Specific and efficient translation inhibition–transcription activation resistance module against viruses in plants

Abstract

Positive-sense RNA viruses using subgenomic RNAs (sgRNAs) cause severe crop losses, yet this conserved feature has rarely been exploited for antiviral engineering. Here, we constructed a virus-activated synthetic immune circuit, wherein viral sgRNA production drives host defense responses via a translation inhibition–transcription activation (TI-TA) module. An optimized translational repression element (dsBC) minimized basal expression while enabling virus-specific activation of resistance-inducing proteins (RIPs) by cognate viral sgRNA promoters. This circuit functioned against three major sgRNA-producing viruses—pepino mosaic virus (PepMV), tobacco mosaic virus (TMV), and cucumber mosaic virus (CMV)—in Nicotiana benthamiana, tomato, and Arabidopsis thaliana, with cross-resistance to related viruses. RIP expression was strictly triggered by viral infection, conferring robust resistance without detectable growth penalties. This programmable strategy provides a versatile platform for engineering resistance against diverse sgRNA-producing RNA viruses and offers a conceptual framework for developing synthetic antiviral immunity in plants.

INTRODUCTION

During the long-term coevolution of plants and pathogens, plants have developed two major layers of innate immunity: pattern-triggered immunity (PTI) and effector-triggered immunity (ETI) (1, 2). PTI is triggered by plasma membrane–localized pattern recognition receptors, which perceive conserved microbial patterns, including flagellin, elongation factor Tu, peptidoglycan, chitin, and lipopolysaccharides, thereby conferring basal immunity against a broad spectrum of nonadapted pathogens (3, 4). In contrast, ETI represents a more robust and specific immune response mediated by resistance (R) proteins. Most proteins belong to the family of nucleotide-binding leucine-rich repeat receptors (NLRs), which recognize pathogen effectors and initiate potent defense responses, typically accompanied by a hypersensitive response (HR) to restrict pathogen proliferation (5, 6). Recent studies have revealed that PTI and ETI function interdependently rather than independently, resulting in largely overlapping and integrated immune responses (7, 8). NLR-mediated plant immunity induces HR, which serves as an effective defense against viral infections. However, only a small number of viral proteins can be recognized by limited NLRs to initiate ETI signaling (9). Several exogenous proteins, including BAX, INF1, and NLP_Pya are capable of rapidly inducing plant HR through PTI/ETI-dependent or independent pathways. As a mammalian Bcl-2 family protein, BAX localizes to mitochondria and triggers cell death upon heterologous expression in plants, followed by the up-regulation of defense-related genes and reactive oxygen species (ROS) accumulation (10, 11). INF1, a 10-kDa extracellular elicitor widely distributed across Phytophthora species, induces HR in a restricted number of plants, particularly in the genus Nicotiana. INF1-triggered immune responses include PTI activation, ROS burst, and elevated expression of ethylene (ET)– and jasmonic acid (JA)–responsive genes (12–16). NLP_Pya derived from the phytopathogenic oomycete Pythium aphanidermatum, triggers leaf necrosis in diverse plant species. It functions by forming transient small pores on lipid membranes, leading to plasma membrane destruction, cytolysis, and the activation of plant defense–related genes. This cellular damage–triggered immune activation is analogous to inflammasome activation in vertebrates (17, 18). Whether resistance-inducing proteins (RIPs) from different biological sources, such as INF1, BAX, and NLP_Pya can be harnessed to combat viral infections remains an intriguing scientific question.

Positive-sense single-stranded RNA (+ssRNA) viruses constitute the vast majority of plant viruses and impose severe threats to global crop production. A large proportion of these viruses, encompassing over 1000 characterized plant +ssRNA viral species, use subgenomic RNAs (sgRNAs) promoters to generate sgRNAs for the expression of internal and 3′-proximal open reading frames (ORFs) (19, 20). The RNA-dependent RNA polymerase (RdRP) of +ssRNA viruses recognizes the 3′ untranslated region (UTR) of the viral genome and synthesizes a complementary negative-strand viral RNA genome. Subsequently, RdRP recognizes specific internal subgenomic promoters (SPs) on the negative-strand RNA template to initiate the synthesis of truncated positive-strand RNA transcripts, termed sgRNAs. Through sgRNA biogenesis, viruses efficiently orchestrate viral gene expression and ensure the timely synthesis of structural and nonstructural proteins during infection, thereby expanding their genomic coding capacity (21). Most economically destructive plant viruses fall into this category, including tobacco mosaic virus (TMV), cucumber mosaic virus (CMV), brome mosaic virus (BMV), and potato virus X (PVX) (22). Additionally, multiple emerging crop-damaging viruses, such as pepino mosaic virus (PepMV), tomato brown rugose fruit virus (ToBRFV), cucumber green mottle mosaic virus (CGMMV), and maize chlorotic mottle virus (MCMV), also use the sgRNA-dependent gene expression strategy (20, 23). Due to the intrinsically high mutation rates of +ssRNA viruses, which primarily result from the lack of proofreading activity in viral RdRP (24, 25), viral populations exist as dynamic quasispecies capable of rapid adaptation to diverse selective pressures. This potent evolutionary plasticity allows viruses to evade NLR-mediated host resistance. For example, resistance-breaking strains have been documented to overcome host resistance conferred by tomato mosaic virus (ToMV) and ToBRFV resistance genes (26, 27). Recent advances in plant immune engineering have facilitated the optimization of NLR-mediated resistance through multiple strategies, including random mutagenesis, structure-guided receptor engineering, integrated decoy engineering, and effector-mediated derepression of autoactive NLRs (28, 29). Despite these advances, durable and broadly applicable resistance approaches specifically targeting sgRNA-producing RNA viruses remain largely undeveloped.

Plants maintain a dynamic balance between growth and immunity to optimize environmental adaptation. Robust immune responses strengthen disease resistance but frequently penalize plant growth and yield, a trade-off universally recognized as the “growth-defense trade-off” (30). Fortunately, multiple strategies have been developed to boost plant immunity without incurring substantial fitness costs (31). A typical paradigm is the “TBF1-cassette,” which exploits pathogen-responsive translational regulation mediated by upstream open reading frames (uORFs) within the 5′ leader sequence of the TBF1 (TL1-binding transcription factor 1). TBF1 acts as a core transcription factor that regulates the transition from growth to defense upon immune activation. Although its translation is constitutively repressed by two conserved uORFs under normal conditions, it is rapidly activated in response to pathogen challenge, ensuring the timely accumulation of defense regulators. Engineering this regulatory module to control the expression of immune genes, such as NPR1, confers broad-spectrum disease resistance without compromising plant growth (32). Further studies have revealed that uORF-dependent translational regulation is tightly associated with RNA secondary structures. Specifically, double-stranded RNA (dsRNA) hairpin structures immediately downstream of the upstream AUG codons (uAUGs) of uORF, termed uAUG-ds, are critical for selective translation initiation. These structural motifs facilitate ribosome recognition of uAUGs, thereby promoting uORFs translation and repressing the translation of downstream main ORF under normal growth conditions. Upon immune activation, RNA helicases unwind these dsRNA structures, enabling ribosomes to bypass uAUGs and initiate translation of defense-related proteins. Notably, uAUG-ds serves as a functional structural partner of uORFs rather than being functionally equivalent to uORFs. While uAUGs define the start codon of uORFs, the downstream dsRNA structures modulate translational efficiency, and, together, they constitute a dynamic regulatory module that fine-tunes protein abundance (33). Previous studies have reported advanced pathogen-responsive resistance strategies. Mei et al. exploited uAUG-mediated translational regulation to modulate the expression of plant immune genes, successfully enhancing resistance against fungal, bacterial, and DNA viral pathogens with negligible fitness penalties (34). Liu’s group used the pathogen-responsive TBF1 promoter coupled with its endogenous uORFs to drive MCD1 expression in Nicotiana benthamiana, achieving broad-spectrum resistance against cross-kingdom pathogens without affecting normal plant growth (35). Engineered NLRs or elicitor peptides also confer potent resistance to potyviruses without growth penalties by activating postinfection defense responses (29, 36). Nevertheless, efficient virus-responsive resistance systems specifically tailored for sgRNA-producing RNA viruses remain largely underexplored to date.

In this study, we developed a dual-layer translation inhibition–transcription activation (TI-TA) strategy for engineering resistance against sgRNA-producing +ssRNA viruses. This design is built upon two core biological principles. First, uAUG-ds elements robustly repress downstream protein translation under normal growth conditions, thereby minimizing constitutive expression-driven growth penalties typically induced by RIPs. Second, sgRNA-producing viruses specifically activate viral subgenomic promoters during infection, providing a virus-dependent transcriptional activation (TA) signal. Based on these two principles, we used uAUG-ds elements as translational inhibition (TI) modules to repress RIP expression in the absence of viral infection, while viral subgenomic promoters were used as TA modules to drive RIP expression specifically upon viral infection. For systematic evaluation, uAUG-ds motifs derived from the 5′ UTRs of TUB7 and TBF1 were designated as dsA and dsB, respectively, and a synthetic 54–nucleotide (nt) class I uAUG-ds element was named dsC and generated the enhanced translational repression module dsBC. Using this TI-TA framework, we demonstrated that the combination of dsBC with viral subgenomic promoter modules confers potent basal repression and robust virus-responsive expression. Substituting green fluorescent protein (GFP) reporters with RIP effectors (e.g., BAX) conferred effective resistance against PepMV in N. benthamiana and tomato. Furthermore, the strategy exhibits partial broad applicability toward diverse sgRNA-producing viruses, including TMV and CMV, with protective efficacy extending to other closely related viruses within the corresponding genera. Collectively, this synthetic system integrates three advantageous features: (i) potent basal translational repression, (ii) strict virus-specific activation, and (iii) partial broad-spectrum antiviral activity without detectable growth penalty. Our work therefore establishes a programmable framework for engineering durable resistance against sgRNA-producing plant RNA viruses.

