). As a negative control, uninfected cell cultures of equal surface areas and volumes of media were processed by the same methods as viral samples, using the viral bands of PRRSV-positive isolates as a reference for proper density collection. Collection volumes were kept consistent across all samples. The purified viral samples of each PRRSV strain were demonstrated to be infectious and produced typical CPE on MARC-145 cells, with recovery rates of final CsCl-purified virus from supernatant at 12% to 30% by 50% tissue culture infective dose (TCID₅₀) (Spearman-Kärber method) (41, 42) (Table 2).

Fig 6.

nsp2 packaging is maintained across genetically diverse strains of PRRSV. MARC-145-derived viral supernatants of PRRSV isolates of broad genetic heterogeneity, including the Lelystad virus (EU prototype), VR-2332 (NA prototype), rSRV07 and rJXwn06 (HP-PRRSV), and NA regional MN184 isolate, as well as the uninfected control, were individually collected and subjected to virus purification through the use of semidiscontinuous sucrose gradients as outlined in Materials and Methods. Samples were separated by SDS-PAGE under denaturing/reducing conditions prior to transfer to nitrocellulose and subjected to Western blot analysis using the α-nsp2-OTU primary antibody followed by detection with the goat α-rabbit HRP secondary. Blots were exposed using the chemiluminescence/chemifluorescence ECL+ substrate and imaged by digital capture via the Syngene G:Box.

Table 2.

Infectivity of viral strains before and after purification based on 50% tissue culture infective dose assay^a

Viral strain	Total virus in supernatant	Total purified virus	% recovery
Lelystad	4.27 × 108	1.26 × 108	30
VR-2332	9.13 × 107	2.25 × 107	25
MN184	1.95 × 108	2.25 × 107	12
rJXwn06	9.13 × 107	2.25 × 107	25
rSRV07	1.95 × 108	4.00 × 107	20

^aLow-passage-number MARC-145 cells were either mock infected or inoculated with one of five strains of PRRSV (VR-2332, rJXwn06, rSRV07, MN184, or Lelystad virus), and the infection was allowed to progress to 60% to 80% CPE (virally infected) or 2 to 3 days p.i. (uninfected cell control). At the time of collection, supernatants were clarified at 5,000 × g (4°C, 1 h) and an aliquot was taken for titration prior to downstream purification. The titers of the initial clarified supernatant as well as the final purified isolates were determined by a 50% tissue culture infective dose assay (TCID₅₀) for each strain purified. Initial and final quantities of total virus were calculated (TCID₅₀/ml × volume [ml]) from the supernatant and purified isolates, respectively. Recovery was defined as the percentage (%) of total virus remaining after purification in relation to the starting initial amount, i.e., [total virus (purified)/total virus (supernatant)] × 100.

nsp2 associates with highly purified PRRSV particles as shown by immunoelectron microscopy.

To determine if nsp2 was associated with the PRRSV particle, purified virus and uninfected control samples were assessed by immunoelectron microscopy using antibodies against nsp2. Electron micrographs of the purified uninfected cell control were generally free of observable contaminants, and no vesicles or subcellular structures were noted; however, occasional protein aggregates were observed (Fig. 4A). Purified viral isolates were found to contain pleomorphic, predominantly spherical virions of regular size (57.9 nm ± 8.1 nm diameter, n = 50) consistent with published results (43) (Fig. 4B to E). Immunoelectron microscopy (IEM) analysis using both α-nsp2-OTU and α-nsp2-HV detected nsp2 associated with particles from all purified viral isolates tested, as shown for purified rSRV07 (Fig. 4B and C) and VR-2332 (Fig. 4E). Immunoelectron microscopy was initially conducted with the Lelystad strain; however, the low virion load and high degree of background on the sample grid were not overcome. Detection of the OTU epitope (Fig. 4B and E) was greater than that of the HV epitope (Fig. 4C), and the transmembrane (TM) and C-terminal (C) epitopes were not detectable in the IEM protocol. Detection of virion-associated nsp2 by IEM was not consistent across all viral particles (Fig. 4B and C), which may be at least partly explained by the highly stringent nature of the assay conditions.

