Viperins, a class of proteins, have been reported to confer an inhibitory strategy against viruses in eukaryotes and prokaryotes. Structural alignments have indicated that Viperin shares structural similarities with radical SAM proteins, including the molybdenum cofactor biosynthetic enzyme MoaA (Fenwick et al., 2017). Building on these findings, we selected Viperins from Homo sapiens and other organisms, as well as MoaA proteins from different species, to assess their antiviral activity in plants. Considering their prevalence across diverse organisms, we chose Viperins and Viperin-like proteins from bacteria, algae, plants, and archaea to be tested in our study. Our selection of protein sequences was based on the work done by Audi Bernheim et al. while others were chosen based on their homology to the selected proteins as identified by the BLAST analysis ( Figure 1 ; Supplementary Figure 1 ; Supplementary Table 1 ). The identified sequences were codon-optimized for expression in the eukaryotic system and cloned into a binary vector for subsequent expression in planta. Subsequently, we generated N. benthamiana stable lines through tissue culture and confirmed the presence of protein via western blot ( Supplementary Figure 2 ). The confirmed lines were challenged with GFP-tagged plant RNA viruses TMV-GFP, TuMV-GFP and PVX-GFP. Plants were observed for GFP fluorescence under UV 5 to 7 days after infiltration. Following phenotypic analysis, systemic leaf samples were harvested for any downstream molecular analysis.

To expedite the screening of candidate Viperin and Viperin-like proteins, we first conducted transient assays using a tobacco mosaic virus (TMV)-based overexpression system, known as the TRBO-G system, which expresses the GFP gene as a reporter (Lindbo, 2007). In the TRBO-G system, the TMV has been engineered to replace the CP encoding sequence with a sequence encoding green fluoresent protein (GFP). The TRBO-G virus cannot move systematically in the infected plants but can efficiently replicate and produce GFP in the infected leaves. The utilization of this reporter system provided an easy and straightforward means to monitor virus infection and replication, thus accelerating the screening process.

To assess the interference activity of Viperin and Viperin-like proteins, we co-delivered binary vector harboring a codon-optimized coding region for the selected protein and the TRBO-G expressing construct (pJL-TRBO) into the leaves of wild-type N.benthamiana plants via Agro-infiltration ( Figure 2A ). Plants infiltrated with TRBO-G alone or TRBO-G in combination with an empty binary vector (EV) were used as control groups. At 3 dpi, we observed varied levels of interference among the identified proteins when challenged with the TMV virus ( Figure 2B ). Our data indicated a significant reduction in GFP signals for Viperin (from Homo sapiens), Molybdenum cofactor biosynthesis protein 1 MOCS1A (from Homo sapiens), Viperin (from Shewanella sp. cp20), Viperin (from Fibrobacter sp. UWT3), MoaA (from Physcomitrium patens), and MoaA (from Zostera marina) compared to the control, providing strong evidence of their antiviral activity against TMV in plants.

To validate the observed reduction in GFP fluorescence, we measured GFP protein in the infiltrated leaves by western blot. Total proteins were extracted from the leaves, and western blot was performed using anti-HA and anti-GFP antibodies to detect Viperin and Viperin-like proteins and the virus protein, respectively. In agreement with the observed low GFP fluorescence, GFP protein levels were also suppressed by some of the Viperin and Viperin-like proteins ( Figures 2C, D ). Notably, Viperin from Homo sapiens, and Viperin from Fibrobacter sp. UWT3 exhibited robust antiviral activity against TMV, as indicated by the presence of less GFP protein. The graphical representation of the GFP protein abundance further corroborates our observation ( Figures 2E, F ). These data indicate that Viperin and Viperin-like proteins exhibit an antiviral role against plant viruses in planta.

To explore the potential of antiviral proteins in conferring heritable resistance against RNA viruses in plants, we created transgenic N. benthamiana plants heterologously overexpressing Viperin and Viperin-like proteins driven by the cauliflower mosaic virus 35S promoter. Protein presence was validated through western blot analysis, and the confirmed lines were used in the following experiments ( Supplementary Figure 2 ).