RESULTS

Stepwise optimization of TI and TA modules yields a PepMV-responsive expression system

Given the persistent and severe threat to crop production posed by +ssRNA viruses that produce sgRNAs, we sought to develop a virus-responsive system capable of specifically activating the expression of RIPs upon viral infection. To achieve this goal, two requirements needed to be fulfilled: (i) RIP expression should remain strongly suppressed in the absence of viral infection through TI; and (ii) RIP expression should be specifically activated upon viral infection through TA. To establish an efficient TI module, we evaluated several uAUG-ds elements. The 5′ UTRs of TUB7 and TBF1 contain naturally occurring uAUG-ds elements, designated dsA (144 nt) and dsB (105 nt), respectively, whereas a 54-nt class I uAUG-ds element was designated dsC (33). The PepMV coat protein (CP) subgenomic promoter (PepMV^SP1) was selected as the initial TA module (Fig. 1A). To assess the translational inhibitory activities of these elements, GFP was inserted between PepMV^SP1 and the 3′ UTR to generate PepMV^SP1-GFP (Fig. 1B). Individual or tandem uAUG-ds elements were subsequently fused to the 5′ end of PepMV^SP1-GFP, generating dsA-, dsB-, dsC-, dsAC-, dsBC-, and ds2C-PepMV^SP1-GFP constructs (Fig. 1B). Transient expression assays in N. benthamiana revealed that tandem uAUG-ds elements (dsAC, dsBC, and ds2C) conferred substantially stronger translational repression than individual elements, as evidenced by reduced GFP fluorescence and protein accumulation (Fig. 1, C and D). Among them, dsBC exhibited the strongest inhibitory effect on GFP protein accumulation (Fig. 1, C and D). Quantitative reverse transcription polymerase chain reaction (qRT-PCR) analysis showed that dsBC did not substantially affect GFP transcript abundance (Fig. 1E), indicating that its inhibitory effect primarily occurred at the translational level. In addition, independent verification assays confirmed that uAUG-ds elements, in the absence of PepMV^SP1, could significantly reduce GFP protein accumulation without altering mRNA levels, solidifying the translational inhibitory function of uAUG-ds modules (fig. S1). Collectively, dsBC was selected as the optimal TI element for subsequent system construction due to its strong and specific translational repression activity.

Fig. 1. PepMV strongly restores GFP expression in the dsBC-PepMV^mini-GFP system.

(A) Schematic of the PepMV genome. TGB1–3, triple gene block proteins; CP, coat protein. (B) Schematic of uAUG-ds, PepMV^SP1, and GFP constructs. The uAUG-ds elements include 144-nt dsA from the TUB7 5′ UTR, 105-nt dsB from the TBF1 5′ UTR, and a 54-nt class I element dsC (33). Tandem fusion of these elements with dsC produced dsAC, dsBC, and ds2C. Blue line indicates PepMV^SP1 and 3′ UTR; pink box indicates the first 36 nt of the CP N terminus. (C to E) GFP fluorescence, protein accumulation, and mRNA levels in N. benthamiana leaves infiltrated with Agrobacterium (OD₆₀₀ = 0.5) carrying GFP, PepMV^SP1-GFP, or uAUG-ds-PepMV^SP1-GFP constructs. EV served as the negative control. Fluorescence was observed at 2 dpi under UV light, GFP accumulation was detected by Western blot using anti-GFP and anti-actin antibodies, and GFP mRNA levels were analyzed by qRT-PCR. (F) Schematic of PepMV^mini and derivatives. PepMV^mini contains the 5′ UTR, 45 nt of the RdRP N terminus, PepMV^SGP192, PepMV^SP1, the 36-nt CP C-terminal sequence, and the 3′ UTR (648 nt total). Kpn I and Pst I sites are located downstream of PepMV^SGP192 and PepMV^SP1, respectively. (G to I) PepMV infection restores GFP expression from dsBC-PepMV^mini-GFP. Agrobacterium carrying dsBC-PepMV^mini-GFP (OD₆₀₀ = 0.5) was co-infiltrated with EV or PepMV (OD₆₀₀ = 0.5). GFP fluorescence (G), protein accumulation (H), and mRNA levels (I) were analyzed at 2 dpi. GFP and PepMV^mini-GFP served as controls. Data are presented as means ± SD (n = 3). Statistical significance was determined by one-way ANOVA followed by Tukey’s HSD test (P < 0.05) using IBM SPSS Statistics 27; different letters indicate significant differences among treatments. Selected pairwise comparisons were performed using Student’s t test in GraphPad Prism 8 when appropriate.

Having identified dsBC as the most effective TI module, we next evaluated whether PepMV^SP1 could function as a virus-responsive TA module. To this end, dsBC-PepMV^SP1-GFP was co-infiltrated with either an empty vector (EV) or PepMV into N. benthamiana leaves. Unexpectedly, PepMV infection failed to restore GFP fluorescence, protein accumulation, or transcript levels relative to the EV control (fig. S2, A to C). These results indicated that PepMV^SP1 and 3′ UTR were insufficient to support efficient virus-dependent activation, revealing a critical limitation of the initial system design. To identify additional cis-regulatory elements required for PepMV-dependent activation, we considered previous reports showing that efficient transcription from the PepMV CP subgenomic promoter requires the N-terminal 36-nt sequence of the CP coding region (37). This observation suggested that sequences adjacent to the CP transcription unit may contain additional cis-elements involved in promoter recognition or activity. Therefore, we tested whether incorporation of the 36-nt sequence immediately upstream of the 3′ UTR could improve PepMV-dependent activation of the dsBC-PepMV^SP1-GFP construct. However, this modification resulted in only weak GFP activation following PepMV infection (fig. S2, D to G), suggesting that partial sequence optimization was insufficient to restore effective virus-responsive expression activation. These results suggested that additional viral cis-elements were required. We therefore reconstructed a larger PepMV-derived mini-replicon, designated PepMV^mini, consisting of the 5′ UTR, the first 45 nt of the RdRP coding sequence, the TGB1 subgenomic promoter (SGP192), the CP subgenomic promoter (SP1), the 36-nt CP-terminal sequence, and the 3′ UTR (Fig. 1F). GFP was fused downstream of PepMV^mini to generate PepMV^mini-GFP, and dsBC was subsequently introduced to generate dsBC-PepMV^mini-GFP (Fig. 1F). When co-infiltrated with PepMV, dsBC-PepMV^mini-GFP produced strong GFP fluorescence, whereas GFP expression was nearly undetectable in EV-treated controls (Fig. 1G). Consistently, Western blot and qRT-PCR analyses showed substantial increases in GFP protein and transcript accumulation following PepMV infection (Fig. 1, H and I). Furthermore, PepMV^mini reduced GFP accumulation primarily at the transcript level, whereas dsBC imposed additional translational repression without markedly affecting GFP mRNA abundance (Fig. 1, G to I).

In addition, to determine whether reporter position influences activation efficiency, GFP was inserted into the Kpn I site of PepMV^mini and dsBC-PepMV^mini, generating PepMV^mini-192-GFP and dsBC-PepMV^mini-192-GFP, respectively (Fig. 1F). Unlike dsBC-PepMV^mini-GFP, dsBC-PepMV^mini-192-GFP showed little GFP activation following PepMV infection (fig. S3, A to C), indicating that reporter position is critical for efficient PepMV-responsive activation. Together, these results demonstrate that the optimized TI element (dsBC) and the reconstructed TA module (PepMV^mini) function cooperatively to establish a tightly repressed yet highly virus-responsive expression system.

dsBC-PepMV^mini-BAX confers PepMV resistance in planta

Because dsBC-PepMV^mini-GFP meets our design criteria—minimal GFP accumulation under virus-free conditions and robust activation upon PepMV infection—we therefore engineered RIP effector genes into this backbone. This strategy aims to achieve PepMV-specific activation of RIPs for enhanced antiviral defense. We selected BAX, INF1, and NLP_Pya, which have been reported to induce immune response and potential resistance to plant pathogens (10, 13, 17). First, we constructed these RIPs into transient expression vectors with 3×HA tags at both N or C termini and then transiently expressed them in N. benthamiana leaves. At 3 to 5 days postinoculation (dpi), we observed that these proteins could markedly induce cell death, and trypan blue staining further confirmed this result (fig. S4). Subsequently, we constructed these RIPs into dsBC-PepMV^mini. Then, dsBC-PepMV^mini-RIPs were coexpressed with either EV or PepMV in N. benthamiana leaves. At 3 to 5 dpi, we observed that the areas coexpressing EV and dsBC-PepMV^mini-RIPs did not trigger HR, whereas the areas coexpressing PepMV and dsBC-PepMV^mini-RIPs did (Fig. 2, A to C). This result indicates that we successfully achieved PepMV-responsive activation of resistance.

Fig. 2. dsBC-PepMV^mini-BAX enhances resistance against PepMV in N. benthamiana and tomato.

(A to C) PepMV infection specifically triggers RIP expression in N. benthamiana. Agrobacterium carrying the indicated constructs (OD₆₀₀ = 0.5) was co-infiltrated with EV or PepMV (OD₆₀₀ = 0.5) into N. benthamiana leaves. Leaves were photographed and trypan blue–stained at 3.5, 2.5, and 5 dpi, respectively, with GFP and transient RIPs as controls. (D and E) dsBC-PepMV^mini-BAX alleviates PepMV symptoms and viral accumulation in N. benthamiana. WT and transgenic plants were rub-inoculated with PepMV sap or healthy mock sap. Systemic leaves were imaged at 20 dpi (scale bar, 2.5 cm), and viral accumulation was quantified via qRT-PCR (n = 6). (F and G) dsBC-PepMV^mini-BAX activates defense-related NbPR1 expression upon PepMV infection. Transient and transgenic plant materials were sampled at 2 and 3 dpi, respectively, for NbPR1 transcript detection (n = 3). For transient assays, Agrobacterium carrying the indicated constructs (OD₆₀₀ = 0.5) was co-infiltrated with PepMV (OD₆₀₀ = 0.5); for transgenic plants, leaves were sampled at 3 dpi after Agrobacterium-mediated PepMV inoculation (OD₆₀₀ = 0.1). (H and I) dsBC-PepMV^mini-BAX improves PepMV tolerance in tomato. Transgenic and WT tomato plants were rub-inoculated with PepMV sap or healthy mock sap; systemic symptoms were photographed at 23 dpi (scale bar, 2.5 cm), and viral accumulation was quantified by qRT-PCR (n = 6). All data are presented as means ± SD. Statistical differences were determined by one-way ANOVA followed by Tukey’s HSD test (IBM SPSS Statistics 27; P < 0.05); different letters indicate significant differences among treatments. Selected pairwise comparisons were performed using Student’s t test in GraphPad Prism 8 when appropriate.