Fig 4.

IEM detection of nsp2 on PRRSV virions. Purified virions of PRRSV isolates rSRV07 (HP-PRRSV) and VR-2332 (NA prototype) or the purified product of the uninfected cell control (referred to as purified cell [−]) were labeled by α-nsp2 and imaged by electron microscopy between ×32,000 and ×150,000 magnification. (A) Purified cell (−) labeled with the α-nsp2-OTU (1:100) plus secondary goat α-rabbit conjugated to 1 nm CG (1:10) plus silver enhancement. (B and C) Purified virions of HP-PRRSV strain rSRV07 labeled with either the α-nsp2-OTU primary (1:100) plus goat α-rabbit–1 nm CG (1:10) plus silver enhancement (B) or the α-nsp2-HV primary (1:100) plus goat α-rabbit–6 nm CG (1:10) (no enhancement) (C). (D) Unlabeled VR-2332 purified virions isolated using the “standard” purification technique. (E) Immunodetection of nsp2 (α-nsp2-OTU) on VR-2332 purified virions isolated using an elongated purification protocol.

To further address the possibility that residual unbound nsp2 was being copurified through a nonspecific interaction with the virion surface, viral particles were subjected to an elongated purification method where banding times within the cesium chloride purification gradients were increased to greater than 48 h, with the intent to strip residual protein material closely associated to the virions (noted in Fig. 4B and C as a dark halo around the outer side of the virions). Upon centrifugation for this elongated period, a significant fraction of the sample was lost due to additional osmotic and mechanical stresses placed on the virion during this protracted purification. The resulting virions displayed increased virion diameter and pleomorphism but had lost the dark halo surrounding the outer side of the virion (Fig. 4D and E). Specific nsp2 staining remained detectable on these extensively purified viral particles, indicating that nsp2 is strongly associated with, or most likely incorporated in or on, the PRRSV virion (Fig. 4E).

Multiple isoforms of nsp2 were present within purified PRRSV particles.

To characterize the anti-peptide antibodies to nsp2 in Western blot assays, increasing amounts (10 to 40 μg) of VR-2332- and rJXwn06-infected as well as uninfected MARC-145 cell lysates were separated by SDS-PAGE and transferred to nitrocellulose followed by probing with each α-nsp2 antibody. Consistent with previous observations (38, 44), multiple bands were seen in VR-2332- and rJXwn06-infected cell lysates as detected by α-nsp2-OTU, α-nsp2-HV, and α-nsp2-C (Fig. 5). The uppermost band (∼130 kDa) migrated similarly using all three antibodies compared with the Super Signal ladder (Thermo Scientific) (Fig. 5), whereas the two bands directly below the uppermost band had similar migration rates when probed with α-nsp2-OTU and α-nsp2-HV (Fig. 5A and B). Lower-molecular-mass protein bands were also seen, some of similar migration. As noted in the IFA data, no PRRSV-specific proteins were recognized by α-nsp2-TM (data not shown). Detection of rJXwn06 nsp2 proteins exceeded that of VR-2332, most likely due to the fact that rJXwn06 replicates to 10-to-100-fold-higher titers than VR-2332 in MARC-145 cells (2).

Fig 5.

Western blot detection of nsp2 in cell lysates by custom antibodies. MARC-145 cell lysates of VR-2332 or rJXwn06 or of uninfected negative controls harvested at 48 h.p.i. were titrated at 10 μg, 20 μg, 30 μg, and 40 μg (per well) by SDS-PAGE Western blot analysis. Full-length nsp2 as well as lower-molecular-mass isoforms were identified with the primary antibodies α-nsp2-OTU (1:2,500) (A), α-nsp2-HV (1:2,500) (B), and α-nsp2-C (1:2,500) (C). L, Super Signal ladder.

