Antiviral RNA interference in plants: Increasing complexity and integration with other biological processes

Abstract

RNA interference (RNAi, also known as RNA silencing) is one of the most important plant defense responses against viral invasion. Although major components of the RNAi pathway, steps leading to viral small interfering RNA biogenesis, and viral counterdefense strategies via RNAi suppressors have been well studied, the broader roles of RNAi in viral infection and seed transmission remain less thoroughly characterized. In particular, the increasing complexity of RNAi-associated mechanisms and their integration with other biological processes have not been comprehensively summarized. Increasing numbers of studies have identified non-canonical RNAi pathways, novel host factors involved in RNAi, and the possibility of small RNAs acting across kingdoms to modulate plant–virus–vector tritrophic interactions. In this review, we provide an overview of the roles of RNAi in plant viral infections and describe recent advances, with emphasis on the discoveries of novel positive and negative RNAi regulators, potential signaling pathways upstream and downstream of antiviral RNAi, and the prospects and challenges of double-stranded RNA applications, either expressed from transgenes or supplied exogenously via spraying. We also discuss how these findings reshape current views on antiviral RNAi, highlight remaining knowledge gaps, and examine how these advances influence plant–virus co-evolution while informing strategies for managing plant virus diseases and reducing their impact.

RNA interference (RNAi) is a key defense mechanism in plants against viruses, but its broader roles in viral infection dynamics and seed transmission remain understudied. Recent advances have revealed non-canonical RNAi pathways, novel host factors, and cross-kingdom small RNA interactions, reshaping the understanding of plant–virus–vector interactions. This review provides an overview of recent advances in elucidating the roles of RNAi in plant viral infections, highlighting newly identified RNAi regulators, signaling pathways, and applications of dsRNA, while also addressing knowledge gaps and their implications for viral disease management.

Classical antiviral RNAi pathway in plants

Plants utilize multilayered innate immune systems to defend against diverse abiotic and biotic stresses, including pathogenic microbes. RNA interference (RNAi, also known as RNA silencing), RNA decay, and quality control–coupled RNA decay are major RNA-mediated immune responses that regulate mRNA transcription, translation, and degradation during plant growth, development, and stress conditions (Guo et al., 2019; Li et al., 2019; Li and Wang, 2019; Baulcombe, 2022). Among these RNA-mediated immune responses, RNAi restricts viral invasion by inactivating viral DNA transcription, cleaving viral RNA transcripts, and repressing viral protein translation (Figure 1) (Baulcombe, 2022). RNAi is initially triggered in primary virus-infected plant cells. In these cells, viral RNAs and their intermediates form double-stranded RNAs (dsRNAs) through inter- or intramolecular base pairing. Viral dsRNAs are recognized and cleaved by DICER-like proteins (DCLs) from the RNase III family to generate virus-derived small interfering RNAs (vsiRNAs) 21–24 nt in length. These vsiRNAs are loaded onto Argonaute proteins (AGOs) to assemble into RNA-induced silencing complexes (RISCs), which then guide the degradation of complementary viral RNAs. This cascade of steps constitutes the post-transcriptional gene silencing (PTGS) branch of RNAi. In contrast, the methylation of viral DNA through the RNA-directed DNA methylation (RdDM) pathway, which involves 24-nt vsiRNAs and depends on DCL3 and AGO4, represents the transcriptional gene silencing (TGS) branch (Figure 1) (Guo et al., 2022). In addition, plant endogenous RNA-dependent RNA polymerases (RDRs), with the assistance of Suppressor of Gene Silencing 3 (SGS3), synthesize dsRNAs using aberrant viral single-stranded RNAs produced during primary cleavage of viral RNA as templates. After DCL processing, these dsRNAs generate pools of secondary vsiRNAs that amplify antiviral RNAi and strengthen the overall defense response (Wang et al., 2010, 2011) (Figure 1). Antiviral RNAi signaling molecules can also move from cell to cell and leaf to leaf, thereby initiating systemic antiviral RNAi in distant, initially uninfected plant tissues (Rosas-Diaz et al., 2018).

Figure 1.

A model illustrating the initiation and amplification of antiviral RNA interference in virus-infected plant cells.

In virus-infected plant cells, viral dsDNA molecules—either from dsDNA viruses or generated during the rolling-circle replication of ssDNA viruses—serve as templates for Pol IV and RDR2 to generate 24-nt viral dsRNAs. RNA virus genomes and their replication intermediates also form dsRNAs. These dsRNAs are recognized and cleaved by RNase III family DCL enzymes (DCL4 and DCL2) into 21- to 22-nt vsiRNAs. From these vsiRNA duplexes, one strand, the guide strand, is loaded into the core AGO component of RISC, which is then activated to target complementary viral RNAs for degradation. This pathway constitutes the PTGS branch of RNA interference (RNAi). Endogenous RDRs, assisted by SGS3, use aberrant viral ssRNAs—products of initial RISC cleavage—as templates to synthesize additional dsRNAs. Processing of these dsRNAs by DCLs generates a pool of secondary vsiRNAs, thereby amplifying the antiviral RNAi response and forming a positive feedback loop. A subset of 24-nt vsiRNAs, generated mainly by DCL3, is incorporated into AGO4 and guides DNA methylation of viral genomes through the RdDM pathway, establishing the TGS branch of RNAi. AGO, Argonaute; DCL, Dicer-like protein; dsRNA, double-stranded RNA; dsDNA, double-stranded DNA; Pol IV, RNA polymerase IV; PTGS, post-transcriptional gene silencing; RDR, RNA-dependent RNA polymerase; RISC, RNA-induced silencing complex; RdDM, RNA-directed DNA methylation; SGS3, Suppressor of gene silencing 3; ssDNA, single-stranded DNA; ssRNA, single-stranded RNA; TGS, transcriptional gene silencing; vsiRNA, virus-derived small interfering RNA.

Although some plant-RNA–derived small interfering RNAs (siRNAs) mediate host defense against viral infection by regulating stress-associated gene expression, their reported roles in antiviral defense are more general and less specific than those of vsiRNAs. During plant–virus co-evolution, multiple counterdefense strategies have emerged to antagonize host antiviral immunity. For example, nearly all known plant viruses encode at least one RNA silencing suppressor (RSS) protein; these RSSs play critical roles in viral adaptation to plants and in the development of disease symptoms (Guo et al., 2019; Li et al., 2019; Jin et al., 2022). Many research laboratories have extensively investigated the mechanisms underlying host antiviral immunity and RSS-mediated viral counterdefense. These topics have been reviewed in detail in several excellent articles that address antiviral RNA components (Carbonell and Carrington, 2015; Yang and Li, 2018; Baulcombe, 2022). Accordingly, this review will not repeat that content but will instead focus on recent discoveries and developments in non-canonical antiviral RNAi in plants, novel RNAi regulators, signaling pathways, as well as the broader scope of antiviral RNAi; it will also discuss novel findings and applications to combat viral diseases.