Compared with dsBC-PepMV^mini-NLP_Pya, dsBC-PepMV^mini-BAX and dsBC-PepMV^mini-INF1 leaves exhibited faster necrosis in response to PepMV infection, indicating more robust activation of plant immunity. Therefore, we generated dsBC-PepMV^mini-BAX and dsBC-PepMV^mini-INF1 transgenic N. benthamiana plants to analyze whether these transgenic plants could confer resistance against PepMV infection. Through qRT-PCR screening, we obtained two independent transgenic lines for dsBC-PepMV^mini-RIPs, and these transgenic lines did not affect plant growth (fig. S5). PepMV was then rub-inoculated onto wild-type (WT), dsBC-PepMV^mini-RIP transgenic, and dsBC-PepMV^mini-GFP (control) transgenic T2 plants. At 4 dpi, we observed that systemic leaves of WT, dsBC-PepMV^mini-GFP, and dsBC-PepMV^mini-INF1 plants developed viral mosaic symptoms, while dsBC-PepMV^mini-BAX plants showed almost no such symptoms and lower accumulation of viral RNA (fig. S6). At 20 dpi, PepMV infection caused severe mosaic patterns and malformation in systemic leaves of WT, dsBC-PepMV^mini-GFP, and BC-dsBC-PepMV^mini-INF1 plants, while only mild symptoms appeared on dsBC-PepMV^mini-BAX plants (Fig. 2D). qRT-PCR results showed that PepMV accumulation was significantly higher in WT, dsBC-PepMV^mini-GFP, and BC-dsBC-PepMV^mini-INF1 plants than that in dsBC-PepMV^mini-BAX plants (Fig. 2E). These results demonstrate that dsBC-PepMV^mini-BAX can confer resistance against PepMV infection in N. benthamiana plants.

PR1 is the classic marker gene downstream of the salicylic acid signaling pathway. As a key component of the plant innate immune system, particularly in systemic acquired resistance, it is widely used as a diagnostic molecular marker for defense signaling activation (38, 39). In transient expression assays, co-infiltration of PepMV with either dsBC-PepMV^mini-BAX or dsBC-PepMV^mini-INF1 into N. benthamiana leaves significantly increased NbPR1 mRNA levels compared to the control (Fig. 2F). This result suggests that PepMV infection can trigger the expression of RIPs under the control of the dsBC-PepMV^mini system, thereby activating defense responses. In transgenic plants, PepMV infection significantly promoted the up-regulation of NbPR1 mRNA levels in dsBC-PepMV^mini-BAX plants compared to the control (Fig. 2G). In contrast, PepMV infection failed to significantly increase NbPR1 mRNA levels in dsBC-PepMV^mini-INF1 plants (Fig. 2G). This lack of immune marker gene activation likely accounts for the inability of dsBC-PepMV^mini-INF1 transgenic plants to resist viral infection.

We further sought to determine whether dsBC-PepMV^mini-BAX could confer resistance to PepMV infection in its natural host, tomato. Thus, we obtained dsBC-PepMV^mini-BAX transgenic tomato plants, in which the expression of the dsBC-PepMV^mini-BAX construct did not affect tomato growth either (fig. S7). PepMV was then rub-inoculated onto WT, dsBC-PepMV^mini-GFP (control), and dsBC-PepMV^mini-BAX transgenic T2 plants. At 23 dpi, compared with WT and dsBC-PepMV^mini-GFP plants with obvious viral mosaic symptoms, the dsBC-PepMV^mini-BAX transgenic plants exhibited very mild symptoms and lower PepMV accumulation levels (Fig. 2, H to I). These results confirm that the artificially designed dsBC-PepMV^mini-BAX construct confers effective resistance against PepMV infection in planta. To further evaluate the specificity of the TI-TA system, dsBC-PepMV^mini-BAX transgenic plants were challenged with unrelated viruses [(tomato yellow leaf curl virus (TYLCV) and turnip mosaic virus (TuMV)], bacterial pathogen (Pseudomonas syringae pv. tomato DC3000) and fungal pathogen (Botrytis cinerea), as well as heat and drought stresses. Compared with WT plants, dsBC-PepMV^mini-BAX plants exhibited neither HR nor enhanced resistance under any of these nontarget biotic or abiotic conditions (figs. S8, A to G, and S9, A to D). These results demonstrate that the TI-TA system is specifically activated by its cognate virus rather than by unrelated pathogens or environmental stresses, indicating high recognition specificity and minimizing the risk of unintended immune activation.

dsBC-TMV^mini-RIPs confer resistance against TMV in plants

To evaluate strategy universality, we selected TMV for testing—a highly destructive sgRNA-producing pathogen that causes catastrophic losses in global crop production. The 5′ UTR, the 60-nt sequence from the N terminus of ORF1 protein, the subgenomic promoter of MP, the subgenomic promoter of the CP protein, the C-terminal 36-nt sequence of the CP protein, and the 3′ UTR were concatenated to form TMV^mini. GFP was inserted downstream of the subgenomic promoter of the CP protein in TMV^mini, named TMV^mini-GFP. dsBC was constructed into the 5′ end of TMV^mini-GFP and named dsBC-TMV^mini-GFP (Fig. 3A). Agrobacterium cultures harboring TMV^mini-GFP or dsBC-TMV^mini-GFP, along with those carrying EV or TMV, were co-infiltrated into N. benthamiana leaves. At 2 dpi, we observed bright GFP fluorescence in the areas co-inoculated with TMV and dsBC-TMV^mini-GFP compared to those co-infiltrated with EV and dsBC-TMV^mini-GFP (Fig. 3B). The Western blot results were consistent with the fluorescence intensity trends. The GFP protein was almost undetectable in the leaves co-infiltrated with EV and dsBC-TMV^mini-GFP, while a strong protein band was detected in the leaves co-infiltrated with TMV and dsBC-TMV^mini-GFP (Fig. 3C). The qRT-PCR results showed that TMV efficiently activated the expression of GFP mRNA under the control of dsBC-TMV^mini-GFP (Fig. 3D). At the same time, we also found that, compared to GFP, dsBC-TMV^mini-GFP significantly reduced the fluorescence intensity of GFP and the accumulation of the GFP protein (Fig. 3, B and C). qRT-PCR analysis revealed that TMV^mini-GFP significantly decreased the mRNA levels of GFP (Fig. 3D). In the absence of TMV infection, the dsBC tandem TMV^mini-GFP construct did not further decrease GFP mRNA levels, but it dramatically suppressed GFP protein accumulation, reducing it to nearly undetectable levels (Fig. 3, C and D). We then constructed the RIPs into dsBC-TMV^mini and analyzed the HR activated by dsBC-TMV^mini-RIPs with or without TMV infection. The results showed that at 2 to 5 dpi, the leaves co-inoculated with dsBC-TMV^mini-RIPs and EV did not show necrosis, while the leaves co-inoculated with dsBC-TMV^mini-RIPs and TMV showed pronounced HR (Fig. 3, E and F). Notably, leaves co-infiltrated with dsBC-TMV^mini-INF1 and EV also developed HR, albeit with far milder necrotic lesions than those coexpressing dsBC-TMV^mini-INF1 and TMV (Fig. 3G). Given that dsBC-TMV^mini-GFP can reduce GFP to nearly undetectable levels (Fig. 3, B and C), we fused GFP-2A or 2A-GFP to the N-terminal or C-terminal of INF1 and named them dsBC-TMV^mini-GFP-2A-INF1 and dsBC-TMV^mini-INF1-2A-GFP, respectively. At 2.5 dpi, dsBC-TMV^mini-GFP-2A-INF1 still showed HR in leaves without TMV inoculation, whereas dsBC-TMV^mini-INF1-2A-GFP failed to elicit HR unless leaves were infected with TMV (Fig. 3H).

Fig. 3. dsBC-TMV^mini-RIPs enhances resistance against TMV in N. benthamiana.

(A) Schematic of TMV^mini and derivatives, containing the 5′ UTR, 60 nt of ORF1 N terminus, MP and CP subgenomic promoters, CP C-terminal 36 nt, and 3′ UTR (total of 827 nt). Pst I site is located downstream of the CP promoter. (B to D) TMV infection activates GFP expression from dsBC-TMV^mini-GFP. Co-infiltration of dsBC-TMV^mini-GFP with TMV (OD₆₀₀ = 0.5 each; EV as control) activated GFP fluorescence (B), protein accumulation (C), and GFP mRNA (D) at 2 dpi. GFP and TMV^mini-GFP served as controls. GFP protein levels were detected via anti-GFP immunoblotting, with actin as loading control; band intensity was quantified by ImageJ. Quantification: mean ± SD (n = 3). (E to H) TMV infection specifically activates RIP expression. Co-infiltration of dsBC-TMV^mini-RIPs with TMV (OD₆₀₀ = 0.5 each; EV as control) triggered necrosis at 5, 3, 2.5, and 2.5 dpi for BAX, NLP_Pya, INF1, GFP-2A-INF1, and INF1-2A-GFP, respectively. GFP and RIPs served as controls. The 2A peptide is a self-cleaving sequence. (I) Transgenic dsBC-TMV^mini-INF1-2A-GFP and NLP_Pya lines show reduced TMV symptoms at 4 dpi. Scale bar, 2.5 cm. (J) qRT-PCR confirms reduced TMV accumulation in these lines (mean ± SD, n = 6). (K) At 20 dpi, transgenic lines exhibit milder late-stage symptoms than controls. (L) TMV co-infiltration activates NbPR1 expression in RIP lines at 2 dpi. (M) In transgenic plants, TMV up-regulates NbPR1 in RIP lines at 3 dpi. Error bars: mean ± SD (n = 3 for L and M). Statistical significance was determined by one-way ANOVA followed by Tukey’s HSD test (P < 0.05) using IBM SPSS Statistics 27; different letters indicate significant differences among treatments. Selected pairwise comparisons were performed using Student’s t test in GraphPad Prism 8 when appropriate.

