The improved BaMV-based expression vector developed in this study was modified from a previously constructed chimeric BaMV vector, pKB19 (Muthamilselvan et al., 2016; Jiang et al., 2019), which harbors BaMV RNA replicase, the silencing suppressor P19 of Tomato bushy stunt virus (TBSV) (519 bp, GenBank Accession No. AJ288926) under the control of a dual constitutive 35S promoter of Cauliflower mosaic virus (CaMV), and an Agrobacterium nopaline synthase (nos) terminator (Figure 1A). In addition, the SS^Ext derived from the N. benthamiana extensin protein with the fusion of the secretory booster, (SP)₁₀, was incorporated into the expression vector to generate the secretory expression vector, pKB19SS^ExtHis(SP)₁₀, which was further used to construct the plasmid pKB19SS^ExtHis(SP)_10-mIFNγ for the production of human mIFNγ glycoproteins (amino acid positions 21–166, GenBank accession number AY121833.1) with enhanced solubility and secretion (Jiang et al., 2020). However, such fusion tags may affect the conformation and biological activity of the recombinant mIFNγ glycoproteins by hampering the proper folding. Thus, in this study, we have improved the design of the BaMV-based protein expression system by the insertion of cleavable linkers, including L_SrtA, L_TEV, and L_SNAC, positioned in between the combinational secretory-affinity tag SS^Ext His(SP)₁₀ [containing a 6X His-tag between SS^Ext and (SP)₁₀] and mIFNγ to generate plasmids designated as pKB19SS^ExtHis(SP)₁₀L_SrtA-mIFNγ, pKB19SS^ExtHis(SP)₁₀L_TEV-mIFNγ, and pKB19SS^ExtHis(SP)₁₀L_SNAC-mIFNγ, respectively (Figure 1A). The construction process of the above plasmids is described in brief as follows. The fragment of 6X His-tag and (SP)₁₀L_SrtA-mIFNγ were synthesized by overlapping PCR with specific primers, including F-XbaI-LPETG-mIFNγ, F-(SP)₁₀-XbaI-LPETG, F-MluI-(SP)₁₀, and F-MluI-6XHis-GG-(SP)₁₀, individually plus R-SpeI-TGA-mIFNγ (Supplementary Table S1). The amplified PCR fragment of His(SP)₁₀L_SrtA-mIFNγ, containing 6X His-tag at the N-terminus, was digested with MluI and SpeI and ligated into the respective sites in the pKB19 expression vector to generate the chimeric plasmid pKB19His(SP)₁₀L_SrtA-mIFNγ. The SS^Ext signal was constructed by primer extension of two mutually complementary primers, F-MluI-SS^Ext and R-MluI-SS^Ext, followed by digestion with MluI. The digested product of the SS^Ext fragment was cloned into pKB19His(SP)₁₀L_SrtA-mIFNγ to generate the final chimeric plasmid pKB19SS^ExtHis(SP)₁₀L_SrtA-mIFNγ. The fragments of L_TEV-mIFNγ and L_SNAC-mIFNγ were amplified with the mIFNγ gene as a template using gene-specific forward primers, namely, F-XbaI-ENLYFQG-mIFNγ and F-XbaI-GSHHW-mIFNγ, respectively, plus the reverse primer R-SpeI-TGA-mIFNγ (Supplementary Table S1). The amplified PCR product fragments of L_TEV-mIFNγ and L_SNAC-mIFNγ were digested with XbaI and SpeI and ligated into the pKB19SS^ExtHis(SP)₁₀ secretory expression vector cut with cognate enzymes to generate the chimeric plasmids pKB19SS^ExtHis(SP)₁₀L_TEV-mIFNγ and pKB19SS^Ext His(SP)₁₀L_SNAC-mIFNγ, respectively.