We then chose the potyvirus turnip mosaic virus (TuMV) as our target RNA virus and evaluated its pathogenicity on our stable plant lines. T2-generation data revealed varying interference levels among the tested lines challenged with TuMV-GFP ( Supplementary Figure 3 ). Notably, N. benthamiana lines overexpressing molybdenum cofactor biosynthesis protein A (MOCS1A) from Homo sapiens, GTP 3’,8-cyclase (MoaA) from Zostera marina and Dunaliella Salina exhibited interference against TuMV-GFP. The identified resistant lines were then challenged with two TuMV-GFP titers (0.03 and 0.05 OD, based on the OD₆₀₀ of the TuMV-carrying Agrobacteria used for inoculation). The results confirmed the resistance of these lines, as indicated by the reduced GFP fluorescence signal compared to the non-transgenic line ( Figure 3A ). RT-qPCR further confirmed the accumulation of less TuMV-GFP genomic RNA in the transgenic lines ( Figures 3B–D ). Additionally, a difference compared to the control was observed even when the viral titer was increased to 0.05, further confirming the resistance of these lines against TuMV-GFP.

To ensure there is no interference from the 35s promoter between different constructs, we tested our resistant lines using sap from wild-type plants infected with the TuMV-GFP virus. Despite using sap, our lines continued to exhibit resistance, indicating that the observed resistance is due to Viperin-like proteins and not because the 35s promoter in the transgenic lines is silencing the viral construct. Moreover, using a non-specific control supports the fact that the resistance observed is indeed due to Viperin-like proteins ( Supplementary Figure 4 ).

To evaluate the effectiveness of Viperin and Viperin-like proteins in interfering with other RNA viruses, we challenged our N. benthamiana stable lines with potato virus x (PVX) ( Supplementary Figure 5 ). Like TuMV-GFP, we used a GFP-tagged PVX virus to facilitate visual detection. Our data from the T2 generation revealed varied levels of viral interference among the tested lines. N. benthamiana lines overexpressing CNX2 from Nicotiana sylvestris, MOCS1A from Homo sapiens and MoaA from Xanthomonas oryzae showed strong interference with PVX compared to the non-transgenic and empty vector (EV) controls.

To validate our findings, the identified resistant lines were challenged with two titers of PVX-GFP (0.03 and 0.05, based on the OD₆₀₀ of the Agrobacteria) ( Figure 4A ). A visual decrease in GFP fluorescence observed in N. benthamiana lines overexpressing the mentioned proteins confirmed their role in virus interference. Notably, a strong reduction in GFP signal was observed in plants overexpressing CNX2 from Nicotiana sylvestris, indicating robust interference provided by this protein. Substantiating these findings, RT-qPCR analysis also showed a prominent decrease in PVX-GFP accumulation in these plants, especially those overexpressing CNX2. This aligns well with the observed phenotypic data ( Figures 4B–D ). We further confirmed our transgenic lines for resistance by challenging them with sap from wild-type plants previously infected with PVX-GFP and found that our lines remained resistant. The data confirms that the observed resistance is only due to Viperin-like proteins and not from the cross-interference mechanism, if any. Non-specific control further confirmed the interference of Viperin-like proteins ( Supplementary Figure 6 ).

To engineer plants expressing Viperin and Viperin-like proteins, we mammalian codon-optimized their coding sequences for expression in the eukaryotic system and ordered them in intermediate vectors with a 3x-HA tag fused to the C-terminus of the sequence. All sequences were ordered as flanked by attL1 and attL2 recombination sites to facilitate gateway cloning. Cloning into the gateway binary vector pK2GW7 was performed by LR recombination to generate final clones harbouring Viperin and Viperin-like sequences driven by the cauliflower mosaic virus (CaMV) 35S promoter. Subsequently, the clones were introduced into Agrobacterium tumefaciens GV3101 via electroporation. Next, N.benthamiana leaf discs were treated with the agrobacterium cultures harboring the plasmids of interests and placed on regeneration media. Shoots from the regeneration media were then transferred to rooting media for root induction. The resulting plantlets were then transferred to soil, and the protein expression was confirmed by western blotting using anti-HA antibody. The confirmed lines were used in the subsequent experiments.