To further investigate nsp2 virion association, Western blot analysis was performed under denaturing, reducing conditions on purified viruses rJXwn06, VR-2332, rSRV07, MN184, and Lelystad, representing diverse genotypes, and on gradient-purified uninfected cell controls (Fig. 6). Multiple bands of different molecular masses (nsp2 isoforms) were detected within PRRSV-positive samples but not within the purified uninfected and gradient-banded cell control. Full-length nsp2 proteins of 126 kDa (rSRV07) or 129.6 kDa (VR-2332) were expected, as well as additional protein bands that had been observed previously (38, 44). Initially, multiple isoforms of nsp2 were detected between 181 and 48 kDa for all PRRSV strains examined at low concentrations of virus, consisting of a central set of similarly migrating products detected by α-nsp2-OTU, as measured in reference to the Benchmark prestained molecular mass marker (Invitrogen, Carlsbad, CA), different from the Super Signal Ladder used previously. This initial purification strategy employing a semidiscontinuous sucrose gradient yielded positive identification of nsp2 within the viral fraction (Fig. 6). However, the dual rate-zonal and isopycnic centrifugation properties of sucrose gradients resulted in diffuse density collection ranges and, with that, an increased probability of cocollection of nonvirion protein contaminants. In order to better assess the relationship of nsp2 with the virion, a stringent purification procedure was established by pelleting through a sucrose cushion followed by multiple rounds of banding within a continuous cesium chloride (CsCl) gradient at high centrifugal forces (Fig. 7 and 8). Western blot detection of 2-fold dilutions of CsCl-purified rSRV07, VR-2332, and the uninfected cellular control was completed to facilitate observation of all bands at optimal intensities and to distinguish dominant products for each strain of PRRSV. This more detailed analysis at higher concentrations of these two viruses revealed the same central set of nsp2 bands, indicated by the arrows; however, different intensities were noted between strains for some of the similarly migrating nsp2 isoforms, as shown by the star symbol (Fig. 7A and B). Detection of virion-associated nsp2 products by α-nsp2-HV resulted in similar banding patterns as noted by the use of α-nsp2-OTU (Fig. 7B [OTU] and C [HV]); however, different intensities of many bands (i.e., 87 kDa, 76 kDa, and 30 kDa) were apparent. It is of note that the total number of bands detected was more than had been previously reported (38). These differences could be partially due to the different experimental conditions (direct detection of purified virions versus immunoprecipitation of cell lysates) and sensitivities of detection as defined by the assays and antibodies used. Strain-specific bands were also observed; additional products migrating at 181, 174, 30, 24, and 18 kDa were seen in purified VR-2332 particles that were not present in rSRV07 particles (Fig. 7). The noted high-molecular-mass products larger than the predicted full-length nsp2 may be the result of posttranslational modifications or of incomplete replicase polypeptide processing.

Fig 7.

Identification of dominant nsp2 products in purified virions. Highly concentrated purified viral samples were generated to identify dominant nsp2 protein products associated with the virion. (A and B) Two-fold dilutions (lanes 1:1 to 1:64) of the purified rSRV07 and VR-2332 viral isolates or of the purified cell (−; lanes C) (only the highest concentration of 1:1 is shown) of identical surface areas, volumes of media, purification strategies, and collection volumes were denatured and reduced prior to electrophoretic separation by SDS-PAGE to differentiate closely migrating high-intensity minor bands. Black arrows: a consistent core set of nsp2 products was noted between strains. Black star: repetitive similarly migrating bands of differing intensities were noted between strains. (B and C) Similar banding patterns were noted for the core set of nsp2 isoforms between VR-2332 as detected by the α-nsp2-OTU (B) and α-nsp2-HV (C), although the ultimate intensities of these bands are noted to differ.

Fig 8.