Non-canonical antiviral RNAi in plants

Although much is known about the canonical RNAi pathways, studies in past decades have indicated the existence of pathways that differ from these classical mechanisms. For example, in addition to the production of 24-nt siRNAs by RNA polymerase IV–RDR2–DCL3, several additional small RNA (sRNA) pathways have been discovered that also direct RdDM and are referred to as non-canonical RdDM mechanisms (Figure 1) (Cuerda-Gil and Slotkin, 2016). In this review, “non-canonical RNAi” is defined as a potential antiviral RNA silencing pathway. Non-canonical TGS–RdDM refers to pathways that deviate from the classical DCL3–AGO4–RDR2 axis; non-canonical PTGS refers to pathways that deviate from the classical DCL2/4–AGO1/2–RDR6-mediated cascades, including those involving atypical siRNA sizes and AGO-incompatible sRNAs. Several examples are known. The first involves inverted repeat- and microRNA (miRNA)-directed DNA methylation, in which RNA polymerase II transcripts are directly cleaved by DCL3 into 24-nt sRNAs that participate in RdDM (Panda et al., 2016). The second is the RDR6 RdDM pathway, in which 21- to 22-nt siRNAs produced during PTGS from RNA polymerase II transcripts and processed by either DCL2 or DCL4 activate RISCs that engage in RdDM (McCue et al., 2015; Cuerda-Gil and Slotkin, 2016). The third is the RDR6–DCL3 RdDM pathway, in which RDR6-mediated dsRNA synthesis is followed by DCL3 processing into 24-nt siRNAs that associate with the RdDM pathway (Marí-Ordóñez et al., 2013). These non-canonical RdDM pathways are mainly distinguished by variations in the steps leading to siRNA biogenesis. Compared with canonical RdDM, these pathways may have lesser importance but remain poorly characterized and insufficiently understood (Cuerda-Gil and Slotkin, 2016). Whether these non-canonical RdDM mechanisms contribute to viral infection in plants remains unclear. Given the importance of RDR6 and DCL3 in canonical antiviral RNAi, it is plausible that non-canonical RdDM mechanisms also have antiviral functions during the establishment of plant virus infection.

In addition to non-canonical RdDM/TGS mechanisms, the existence of non-canonical PTGS pathways in plants has also been suggested. Evidence for such pathways was recently provided by Samarskaya et al. (2023), who conducted a comparative analysis of siRNAs in the context of potato virus Y (PVY) infection versus spraying of dsRNA corresponding to a fragment of the PVY genome. In contrast to PVY-induced production of discrete 21- and 22-nt siRNA species, the externally applied PVY dsRNA fragment generated a non-canonical pool of sRNAs, appearing as ladders of ∼18–30 nt in length, which indicated an unexpected sRNA biogenesis pathway (Samarskaya et al., 2023). It remains unclear whether, or which, AGO proteins can load non-canonical sRNAs of 18–30 nt that are not normally associated with AGO to form a functionally active RISC. Although speculative, one possibility is that only a small fraction of external dsRNA processed by DCL2 and DCL4 produces AGO2- and AGO1-compatible 22- and 21-nt siRNA duplexes, respectively, and that only these duplexes are responsible for virus suppression. Intriguingly, Frascati et al. reported the absence of a peak at 21- and 22-nt siRNA duplexes after exogenous dsRNA application for protection against tomato leaf curl New Delhi virus (ToLCNDV) (Frascati et al., 2024). Instead, they only observed enrichment of 24-nt siRNAs derived from sprayed dsRNA, suggesting that exogenous dsRNAs were not processed into siRNAs through the canonical PTGS pathway. Caution is warranted, however, because the observed enrichment of 24-nt siRNAs derived from sprayed dsRNA may alternatively have resulted from the initial production of a small pool of 21- and 22-nt vsiRNAs that subsequently contributed to the amplification of predominantly 24-nt siRNAs.

Thus, the vast majority of siRNAs generated from external dsRNAs fall within non-canonical AGO-incompatible size ranges, raising questions about whether external dsRNAs directly contribute to reductions in viral RNA. Nevertheless, potential non-canonical PTGS mechanisms for processing exogenous and endogenous dsRNAs and for loading non-canonical siRNAs (18–30 nt) require further investigation.

Positive and negative modulators of antiviral RNAi

Mechanisms and compartments involved in vsiRNA biogenesis and amplification

Despite extensive research on core effectors of the antiviral RNAi pathway, such as DCL, AGO, and RDR proteins, the mechanisms involving many host factors in antiviral RNAi remain incompletely understood. Cucumber mosaic virus (CMV), turnip mosaic virus (TuMV), potato virus X (PVX), and Arabidopsis thaliana have been widely used to investigate plant–virus interactions and antiviral RNAi. These studies have revealed multiple novel host factors that play crucial roles in antiviral RNAi (Guo et al., 2017, 2018; Gao et al., 2018; Rosas-Diaz et al., 2018; Rubio et al., 2019; Brosseau et al., 2020; Liu et al., 2022). For example, four novel RNAi regulators were identified through forward genetic screens using A. thaliana and CMV as an experimental model (Table 1 and Figure 2). These four regulators—two phospholipid flippases and aminophospholipid-transporting ATPases 1 and 2—are multispan transmembrane magnesium transporters that reinforce the amplification of RDR1- and RDR6-generated secondary vsiRNAs (Guo et al., 2017) (Figure 2). ANTIVIRAL RNAI-DEFECTIVE 2 (AVI2) has also been shown to positively regulate antiviral RNAi through participation in vsiRNA biogenesis (Guo et al., 2018). In addition, Enhancer of RDR6 3 (ENOR3) has been reported to activate host defense against CMV infection through additive interactions with RDR1, RDR6, DCL2, and DCL4 (Gao et al., 2018). Because phospholipid production at RNA virus replication sites is often increased, it has been suggested that these phospholipids are recruited into membrane-bound RDR1/6-associated RNA synthesis compartments to support robust siRNA biogenesis.

Table 1.

Plant RNAi regulators and their functions during viral infections.