Two- to three-week-old wild-type Nicotiana benthamiana plants, cultivated under long-day conditions (16 hours light, 8 hours dark at 25°C), served as the experimental subjects for all transient assays. In transient experiments, wild-type Nicotiana benthamiana plants were initially grown on MS (Murashige & Skoog) media for 10 days and subsequently transplanted into soil. In experiments utilizing permanent lines (which stably expresses Viperin and Viperin-like proteins), T2 generation seeds were germinated on MS media supplemented with kanamycin, and the surviving seedlings were used in virus-targeting experiments.

The constructs containing pJL-TRBO (TMV-GFP), TuMV-GFP, and PVX-GFP were separately introduced into Agrobacterium tumefaciens strain GV3101 through electroporation. Single colonies grown overnight on a selective medium were centrifuged and resuspended in an infiltration medium (10 mM MES [pH 5.7], 10 mM CaCl2, and 200 μM acetosyringone) and incubated at ambient temperature for 2 hours. Infiltration into plants overexpressing Viperin and Viperin-like proteins involved using a cell density of 0.03 or 0.05 (OD600) for PVX-GFP and TuMV-GFP. A cell density of 0.005 was used in case of pJL-TRBO. The infiltration was performed on the abaxial side of leaves using a needle-less syringe. GFP expression was examined at 5-7 days after infiltration (dai) using a handheld UV light, and photographs were captured with a Nikon camera under UV light. Leaf samples were collected at 7 and 10 dai for subsequent molecular and transcriptomic analyses.

To perform sap inoculation, 3-week-old wild-type Nicotiana benthamiana plants were infiltrated with Agrobacterium strain GV3101 carrying infectious clones of either TuMV-GFP or PVX-GFP. After 7 to 10 dpi (days post infiltration), systemic leaves were harvested and effectively crushed in phosphate buffer with a pH of 7.4 to collect sap. Next, individual leaves from 3-week-old wild-type N. benthamiana plants and transgenic lines overexpressing Viperin or Viperin-like proteins were dusted with 200–450 mesh-sized carborundum particles and applied with equal amount of sap containing either TuMV-GFP or PVX-GFP. Virus was spread uniformly on the leaf surface and the plants were kept in a greenhouse for another 7 to 10 days. To monitor virus spread, the plants were examined in darkness using a handheld ultraviolet (UV) light, and photographs were taken to record the results.

Total proteins were extracted from 100 mg of the sample using an extraction buffer (100 mM Tris-Cl, pH 8, 150 mM NaCl, 0.6% IGEPAL, 1 mM EDTA, and 3 mM DTT) supplemented with protease inhibitors (PMSF, leupeptin, aprotinin, pepstatin, antipain, chymostatin, Na2VO3, NaF, MG132, and MG115). The protein extracts were separated on a 10% polyacrylamide gel. Immunoblot analysis was performed using a mouse α-GFP antibody (1:3000, Invitrogen) for GFP expressed by the virus and a rat α-HA (1:1000) antibody to detect viperin-like proteins. Antigens were visualized through chemiluminescence using an ECL-detecting reagent (Thermo Scientific). Quantitative analysis was conducted by calculating densitometric data from the relative quantification of protein bands on immunoblot membranes using ImageJ software, and the average values from three independent biological replicates were graphically summarized.

Total RNA was extracted from systemic leaves using Direct-zol RNA Miniprep Kits (Zymo Research) following the manufacturer’s instructions. Viral RNA quantification was conducted through one-step RT-qPCR employing the iTaq Universal SYBR Green One-Step Kit (Bio-Rad). RT-qPCR reactions were executed in 10µl volumes using primers amplifying the GFP transcript ( Supplementary Table 2 ) with three technical replicates in a 96-well format and analyzed using a StepOnePlusTM Real-Time PCR System (Applied Biosystems) for the 96-well plates. To normalize for total input, expression levels were determined by subtracting the cycle threshold (Ct) values of the housekeeping reference gene (tobacco PP2A) from the target Ct values, yielding ΔCt levels. The relative transcript abundance was calculated as 2−ΔΔCt. All analyses were performed with three biological replicates.