We thus introduced dsBC-TMV^mini-INF1-2A-GFP and dsBC-TMV^mini-NLP_Pya, which rapidly triggered a necrotic response under TMV infection conditions, into N. benthamiana and obtained transgenic dsBC-TMV^mini-RIP–positive plants, which did not affect plant growth (fig. S10). At 4 dpi, compared with healthy plants, the leaves of the WT and dsBC-TMV^mini-GFP transgenic N. benthamiana systems inoculated with TMV showed obvious mosaic symptoms and curling, while the dsBC-TMV^mini-RIP transgenic N. benthamiana inoculated with TMV exhibited nearly invisible symptoms in systemic leaves (Fig. 3I). qRT-PCR analysis revealed that the TMV RNA levels in the dsBC-TMV^mini-RIP transgenic plants were significantly lower than those in the WT and dsBC-TMV^mini-GFP transgenic plants (Fig. 3J), indicating that dsBC-TMV^mini-RIPs confer resistance to TMV infection in plants. At the late stage of TMV infection, TMV caused severe damage to WT and dsBC-TMV^mini-GFP transgenic N. benthamiana, almost completely killing these plants. Although TMV was still able to systemically infect dsBC-TMV^mini-RIP transgenic N. benthamiana at its later infection stages, the damage of TMV to dsBC-TMV^mini-RIP transgenic plants was milder than that in the WT and dsBC-TMV^mini-GFP transgenic plants (Fig. 3K). We further examined the NbPR1 expression activated by these dsBC-TMV^mini-RIPs. In transient expression assays, co-infiltration of TMV with either dsBC-TMV^mini-INF1-2A-GFP or dsBC-TMV^mini-NLP_Pya significantly up-regulated NbPR1 mRNA levels in N. benthamiana leaves compared to the control (Fig. 3L). Consistently, upon TMV infection, transgenic lines harboring these constructs also exhibited significantly elevated NbPR1 expression relative to controls (Fig. 3M). These results thus indicate that TMV infection activates RIP expression through the dsBC-TMV^mini system, which in turn activates defense responses.

To further assess the specificity of the TMV-responsive TI-TA system, dsBC-TMV^mini-RIP transgenic plants were challenged with unrelated viruses (TYLCV and TuMV), together with heat and drought stresses. Similar to the PepMV-responsive system, no ectopic HR or enhanced resistance was observed under any nontarget conditions (figs. S8, A to D, and S9, E and F), indicating that activation of the TI-TA system remained strictly dependent on the TMV infection. To further benchmark the antiviral performance of the TI-TA system, we compared dsBC-TMV^mini-RIP transgenic plants with NN transgenic N. benthamiana plants expressing the TMV resistance gene N (40). Following TMV inoculation, NN plants exhibited stronger resistance than dsBC-TMV^mini-RIP plants, displaying almost no visible disease symptoms and barely detectable viral accumulation (fig. S11). Nevertheless, dsBC-TMV^mini-RIP plants still showed substantially reduced disease severity and viral RNA accumulation compared with WT plants. These results indicate that although the current TI-TA system provides weaker protection than N-mediated resistance, it nevertheless confers effective antiviral activity against TMV infection.

CMV^miniRNA3-dsBC-NLP_Pya blocks systemic CMV infection in Arabidopsis thaliana

To further validate the broad applicability of this strategy across diverse plant viruses producing sgRNAs in different host species, we selected CMV, a tripartite RNA virus generating two sgRNAs, and A. thaliana as our experimental system. The 5′ UTR of CMV RNA3, the N-terminal 102-nt sequence of the MP gene, the subgenomic promoter of CP, and the 3′ UTR were concatenated and named CMV^miniRNA3 (Fig. 4A). GFP was constructed behind the subgenomic promoter of CP of CMV^miniRNA3 and named CMV^miniRNA3-GFP. dsBC was constructed into the 5′ end of CMV^miniRNA3-GFP and named as dsBC-CMV^miniRNA3-GFP (fig. S12A). Agrobacterium containing CMV^miniRNA3-GFP or dsBC-CMV^miniRNA3-GFP and Agrobacterium containing EV or CMV were co-infiltrated into N. benthamiana leaves. At 2 dpi, we observed that the areas inoculated with CMV and dsBC-CMV^miniRNA3-GFP did not show bright GFP fluorescence, compared to the areas injected with EV and dsBC-CMV^miniRNA3-GFP (fig. S12B). The Western blotting analysis revealed that CMV could only weakly activate the GFP protein expression under the control of dsBC-CMV^miniRNA3-GFP (fig. S12C). We also observed that the leaves co-inoculated with EV and CMV^miniRNA3-GFP showed almost no fluorescence, while CMV infection activated bright GFP fluorescence in the areas inoculated with CMV^miniRNA3-GFP (fig. S12B). These results suggest that the dsBC at the 5′ end interferes with CMV-initiated GFP protein expression under the control of dsBC-CMV^miniRNA3-GFP. To solve this problem, we constructed single dsA, dsB, and dsC to the 5′ end of CMV^miniRNA3-GFP and named them dsA-CMV^miniRNA3-GFP, dsB-CMV^miniRNA3-GFP, and dsC-CMV^miniRNA3-GFP, respectively (fig. S12A). Then, the CMV-initiated GFP protein expression under the control of dsA-CMV^miniRNA3-GFP, dsB-CMV^miniRNA3-GFP, and dsC-CMV^miniRNA3-GFP was analyzed. The results showed that both fluorescence intensity and Western blot results indicated that dsB-CMV^miniRNA3-GFP most effectively inhibited the accumulation of the GFP protein under the conditions without CMV infection (fig. S12, E and G). Furthermore, qRT-PCR analysis revealed that, compared to CMV^miniRNA3-GFP, dsB did not reduce the GFP mRNA levels (fig. S12I). Under the conditions of CMV infection, compared to the areas inoculated with dsA-CMV^miniRNA3-GFP or dsC-CMV^miniRNA3-GFP, the areas inoculated with dsB-CMV^miniRNA3-GFP displayed the brightest green fluorescence, which were consistent with the fluorescence intensity of Western blot results (fig. S12, F and H). qRT-PCR analysis showed that CMV infection was able to effectively activate the expression of the GFP mRNA under the control of dsB-CMV^miniRNA3-GFP (fig. S12J). However, we found that compared to the areas inoculated with CMV and CMV^miniRNA3-GFP, the leaves inoculated with CMV and dsB-CMV^miniRNA3-GFP displayed significant weaker fluorescence intensity, GFP protein accumulation, and GFP mRNA levels (fig. S12, F to J). To further solve this problem, we constructed dsBC into the Spe I site of CMV^miniRNA3 and named it CMV^miniRNA3-dsBC-GFP (Fig. 4A). Agrobacterium containing CMV^miniRNA3-GFP or CMV^miniRNA3-dsBC-GFP and Agrobacterium containing EV or CMV were co-infiltrated into N. benthamiana leaves. At 2 dpi, compared to EV, CMV infection activated bright GFP fluorescence and high GFP protein accumulation in the areas inoculated with either CMV^miniRNA3-GFP or CMV^miniRNA3-dsBC-GFP (Fig. 4, B and C), although the GFP mRNA levels of CMV^miniRNA3-dsBC-GFP were lower than those produced by CMV^miniRNA3-GFP (Fig. 4D). The Western blot and qRT-PCR analyses revealed that, in the absence of CMV, CMV^miniRNA3-dsBC-GFP did not further affect GFP mRNA levels compared to CMV^miniRNA3-GFP, but it indeed further reduced the GFP protein accumulation (Fig. 4, B to D).

Fig. 4. CMV^miniRNA3-dsBC-NLP_Pya confers extreme CMV resistance in A. thaliana.

(A) Schematic of CMV genome and CMV^miniRNA3 derivatives. CMV^miniRNA3 contains the RNA3 5′ UTR, MP N-terminal 102 nt, CP subgenomic promoter, and 3′ UTR (total of 1160 nt). Mlu I and Sal I sites are downstream of the CP promoter. Purple box: ribozyme (RZ). (B to D) CMV drives GFP expression from CMV^miniRNA3-dsBC-GFP. Agrobacterium carrying CMV^miniRNA3-dsBC-GFP (OD₆₀₀ = 0.5) was co-infiltrated with EV or CMV (final OD₆₀₀ = 0.2 for each construct, total OD₆₀₀ = 0.6) into N. benthamiana leaves. Co-infiltration of CMV^miniRNA3-dsBC-GFP with CMV (not EV) activated GFP fluorescence (B), protein accumulation (C), and GFP mRNA (D) at 2 dpi. GFP and CMV^miniRNA3-GFP served as controls. Quantification: mean ± SD (n = 3). (E to G) CMV infection specifically activates RIP expression. Agrobacterium carrying CMV^miniRNA3-dsBC-RIPs (OD₆₀₀ = 0.5) was co-infiltrated with EV or CMV (final OD₆₀₀ = 0.2 for each construct, total OD₆₀₀ = 0.6) into N. benthamiana leaves. Co-infiltration of CMV^miniRNA3-dsBC-RIPs with CMV (not EV) triggered necrosis at 2.5 dpi; GFP and RIPs served as controls. (H and I) CMV^miniRNA3-dsBC-NLP_Pya blocks systemic CMV infection. CMV sap rub-inoculation was applied, and systemic tissues were photographed at 20 dpi. Scale bar, 2.5 cm. Viral accumulation was quantified by qRT-PCR (n = 6). (J and K) Defense marker NbPR1 (transient N. benthamiana, Agrobacterium carrying CMV^miniRNA3-dsBC-RIPs (OD₆₀₀ = 0.5) was co-infiltrated with CMV (final OD₆₀₀ = 0.2 for each construct, total OD₆₀₀ = 0.6), 2 dpi), and AtPR1 (transgenic A. thaliana, CMV sap inoculation, 3 dpi) were detected via qRT-PCR (n = 3, mean ± SD). Statistical differences were determined by one-way ANOVA followed by Tukey’s HSD test (IBM SPSS Statistics 27; P < 0.05); different letters indicate significant differences among treatments. Selected pairwise comparisons were performed using Student’s t test in GraphPad Prism 8 when appropriate.