RNAi regulator	Virus	Host	Main functions	References
CAMTA3	CLCuMuV, CMV, PPV	Nb	Positive transcriptional activator of RDR6 and BN2 to initiate the RNAi-mediated antiviral response	Wang et al. (2021); Wang et al. (2022)
BN2	CLCuMuV, CMV, PPV	Nb	Potential sRNA degradation enzyme that post-transcriptionally and indirectly regulates the levels of DCL1 and AGO1/2 mRNAs by degrading their cognate microRNAs, thus conferring resistance to plant viruses	Wang et al. (2021); Wang et al. (2022)
rgsCaM	TEV	Nt	Potential proviral function induced by HC-Pro; suppression of PTGS	Anandalakshmi et al. (1998)
	TYLCCNV/TYLCCNB	Nb	Proviral role repressing RDR6 expression and promoting autophagic degradation of SGS3	Li et al. (2014); Li et al. (2017)
	TGMV	At	Proviral role with an unclear mechanism	Chung et al. (2014)
ALA1/2	CMV	At	Antiviral role enhancing amplification of vsiRNAs by RDR1 and RDR6	Guo et al. (2017)
ENOR3			Antiviral role participating in the biogenesis of highly abundant vsiRNAs	Gao et al. (2018)
AVI2			Antiviral role acting additively with RDR1, RDR6, DCL2, and DCL4	Gao et al. (2018)
VIR1			Proviral role possibly restricting transcriptional induction of DCL4	Liu et al. (2022)
RDO5			Antiviral role promoting amplification of vsiRNAs specifically through RDR6	Liu et al. (2022)
BAM1/2	TYLCV	At	Potential antiviral role through cell-to-cell spread of RNAi	Rosas-Diaz et al. (2018)
miR168, miR444	RSV	Os	Proviral roles inhibiting AGO1 and RDR1, respectively	Wu et al. (2015); Wang et al. (2016b)
miR168a, miR403a, miR162b, miR1515a	SMV	Gm	Potential proviral functions inhibiting AGO1, AGO2, DCL1, and DCL2, respectively	Bao et al. (2018)
miR403a	ToMV	Nb	Proviral role inhibiting AGO2	Diao et al. (2019)
ERF2	CMV	Ph	Antiviral role positively regulating expression of RDR2, RDR6, DCL2, and AGO2	Sun et al. (2016)
Importin, α2	BaMV	Nb	Antiviral role regulating levels of RDR6-dependent secondary vsiRNAs	Wang et al. (2024)

The important plant RNAi regulators in plant–virus interactions are summarized, and their known functions are described based on research in various plant species, including At, Arabidopsis thaliana; Gm, Glycine max; Nb, Nicotiana benthamiana; Os, Oryza sativa; and Ph, Petunia hybrida. BaMV, bamboo mosaic virus; CLCuMuV, cotton leaf curl Multan virus; CMV, cucumber mosaic virus; PPV, plum pox virus; RSV, rice stripe virus; SMV, soybean mosaic virus; TEV, tobacco etch virus; TGMV, tomato golden mosaic virus; ToMV, tomato mosaic virus; TYLCCNV, tomato yellow leaf curl China virus; TYLCCNB, tomato yellow leaf curl betasatellite; TYLCV, tomato yellow leaf curl virus.

Figure 2.

A schematic presentation of the roles of DCLs, RDRs, and RNAi regulators in plant antiviral defense.

In virus-infected plant cells, viral dsRNA precursors are generated through direct binding of two DNA virus transcripts, replication of RNA virus genomes, or folding of viral single-stranded RNAs. The resulting dsRNAs are recognized by DCL enzymes and processed into 21- to 24-nt vsiRNAs. To inhibit viral infection, these vsiRNAs bind to AGO to form a RISC that targets DNA virus transcripts and RNA virus RNAs through TGS and PTGS. Viral RNAs can also be recognized and recruited by RDR enzymes, together with the chaperone protein SGS3, to synthesize dsRNAs. Secondary vsiRNAs are then cleaved from these dsRNAs to enhance antiviral RNAi. During antiviral RNAi, DCL3, RDR2, and AGO4 function in TGS; DCL2, DCL4, RDR1, RDR6, SGS3, and AGO1/2 are involved in PTGS. Several RNAi regulators have recently been identified, including CaM, CAMTA3, rgsCaM, AVI2, ALA1/2, ENOR3, RDO5, VIR1, RTL1, VIM5, miR168, miR403, miR444, and miR1515a. These regulators can modulate host antiviral RNAi by influencing RNAi core components during viral infection in plants. In addition, antiviral RNAi has been shown to inhibit viral seed transmission, although the underlying mechanism remains unexplored. The knowledge summarized in this figure is mainly based on studies using A. thaliana as a model plant. AGO, Argonaute; ALA, aminophospholipid-transporting ATPase; AVI2, Antiviral RNAi-defective 2; BAM1/2, Barely any meristem 1/2; DCL, Dicer-like enzyme; DRB, double-RNA binding protein; DRM2, Domains rearranged methylase 2; ENOR3, Enhancer of RDR6 3; Pol V, RNA polymerase V; PTGS, post-transcriptional gene silencing; RdDM, RNA-directed DNA methylation; RDO5, Reduced dormancy 5; RDR, RNA-dependent RNA polymerase; RISC, RNA-induced silencing complex; SGS3, Suppressor of gene silencing 3; TGS, transcriptional gene silencing; VIR1, Antiviral RNAi regulator 1.

Recently, a CMV-induced small peptide of 71 amino acids, named “virus-induced small peptide 1” (VISP1), was identified as a selective autophagy receptor that interacts with SGS3 and mediates autophagic degradation of SGS3/RDR6 bodies. Overexpression of VISP1 impairs vsiRNA amplification and promotes viral infection (Tong et al., 2021). At later stages of viral infection, as viral replication and RSS content significantly increase, VISP1 primarily targets RSSs for autophagic degradation to inhibit viral infection (Tong et al., 2023), as demonstrated by its interaction with the RSS C2/AC2 protein of two geminiviruses. This finding suggests that VISP1 has a conserved role as a selective autophagy receptor in immune responses to a variety of viruses (Tong et al., 2023). In Solanum lycopersicum, PRMT6 was identified as a protein interacting with and targeting viral RSS. PRMT6 inhibits the RSS activity of tomato bushy stunt virus (TBSV) P19 by methylating its key arginine residues R43 and R115, thus reducing its dimerization and sRNA-binding activities and enhancing plant resistance to TBSV infection (Zhu et al., 2024).

Natural variation and viral targeting of RNAi regulators

Plant RNA viruses, due to their error-prone replication mechanisms, generate substantial genetic variability that enables rapid adaptation to changing environmental conditions. Less is known about the natural variation exhibited by wild host plants in response to viral infection. Through genome-wide association studies, Reduced Dormancy 5 (RDO5)/Delay of Germination 18 (DOG18) and Antiviral RNAi Regulator 1 (VIR1) were cloned and later identified as positive and negative regulators, respectively, of antiviral RNAi in A. thaliana. Molecular and genetic analyses have shown that RDO5 enhances siRNA amplification mainly through the RDR6-dependent pathway, whereas VIR1 represses antiviral RNAi, possibly by restricting DCL4 transcription (Li et al., 2022; Liu et al., 2022). A natural variant of A. thaliana AGO2 has also been reported to contribute to non-host resistance to PVX infection (Brosseau et al., 2020), highlighting plant–virus co-evolution and the positive selection of RNAi pathway variants. Notably, this AGO2 variant carries a non-synonymous SNP that enhances antiviral activity, exemplifying how natural genetic variation can modulate RNAi efficacy. Such favorable alleles represent valuable genetic resources for breeding virus-resistant crops, whether through marker-assisted selection or transgenic approaches. With the increasing availability of population-scale resequencing data, additional antiviral RNAi-related polymorphisms are likely to be identified in diverse plant species, providing opportunities for translational application in sustainable crop protection. The discovery of further RNAi regulators and effectors not only would advance understanding of the mechanisms and complexity underlying antiviral RNAi but also could facilitate the development of new strategies for engineering virus resistance.

