MARC-145 cells were maintained in Dulbecco’s modified Eagle’s medium (HyClone) supplemented with 10% fetal bovine serum (HyClone) at 37°C with 5% CO₂. TA-12 (GenBank accession no. HQ416720.1), a highly pathogenic PRRSV isolate, was used in virus infection studies.

The web tool of PDBeMotif¹ was used for analyzing the motifs which may be functional in NSP7α. The alignment of the protein sequences were completed using DNAMAN software. PRISM2.0² was used for the prediction of bonding sites between NSP7α and NSP9 (Tuncbag et al., 2011).

The full-length PRRSV cDNA clone FL12 of NVSL 97-7895 (GenBank accession no. AY545985.1) was used as the templates for the amplification of the cDNAs encoding NSP7α and NSP7β. The fragment of nsp7α or nsp7β was then inserted into a vector as described in a previous study (Chen et al., 2016), for the expression of NSP7α or NSP7β with an N-terminal HIS-tag and a TEV protease cleavage site between the tag and the fused protein.

HIS-TEV-NSP7α and HIS-TEV-NSP7β were expressed in E. coli BL21 (DE3), and purified using Ni-NTA column (CWBIO, China). The 200 μg of each purified protein was mixed with equal volume of Freund’s complete adjuvant, and injected into rabbits in primary immunization. Rabbits were boosted three times at 2 weeks intervals, with mixture of 200 μg antigen in Freund’s incomplete adjuvant. The rabbit antiserums were collected 2 weeks after the last injection.

MARC-145 cells were infected with a MOI of 0.1, and lysed in SDS loading buffer at 8, 16, 24, 48, 72, and 96 h post infection (hpi). The proteins in cell lysis solution were separated by SDS-PAGE, and transferred onto a PVDF membrane. After blocking with 5% skimmed milk overnight at 4°C, the membranes were, respectively, incubated with anti-NSP7α and anti-NSP7β polyclonal antibodies, followed by incubation with goat-anti-rabbit IgG (CWBIO, China) conjugated horseradish peroxidase as secondary antibody. Immunodetection was performed using enhanced chemiluminescence (ECL) reagents (CWBIO, China) according to the supplier’s instruction.

Porcine reproductive and respiratory syndrome virus-infected MARC-145 cells were fixed at 8, 16, 24, 48, and 72 hpi using 4% paraformaldehyde, permeabilized using 0.2% Triton X-100 for 10 min at room temperature, then blocked with 5% skimmed milk overnight at 4°C. Subsequently, the cells were, respectively, incubated with polyclonal antibodies against NSP7α and NSP7β for 1 h. After incubated with FITC-conjugated goat-anti-rabbit IgG (CWBIO, China) for 1 h, the cells were stained with PI for 5 min. The fluorescence images were captured using a confocal microscope (Nikon, Japan).

Uniformly labeled NSP7α was expressed in E. coli BL21 (DE3) grown in M9 minimal media containing 0.2% (w/v) ¹³C Glucose and 0.1% (w/v) ¹⁵N NH₄Cl as the only carbon and nitrogen sources. HIS-tagged NSP7α was purified by affinity chromatography on a Ni-NTA column. The HIS-tag at the N terminus of NSP7α was removed overnight by TEV protease, subsequently. The final NMR sample was concentrated to about 0.8 mM in 20 mM KH₂PO₄ buffer (pH 7.5) with 92% H₂O/8% D₂O, 100 mM NaCl, and 1 mM TCEP. The residues of NSP7α were renumbered as residues 2–150 (the glycine from the C-terminus of TEV protease cleavage site was numbered as residue 1) in structural analysis.

NMR assignment of ¹⁵N, ¹³C labeled NSP7α was described in our previous report, and the ¹H, ¹³C, and ¹⁵N chemical shifts for NSP7a were deposited in the BMRB under BMRB accession number 30014 (Chen et al., 2016). All the NMR experiments were carried out at 298K on a Bruker Avance 600 MHz spectrometer. All spectra were processed using NMRPipe/Draw (Delaglio et al., 1995) and analyzed using CARA³ (Keller, 2004).

The structure of NSP7α was refined by CNS (Brünger et al., 1998) based on the chemical shift assignments and NOE distance restraints. Backbone φ and ψ dihedral angle restraints obtained using TALOS+ (Shen et al., 2009). H-bond restraints were determined by secondary structure elements in initial structure. The final 20 structures with lowest energies were selected for representation and analyzed using PROCHECK_NMR (Laskowski et al., 1996). All representing structure figures were generated using PyMOL⁴.