The transient expression of recombinant proteins in N. benthamiana was achieved using Agrobacterium-mediated infiltration. A. tumefaciens (PGV3850 strain) harboring the expression constructs of different FGs were cultured, harvested by centrifugation at 12,000 rpm for 1 min, and recovered in agro-infiltration buffer (10 mM MES and 10 mM MgCl₂, pH 5.5) to reach appropriate OD₆₀₀ for each construct. The culture solutions were then infiltrated into 6-week-old N. benthamiana plants by a syringe- or vacuum-infiltration process (Jiang et al., 2020). The infiltrated plants were individually maintained in the culture room at 28°C for 16-h light/8-h dark intervals.

The infiltrated leaves were harvested at 5 days post-infiltration (dpi), and the total proteins extracted with 1:2.5 (w/v) protein extraction buffer (50 mM Tris-HCl, pH 8.0, 10 mM MgCl₂, 10 mM KCl, 1 mM EDTA, 20% glycerol, 10% β-mercaptoethanol, and 2% SDS). The extracted total proteins were analyzed by electrophoresis through a 12% polyacrylamide gel containing 1% sodium dodecyl sulfate (SDS-PAGE), followed by visualization using Coomassie blue staining (CBS) and analysis by immunoblotting (IB). The rabbit primary antibodies against His-tag antibodies (1:5000 dilution) or mIFNγ (1:5000 dilution) were used in IB for the determination of FGs or mIFNγ glycoprotein accumulations, as described previously (Jiang et al., 2019).

For isolation of the soluble FGs (SS^ExtHis(SP)₁₀L_SrtA-mIFNγ or SS^ExtHis(SP)₁₀L_TEV-mIFNγ), the infiltrated leaves were harvested at 5 dpi. Leaf homogenates were prepared with the use of 1:2 (w/v) extraction buffer A [50 mM Tris–HCl, pH 7.6, 120 mM KCl, 15 mM MgCl₂, 0.1 mM PMSF, 20% glycerol, 0.1% β-mercaptoethanol, and one tablet of cOmplete™ EDTA-free proteinase inhibitor cocktail (Roche Life Science, Penzberg, Germany)] (Osman and Buck, 1996). The homogenized extracts were filtered through a layer of Miracloth and centrifuged at 30,000 × g for 30 min to separate the pellet (designated P30) and supernatant (designated S30) fractions. To remove the abundant RuBisCO protein contaminations, the S30 fraction was treated with ultrapure acetic acids to achieve a pH value of 5.1 and then subjected to centrifugation at 30,000 × g for 30 min to obtain the P30-treated (P30t) and S30-treated (S30t) fractions as described previously (Park et al., 2015; Jiang et al., 2020). After the removal of most RuBisCO protein contaminations, the pH of soluble FGs in the S30t fraction was adjusted to 7.0 by adding 1 M NaOH to avoid protein degradation.

To purify the FGs (SS^ExtHis(SP)₁₀L_SrtA-mIFNγ or SS^ExtHis(SP)₁₀L_TEV-mIFNγ glycoproteins) from the S30t fraction, the soluble FGs were vacuum-filtered through a 0.22-μm Millipore membrane and applied to a Ni²⁺-NTA column according to the manufacturer’s instructions. All fractions were collected and brought to equal volume with protein sample buffer (with 2% SDS) and assayed by SDS-PAGE, followed by visualization with CBS and IB with specific antiserum. The fractions containing the target FGs were pooled and subjected to further processing.

For the removal of the L_SrtA-linked secretory-affinity tag (SS^ExtHis(SP)₁₀) from the target protein, 1 μg of FG [SS^ExtHis(SP)₁₀L_SrtA-mIFNγ] was treated with SrtA, produced in the laboratory from E. coli harboring the SrtA overexpressing plasmid purchased from Addgene (Watertown, MA, United States). The multifunctional SrtA exhibits cysteine protease, protein ligase, and transpeptidase activities, specifically recognizing the LPET/G motif (the “/” sign denotes the cleavage site) in proteins (Ton-That et al., 1999; Tsukiji and Nagamune, 2009). To determine the optimal reaction condition, two-fold diluted SrtA (from 4 to 0.25 μg/μL) was incubated with 1 μg of the purified SS^ExtHis(SP)₁₀L_SrtA-mIFNγ in reaction buffer (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 10 mM CaCl₂, and 4 mM β-mercaptoethanol) at 28°C overnight. The processed proteins were subjected to further purification as described below.