Recently, Wen et al. (2024) employed proximity labeling using the barley stripe mosaic hordeivirus RSS protein as bait to identify host factors required for antiviral RNAi in Nicotiana benthamiana. The DEAD-box RNA helicase 20 (RH20), a conserved eukaryotic protein with a human homolog known as DDX5, was identified (Wen et al., 2024). RH20 was found to negatively regulate the infection of several positive-sense RNA viruses and to interact with SGS3/RDR6 to execute this antiviral function, identifying RH20 as a new component of RNAi-amplifying cytoplasmic SGS3/RDR6 bodies. Similarly, Huang et al. independently observed a role for RH20 in plant virus resistance. In contrast to the first study, RH20 was characterized as a phase separation protein that interacted with AGO2 to promote the accumulation of both endogenous and exogenous sRNAs (Huang et al., 2024). Li et al. (2021) further identified RH6, RH8, and RH12 as components of nuclear dicing (D) bodies, which are involved in miRNA processing and antiviral defense. These helicases interacted with SERRATE, a key component of D bodies, to promote its phase separation and the formation of D bodies. Infection by TuMV counteracted this process by decreasing the accumulation of these helicases as well as the accumulation of D bodies (Li et al., 2021). Another helicase, the chloroplast-localized RH3, interacts with AGO2 and facilitates the loading of siRNA strands into AGO2 at endoplasmic reticulum–chloroplast membrane contact sites (Huang et al., 2025a). These contact sites also serve as replication sites for certain potyviruses and tymoviruses. Altogether, these results underscore the importance of RHs and their functional diversification in the antiviral RNAi pathway. Interestingly, in a separate study, Li et al. (2020) demonstrated that apple SERRATE negatively affects drought resistance by regulating the expression of MYB88 and MYB129 as well as by influencing miRNA biogenesis. Whether and how this regulation is affected during plant virus infection has not been studied (see also “The role of vsiRNAs in plant drought tolerance” on viruses and plant drought resistance) (Li et al., 2020).

Given efforts to dissect the interplay between virus and host during antiviral RNAi and viral counterdefense by RSS proteins, complexity and integration with other host processes have continued to increase, leading to the identification of additional host proteins involved in these pathways. Whereas RNA viruses are primarily subject to PTGS, the cytoplasmic branch of antiviral RNAi, reports have indicated the involvement of more complex nucleocytoplasmic communication in antiviral RNAi and viral counterdefense. For example, earlier studies revealed that the CMV 2b RSS protein compromises not only AGO1 of the cytoplasmic RISC in PTGS but also AGO4 of the nuclear RNA-induced transcriptional silencing complex, which is involved in TGS (Hamera et al., 2012). RNA potyviruses suppress antiviral RNAi through the helper component–proteinase (HC-Pro) RSS protein; their viral protein genome-linked (VPg) RSS protein interferes with the RDR6 cofactor SGS3, localizes to Cajal bodies and nuclear condensates, and associates with numerous TGS components (Rajamäki and Valkonen, 2009; Rajamäki et al., 2014). During viral infection, transport of viral proteins into the nucleus by members of the importin α family plays an important role, and several studies have highlighted a role for these proteins in antiviral RNAi. A recent study on bamboo mosaic virus (BaMV) demonstrated that importin α proteins negatively regulate BaMV infection. Silencing of importin α2 increased BaMV levels in local and systemic leaves due to reduced accumulation of RDR6-dependent secondary vsiRNAs (Wang et al., 2024). An earlier study by Huang et al. (2019) showed that AGO10a sequesters vsiRNAs during BaMV infection. Furthermore, the BaMV triple-gene blockp1 protein, a component of the viral movement protein complex and the viral RSS, interacts with and upregulates AGO10a. Although the proviral role of AGO10a primarily involves cytoplasmic functions in PTGS, the protein is mainly localized in the nucleus. Silencing of importin α2 also upregulated AGO10a but inhibited its nuclear import. Therefore, importin α2 acts as a negative regulator of RNAi by controlling nucleocytoplasmic shuttling of AGO10a and promoting vsiRNA degradation through a triple-gene block p1–AGO10a–SDN1 exonuclease pathway (Huang et al., 2019; Wang et al., 2024).

miRNAs orchestrating the antiviral RNAi pathway

Unlike siRNAs, miRNAs are a class of small, single-stranded RNAs 18–24 nt in length generated from host-encoded non-coding RNA transcripts. miRNAs primarily regulate gene expression at the PTGS level through degradation of target mRNAs or inhibition of protein translation (Zhang et al., 2019; Millar, 2020). miRNAs can also regulate RNAi effectors via feedback mechanisms that influence host defenses against viral infection. Through microarray analyses of soybean mosaic virus–infected susceptible soybean plants, Bao and colleagues found that the expression levels of miR168a, miR403a, miR162b, and miR1515a were significantly upregulated, whereas the expression of their target genes AGO1, AGO2, DCL1, and DCL2 was inhibited (Bao et al., 2018), resulting in repression of RNAi-mediated disease resistance. Consistent with this finding, miR403a has been shown to activate N. benthamiana defense against tomato mosaic virus infection by regulating AGO2 expression (Diao et al., 2019). In rice, AGO18 enhances antiviral RNAi by sequestering miR168, which represses expression of AGO1, a principal antiviral effector (Wu et al., 2015). Rice stripe virus (RSV) infection induces miR444 expression, which targets the MIKC(C)-type MADS-box proteins (OsMADS23, OsMADS27a, and OsMADS57) to activate the RDR1-dependent antiviral silencing pathway (Wang et al., 2016b). In addition to their canonical roles in host immunity, recent findings highlight more diverse and complex functions of miRNAs in plant–virus interactions. For instance, insect-derived miR263a acts as a “double agent” during RSV infection, facilitating viral replication in insect vectors while activating antiviral responses in rice hosts, thereby mediating a cross-kingdom defense regulation mechanism (Zhao et al., 2025a). In parallel, RSV impairs host miRNA biogenesis by disrupting the phase separation of SERRATE, a core component of D bodies, thus interfering with miRNA processing and contributing to viral pathogenesis (Zou et al., 2025). These findings emphasize intricate connections among miRNAs, viral infection, and liquid–liquid phase separation, offering new insights into sRNA-mediated antiviral strategies.

Beyond sRNA regulation, the post-translational modification (PTM) landscape of RNAi machinery components is increasingly recognized as a regulatory axis. Plant viruses may modulate or disrupt PTM pathways—such as phosphorylation or ubiquitination—of DCL, AGO, or RDR proteins to fine-tune or suppress RNAi activity. For example, the P0 protein of polerovirus has been shown to trigger degradation of AGO1 and prevent assembly of sRNA-loaded RISCs, likely by interfering with host ubiquitination machinery, effectively dismantling antiviral RNA silencing (Csorba et al., 2010). Although direct evidence in plants remains limited, viral interference with host PTM systems represents a promising research direction and may reveal novel viral strategies for overcoming RNA-based immunity.