To investigate the interaction of NSP7α with the other NSPs, yeast two-hybrid assays were performed. NSP7α cDNA fragment of NVSL 97-7895 was cloned into pGADT7 and the other NSPs’ cDNA fragments were inserted into pGBKT7. Primer sequences for nsp7α cloning were listed in Table 1. The pGBKT7 constructs were, respectively, co-transformed with pGADT7-NSP7α into Saccharomyces cerevisiae (yeast) strain Gold using the Yeastmaker Yeast Transformation System kit (Clontech, USA). Yeast co-transformed with pGBK7-P53 and pGADT7-T was used as a positive interaction control, and yeast co-transformed with pGBKT7-Lam and pGADT7-T was used as a negative control.

The cDNA fragment encoding NSP7α was amplified by PCR using primer pairs BamHINSP7α-F/XhoINSP7α-R. The fragment of nsp7α was inserted into pTriEx-FLAG for the expression of FLAG-tagged NSP7α. The site-directed mutants of nsp7α were amplified using primers listed on Table 1 and cloned into pTriEx-FLAG. NSP9 cDNA fragment was cloned into pTriEx1.1 vector to make pTriEx-NSP9 construct for the expression of HIS-tagged NSP9 (Liu et al., 2016).

Protein mixtures of FLAG-tagged NSP7α and HIS-tagged NSP9 were incubated for 1 h at room temperature. The solution was incubated with nickel magnetic beads (Millipore, USA) for 0.5 h at room temperature. After being washed with wash buffer (20 mM TRIS-HCl [pH 7.9], 300 mM NaCl, and 10 mM imidazole) for three times, the bound proteins were eluted with elution buffer (20 mM TRIS-HCl [pH 7.9], 300 mM NaCl, and 300 mM imidazole). Eluted proteins were separated by SDS-PAGE, and then detected by Western blot. FLAG-tagged and HIS-tagged proteins were, respectively, probed by mouse anti-FLAG antibody (CWBIO, China) and mouse anti-HIS (CWBIO, China) antibody, followed by incubation with goat-anti-mouse IgG (CWBIO, China) conjugated with horseradish peroxidase as secondary antibody. The protein mixture containing NSP7α and HIS-tagged TEV protease was used as negative control.

Antisera against NSP7α and NSP7β were used for the detection of NSP7α and NSP7β expression in PRRSV infected MARC-145 cells. By Western blot, anti-NSP7α antibodies detected a band of approximately 16 kDa, corresponding to the predicted size of NSP7α, by 16 hpi (Figure 1A). The production of this protein reached a peak level at 48 hpi, and the protein amount gradually decreased from 48 to 96 hpi. A series of protein bands at higher molecular weights were also recognized. These bands probably represented the polyprotein precursors of NSP7, and the products in accordance with the predicted size of NSP3-8 and NSP5-7 were marked with arrows in the figure. However, no protein with the predicted size of NSP7 were detected in the PRRSV infected cells.

The antiserum against NSP7β also recognized the polyprotein precursors NSP3-8 and NSP5-7 (Figure 1B). It detected neither NSP7 nor NSP7β in the PRRSV infected cell lysates, probably due to the instability characteristics of the proteins. These results were consisted with the previous study that was carried out using PRRSV monoclonal antibodies (Li et al., 2012).

To visualize the intracellular expression and localization of NSP7 and its precursors, PRRSV infected MARC-145 cells were stained with the antisera and observed by immunofluorescence microscopy. Although the antiserum against NSP7β did not detect the cleaved form of NSP7β in Western blot, the two sera gave similar results in immunofluorescence assay, probably because the antiserum against NSP7β could detect the NSP7 precursors as well as the antiserum against NSP7α. As shown in Figure 2, the signal for NSP7α and NSP7β could be detected in the infected cells as early as 8 hpi and the fluorescent foci were small and mostly concentrated on one side of the nucleus. After 16 hpi, the fluorescence signals for both NSP7α and NSP7β were more diffused in the cytoplasm of the infected cells. By 72 hpi, both the number of fluorescent cells and the intensity of the fluorescence obviously decreased. Protein degradation, and partially loss of the infected cells during the experimental process could both contribute to the reduction of the NSP proteins.