To remove the L_TEV-linked tag, 10 mg of SS^ExtHis(SP)₁₀L_TEV-mIFNγ was treated with 0.2 mg of TEV protease as described previously (Kapust and Waugh, 1999). The TEV protease exhibits high cysteine protease activity under a wide range of reaction conditions (pH, salt, and temperature), which recognizes the specific ENLUFQG/S motif. The TEV protease concentrations were diluted either two-fold from 4 to 0.25 μg/μL or five-fold from 1 to 0.008 μg/μL to determine the optimal reaction condition. Each diluted TEV protease was added to 1 μg of the purified SS^ExtHis(SP)₁₀L_TEV-mIFNγ in reaction buffer (50 mM Tris-HCl, pH 8.0, 5 mM DTT, and 0.5 mM EDTA) at 4°C overnight. The processed proteins were subjected to further purification to extract the target protein.

The above protease-treated FGs in the reaction mixture were passed through a second Ni²⁺-NTA agarose column for three times, and the target mIFNγ glycoproteins were collected in the unbound fractions (flow-through). The Ni²⁺-NTA agarose was washed once with 10 mL washing buffer (20 mM 50 mM Tris–HCl, pH 8.0, 150 mM NaCl, and 20 mM imidazole), and the 6x His-tagged proteases (sortase A or TEV protease) bound to the Ni²⁺-NTA agarose were subsequently eluted for recycling with 10 mL of elution buffer A (20 mM 50 mM Tris–HCl, pH 8.0, 150 mM NaCl, and 100 mM imidazole) and 10 mL of elution buffer B (20 mM 50 mM Tris–HCl, pH 8.0, 150 mM NaCl, and 300 mM imidazole). The flow-through fractions containing tag-free mIFNγ glycoproteins were collected and then loaded on a HiLoad™ 16/60 Superdex™ 200 pg (S200) using an FPLC system (AKTA Purifier, GE Healthcare, IL, United States) in S200 buffer (50 mM Tris-HCl, pH 8.0, 200 mM NaCl, and 5 mM β-mercaptoethanol) at a flow rate of 0.5–1 mL/min. The FG-containing fractions were pooled and further concentrated using a 10-kDa NMWL Centricon filter (GE Healthcare, IL, US). The concentrated mIFNγ glycoproteins were dialyzed with PBS buffer for further biological activity assay.

The quantification of all purified proteins was confirmed by using different methods, as followed by a previous study (Jiang et al., 2019; Jiang et al., 2020), including Bradford colorimetric assay, ELISA, and Coomassie blue staining (CBS) following SDS-PAGE.

To investigate the biological activity of recombinant IFN-γ variants, an anti-SINV assay was performed, as described previously (Jiang et al., 2020), using the reporter Sindbis virus carrying an enhanced green fluorescent protein expression cassette (designated as SINV-eGFP) kindly provided by Professor Lih-Hwa Hwang (Graduate Institute of Microbiology and Immunology, National Yang-Ming University, Taipei, Taiwan). In brief, HEK293 cells (2.5 × 10⁵) were seeded in a 24-well plate 8 h prior to treatment. The cells were washed twice with PBS and treated with DMEM (Mock), the extract from healthy leaves of N. benthamiana as the negative control (NC), or five-fold serial dilutions (from 50 ng/mL to 0.08 ng/mL) of commercial mIFN-γ (Accession No. CAA31639) produced from E. coli as a positive control (PC-IFN-γ), another PC-IFN-γ (Accession No. NP_000610.2) produced from HEK293 (R&D Systems, Inc., MN, United States), or three IFN-γ variants (SSmIFNγ(SP)₁₀, SSHis(SP)₁₀L_TEV-mIFNγ, and Tag-free mIFNγ) produced from infiltrated N. benthamiana for 12 h at 37°C with 5% CO₂ (Busnadiego et al., 2020). Subsequently, the cell monolayer was infected with SINV-eGFP at a multiplicity of infection (MOI) of 1 (Tseng et al., 2015; Jiang et al., 2020). At 24 h post-infection (hpi), the levels of eGFP signals and viral proteins were detected by fluorescence microscopy and IB analysis with specific rabbit primary antibodies. The inhibition ratio of SINV was further determined by ELISA with eGFP-specific antibodies.