Signaling pathways and scope of antiviral RNAi

Signaling upstream of antiviral RNAi

Although RNAi is a major plant antiviral defense that activates the RISC upon the production of viral dsRNA and its subsequent processing into vsiRNAs, it remains largely unclear how RNAi is transcriptionally initiated and reinforced, and which signals are required to prime RNAi in virus-infected cells. Recently, Wang and colleagues reported that in virus-infected plants wounded by mechanical force or insect feeding, the wounding-triggered Ca²⁺ flux activated calmodulin-dependent transcription, resulting in upregulation of multiple RNAi core genes to intensify antiviral defense. These findings indicate that Ca²⁺ flux functions upstream of RNAi (DeMell and Dinesh-Kumar, 2021; Wang et al., 2021). Because up to 80% of known plant viruses are transmitted by insect vectors that feed by piercing–sucking, biting, or chewing (Wu et al., 2024a), invasion by nearly all plant viruses is associated with wounding and Ca²⁺ signaling activation, suggesting that RNAi is primed before virus entry into host cells. Wang and colleagues further demonstrated that the wounding-induced Ca²⁺ flux activates calmodulin-binding transcription activator 3 (CAMTA3), which in turn induces expression of RDR6 and bifunctional nuclease-2 (BN2) via binding to their promoters. BN2 acts as a ribonuclease that stabilizes DCL1 and AGO1/2 mRNAs by degrading their cognate miRNAs. Therefore, CAMTA3 and BN2 can be considered positive RNAi regulators. In N. benthamiana, disruption of CAMTA3 or BN2 function increases susceptibility to geminivirus, cucumovirus, and potyvirus infections. As a counterdefense, the geminivirus-encoded V2 protein suppresses RNAi-based host defense by disrupting the interaction between calmodulin and CAMTA3 (Wang et al., 2021). Similarly, P1—the RSS of wheat yellow mosaic virus (WYMV)—interferes with the interaction between NbCaM and NbCAMTA3, thereby suppressing RNA silencing and facilitating viral infection in wheat (Chen et al., 2023). These findings implicate Ca²⁺ signaling in the RNAi pathway, although Ca²⁺-signaling-activated antiviral RNAi can be subverted by viral RSSs. Recently, Huang et al. identified a RING1–IBR–RING2-type ubiquitin ligase in rice that directly recognizes viral coat proteins delivered by insect vectors, triggering degradation of the JAZ3 repressor, induction of jasmonic acid (JA) signaling, and subsequent activation of AGO18-mediated antiviral RNAi at the early stage of infection (Huang et al., 2025b).

In addition, JA signaling has been proposed as an upstream regulator of RNAi, enhancing antiviral RNAi in rice (Yang et al., 2020b). JA and its derivatives are critical plant hormones involved in development, immunity, and responses to abiotic stress and pathogens. Yang and colleagues demonstrated that JA signaling transcriptionally activates AGO18, a core RNA silencing component that promotes rice antiviral defense by sequestering miR168 and miR528, which repress key antiviral defense proteins. JA signaling and RNA silencing synergistically strengthen rice antiviral defense; the JA-responsive transcription factor JAMYB directly binds to the AGO18 promoter to upregulate AGO18 expression. Furthermore, the authors showed that the RSV-encoded coat protein triggered JA accumulation and activated JAMYB, initiating this host defense network. Plant–virus interactions involve diverse and dynamic regulatory networks. Although studies regarding the crosstalk of Ca²⁺, JA, and other hormone signaling pathways with RNAi remain limited, the conserved nature of these pathways suggests that varying degrees of such crosstalk occur in all plant–virus interactions. Elucidating how plants perceive viral infections and how this interplay activates antiviral RNAi remains a key challenge for the future.

Downstream effects of antiviral RNAi: Viral seed transmission and stem cell antiviral immunity

Antiviral RNAi not only directly affects viral titers and concomitant infection throughout the plant but also influences virus dissemination both horizontally, to other plants, and vertically, to offspring through seed transmission. The impact of final virus titers, shaped by viral replication and regulated by antiviral RNAi, on the success of horizontal transmission by insect vectors is evident. Viruses strongly targeted by antiviral RNAi that fail to counterdefend through viral silencing suppressors or other evasion strategies accumulate at low titers, reducing their acquisition and subsequent transmission rates by insect vectors during feeding. In contrast, vertical transmission via seeds does not appear to depend strictly on viral titers. Seed transmission of plant viruses strongly promotes dissemination and spread across continents; it serves as a primary infection source capable of causing rapid and severe epidemics. Nevertheless, seed transmission is not a universal trait among plant viruses. For many years, the mechanisms determining vertical transmission via seeds remained unclear. However, recently, Liu and Ding demonstrated that RNAi is required to suppress seed transmission (Liu and Ding, 2024; Wang and Liu, 2024). In Arabidopsis dcl2 dcl4 knockout lines, the average seed transmission rate of CMV increased by more than 10-fold compared with wild-type Col-0 plants. A mild but significant increase in seed transmission was also observed in rdr1 rdr6 mutant plants infected with CMV-2aTΔ2b (lacking the viral 2b RSS) but not with CMV. Furthermore, AGO5 participates in antiviral RNAi against the RSS-deficient mutant virus during systemic infection of progeny seeds, but not of whole plants. These findings indicate a tissue-specific difference in the requirement of RNAi components to inhibit systemic viral infection of vegetative tissues versus progeny seed tissues; they demonstrated a defined RNA-based effector mechanism as an immune pathway to combat viral seed transmission. Many plant viruses have not been reported to transmit vertically by seed; however, it has not been determined whether this is an absolute trait, or whether these viruses are more strongly targeted by RNAi core effectors—AGO1, AGO2, AGO5, DCL2/4, and RDR1/6—or fail to adequately counteract these defenses with their RSSs, thereby blocking systemic progeny seed infection and transmission.

Besides seeds, shoot apical meristem (SAM) tissues often remain protected against viral infection and play an important role in restricting vertical virus transmission to progeny. Recent studies have indicated a key role for RDR1; DCL2, -3, and -4; and salicylic acid (SA) in preventing TuMV from invading SAM stem cells (Incarbone et al., 2023). During TuMV infection, SA induction was observed, which subsequently upregulated RDR1 expression. Without this concerted action, TuMV was able to invade stem cells. Overexpression of RDR1 in the absence of SA did not restore stem cell exclusion of TuMV, indicating an additional role for SA in SAM stem cell protection. Other viruses that induced SA were also excluded from stem cells; in contrast, tobacco rattle virus did not induce SA and was able to invade stem cells, strengthening the correlation between SA activation and virus-free SAM stem cells. Further experiments showed that RDR1 is required to remotely supply vsiRNAs for activation of antiviral RNAi but not the production of host-derived virus-activated siRNAs (vasiRNAs) (see also “The roles of vasiRNAs in plant–virus interaction”) (Incarbone et al., 2023). In addition to RDR1 and SA, WUSCHEL, a meristem-specific transcription factor, has been shown to mediate virus exclusion from SAM stem cells (Wu et al., 2020) through a mechanism involving translational inhibition.