Because CMV^miniRNA3-dsBC-GFP can efficiently inhibit the accumulation of GFP protein and initiate GFP expression by CMV infection, RIPs were cloned into CMV^miniRNA3-dsBC and named CMV^miniRNA3-dsBC-RIPs. Agrobacterium containing the constructs and Agrobacterium containing EV or CMV were inoculated into N. benthamiana leaves. The results showed that, at 2.5 dpi, the leaves co-inoculated with CMV^miniRNA3-dsBC-RIPs and EV did not elicit HR, while the leaves co-inoculated with CMV^miniRNA3-dsBC-RIPs and CMV showed obvious cell necrosis (Fig. 4, E to G), indicating that the RIPs under the control of dsBC-CMV^miniRNA3 were specifically activated by the CMV infection. We then transformed CMV^miniRNA3-dsBC-RIPs into A. thaliana to prepare transgenic plants and obtained positive transgenic plants of CMV^miniRNA3-dsBC-RIPs, which did not affect plant growth (fig. S13). CMV was then mechanically inoculated into CMV^miniRNA3-dsBC-RIP transgenic A. thaliana, Col-0, and CMV^miniRNA3-dsBC-GFP transgenic A. thaliana. At 20 dpi, compared with healthy plants, the systemic leaves of Col-0, CMV^miniRNA3-dsBC-GFP, and CMV^miniRNA3-dsBC-INF1 transgenic A. thaliana inoculated with CMV exhibited obvious CMV symptoms, such as mosaic and stunting, while the systemic leaves of CMV-inoculated CMV^miniRNA3-dsBC-NLP_Pya transgenic A. thaliana showed no obvious viral symptoms, and their phenotype was the same as that of healthy Col-0 plants (Fig. 4H). Notably, CMV-inoculated CMV^miniRNA3-dsBC-BAX transgenic A. thaliana plants exhibited pronounced cellular necrosis (Fig. 4H). qRT-PCR analysis revealed that the CMV RNA levels in Col-0, CMV^miniRNA3-dsBC-GFP, and CMV^miniRNA3-dsBC-INF1 transgenic A. thaliana were significantly high, while CMV RNA was undetectable in CMV^miniRNA3-dsBC-NLP_Pya A. thaliana plants inoculated with CMV (Fig. 4I). These results indicate that CMV^miniRNA3-dsBC-NLP_Pya can confer immunity to CMV infection in A. thaliana without any growth penalty.

Furthermore, in transient expression assays, co-infiltration of CMV with either CMV^miniRNA3-dsBC-INF1 or CMV^miniRNA3-dsBC-NLP_Pya into N. benthamiana leaves significantly increased NbPR1 mRNA levels compared to the control. This result suggests that CMV infection can trigger the expression of RIPs under the control of the CMV^miniRNA3-dsBC system, thereby activating the defense response (Fig. 4J). In transgenic plants, CMV infection significantly promoted the up-regulation of AtPR1 mRNA levels in CMV^miniRNA3-dsBC-NLP_Pya plants compared to the control (Fig. 4K). However, CMV infection failed to significantly increase AtPR1 mRNA levels in CMV^miniRNA3-dsBC-INF1 plants (Fig. 4K). This lack of immune marker gene activation likely accounts for the inability of CMV^miniRNA3-dsBC-INF1 transgenic plants to resist CMV infection. Last, to determine whether this high specificity is conserved across different virus-host systems, CMV^miniRNA3-dsBC-RIP transgenic plants were challenged with unrelated viruses [beet severe curly top virus (BSCTV) and TuMV], bacterial pathogen (Pst DC3000), and fungal pathogen (B. cinerea), as well as heat and drought stresses. Consistent with the results obtained for the PepMV- and TMV-responsive systems, no ectopic HR or enhanced resistance was detected under any nontarget conditions (figs. S8, H to N, and S9, G and H). These findings further demonstrate that specific activation by the cognate virus is a general property of the TI-TA platform rather than a virus-specific phenomenon.

The TI-TA strategy for controlled RIP expression confers cross-resistance against related viruses within the same genus

To analyze whether this strategy could help plants acquire cross-resistance against related viruses within the same genus, transgenic dsBC-Virus^mini-RIP materials were inoculated with PVX (41), a virus belonging to the same genus as PepMV, and with ToBRFV (42, 43), a virus belonging to the same genus as TMV. The results showed that systemic leaves of control plants (WT and dsBC-PepMV^mini-GFP) infected with PVX exhibited obvious mosaic symptoms at 20 dpi, while symptoms on systemic leaves of dsBC-PepMV^mini-BAX plants were markedly milder (Fig. 5A). qRT-PCR analysis confirmed that PVX accumulation in dsBC-PepMV^mini-BAX plants was significantly lower than in control plants (Fig. 5B). Furthermore, under PepMV infection, the NbNPR1 mRNA level in dsBC-PepMV^mini-BAX plants was higher than in the control plants (Fig. 5C). At 4 dpi with ToBRFV, although dsBC-TMV^mini-RIP plants could still be infected by ToBRFV, the severity of systemic symptoms, such as mosaic and leaf curling, was slightly reduced compared to control WT and dsBC-TMV^mini-GFP plants (Fig. 5D). qRT-PCR results indicated lower ToBRFV accumulation in dsBC-TMV^mini-RIP plants than in controls (Fig. 5E). Correspondingly, the NbNPR1 mRNA level in dsBC-TMV^mini-RIP plants was elevated (Fig. 5F). These data demonstrate that the dsBC-Virus^mini-RIP strategy can provide partial cross-resistance to two representative viruses within the same genus as the targeted viruses, with reduced viral accumulation and alleviated disease symptoms in transgenic plants.

Fig. 5. dsBC-Virus^mini-RIPs or Virus^mini-dsBC-RIPs confer cross-resistance against related viruses within the same genus.

(A) PVX symptoms are attenuated in dsBC-PepMV^mini-BAX transgenic N. benthamiana at 20 dpi. PVX sap was rub-inoculated onto transgenic and WT plants. Scale bar, 2.5 cm. (B) qRT-PCR confirms reduced PVX accumulation in dsBC-PepMV^mini-BAX lines (mean ± SD, n = 3). (C) PVX increases NbPR1 expression in dsBC-PepMV^mini-BAX lines at 3 dpi. (D) ToBRFV symptoms are milder in dsBC-TMV^mini-INF1-2A-GFP and NLP_Pya lines at 4 dpi. ToBRFV sap was rub-inoculated onto transgenic and WT plants. Scale bar, 2.5 cm. (E) qRT-PCR shows reduced ToBRFV accumulation in these lines (mean ± SD, n = 3). (F) ToBRFV increases NbPR1 expression in dsBC-TMV^mini-RIP lines at 3 dpi. (G) CMV^miniRNA3-dsBC-NLP_Pya blocks systemic CMV-Q infection in A. thaliana. Transgenic Arabidopsis lines were rub-inoculated with CMV-Q sap; Col-0 with healthy sap served as mock. At 20 dpi, systemic leaves were photographed. Scale bar, 2.5 cm. (H) qRT-PCR shows undetectable CMV-Q RNA in CMV^miniRNA3-dsBC-NLP_Pya lines (mean ± SD, n = 3). (I) CMV-Q increases AtPR1 expression in CMV^miniRNA3-dsBC-NLP_Pya lines at 3 dpi. Error bars: mean ± SD (n = 3 for B, C, E, F, H, and I). Statistical differences were determined by one-way ANOVA followed by Tukey’s HSD test (IBM SPSS Statistics 27; P < 0.05); different letters indicate significant differences among treatments. Selected pairwise comparisons were performed using Student’s t test in GraphPad Prism 8 when appropriate.

Considering the limited host range and economic impact of other viruses within the genus Cucumovirus compared to CMV, which infects over a thousand plant species, CMV stands out. Its diverse strains cause a wide array of symptoms and pose a major threat to many important economic crops (44, 45). We further evaluated the resistance of CMV^miniRNA3-dsBC-NLP_Pya materials against an additional CMV strain, CMV-Q, to verify the intraspecies cross-resistance of this strategy. At 20 dpi, Col-0 and CMV^miniRNA3-dsBC-GFP plants inoculated with CMV-Q exhibited obvious stunting and mild mosaic symptoms, whereas CMV^miniRNA3-dsBC-NLP_Pya plants showed a growth phenotype similar to uninoculated Col-0 after CMV-Q challenge (Fig. 5G). qRT-PCR analysis revealed that CMV accumulation in CMV^miniRNA3-dsBC-NLP_Pya plants was nearly undetectable compared to controls (Fig. 5H). Additionally, the AtNPR1 mRNA level in CMV^miniRNA3-dsBC-NLP_Pya plants was significantly higher than in the control group (Fig. 5I). These results indicate that the CMV^miniRNA3-dsBC-NLP_Pya strategy can confer effective cross-resistance against different strains of CMV.

DISCUSSION

Developing durable resistance against plant viruses remains a major challenge because of the rapid evolution, high mutation rates, and remarkable adaptability of viral pathogens (24, 25). To date, several antiviral engineering strategies have been developed, including dominant resistance (R) genes, pathogen-derived resistance (PDR), RNA interference (RNAi), CRISPR-Cas–based genome editing, host susceptibility factor engineering, and immune signaling enhancement. Among these approaches, NLR-mediated resistance can provide strong protection against viral infection (9); however, resistance-breaking viral variants frequently emerge and compromise its durability (27, 43, 46, 47). Classical PDR and cross-protection approaches exhibit favorable biosafety and practical applicability but suffer from narrow antiviral spectra and potential risks of viral recombination and virulence reversion under field conditions (48, 49). RNAi-based strategies enable flexible and sequence-specific viral suppression; however, their efficacy can be compromised by viral sequence variation and diverse viral suppressors of RNA silencing that antagonize host silencing pathways (50, 51). CRISPR-Cas systems provide high-precision and multiplex editing for viral and host target modification, yet mutations within gRNA target sites frequently generate viral escape variants, and constitutive Cas expression may raise biosafety and regulatory concerns (52–54). Engineering host susceptibility factors, such as eIF4E family proteins, can confer durable resistance against specific groups of viruses, but their effectiveness is often limited to particular virus taxa (55–58). Likewise, manipulation of host proteins involved in virus infection through overexpression or knockout can enhance antiviral resistance, although such modifications rarely provide complete immunity (59). In addition, constitutive activation of immune pathways can confer broad-spectrum disease resistance but often results in notable growth penalties because of the well-known growth-defense trade-off (60, 61). To address these limitations, pathogen-responsive resistance systems that activate defense responses only upon pathogen infection have attracted increasing attention. Recent studies have demonstrated that translational regulation mediated by uORFs can effectively control the expression of immune-related genes and enhance resistance against fungal, bacterial, and DNA viral pathogens without obvious fitness costs (34, 35). Likewise, engineered NLRs and inducible elicitor peptide systems can provide strong resistance to potyviruses while minimizing growth penalties by activating defense responses only after infection (29, 36). These findings collectively highlight the potential of pathogen-responsive defense systems to overcome the limitations associated with constitutive resistance. Nevertheless, virus-responsive antiviral systems specifically designed for +ssRNA viruses that produce sgRNAs remain limited.