Statistical analysis was performed using the standard Student’s t-test (SPSS version 20, IBM Corp, Armonk, NY, United States). Mean expression ratio (%) and standard deviation (SD) from three independent experiments with technical triplicates are presented. p values < 0.001 were considered significant.

Previous studies have developed several effective fusion tags to increase the productivity and ease of downstream processing (Xu et al., 2007; Xu et al., 2010; Zhang et al., 2016; Ramos-Martinez et al., 2017; Zhang et al., 2019). However, the presence of fusion tags may negatively influence the biological functions of the target proteins, and an effective approach for the removal of such tags is highly desirable. In this study, we designed a series of BaMV-based viral vectors in which cleavable linkers, including L_SrtA, L_TEV, and L_SNAC, were inserted between the N-terminal SS^ExtHis(SP)₁₀ tag and mIFNγ, as shown in Figure 1A. By using such a BaMV-based secretory expression system, we pursue the possibility of efficient production of near-native recombinant glycoproteins without any foreign fusion tags. To compare the processing efficiency of various cleavable linkers in our BaMV-based vectors, these constructions were infiltrated into N. benthamiana leaves through Agrobacterium-mediated inoculation, and the expression profiles were further examined. The previously reported constructs of pKB19mIFNγ (Jiang et al., 2019) and pKB19SS^ExtmIFNγ(SP)₁₀ (Jiang et al., 2020) were also included in the experiment, serving as the bases for comparison.

To investigate whether inserted cleavable linkers affect the expression of mIFNγ fusion glycoproteins in the BaMV-based expression system, total proteins extracted from the infiltrated N. benthamiana leaf samples were examined at 5 dpi by IB analysis with specific antibodies against mIFNγ or His-tag (Figure 1B). The result of IB analysis using anti-His-tag antibodies revealed the presence of unprocessed SS^ExtHis(SP)₁₀L-mIFNγ protein (FP, Figure 1B) and SS^ExtHis(SP)₁₀L-mIFNγ glycoprotein (FG, Figure 1B) with relative molecular masses (M _r ) of 25 kDa and 32–40 kDa, respectively, which are similar to those observed in the control sample SS^ExtmIFNγ(SP)₁₀. The product of FGs (32–40 kDa) was possibly produced by the partial Hyp-O-glycosylation of FPs (25 kDa) via the N. benthamiana secretory pathway, as observed previously (Dolan et al., 2014; Jiang et al., 2020). By IB analysis with mIFNγ-specific antibodies, three extra protein bands were detected in leaf extracts of SS^ExtHis(SP)₁₀L_TEV-mIFNγ, with apparent M _r of 16 kDa (Mγ), 18 kDa (1N-MG), and 20 kDa (2N-MG) (indicated at the right of the panel, Figure 1B), as compared to those expressing SS^ExtHis(SP)₁₀L_SrtA-mIFNγ or SS^ExtHis(SP)₁₀L_SNAC-mIFNγ. These extra protein banding patterns were also observed in the control sample SS^ExtmIFNγ(SP)₁₀ (Jiang et al., 2020), which also exhibit the same molecular weight, indicating SS^ExtHis(SP)₁₀L_TEV-mIFNγ glycoproteins might be post-translationally modified by the proteolytic cleavage of the combinational secretory-affinity tag, SS^ExtHis(SP)₁₀, with varying degrees of N-glycosylation (including 0N, 1N, and 2N glycosylation, corresponding to the molecular weights of 16 kDa, 18 kDa, and 20 kDa, respectively). It also implied that plant endogenous protease activity could partially digest the FGs to form the various cleaved mIFNγ glycoproteins during the secretory process.