The role of vsiRNAs in plant drought tolerance

Viral siRNAs have also been implicated in drought tolerance. A 21-nt vsiRNA from the cowpea mild mottle virus triple gene block 1 targets a host transcription factor gene (PvTCP2) encoding a transcriptional repressor. Its attenuation led to transcriptional upregulation of the core autophagy-related gene PvATG8c and activation of autophagy to levels greater than those induced by drought stress or viral infection alone (Wu et al., 2024b). In another recent study, Lei et al. (2025) demonstrated that silencing of miRNA169a promoted vascular tissue formation and increased water use efficiency. Moreover, photosynthesis rates and enzymatic antioxidant activity were elevated, whereas levels of reactive oxygen species (ROS) were reduced, collectively strengthening drought resistance. Considering that many plant viruses encode RSSs that act by sequestering siRNAs and may also bind structurally similar miRNAs, it remains unknown whether plant viruses commonly promote drought resistance by sequestering miRNA169a.

The roles of vsiRNAs in plant–virus–vector interaction

vsiRNAs are the primary executors of antiviral RNAi in plant–virus interactions. Recent research has shown that, in addition to targeting viral RNA to inhibit infection, vsiRNAs can act in trans to target host RNAs, thereby inducing viral symptoms, altering plant biological processes, and modulating plant immunity or fitness during infection (Figure 2) (Matsumura and Kormelink, 2023). For example, silencing of chloroplast-related genes by vsiRNAs from CMV Y satellite RNA induces yellow symptoms in tobacco by targeting a gene involved in chlorophyll biosynthesis (Shimura et al., 2011), which might increase plant attractiveness to insect vectors. Silencing of stress-related genes, including the abscisic acid-related 7-like protein gene, may contribute to enhanced host susceptibility to viral infection (Guo and Wong, 2020). Silencing of eIF4A, a regulator of autophagy and a target of an RSV-derived siRNA, activates antiviral autophagy and inhibits viral infection in rice (Zhang et al., 2021). Silencing of vacuolar H+-PPase by WYMV-derived vsiRNA-20 amplifies ROS signaling to promote cell death, thereby suppressing plant defense responses (Yang et al., 2020a). Notably, Liu et al. generated several transgenic wheat lines using four artificial miRNA expression vectors carrying vsiRNAs from WYMV and found that two lines expressing amiRNA1 were highly resistant to WYMV infection. Further analyses revealed that vsiRNA1 modulated the expression of a wheat thioredoxin-like gene (TaAAED1), a negative regulator of ROS production in the chloroplast (Liu et al., 2021). In addition, Xia et al. constructed a degradome library of sugarcane mosaic virus (SCMV)-inoculated maize plants and identified more than 3000 cleavage sites potentially produced by vsiRNAs in positive-strand RNAs of SCMV and 42 in maize transcripts, including some associated with chloroplast functions and biotic or abiotic stress responses (Xia et al., 2018). Despite the limited number of reports, evidence is slowly emerging that viruses, through their vsiRNAs, can alter plant biological processes that determine the outcome of infection. A very recent study revealed the production of vsiRNAs from the highly conserved dsRNA panhandle structure at the genomic RNA termini of rice stripe tenuivirus during replication in its planthopper insect vector. In the planthopper, these vsiRNAs targeted DOPA decarboxylase (DCC), a component of the proteolytic prophenoloxidase (PPO) pathway that restricts or kills pathogens. Knockdown of DCC in planthoppers led to a dramatic increase in viral RNA levels, demonstrating a role for DCC in the prophenoloxidase-mediated insect immune response and its modulation by vsiRNAs (Zhao et al., 2025b). Findings from these studies, which demonstrate extended targeting of host genes by vsiRNAs, highlight the increasingly complex arms race between viruses and their plant and insect hosts, a conflict from which viruses ultimately benefit during replication and dissemination.

In addition to targeting plant host–encoded RNA transcripts, the possibility of in planta–produced vsiRNAs targeting insect vector RNA has been suggested. Such vsiRNAs, once acquired during insect vector feeding on virus-infected plants, could potentially alter insect behavior (Figure 3) (Matsumura and Kormelink, 2023). Trans-kingdom sRNA communication has already been demonstrated between fungi and plant hosts, and a recent study by Han et al. (2025) showed such communication between insects and plants. In that study, a salivary gland–enriched miRNA from whiteflies, conserved across phloem-feeding insects, was released from salivary exosomes into tobacco cells; it hijacked AGO1 and silenced a defense gene that positively regulates the phytohormones SA and JA. Although experimental evidence is limited, these findings by Han et al. support the possibility of trans-kingdom RNAi occurring in the opposite direction—namely, vsiRNAs produced in plants acting on insects, potentially affecting viral vector behavior and facilitating viral spread (Han et al., 2025). In accordance with this possibility, an earlier study profiled miRNAs in the whitefly Bemisia tabaci Middle East–Asia Minor I after acquisition of tomato yellow leaf curl China virus (TYLCCNV) from infected plants (Wang et al., 2016a). However, no vsiRNAs were detected in whiteflies that could have originated from plants after TYLCCNV acquisition, suggesting either that these vsiRNAs were not acquired from plants or that their ingested levels were below the threshold for detection. Alternatively, siRNA processing mechanisms may exist in insects that destabilize or degrade ingested siRNAs after passage through the midgut. Another complicating factor is that many insect vectors, including aphids, whiteflies, leafhoppers, planthoppers, and thrips, preferentially feed on phloem. Whether vsiRNAs accumulate in phloem cells at sufficiently high levels for acquisition remains uncertain. The recent finding of tomato yellow leaf curl virus replication in whiteflies (He et al., 2020) further complicates the identification of trans-kingdom RNAi in the geminivirus pathosystem, given that vsiRNAs detected in insects could originate either from viral replication within the insect or from uptake during acquisition feeding. Accordingly, definitive evidence of trans-kingdom sRNA communication in plant–virus–vector interactions must come from viruses transmitted by insect vectors in a non-persistent or circulative manner, such that viral replication does not occur in the insect. At present, direct evidence for cross-kingdom RNAi in plant–virus–vector systems remains scarce; further functional validation is required to confirm such interactions.

Figure 3.

Possible role of trans-kingdom small RNAs in plant–virus–vector tritrophic communication.

(A) In virus-infected plant cells, virus-generated siRNAs not only target the viral genome but also regulate host gene expression. Viral infection also activates the plant endogenous small RNA (sRNA) pathway, leading to the accumulation of virus-activated host siRNAs (vasiRNAs), virus-activated host miRNAs (vamiRNAs), and trans-acting siRNAs (tasiRNAs). These infections generate altered endogenous sRNA profiles that primarily fine-tune the expression of endogenous genes through RNA-mediated silencing, and some of them may be redirected to target the viral genome (Matsumura and Kormelink, 2023).