Multiple nsp2 isoforms are packaged within recombinant VR-2332 expressing myc-tagged nsp2. rV7-Myc expressing a modified c-myc-tagged nsp2 was amplified in MARC-145 cells and purified as described in Materials and Methods. Purified virions of the nsp2-tagged rV7-Myc or untagged parental VR-2332 virus or the purified uninfected cell control were assessed by Western blot analysis using the α-myc MAb 9E10 detection antibody followed by labeling with the secondary goat α-mouse HRP conjugate and assessed in relation to the Kaleidoscope protein molecular mass marker.

Epitope-tagged nsp2 is incorporated into purified PRRSV particles.

To further confirm nsp2 packaging in PRRSV virions, a recombinant strain of the parental VR-2332 virus expressing an exogenous c-myc epitope within the hypervariable region of nsp2 (rV7-Myc) was purified by multiple rounds of CsCl gradient purification and assessed by Western blot analysis in relation to the similarly purified products of the uninfected control and the untagged parental VR-2332 strain. Myc-tagged nsp2 was detected using the primary α-myc 9E10 monoclonal antibody (MAb). Multiple myc-tagged nsp2 isoforms were identified as present within the rV7-Myc-purified virions but not within the purified uninfected control or the purified virions of the untagged parental VR-2332 strain (Fig. 8). The 3× c-myc epitope tag replaced amino acids 323 to 431 in relation to the parental VR-2332 virus strain. The net molecular mass change of the rV7-Myc nsp2 is therefore expected to be a reduction of approximately 10 kDa compared to the wild-type VR-2332 nsp2. A total of eight nsp2-myc isoforms were detected at 119 (full-length nsp2), 82, 71, 66, 55, 53, 51, and 19 kDa as measured in reference to the Kaleidoscope molecular mass marker (Fig. 8). These results further verify, using a different antibody, that several isomers of nsp2 were packaged as part of the PRRSV virion.

DISCUSSION

The rapid evolution of PRRSV has yielded a highly diverse antigenic population of PRRSV isolates. A substantial fraction of these mutation and recombination events are noted to occur within the coding region of nsp2, the largest and most genetically diverse protein encoded by PRRSV. While nsp2 is encoded by 21% to 23% of the PRRSV genome, it accounts for 23% to 48% of the pairwise nucleotide differences between the five wild-type study strains presented in this report (Table 3 and data not shown). Previous research has demonstrated that multiple isoforms of nsp2 are generated through viral replication within the permissive cell line MARC-145 (38, 44); we have additionally noted the presence of nsp2 isoforms within virally infected porcine alveolar macrophages (data not shown). Transfection of a VR-2332 nsp2-3 construct alone was also shown to be sufficient to yield multiple nsp2 products (18), a subset of which were of lower molecular mass than the recently reported translational isoforms (44). The presence of lower-molecular-mass nsp2 protein products suggests that cleavage isoforms may be generated by the viral PLP2 protease or cellular proteases or both (18, 38). The role(s) of these isoforms in the viral replication cycle is yet to be determined but may facilitate differential localization of nsp2 functional domains within the virally infected cell or allow for preferential packaging of selected isoforms.

Table 3.

Percent pairwise identity between strains^a

Alignment and strain	% pairwise identity with strain:
	VR-2332	MN184C	rJXwn06	rSRV07	Lelystad
Whole genome (nucleotide)
VR-2332	100.0	84.9	89.2	89.1	61.5
MN184C		100.0	83.2	83.2	61.8
rJXwn06			100.0	99.4	61.4
rSRV07				100.0	61.4
Lelystad					100.0
Nsp2 (nucleotide)
VR-2332	100.0	72.8	82.7	82.4	52.1
MN184C		100.0	70.9	70.9	55.5
rJXwn06			100.0	99.4	55.6
rSRV07				100.0	55.4
Lelystad					100.0
Nsp2 (protein)
VR-2332	100.0	70.3	78.8	78.4	37.7
MN184C		100.0	67.4	67.3	38.2
rJXwn06			100.0	99.4	37.2
rSRV07				100.0	37.1
Lelystad					100.0