The overall accumulation of various mIFNγ forms was further quantified by ELISA. The result revealed the average levels of TPs produced in leaves infiltrated with A. tumefaciens harboring constructs for the expression of SS^ExtHis(SP)₁₀L_srtA-mIFNγ and SS^ExtHis(SP)₁₀L_TEV-mIFNγ accumulated up to 595 ± 60.1 and 688 ± 5.9 μg/g fresh weight, accounting for 9.8% and 10.8% of the TSP, respectively (Table 1), which are significantly higher than those observed in the control groups. The result indicated that the fusion of combinational secretory-affinity tag (SS^ExtHis(SP)₁₀L_srtA or SS^ExtHis(SP)₁₀L_TEV) could further improve the yield of TPs. However, the expression of SS^ExtHis(SP)₁₀L_SNAC-mIFNγ glycoprotein in infiltrated N. benthamiana leaves was lower than that of other constructs (Figure 1B; Table 1), and thus it was excluded in the following experiments.

To determine the feasibility of SrtA in removing the combinational secretory-affinity tag from the SS^ExtHis(SP)₁₀L_srtA-mIFNγ fusion glycoproteins (FGs), the following experiment was performed. Ni²⁺-NTA column-purified FGs (32–40 kDa) and SrtA (17.4 kDa) were collected, examined by SDS-PAGE (Supplementary Figure S1), and quantified for their total soluble protein (TSP) levels by using the Bradford colorimetric assay (Sigma-Aldrich, St. Louis, MO, United States). To determine the optimal protein-to-enzyme ratio, the purified FGs (1 μg, SS^ExtHis(SP)₁₀L_SrtA-mIFNγ) were incubated with various concentrations of SrtA (two-fold serial dilutions from 4 to 0.25 μg/μL). Subsequently, each reaction was analyzed by SDS-PAGE, and the FGs or various cleaved mIFNγ forms were visualized by CBS and IB analysis with specific antibodies against mIFNγ or His-tag. As shown in the CBS-stained gel, the cleaved mIFNγ forms with Mγ (16 kDa), 1N-MG (18 kDa), and 2N-MG (20 kDa) were partially released from FGs (SS^ExtHis(SP)₁₀L_SrtA-mIFNγ) following SrtA digestion (Figure 2A). Additionally, the dimeric and heterogeneous mIFNγ glycoproteins (32–40 kDa, DG) were also observed, which might be generated from the subsequent dimerization of two cleaved mIFNγ monomers with different degrees of N-glycosylation (Sareneva et al., 1994; Sareneva et al., 1995). The result of IB analysis with mIFNγ-specific antibodies showed that the maximal efficiency for the removal of combinational secretory-affinity tag was approximately 50%, with high doses (2 and 4 μg/μL) of SrtA digestion. Analysis with His-tag-specific antibodies further verified that the cleavage of FGs was not complete (Figure 2). Based on the above observations, a 1:2 ratio of FG to enzyme was chosen for SrtA-mediated cleavage. The cleaved combinational secretory-affinity tag was subsequently separated from the target proteins by passing through an Ni²⁺-NTA column. However, analysis of the eluted fraction from 300 mM imidazole treatment revealed that the majority of the processed mIFNγ glycoproteins (32–40 kDa), including Mγ (16 kDa), 1N-MG (18 kDa), and 2N-MG (20 kDa), were still accompanied with SrtA (Figures 2B, IB panel, lane 4). This observation indicated that SrtA-mediated digestion could not efficiently process the FGs (SS^ExtHis(SP)₁₀L_SrtA-mIFNγ) and that the contamination of the combinational secretory-affinity tag and SrtA may pose a threat to the downstream process of target proteins.

To test the applicability of L_TEV linker and TEV protease in the BaMV-based system, experiments similar to those described above were performed. Both SS^ExtHis(SP)₁₀L_TEV-mIFNγ and TEV proteins were also initially purified by the Ni²⁺-NTA column. Analysis of the eluates revealed that the FGs (32–40 kDa) could be efficiently obtained, accompanied with a few processed mIFNγ forms, including Mγ (16 kDa), 1N-MG (18 kDa), and 2N-MG (20 kDa) (Supplementary Figure S2).