(B) Most plant viruses depend on insect vectors for transmission. When feeding on infected plants, these vectors can acquire virus- or host-derived sRNAs. If such sRNAs are cross-kingdom transferred to the insect or to other plants, they may manipulate vector behavior to enhance viral spread, simultaneously triggering plant immune responses (trans-kingdom RNAi). Current findings only partially support this phenomenon in plant–virus–vector interactions, highlighting the need for further mechanistic studies. Image created using BioRender.com, with permission.

The roles of virus-activated siRNAs (vasiRNAs) in plant–virus interaction

Plant virus infection triggers antiviral RNAi, which not only generates vsiRNAs but also promotes the accumulation of endogenous plant sRNAs, also referred to as vasiRNAs. Emerging evidence indicates that vasiRNAs are involved in antiviral immunity in plants. Hu et al. first showed that infection of Arabidopsis with oilseed rape mosaic virus causes a global, size-specific enrichment of miRNAs, trans-acting siRNAs (tasiRNAs), and other phased siRNAs. This enrichment is observed primarily for 21-nt RNAs with a 5′-terminal guanine. vasiRNA enrichment is not correlated with defects in miRNA-dependent tasiRNA biogenesis or with global changes in the levels of mRNA and tasiRNA targets (Hu et al., 2011). Cao et al. (2014) noted that activation of antiviral RNAi in Arabidopsis is accompanied by the production of an abundant class of endogenous siRNAs mapped to exon regions of more than 1000 host genes and rRNAs. These vasiRNAs are predominantly 21 nt long and are genetically distinct from known endogenous siRNAs; they resemble viral siRNAs in their requirement for DCL4 and dependence on RDR1 for biogenesis (Cao et al., 2014). Binding of these vasiRNAs to AGO2 directs widespread silencing of host target genes. Notably, production of vasiRNAs is inhibited by the silencing suppressor protein 2b of CMV. The authors hypothesized that vasiRNAs contribute to a broad-spectrum defense response (Cao et al., 2014). RDR1-dependent vasiRNAs similar to those described by Cao et al. were also reported by Leonetti et al. (2021) in a study of three Brassicaceae species infected with a DNA caulimovirus. Recently, Fletcher et al. showed that Capsicum and N. benthamiana WA (an accession containing a functional RDR1 gene) infected with tomato spotted wilt virus or capsicum chlorosis virus produced abundant vsiRNAs; host-specific differences were evident for each pathosystem. Furthermore, both solanaceous host species, except for the widely used N. benthamiana LAB accession lacking a functional RDR1 gene, generated highly abundant vasiRNAs against numerous endogenous transcripts. Targets of these vasiRNAs included ribosomal protein–encoding genes and many genes involved in protein processing in the endoplasmic reticulum, suggesting that co-localization of viral and endogenous transcripts provides a basis for vasiRNA biogenesis (Fletcher et al., 2022). The number of studies on vasiRNAs remains limited, as do investigations into their functions and the underlying mechanisms in plant defense, plant–virus interactions, and plant–virus–insect interplay, whether antiviral or proviral.

In addition to internal signaling networks, an intriguing but underexplored question is whether vasiRNAs, as well as siRNAs and vsiRNAs, influence the plant-associated microbiome through cross-kingdom regulatory mechanisms. The plant microbiota has increasingly been recognized as a “second genome” that plays essential roles in disease resistance and immune modulation. Although direct evidence is lacking in the context of plant antiviral RNAi, recent studies in other systems suggest that sRNAs can move between organisms and affect microbial gene expression. For example, sRNAs have been shown to move from plants into pathogenic fungi to silence virulence genes (Cai et al., 2018), raising the possibility of similar interactions with bacterial or fungal members of the microbiota. It is thus conceivable that vsiRNAs may alter microbial populations and indirectly modulate virus–host interactions. Exploration of this possibility may lead to new perspectives regarding the ecological impact of RNAi-based defense.

Exploitation of antiviral RNAi: Prospects and challenges of virus-specific dsRNA applied by transgene and external spray methods

Advances in RNAi research and its role in antiviral immunity have created opportunities for deploying RNAi-based technologies in crop protection and agricultural production to defend against pests and pathogens (Wu et al., 2024a). Even before RNAi mechanisms were clearly dissected, transgenic expression of homologous RNA was successfully utilized to defend against different plant viruses by initiating antiviral RNAi (Beachy et al., 1990; Guo and Garcia, 1997). Subsequent studies revealed that generation of vsiRNAs is a prerequisite for successful RNA-mediated resistance in plants. Since then, many highly efficient approaches have been developed to engineer virus-resistant transgenic plants, mostly based on different precursor RNAs for siRNA production, including dsRNA and hairpin RNA (Wu et al., 2024a). For example, Li et al. expressed a hairpin RNA construct in which the stem sequence contained sequences complementary to four different rice viruses: rice black-streaked dwarf virus, southern rice black-streaked dwarf virus, RSV, and rice ragged stunt virus. Transformed plants stably expressing these hairpin RNA molecules exhibited strong and broad resistance against all four viruses (Li et al., 2024). A similar strategy had already been applied to successfully generate virus resistance to five different tomato-infecting orthotospoviruses (Hassani-Mehraban et al., 2009). Numerous comparable or simpler strategies have since been employed to defend against different viruses in diverse crop species (Wu et al., 2024a). However, due to public concerns about the use of genetically modified organisms and genetically modified crops in food production, researchers have begun to explore the potential for exogenous application of dsRNA to protect against plant virus infection. The first laboratory studies regarding this approach appeared about two decades ago (Tenllado and Díaz-Ruíz, 2001). External spray application of dsRNA has since received increasing attention and has recently been developed as a non-transgenic strategy for crop protection against pests and pathogens (Zhao et al., 2024). Recent advances in nanomaterial-based delivery systems, such as guanidinium-functionalized siRNA nanoparticles, have shown promising potential in enabling systemic gene silencing across plant tissues and species barriers, thereby broadening the applicability of topical RNAi approaches for antiviral defense (Lin et al., 2025). Advantages of this approach include low toxicity, safety, convenience, and greater environmental sustainability (Hernández-Soto and Chacón-Cerdas, 2021). To date, this technology has been successfully applied to target more than 10 economically important plant viruses in over 10 plant species (Taliansky et al., 2021). Although exogenous dsRNA application is generally considered safe, concerns remain regarding unintended effects on non-target organisms and the regulatory oversight required for large-scale use. Studies on environmental stability and species specificity are crucial for risk assessment.