Compared with existing approaches, the virus-responsive RIP system developed in this study has several notable advantages (Fig. 6). First, antiviral activity is activated only after viral infection, thereby minimizing developmental abnormalities under normal growth conditions. Second, unlike RNAi and CRISPR-based technologies that directly target viral nucleotide sequences, our strategy exploits the conserved process of viral sgRNA synthesis to activate defense responses, potentially reducing the likelihood of resistance breakdown caused by single mutational events. Third, because sgRNA production is a common feature among many +ssRNA viruses, this platform may have broader applicability than virus-specific resistance mechanisms. Additional specificity assays demonstrated that the TI-TA system was not activated by unrelated viruses, bacterial or fungal pathogens, or abiotic stresses including heat and drought (figs. S8 and S9), indicating that viral mini-replicon recognition is highly specific and that the TI-TA system is unlikely to be ectopically activated by nontarget biotic or abiotic stimuli. Such stringent specificity minimizes the risk of unintended HR activation and enhances the biosafety and practical potential of this strategy for agricultural applications. However, several limitations should also be acknowledged. Differences in the abundance, timing, and regulatory mechanisms of sgRNA accumulation among viruses may substantially influence the activation efficiency of the TI-TA system. In addition, plant species may differ in their sensitivity to RIP-induced immune responses, which could further affect system performance. Moreover, all experiments in the present study were conducted under controlled greenhouse conditions using laboratory-maintained viral isolates. Therefore, the stability, durability, and practical performance of the TI-TA system under field conditions remain to be evaluated. In agricultural environments, fluctuating temperatures, mixed infections involving multiple pathogens, and genetically diverse viral populations may influence viral replication dynamics, sgRNA production, and consequently the activation efficiency of the TI-TA circuit. Field trials using diverse viral isolates and representative environmental conditions will therefore be essential to assess the robustness and long-term effectiveness of this strategy. Another limitation is that the current study used constitutive transgenic expression of TI-TA components, which facilitated proof-of-concept validation but may not fully satisfy future regulatory requirements for commercial crop deployment. Future optimization could explore tissue-specific or infection-responsive promoters to restrict transgene expression to relevant tissues or infection sites, as well as minimized expression cassettes to simplify genetic constructs and improve biosafety and regulatory compatibility. The current TI-TA system is designed to target a specific virus or closely related viruses within the same genus. Therefore, achieving resistance against multiple unrelated viruses will likely require stacking several TI-TA modules using strategies such as tRNA- or ribozyme-mediated polycistronic expression. Although the system conferred effective immunity resistance against CMV, it did not provide complete protection against PepMV or TMV. Accordingly, further efforts are required to improve the virus-responsive activation efficiency of TI-TA upon viral challenge. Artificial intelligence–assisted screening can be adopted to identify more efficient RIPs that can potently trigger immune responses across diverse plant species. Furthermore, combining the TI-TA platform with complementary antiviral strategies, including host susceptibility factor engineering, RNAi, and CRISPR-Cas–based technologies, may provide broader-spectrum and more durable resistance against emerging plant viruses.

Fig. 6. Schematic working model of the TI-TA synthetic immune module conferring plant resistance to viral infection.

Our innovative system integrates three key components that combine uAUG-ds as TI, Virus^mini as virus-activated transcription (TA), and RIPs to create a targeted defense that only activates upon infection. The virus-responsive resistance system conferred robust resistance in N. benthamiana, tomato, and A. thaliana plants against three economically important viruses—PepMV, TMV, and CMV—with effectiveness extending to other viruses within the respective genera. This adaptable strategy enables rapid development of virus-resistant crops and may inform new approaches against animal RNA viruses, advancing plant and animal disease control. Created in BioRender. Yang, X (2026) https://BioRender.com/7z4l4rd.

In this study, we found distinct BAX-mediated responses in different plant hosts, with A. thaliana exhibiting markedly stronger responses than N. benthamiana (Figs. 2D and 4H and fig. S14). This interspecific difference might be attributed to a combination of intrinsic host-specific biological characteristics and virus-dependent variations in the activation efficiency of the TI-TA system. On one hand, interspecific variations in cell death signaling pathways, translational regulatory mechanisms, and immune response networks lead to divergent host sensitivity to BAX-induced cytotoxicity. On the other hand, the accumulation level of BAX effector protein is governed by the activation efficiency of the TA module triggered by different viral infections. Transient expression assays verified that CMV infection increased GFP transcript accumulation by ∼336%, while PepMV and TMV resulted in only 13 and 47% increases in transcript accumulation, respectively, demonstrating that CMV activates the TI-TA system far more efficiently than PepMV and TMV under the tested conditions (Figs. 4D, 1I, and 3D). Accordingly, CMV-infected A. thaliana accumulates higher levels of BAX than PepMV-infected N. benthamiana, which might be a key factor contributing to the more intense phenotypic responses in A. thaliana. The differential activation efficiency among the three viral systems may result from multiple virus-specific factors. One possibility is that differences in mini-replicon architecture, including viral UTRs, sgRNA promoter sequences, and regulatory elements required for replication complex assembly, affect the efficiency of viral activation. In addition, the ability of viral RdRps to recognize and use heterologous mini-replicons may vary among viruses, leading to differences in sgRNA production and subsequent TI-TA activation. Furthermore, virus-host compatibility and replication dynamics in different plant species may influence the abundance and timing of viral signals available for system activation. Therefore, the stronger resistance observed in the CMV system likely reflects more efficient activation of the corresponding mini-replicon rather than an inherent superiority of the TI-TA framework. Future optimization of viral regulatory elements and host-virus combinations may further enhance the performance of PepMV- and TMV-based TI-TA systems.

Beyond interspecific response differences, the cross-species transferability of this TI-TA experimental platform is another critical issue worthy of discussion. Mechanistically, the platform consists of three core components: the TI module, the TA module, and the RIPs, which differ greatly in their cross-species adaptability. The TI module is based on uAUG-ds element, a mechanism that is expected to function broadly across plant species because it relies on conserved translational machinery (33). The TA module depends on viral RdRP activity and therefore should be functional in any compatible host-virus interaction (21). In contrast, the RIPs are the primary limiting factor for cross-species application. All RIPs tested in this study were only validated in dicot plants, and their functional activity in monocot crops remains uncharacterized. Meanwhile, our INF1-related experiments further confirmed the host specificity of effector proteins. In transient expression assays, CMV infection efficiently activated CMV^miniRNA3-INF1 and triggered a typical HR in N. benthamiana (Fig. 4F); however, transgenic A. thaliana stably expressing CMV^miniRNA3-INF1 showed no enhanced CMV resistance compared with the WT Col-0 (Fig. 4H), indicating that the ability of effectors to activate plant defense responses is highly host-dependent. Therefore, the cross-species deployment of this platform requires targeted screening and optimization of effector proteins to ensure stable biological activity in target host species. Notably, different RIPs also exhibit inconsistent antiviral efficacy even within dicot hosts. For instance, transgenic lines expressing INF1 displayed relatively weak resistance to PepMV, which was consistent with the finding that PepMV infection failed to significantly increase PR1 expression in INF1 transgenic plants (Fig. 2, D and G). An important limitation of the current study is that RIP protein accumulation could not be directly quantified after viral infection. Because specific antibodies or epitope-tagged versions of RIP proteins were unavailable in the current experimental system, we cannot exclude the possibility that the observed differences among RIPs result from variation in protein abundance, intrinsic effector activity, or downstream immune activation efficiency. Therefore, the relative contribution of these factors to RIP-mediated antiviral resistance remains unresolved. Future studies incorporating fluorescent protein fusions, optimized epitope-tagging strategies, or RIP-specific antibodies will enable direct monitoring of RIP accumulation and provide mechanistic insights into the differential performance of RIP effectors.

The cross-protection spectrum of the TI-TA system is influenced by the evolutionary relatedness between the target virus and the virus-derived mini-replicon. Although TI-TA transgenic plants exhibited cross-resistance against related viruses, the extent of protection varied among virus pairs (Fig. 5). To explore the potential determinants of this variation, we compared the sequence homology of subgenomic promoters from three virus pairs: TMV/ToBRFV (69.08%), CMV-Fny/CMV-Q (65.3%), and PepMV/PVX (32.99%) (fig. S15, A to C). The relatively high promoter similarity between TMV and ToBRFV, as well as between CMV-Fny and CMV-Q, may increase the likelihood that viral replicases recognize heterologous mini-replicons. Although PepMV and PVX share relatively low promoter homology, conserved sequence elements required for potexvirus replication and sgRNA synthesis may still enable PVX RdRp to recognize the PepMV-derived mini-replicon (fig. S15D) (62, 63). Consistent with the activation efficiencies observed in Figs. 1I, 3D, and 4D, we speculate that related viruses differ in their ability to activate heterologous mini-replicons. CMV-Q most efficiently activated CMV^miniRNA3-dsBC-RIPs, whereas PVX showed the weakest activation of dsBC-PepMV^mini-RIPs, which may explain why CMV^miniRNA3-dsBC-RIPs conferred nearly complete immunity against CMV-Q, while dsBC-TMV^mini-RIPs and dsBC-PepMV^mini-RIPs provided only partial resistance to ToBRFV and PVX, respectively (Fig. 5). Although the binding affinity of viral RdRps to noncognate mini-replicons was not examined in this study, it may contribute to differences in mini-replicon activation. Furthermore, variation in RIP expression levels following heterologous mini-replicon activation may further influence the final cross-resistance outcome.