To determine the optimal reaction condition, the purified SS^ExtHis(SP)₁₀L_TEV-mIFNγ glycoproteins (1 μg) were treated with different concentrations of TEV protease (two-fold serial dilutions ranging from 4 to 0.25 μg/μL) at 4°C overnight. Subsequently, each reaction was analyzed by SDS-PAGE, and the FGs or the cleaved mIFNγ forms were visualized by CBS and IB. The result of IB analysis with His-tag-specific antibodies (Figure 3A, the bottom panel) showed that the SS^ExtHis(SP)₁₀ tag was completely removed by TEV protease from FGs (SS^ExtHis(SP)₁₀L_TEV-mIFNγ). To further fine-tune the reaction condition for the removal of the SS^ExtHis(SP)₁₀ tag from FGs, the purified SS^ExtHis(SP)₁₀L_TEV-mIFNγ glycoproteins (1 μg) were further incubated with five-fold serial dilutions (ranging from 1 μg/μL to 0.008 μg/μL) of TEV protease. Analysis of the digestion result by IB with His-tag-specific antibodies revealed that FGs could be efficiently digested by using 0.2 μg/μL of TEV protease (Figure 3B, the bottom panel). Consequently, a 1:0.2 ratio of FG to enzyme was chosen for TEV protease-mediated cleavage. As expected, following the processing of the fusion tag, the monomeric polypeptide subunit of mIFNγ glycoproteins would assemble into dimeric and heterogeneous mIFNγ glycoproteins (32–40 kDa, DG), as shown by IB analysis with anti-mIFNγ-specific antibodies (Figures 3A, B, the middle panel). These findings indicated that TEV protease may serve as an efficient process in the production of target proteins containing L_TEV-linked fusion tags in the BaMV-based expression vector.

Subsequently, we developed a protocol for obtaining tag-free recombinant glycoproteins from plant-made SS^ExtHis(SP)₁₀L_TEV-mIFNγ, as illustrated in Figure 4A. To examine whether this protocol could be efficiently scaled up for processing the FGs into tag-free mIFNγ, 10 mg of SS^ExtHis(SP)₁₀L_TEV-mIFNγ was incubated with TEV protease (2 mg/mL) in a dialysis tube (MWCO 12–14 kDa) overnight at 4°C. After incubation, the released mIFNγ glycoproteins and SS^ExtHis(SP)₁₀ tag were separated by passing through a second Ni²⁺-NTA column. Analysis of each fraction by SDS-PAGE and IB with mIFNγ-specific antibodies revealed that the released mIFNγ glycoproteins, including Mγ (16 kDa), 1N-MG (18 kDa), 2N-MG (20 kDa), and DG (32–40 kDa), were efficiently recovered by a low concentration of imidazole (25 mM) (Figure 4B, the middle panel, lane 3). The result of IB analysis with His-tag-specific antibodies confirmed that the combinational secretory-affinity tag was efficiently removed from FGs (Figure 4B, the bottom panel, lane 3). Additionally, TEV protease (M _r of 27 kDa) was found in the fractions eluted with 300 mM imidazole (Figure 4B, lane 4), indicating that TEV protease could be separated from the products and recovered for recycling through the Ni²⁺-NTA column. These results showed that the use of TEV protease-mediated processing could be scaled up to efficiently remove the SS^ExtHis(SP)₁₀ tag from SS^ExtHis(SP)₁₀L_TEV-mIFNγ. The observation also suggested that this system could be used as a means to mass-produce near-native mIFNγ glycoproteins and other recombinant therapeutic proteins. To further improve the purity of tag-free mIFNγ, the collected fractions were pooled and applied to gel filtration chromatography (Superdex™ 200 pg, S200). Each step-fraction was collected, and the protein contents were confirmed again by SDS-PAGE stained with CBS and IB with mIFNγ-specific antibodies (Figure 4C). The results indicated that tag-free mIFNγ could be efficiently obtained. The final yield was estimated to be approximately 102 ± 3 mg/kg fresh tissue weight (FW) with a purity higher than 98%, which is comparatively higher than that for SS^ExtmIFNγ(SP)₁₀ (94 ± 7 mg/kg FW), as reported previously (Jiang et al., 2020).