^aNucleotide (whole genome and nsp2 only) and protein (nsp2) sequence pairwise alignments were generated using the Geneious alignment algorithm within the Geneious R6 (v6.1.6; Biomatters Limited) software package for each combination of study strains (VR-2332 [U87392], rJXwn06 [EF641008], rSRV07 [JX512910], MN184C [EF488739], and Lelystad virus [M96262]). Data are presented as percent homology between whole genomes, nsp2 nucleotide sequences, and nsp2 peptide sequences.

nsp2 was positively identified upon the outer side of the viral envelope of the virion by IEM using antibodies to the viral OTU domain and the HV region, but not by antibodies to the predicted TM region or the C-terminal epitope. Failure to detect nsp2 at or downstream of the hydrophobic region by IEM might suggest that nsp2 was packaged as an integral membrane protein, that the native structure precluded detection, or that these antibodies were not well suited for the experimental conditions under which IEM was completed. Previous work has shown that nsp2 and nsp3 proteins of equine arteritis virus (EAV) were necessary and sufficient to induce cellular membrane rearrangements similar in appearance to those typically observed during viral infection (23). PRRSV nsp2 possesses a hydrophobic domain downstream of the HV region that is predicted to function in cellular membrane rearrangement in a manner similar to that seen with EAV nsp2. It is therefore predicted that PRRSV nsp2 associates with the virion during packaging through integration with the membrane of the viral envelope; however, further studies are needed to define the relationship of nsp2 with the virion.

Purified virions possessed multiple isoforms of nsp2, whereas packaged isoforms were similar in size to a subset of the isoforms detected within virally infected cell lysates. Highly concentrated purified isolates were additionally found to possess a “core” set of nsp2 isoforms, as well as strain-specific isoforms. While it was unlikely that the α-nsp2 custom antibody set was cross-reacting with an enriched cellular product copurifying with the virion density fraction (and not identified in the uninfected cellular control), to eliminate this possibility, we verified nsp2 packaging using a recombinant VR-2332 strain expressing a c-myc-tagged nsp2 (rV7-Myc), which further confirmed our initial results (Fig. 8). Reduced growth kinetics for rV7-Myc hindered generation of purified viral stocks to as high a concentration as that of the wild-type viruses, particularly VR-2332 and the HP-PRRSV isolates rJXwn06 and rSRV07. Due to the lower initial viral titers, it is possible that minor nsp2 isoforms were not identified with this virus.

The function(s) and exact boundaries of nsp2 isoforms are yet to be determined; however, it was generally noted that IFA detection of nsp2 by three (OTU, HV, and C) of the four target epitopes did maintain subtle differences in cellular localization (Fig. 2). nsp2 predominantly localized to the perinuclear region of infected cells (VR-2332, 48 h.p.i) in a manner similar to what has been previously described for PRRSV (45) and EAV (46) when assessed by IFA (Fig. 2). Differences in nsp2 localization as detected by targeting the N-terminal protease domain, the central HV region, or the C-terminal domain were generally subtle but were consistent across a range of replicates and antibody dilutions (Fig. 2B). It is postulated that some isoforms, such as those possessing the recognized hydrophobic region, may maintain a constrained localization in and around internal membrane structures whereas isoforms lacking this domain may adopt a more diffuse localization pattern. It is currently unknown if nsp2 isoforms localize preferentially to different regions of the cell but would allow spatial separation of nsp2 functional domains as a mechanism of targeted activity.