In addition to demonstrating the antiviral efficacy of the TI-TA system, our results also uncovered several unexpected RNA-level phenomena related to transcript stability, translational regulation, and construct architecture, which warrant further investigation. We found that PepMV^mini (Fig. 1I), TMV^mini (Fig. 3D), and CMV^miniRNA3 (Fig. 4D) all markedly reduced GFP mRNA accumulation. In plants, RNA quality control is mediated by nonsense-mediated decay (NMD), no-stop decay (NSD), and no-go decay (NGD), with XRN4 serving as a key component of the 5′ → 3′ RNA decay pathway (64, 65). Using TRV-VIGS (66, 67), we showed that silencing NbXRN4 increased GFP mRNA levels in dsBC-PepMV^mini-GFP, dsBC-TMV^mini-GFP, and CMV^miniRNA3-dsBC-GFP (fig. S16), indicating that XRN4-dependent RNA degradation contributes to the reduced transcript abundance of Virus^mini-based constructs. NMD is known to target viral RNAs containing premature termination codons (PTCs), including those of PVX, TCV, and CGMMV (68–71). For example, PVX RNAs carrying extended 3′ UTRs are degraded through NMD (72), whereas the PEMV2 P26 protein counteracts NMD to protect viral RNAs (73). Moreover, m⁶A modification promotes NMD-mediated degradation of PepMV RNA (74). Because Virus^mini constructs contain PTCs, it is plausible that they are recognized by the NMD machinery, thereby contributing to the reduced accumulation of GFP mRNA. Here, we also found that dsA, dsB, and dsC effectively suppressed GFP protein accumulation without reducing GFP transcript levels, suggesting that these elements primarily act through translational repression. However, tandem fusion of dsA/B/C with dsC did not further enhance TI (fig. S1). In the presence of the PepMV subgenomic promoter PepMV^SP1, fusion of dsA/B/C with dsC resulted in an additional reduction in GFP protein accumulation, possibly due to the presence of a termination codon or specific RNA structural features associated with PepMV^SP1. These observations suggest that combining uAUG-ds elements with viral regulatory sequences may provide a practical strategy for fine-tuning the expression of highly toxic proteins to biologically tolerable levels. In addition, unlike dsB and dsC, dsA not only inhibited GFP translation but also significantly increased GFP mRNA accumulation (Fig. 1E and fig. S1D), implying the existence of an as-yet-unknown mechanism linking dsA to transcript regulation. Another unexpected observation was that dsBC-TMV^mini-INF1 lost its ability to suppress INF1 expression and triggered HR in the absence of viral infection (Fig. 3G). This leakage phenotype was eliminated in dsBC-TMV^mini-INF1-2A-GFP but not in dsBC-TMV^mini-GFP-2A-INF1 (Fig. 3H), suggesting that the INF1 coding sequence together with the TMV^mini 3′ UTR may form a specific RNA structure that interferes with dsBC-mediated translational repression. Insertion of GFP between INF1 and the 3′ UTR may disrupt this putative structure and restore repression. Although this hypothesis requires further experimental validation, it highlights the importance of local RNA structural context in determining TI-TA performance. In addition, our results indicate that the optimal positioning of dsBC may vary among different viral mini-replicons. Unlike PepMV^mini and TMV^mini, in which placement of dsBC at the 5′ terminus enabled efficient virus-responsive activation, CMV^miniRNA3 required internal insertion of dsBC to achieve functionality (figs. S12, B to D, and S4, B to D). Therefore, the current design principles should be considered empirical guidelines rather than fully predictive rules, as they were derived from a limited number of viral mini-replicon configurations tested in this study. The requirement for internal dsBC placement in CMV^miniRNA3 highlights the current limitation that the optimal architecture for a different virus cannot yet be reliably predicted solely from viral sequence information or known regulatory elements. Specifically, 5′-terminal dsBC insertion represents a useful initial design strategy because of its successful performance in two independent viral backbones, whereas internal dsBC insertion provides an alternative configuration when 5′-terminal placement compromises mini-replicon activity or virus-responsive expression. Establishing more generalizable and predictive design principles will require systematic evaluation of diverse viral mini-replicons combined with computational analyses of RNA sequence features, secondary structures, and potential interactions between dsBC elements and viral regulatory regions. Such integrative approaches may ultimately enable rational optimization of TI-TA architectures for newly emerging viral pathogens without extensive empirical screening.

In this study, we combined distinct uAUG-ds repressor elements with the viral PepMV^SP1 promoter to achieve tunable, gradient attenuation of target protein expression, a tool that facilitates functional characterization of cytotoxic or growth-suppressive proteins in planta. This work innovatively fuses the dsBC translational repression module with viral mini-replicons to assemble the TI-TA synthetic immune circuit, which drives strict virus-specific activation of RIP effector expression exclusively during viral infection (Fig. 6). The dsBC-PepMV^mini-BAX construct confers robust PepMV resistance in transgenic N. benthamiana and tomato lines. Separately, dsBC-TMV^mini-INF1-2A-GFP and dsBC-TMV^mini-NLP_Pya grant TMV resistance in stably transformed N. benthamiana. CMV^miniRNA3-dsBC-NLP_Pya delivers near-complete immunity against systemic CMV infection. Furthermore, the TI-TA architecture mediates partial cross-resistance against phylogenetically related viruses within the same genus. Collectively, this programmable synthetic immune strategy provides a conceptual framework for engineering resistance against sgRNA-producing +ssRNA viruses and establishes a foundation for future optimization toward agricultural applications. Beyond agricultural deployment, this experimental framework also provides a tractable system to dissect how growth-regulatory cytotoxic proteins modulate +ssRNA virus infection dynamics in plant hosts. Last, the core TI-TA design logic offers conceptual reference and design inspiration for developing RNA-targeted synthetic immunotherapies against animal RNA viral pathogens.

MATERIALS AND METHODS

Plant growth conditions and plant materials

The WT and transgenic N. benthamiana plants were grown in a climate chamber with a 15-hour daylight and 9-hour dark photoperiod at 23° to 25°C as described previously (75). Solanum lycopersicum (microTom) was grown under the same conditions as N. benthamiana. Generation of transgenic N. benthamiana plants and S. lycopersicum was completed by BioRuan (Wuhan, China). T2 transgenic N. benthamiana and S. lycopersicum plants were inoculated with PepMV or TMV by rubbing plant leaves with virus-infected N. benthamiana crude sap (76). A. thaliana seedlings were vernalized in the darkness at 4°C for 2 to 3 days and then grown in soil at 22° to 23°C with a 10/14-hour light/dark cycle (76). All A. thaliana WT and transgenic plants are in the Col-0 background. To generate CMV^miniRNA3-dsBC-RIPs overexpression lines, CMV^miniRNA3-dsBC-RIPs were inserted into pCAMBIA1300 vector and then transformed into Col-0 plants using the flower dip method (77). Progeny seeds were screened on 1/2 Murashige and Skoog medium plates supplemented with hygromycin B (50 ng/ml), 3% sucrose, and 1.2% agar with pH 5.8 to 6.0 (78). T3 transgenic A. thaliana plants were inoculated with CMV by rubbing plant leaves with virus-infected N. benthamiana crude sap. qRT-PCR was conducted to detect these transgenic plants using the corresponding primers listed in table S1.

Plasmid construction

To systematically validate the TI-TA concept and identify optimal regulatory components, we constructed a series of vectors following a stepwise design logic: starting with core TI elements (uAUG-ds), then integrating viral TA elements, and last, assembling full Virus^mini backbones for functional verification.

Construction of core TI element vectors (screening potent uAUG-ds)

We first generated basic vectors to evaluate the inhibitory efficacy of different uAUG-ds elements. The GFP coding sequence was amplified from TMV-GFP and inserted into pCAMBIA1300 at the Sac I and Bam HI sites using Vazyme’s DNA seamless cloning kit (catalog no. C117) to serve as a control (79). The 5′ UTRs of TUB7 and TBF1 contain uAUG-ds elements (designated dsA, 144 nt; dsB, 105 nt), and a synthetic class I uAUG-ds (dsC, 54 nt) was also used (33). To enhance translational repression, we tandemly fused dsC to dsA, dsB, and dsC, generating dsAC, dsBC, and ds2C, respectively. All single/dual uAUG-ds elements were synthesized and cloned into pCAMBIA1300 at Sac I/Bam HI sites, with a Kpn I site inserted between the uAUG-ds and PepMV 3′ UTR. The resulting vectors were named pCAMBIA1300-Sac I-dsA/B/C/AC/BC/2C-Kpn I-3′ UTR-Bam HI. Subsequently, the GFP fragment was inserted into the Kpn I and Bam HI sites of these vectors, yielding dsA/B/C/AC/BC/2C-GFP, which were used to screen for the most potent TI elements (fig. S1A).

Integration of viral TA element (PepMV^SP1) and initial functional test

To examine whether viral subgenomic promoters (TA elements) can drive gene expression under TI control, the PepMV CP subgenomic promoter (PepMV^SP1) was fused with GFP by overlap PCR, and the resulting cassette was inserted into the Kpn I site of the pCAMBIA1300-Sac I-dsA/B/C/AC/BC/2C-Kpn I-3′ UTR-Bam HI vectors, generating dsA/B/C/AC/BC/2C-PepMV^SP1 (Fig. 1B). A control vector containing only PepMV^SP1-GFP without TI elements was also constructed. Compared with dsB, dsBC further reduced GFP protein accumulation and exhibited the strongest inhibitory effect among all tested uAUG-ds elements. Therefore, pCAMBIA1300-Sac I-dsBC-Kpn I-3′ UTR-Bam HI was selected for subsequent experiments. To determine whether the 36-nt sequence at the C terminus of the CP protein enhances viral activation, the PepMV^SP1-GFP fragment was fused with 3′ UTR-B (the 3′ UTR with the CP C-terminal 36 nt appended to its 5′ end) via overlap PCR. The fused product was then seamlessly cloned into pCAMBIA1300-Sac I-dsBC-Kpn I-3′ UTR-Bam HI linearized with Kpn I/Bam HI or Sac I/Bam HI, yielding dsBC-PepMV^SP1-GFP-B and the corresponding control PepMV^SP1-GFP-B, respectively (fig. S2D).