The solution structure of PRRSV NSP7α was determined by NMR spectroscopy, and three dimensional structure was calculated using 2466 NOE distance restraints (inter-residue), 202 dihedral angle restraints and 85 H-bond restraints. Coordinates of the ensemble of the 20 lowest energy structures were deposited in the Protein Data Bank⁵ under PDB accession number 5I65. Parameters of the structure are listed in Table 2. As shown in Figure 3A, the ensemble of structures for the most domains are well superposed. NSP7α contains a bundle of three α-helices (α1 from residue Asp13 to Lys21, α2 from residue Phe41 to Gln56, α3 from residue Gln59 to Ala74) and six-strand mixed β-sheet (β1 containing residues Cys27 to Val29, β2 containing residues Ile86 to Leu90, β3 containing residues Phe99 to Val103, β4 containing residues His108 to Glu114, β5 containing residues Arg116 to Ala119 and β6 containing residues Met123 to Val129 in the order α1-β1-α2-α3-β2-β3-β4-β5-β6. The mixed β-sheet is formed by five antiparallel strands (β2 to β6) and one short parallel strand (β1) (Figure 3B).

In order to search for the functional motifs of NSP7α, the protein structure was analyzed using web tools of PDBeMotif⁶, but no recognizable functional motifs was identified. The surface charge distribution of the protein shows that the protein has some scattered electropositive and electronegative patches, and more of the electropositive patches locate on the surface of the β-sheet (Figure 3C).

Sequence alignment shows that PRRSV (NVSL 97-7895) NSP7α shares 33.33% identity with EAV (Bucyrus strain) NSP7α (Figure 4). It is noteworthy that the C-terminus of PRRSV NSP7α has a predominant proline-rich region (PRR), with 10 prolines out of the last 20 amino acids. In comparison, the C-terminus of EAV NSP7α is shorter and lack the PRR, but its NSP7β N-terminus has four proline residues and five other same amino acids as the C-terminus of PRRSV NSP7α and shares less identity with the N-terminus of PRRSV NSP7β. Therefore, this discrepancy could be attributed to the position change of the cleavage site between NSP7α and NSP7β. It is unfortunate that the structure of the C-terminal PRR was not solved, as prolines lack the amide protons necessary for the standard triple resonance experiment and the backbone of this region was not able to be assigned.

The solution structure of EAV NSP7α (PDB accession number: 2L8K) have previously been reported (Manolaridis et al., 2011). PRRSV NSP7α possesses similar secondary structures as EAV NSP7α (Figures 4, 5A,B). Both of them have three α-helices gathered together in the N-terminal half, and all the β-sheets of the two proteins are located at one side of the α-helices. Strand β4 and strand β5 of PRRSV NSP7α locate at a similar position with an equivalent length to the strand β4 of EAV NSP7α. However, overlay of the 3-D structure of the two proteins showed that the two structures were not well superimposed, and the best superimposed region located from helices α2 to helices α3 (Figure 5C).

Previous studies have reported that arterivirus NSP7α is essential for the viral replication and it plays an important role in viral RNA synthesis (Manolaridis et al., 2011; Zhang et al., 2013). It remains unknown whether the protein is involved in the replication and transcription complex (RTC) formation by interacting with other NSPs. To investigate the interaction of PRRSV NSP7α with other NSPs, yeast two-hybrid assays were carried out, and it was found that NSP7α could interact with NSP9 (Figure 6A), the viral RNA-dependent RNA polymerase. To validate the NSP7α–NSP9 interaction, FLAG-tagged NSP7α and HIS-tagged NSP9 were expressed in E. coli and used for pull-down assay. The data confirmed that NSP7α could bind to NSP9 in vitro (Figure 6B).

To investigate the interaction of NSP7α with NSP9 in more details, the 3-D structure of NSP9 was predicted (Liu et al., 2016), and used for the docking prediction of NSP7α–NSP9 interaction by using PRISM2.0 (Tuncbag et al., 2011). In the predicted interaction model with the lowest energy (-27.86 kcal/mol) among the outputs, three amino acids (K61, L69, and F72) in helix α3 of NSP7α were predicted to be involved in the recognition of NSP9 (Figure 7). Pull-down assay using the NSP7α mutants containing single substitutions at the three sites showed that mutations at L69 and F72 abolished the binding ability of NSP7α to NSP9, but the K61A mutant could interact with NSP9 as well as WT NSP7α (Figure 8). It is worth to note that L69 and F72 were conserved between PRRSV and EAV NSP7α, while K61 in PRRSV NSP7α was replaced by another basic amino acid R in EAV NSP7α (Figure 4). This result confirmed that helix α3 of NSP7α was a critical region involved in the NSP7α–NSP9 interaction, and it also indicated that the two conserved amino acids L69 and F72 were essential for the protein to maintain its ability binding to NSP9.