Interestingly, nsp2 was observed to also localize to cellular projections that appeared to connect two or more cells over distances of multiple cell lengths, consistent between the three target epitopes (OTU, HV, and C terminus) (Fig. 3). Cellular projections were noted to occur in equal frequencies in infected and uninfected cells and therefore are not believed to be virally derived or nsp2 derived; however, nsp2 densely localized to these protrusions. In some occurrences, the cell-to-cell process, extending from an infected cell to an uninfected cell, resulted in a high-intensity point of nsp2 staining at the apparent point of contact with the membrane/cytosol of the uninfected cell (Fig. 2A, panel IV, lower center). Previous work by Cafruny et al. indicated that a cell-to-cell transmission route occurs for PRRSV within MARC-145 cells during the midphase viral growth cycle (>24 h postinfection) (47), as measured by α-N IFA detection of MOI-independent viral focus development. Similarly, in this study, N protein was found to weakly colocalize with nsp2 at distal and proximal regions of nsp2 (+) cellular projections (Fig. 3B, Merge) and with a punctate pattern along the axis of some projections (Fig. 3B, right projection).

Investigations into the role of nsp2 in viral replication and immune evasion have yielded an increasingly complex understanding of nsp2 mechanisms. This has been further compounded by the recent descriptions of translational (44) and cleavage (38) nsp2 isoforms of as-yet-unknown function. nsp2 is a region of substantial diversity at both the genetic and protein levels, and emerging strains often possess novel mutations, insertions, and deletions within this coding region (12, 30, 48). nsp2 is the most genetically diverse protein of PRRSV; however, the mechanism potentiating the higher mutation rate within the nsp2 coding region has not been defined. Curiously, nsp2 elicits a strong humoral immune response in animals experimentally infected with PRRSV, and the magnitude of that antibody response(s) is similar to those generated against the highly immunogenic and highly abundant N protein by 14 days postinfection (27). N protein is a dominant component of the PRRSV virion (20% to 40% of the total virion protein mass) (28, 49), so it was therefore surprising that the nsp2 replicase protein with no known extracellular localization or function could elicit an antibody response commensurate with that generated against the N protein. Likewise, it has been shown that hosts infected or vaccinated with the arterivirus prototype virus EAV established a serological response to nsp2 (50), suggesting that the mechanism by which the anti-nsp2 humoral immune response is elicited may be conserved across the Arteriviridae. Genetic alignment of the nsp2 coding region of PRRSV demonstrates a large central bipartite region of high genetic variability (Fig. 1 and unpublished results), which includes a number of single nucleotide polymorphisms (SNPs), insertions, and deletions between strains (12, 51, 52). Much of this central hypervariable region was found to be nonessential for the North American type 2 prototype strain VR-2332 to replicate within the MARC-145 cell line (37). A viral mutant possessing a large deletion within the hypervariable region of nsp2 (Δ324 to 726 aa) was also found to be replication competent in vivo, yet with reduced kinetics relative to wild-type virus or to other mutants with smaller nsp2 hypervariable deletions (Δ727 to 813 aa, Δ543 to 726 aa, and Δ324 to 523 aa) (53). The major factor(s) potentiating the high evolutionary rate of nsp2 is currently unknown; however, multiple experiments have delineated T-cell and B-cell epitopes within the nsp2 coding region either bioinformatically or experimentally (17, 21, 29, 30, 54). Inclusion of nsp2 within the PRRSV virion suggests that it may function in previously unknown roles related to extracellular function, entry, or immediate-early viral replication events. The known robust humoral immune response generated against nsp2, coupled with the observed high genetic diversity within this coding region, points toward a driving selective pressure against the nsp2 protein acting to alter or hinder one or more requisite functions. nsp2 packaging was conserved across a genetically diverse set of study strains ranging from original outbreak isolates to contemporary highly pathogenic strains, suggesting that this is a wholly conserved feature of PRRSV. These packaged nsp2 isoforms shared a consistent core set between viral strains; however, additional strain-specific products were also identified. These findings suggest that the ultimate antigenic composition of the virion may be more complex than originally predicted. Identification of PRRSV nsp2 packaging establishes a foundation for new considerations relevant to future vaccine design and has important implications for as-yet-unexplored functions of nsp2 as a structural protein.