The inhibitory effect of recombinant IFN-γ has been demonstrated against several viruses (Parvez et al., 2006; Wang et al., 2014; Busnadiego et al., 2020) and cancer cells (Razaghi et al., 2017; Shen et al., 2018). Moreover, as presented in our previous study, plant-made SS^ExtmIFN(SP)₁₀ was shown to induce antiviral activity in a reporter Sindbis virus expressing eGFP (SINV-eGFP) and reduce influenza virus (H1N1 strain) replication to a similar extent, as achieved by the commercial mIFNγ produced in E. coli (Jiang et al., 2020). To evaluate the efficacy of the IFNγ variants, including SS^ExtmIFNγ(SP)₁₀, SS^ExtHis(SP)₁₀L_TEV-mIFNγ, and tag-free mIFNγ purified from N. benthamiana, the recombinant proteins were prepared in PBS buffer and the concentrations of each sample were further confirmed by SDS-PAGE stained with CBS and IB with mIFNγ-specific or His-tag-specific antibodies (Figure 5). The commercial mIFNγ produced in E. coli or mammalian cells, designated PC-mIFNγ (E. coli) and PC-IFNγ (HEK-293), respectively, were used as the positive controls. The antiviral effect was then evaluated using the SINV-eGFP reporter virus as a target pathogen. HEK293 cells were cultured in DMEM with IFN-γ variants at various concentrations, which were 5-fold serially diluted from 50 to 0.08 ng/mL for 12 h, followed by challenging with the SINV-eGFP. The cells cultured in DMEM treated with PBS buffer were used as the mock treatment group (Mock), whereas those treated with the total protein extract of healthy Nicotiana benthamiana served as the negative control (NC). At 24 h post-infection (hpi), the eGFP fluorescence, representing the overall infection status, was examined and recorded by fluorescent microscopy. As shown in Figure 6A, upon treatment with 2–50 ng IFNγ variants, the eGFP fluorescence signals were obviously decreased and mostly confined to a single-cell level, as compared to those in the Mock and NC (50 ng) groups. It is worth noting that cells treated with the tag-free mIFNγ variant at low doses (0.08 ng or 0.4 ng) had lower eGFP signals, demonstrating the antiviral efficacy of tag-free mIFNγ. The results of eGFP fluorescence intensity measurements and IB analysis showed that treatment with the tag-free mIFNγ variant at the lowest dose (0.08 ng) significantly decreased the level of eGFP (p-values <0.001), which is comparable to the effect of commercial PC-IFNγ (HEK293). Treatments with three IFNγ variants at medium (2 ng) or high (50 ng) doses reduced the eGFP levels in a dose-dependent manner (Figure 6B; Supplementary Figure S3). Among them, the tag-free mIFNγ was observed to have a higher efficacy in decreasing SINV-eGFP infection as compared to those of variants with fusion tags, including SS^ExtmIFNγ(SP)₁₀ and SS^ExtHis(SP)₁₀L_TEV-mIFNγ (a significant difference with p-values <0.001). To corroborate this result, the inhibition ratio of SINV-eGFP infection was determined by ELISA for quantification of eGFP levels, and the half-maximal inhibitory concentration (IC₅₀) of each recombinant protein was estimated. As shown in Figure 7; Table 2, tag-free mIFNγ exhibited a higher inhibitory effect on SINV-eGFP infection, which was similar to that of commercial PC-IFNγ (HEK293), with IC₅₀ values of 2.5–2.9 ng/mL. In contrast, the IC₅₀ values of SS^ExtmIFNγ(SP)₁₀ and SS^ExtHis(SP)₁₀L_TEV-mIFNγ were 7.9 and 37.2 ng/mL, respectively, which were notably higher than that of the tag-free protein. The lower IC₅₀ value and similar efficacy comparable with commercial IFNγ (HEK-293) protein indicated that plant-made tag-free mIFNγ exhibits satisfactory antiviral activity, which also suggested that these IFNγ glycoproteins may maintain the native folding suitable for inhibiting viruses.