Assembly of PepMV^mini backbone and dual-site GFP insertion

To further enhance the efficiency of virus-activated target protein expression and recapitulate authentic viral transcriptional regulation, we synthesized PepMV^mini, which comprises the PepMV 5′ UTR, the 45-nt sequence from the N terminus of RdRP, the TGB1 subgenomic promoter (PepMV^SGP192), the CP subgenomic promoter (PepMV^SP1), the 36-nt sequence from the C terminus of CP, and the 3′ UTR-B. PepMV^mini was then cloned into pCAMBIA1300-Sac I-dsBC-Kpn I-3′ UTR-Bam HI linearized with Kpn I/Bam HI or Sac I/Bam HI, generating dsBC-PepMV^mini and PepMV^mini, respectively (Fig. 1F). Two unique restriction sites were engineered in PepMV: Kpn I (downstream of PepMV^SGP192) and Pst I (downstream of PepMV^SP1). GFP was inserted at the Kpn I site to generate PepMV^mini-192-GFP, which tests TGB1 promoter-driven expression (Fig. 1F). GFP was inserted at the Pst I site to generate PepMV^mini-GFP, which tests CP promoter-driven expression and serves as the core TA reporter. GFP was further inserted into Kpn I– or Pst I–linearized dsBC-PepMV^mini vectors, yielding dsBC-PepMV^mini-192-GFP and dsBC-PepMV^mini-GFP, respectively, to validate TI-TA synergy (Fig. 1F).

RIP expression vectors for PepMV resistance

To test antiviral function, we replaced GFP in dsBC-PepMV^mini with RIPs (BAX, INF1, and NLP_Pya), generating dsBC-PepMV^mini-RIPs. For transient cell death assays, RIPs with N/C-terminal 3×HA tags were constructed into pCAMBIA1300 (fig. S4).

Universal TI-TA vectors for TMV and CMV

To establish a TI-TA system for TMV-activated expression, we synthesized TMV^mini, a mini-replicon containing the TMV 5′ UTR, the N-terminal 60 nt of ORF1, the subgenomic promoters of MP and CP, the CP C-terminal 36 nt, and the TMV 3′ UTR. A unique Pst I site was engineered immediately downstream of the CP subgenomic promoter to allow reporter/effector insertion. The TMV^mini fragment was cloned into the pCAMBIA1300-Sac I-dsBC-Kpn I-3′ UTR-Bam HI backbone at two different positions: at Sac I/Bam HI sites to generate the control vector pCAMBIA1300-TMV^mini (without the TI module) and at Kpn I/Bam HI sites to generate the TI-containing vector pCAMBIA1300-dsBC-TMV^mini (Fig. 3A). GFP was inserted into the Pst I site of both vectors to yield pCAMBIA1300-TMV^mini-GFP and pCAMBIA1300-dsBC-TMV^mini-GFP, respectively. For antiviral assays, RIPs (including BAX, INF1, and NLPPya) were cloned into the Pst I site of pCAMBIA1300-dsBC-TMV^mini to produce pCAMBIA1300-dsBC-TMV^mini-RIPs. To monitor INF1 expression and leakage, we also constructed pCAMBIA1300-dsBC-TMV^mini-GFP-2A-INF1 and pCAMBIA1300-dsBC-TMV^mini-INF1-2A-GFP using an overlapping PCR strategy with the 2A self-cleaving peptide (Fig. 3H).

To develop a TI-TA system responsive to CMV, we constructed CMV^miniRNA3-RZ, a mini-replicon derived from CMV RNA3 (1 to 221 nt, 639 to 1256 nt, and 1914 to 2216 nt) with a ribozyme (RZ) at the 3′ end, containing Spe I, Mlu I, and Xba I restriction sites (Fig. 4A). First, GFP was inserted into the Mlu I/Xba I sites to generate pCAMBIA1300-CMV^miniRNA3-GFP-RZ (Fig. 4A). To test different TI modules, the full CMV^miniRNA3-GFP-RZ cassette was cloned into the Kpn I/Bam HI sites of pCAMBIA1300-dsBC/A/B/C-Kpn I-3′ UTR-Bam HI, yielding a series of TI-containing reporter vectors (fig. S12A). To position dsBC upstream of the viral regulatory elements without disrupting promoter activity, we inserted dsBC into the Spe I site of pCAMBIA1300-CMV^miniRNA3-GFP-RZ, generating pCAMBIA1300-CMV^miniRNA3-dsBC-GFP-RZ (Fig. 4A). Last, RIPs were cloned into the Mlu I/Xba I sites to produce the effector vectors pCAMBIA1300-CMV^miniRNA3-dsBC-RIP-RZ.

All constructs were verified by DNA sequencing. Primers and sequences are listed in tables S1 and S2.

Agrobacterium infiltration and GFP imaging

Agrobacterium tumefaciens cultures [optical density at 600 nm (OD₆₀₀) = 0.5] containing pCAMBIA1300-GFP, pCAMBIA1300-uAUG-ds-GFP, and pCAMBIA1300-uAUG-ds-PepMV^SP1-GFP plasmids were infiltrated into 5- to 6-week-old N. benthamiana leaves. A. tumefaciens (OD₆₀₀ = 0.5) harboring pCAMBIA1300-dsBC-PepMV^SP1-GFP, pCAMBIA1300-dsBC-PepMV^mini-GFP, and pCAMBIA1300-dsBC-PepMV^mini-192-GFP plasmids were separately co-infiltrated with A. tumefaciens containing EV or PepMV plasmids (OD₆₀₀ = 0.5) into N. benthamiana leaves. A. tumefaciens (OD₆₀₀ = 0.5) containing the pCAMBIA1300-dsBC-TMV^mini-GFP plasmid was separately co-infiltrated with A. tumefaciens (OD₆₀₀ = 0.5) harboring EV or TMV into N. benthamiana leaves. Similarly, Agrobacterium strains containing GFP controlled by uAUG-ds and CMV^miniRNA3 were separately co-infiltrated with A. tumefaciens containing EV or CMV (pCB301-CMV RNA1, pCB301-CMV RNA2, and pCB301-CMV RNA3 plasmids, with a final OD₆₀₀ of 0.2 for each plasmid) and co-infiltrated into N. benthamiana leaves. At 2 dpi, the agroinfiltrated leaves were illuminated under a long-wavelength ultraviolet lamp (UVP) and photographed under a yellow filter.

Western blot analysis

Total protein was extracted from infiltrated N. benthamiana leaves as described previously (76). Immunoblotting was performed with the primary mouse polyclonal antibodies: anti-GFP polyclonal antibodies from ROCHE (1:5000, catalog no. 11814460001), anti-actin polyclonal antibodies from Abclonal (1:5000, catalog no. AC009), anti-PepMV CP polyclonal antibodies (1:5000, custom-made antibodies) (74), and anti-TMV CP and anti-CMV CP polyclonal antibodies (1:5000, custom-made antibodies). The membranes were then incubated with anti-mouse immunoglobulin G horseradish peroxidase–linked antibodies obtained from Easybio (1:5000, catalog no. BE0102).

RNA extraction and qRT-PCR

RNA extraction was performed as described previously (76), and qRT-PCR was conducted according to previously described methods (80). Total RNA was extracted from the collected leaf samples, and qRT-PCR was performed using gene-specific primers listed in table S1. The NbActin2, AtActin2, and SlActin genes served as endogenous controls (74). For PR1 expression analysis in N. benthamiana, two experimental setups were used. In transient expression assays, A. tumefaciens cultures (OD₆₀₀ = 0.5) carrying the indicated constructs (EV, dsBC-Virus^mini-GFP, or dsBC-Virus^mini-RIPs) were co-infiltrated with Agrobacterium carrying the corresponding virus (PepMV, TMV, or CMV) at OD₆₀₀ = 0.5 (for PepMV and TMV) or final OD₆₀₀ = 0.2 per CMV component (total OD₆₀₀ = 0.6). Inoculated leaves were collected at 2 dpi for qRT-PCR analysis. In transgenic plant assays, Agrobacterium suspensions harboring the corresponding virus (OD₆₀₀ = 0.1) were infiltrated into leaves of WT, dsBC-Virus^mini-GFP, and dsBC-Virus^mini-RIP transgenic N. benthamiana plants. For CMV, virus sap was rub-inoculated onto transgenic A. thaliana lines and Col-0 WT plants. Inoculated leaves were collected at 3 dpi for qRT-PCR analysis. Gene-specific primers for NbPR1 (in N. benthamiana) or AtPR1 (in A. thaliana) were used, with the corresponding Actin genes as endogenous controls. Relative expression levels were calculated using the 2^−ΔΔCT method. All qRT-PCR experiments were performed with at least three biological replicates. In all qRT-PCR graphs, error bars represent means ± SD, and detailed information on biological replicates is specified in the corresponding figure legends. Statistical analyses were performed according to the experimental design. For comparisons among multiple groups, one-way analysis of variance (ANOVA) followed by Tukey’s honestly significant difference (HSD) test was performed using IBM SPSS Statistics 27 (P < 0.05). For selected pairwise comparisons between two groups, Student’s t test was performed using GraphPad Prism 8 when appropriate. Different lowercase letters indicate statistically significant differences among treatments. Bar charts were generated using GraphPad Prism 8. All primers used for qRT-PCR are listed in table S1.

Trypan blue staining

Trypan blue staining was conducted using a previously described method (81).

Viral inoculation

The Agrobacterium (OD₆₀₀ = 0.1) containing infective clones of PepMV, TMV, PVX, or ToBRFV was injected into 4- to 5-week-old N. benthamiana seedlings at the six- to seven-leaf stage. For PepMV, TMV, and PVX, systemic leaves were collected at 7 to 10 dpi, whereas for ToBRFV, systemic leaves were collected at 5 dpi. One gram of infected systemic leaves was homogenized in 100 ml of 10 mM phosphate buffer (pH 7.2) and centrifuged at 3000 rpm for 5 min at 4°C. A total of 10 μl of the virus-containing supernatant was mechanically inoculated onto the leaves of 4- to 5-week-old N. benthamiana seedlings. A. tumefaciens containing plasmids pCB301-CMV RNA1, pCB301-CMV RNA2, and pCB301-CMV RNA3 were mixed at equal concentration to a final OD₆₀₀ = 0.3 and infiltrated into six- to seven-leaf-stage N. benthamiana plants as described previously (78). Subsequently, 1 g of CMV-infected systemic leaves was ground in 100 ml of 10 mM phosphate buffer (pH 7.2) supplemented with 2% PVP-40. The supernatant-containing CMV obtained from the same process was mechanically inoculated into two expanded leaves of 4-week-old A. thaliana seedlings (78).

Supplementary Materials

This PDF file includes:

Figs. S1 to S16

Tables S1 and S2