Despite increasing understanding of RNAi and innate immunity mechanisms in plants (Niehl et al., 2016; Liu and Ding, 2024; Liu et al., 2024; Zhao et al., 2024), many questions remain regarding the application of exogenous dsRNA in the field to protect against viral infections. These questions concern not only technological limitations and risks associated with field-applied dsRNA (Zhao et al., 2024) but also fundamental scientific issues related to the underlying mechanisms (Figure 4). For example: (1) How does exogenous dsRNA enter plant recipient cells, and what occurs after entry? Studies have indicated that receptor proteins localized in the plasma membrane may recognize exogenous dsRNA and trigger plant immunity. In such cases, it may be difficult to determine whether inhibition of viral disease results from RNAi-mediated or immunity-induced effects. Niehl et al. showed that in vitro–generated dsRNAs induce pattern-triggered immunity (PTI) responses dependent on the co-receptor SERK1 but independent of DCL proteins in Arabidopsis (Niehl et al., 2016). Moreover, dsRNA treatment of Arabidopsis induced SERK1-dependent antiviral resistance, suggesting that dsRNA-mediated resistance involves membrane-associated processes and functions independent of RNA silencing (Niehl et al., 2016). Thus, dsRNA sensitivity may represent a useful trait to increase antiviral resistance in crop protection by activating PTI rather than RNAi. Accordingly, some ambiguity may exist in attributing dsRNA spray–based antiviral protection solely to the RNAi pathway. (2) Where and how are topically applied dsRNAs, once inside the cell, further processed into siRNAs and loaded into AGO–RISCs? It is generally assumed that the functional mechanism of sprayed dsRNA-induced antiviral RNAi is similar, or even identical, to that of in vivo RNAi-based defense against plant viruses. However, two recent studies revealed that externally applied dsRNA fragments are processed into a non-canonical pool of sRNAs, which appear either as ladders ∼18–30 nt in length from PVY dsRNA or as an enrichment of 24-nt siRNAs from ToLCNDV dsRNA (Samarskaya et al., 2023; Frascati et al., 2024). These findings indicate an unexpected sRNA biogenesis pathway. Notably, these non-canonical sRNAs do not move efficiently and fail to induce transitive amplification (i.e., spreading of the RNAi signal to sequences flanking the siRNA target site). Further research on this topic is required and may not only clarify the underlying mechanisms but also contribute to improving sprayed dsRNA-mediated antiviral control and crop protection against other pests and pathogens. (3) Does dsRNA always induce RNAi in plants? Recent evidence indicates that this is not necessarily the case. In plants, nitrate reductase is a key enzyme for NO biosynthesis, encoded in Arabidopsis by two nitrate reductase genes, NIA1 and NIA2. Li et al. recently showed that a natural antisense transcript, as-NIA1, transcribed from the 3′ untranslated region of NIA1 stabilizes NIA1 mRNA to maintain its circadian oscillation. Detailed analysis revealed that mRNA adenosine methylase deposits N6-methyladenosine on as-NIA1, enabling direct binding of polypyrimidine tract-binding protein 3 to UC-rich elements in as-NIA1, which subsequently facilitates stabilization of the NIA1 mRNA target in vivo (Li et al., 2025). These findings suggest that dsRNA formation in the case of as-NIA1 and NIA1, rather than triggering RNAi, promotes stabilization of the latter and prevents RNA decay. Such observations indicate that not all dsRNAs necessarily trigger RNAi. Substantial uncertainties remain regarding the entry of sprayed dsRNAs into plant cells, the processing of exogenous dsRNA, the functionality of non-canonical siRNAs derived from sprayed dsRNAs, and whether these are incorporated into AGO complexes or act through alternative mechanisms.

Figure 4.

A model showing exogenous application of dsRNA-induced antiviral pattern-triggered immunity (PTI) and antiviral RNA silencing.

After spraying, dsRNA may be perceived directly by an unidentified plasma membrane receptor (PRR) and its co-receptor kinase SERK1, or it may enter cells through vesicle-mediated transport or damaged cells, where it is recognized by cytosolic PRR-related proteins. Cytosolic PRRs with the dsRNA-priming complex then phosphorylate downstream signaling components and induce typical pattern-triggered immunity (PTI) responses, including activation of mitogen-activated protein kinase (MAPK) cascades, induction of hormone signaling, and transcriptional reprogramming of defense gene expression through the activation of key transcription factors (TFs) (Niehl et al., 2016). In addition to membrane-associated dsRNA recognition events, dsRNA is generally presumed to be delivered into the cytoplasm, either by vesicle-mediated transport or through damaged cells, to trigger antiviral RNAi. Although endogenous dsRNA recognition by DCL proteins during antiviral RNAi has been extensively studied, the functional mechanisms of exogenous dsRNA-induced antiviral RNAi in recipient cells remain obscure and require further investigation. Exogenous application of dsRNA may also induce antiviral PTI and RNAi in neighboring cells through plasmodesmata (PD), vesicle-mediated transport, damaged cells, or other unknown mechanisms. dsRNA, double-stranded RNA; MAPK, mitogen-activated protein kinase; PRR, pattern recognition receptor; SERK1, somatic embryogenesis receptor kinase 1; TFs, transcription factors; PD, plasmodesmata; VRC, viral replication complex. Image created using BioRender.com, with permission.

Despite many limitations and gaps in knowledge of RNAi, increasing evidence supports the successful application of foliar RNAi-based strategies for plant virus disease control (Mitter et al., 2017; Tabein et al., 2020; Delgado-Martín et al., 2022). Considering that dsRNA is the trigger of antiviral RNAi and can also function as a potent PTI elicitor, Necira et al. recently analyzed the impact of exogenously applied dsRNA on both layers of defense against PVX expressing GFP. They found that sequence-specific and non-specific dsRNA reduced virus accumulation in both inoculated and systemic leaves; it induced typical antimicrobial PTI pathways, including calcium, ethylene, and mitogen-activated protein kinase signaling (Necira et al., 2024). However, sequence-specific dsRNA not only triggered a specific RNAi response but also exerted a greater impact on defense than non-specific dsRNA, suggesting that sequence-specific RNAi and non-specific PTI pathways synergistically help to protect plants against viruses. Although the functionality of non-canonical sRNAs generated from exogenous dsRNA has not been fully elucidated, it is possible that only a small subset enters AGO complexes, and its antiviral activity differs from that of canonical vsiRNAs. Thus, for further technological development and successful lab-to-field transition, additional research is needed to clarify the precise mechanisms of external dsRNA-triggered antiviral RNAi and immunity in plants.

Concluding remarks

Plants have evolved unique strategies to combat DNA and RNA virus invasion (Wang et al., 2022). Viral infections in plants activate multiple layers of RNA-based host immunity (Chen et al., 2024a, 2024b; Ge et al., 2024; Yang et al., 2024). Among these immune responses, RNAi plays a predominant role in restricting virus proliferation through the inactivation of viral DNA transcription (TGS), cleavage of viral RNAs, and repression of viral protein translation (PTGS) (Wu et al., 2024a; Lozano-Durán, 2024). Although many plant RNAi regulators and virus-encoded RSSs have been identified and cloned over the past two decades, the molecular mechanisms underlying antiviral RNAi continue to increase in complexity. This is due not only to its integration with multiple processes—such as JA, SA, and Ca²⁺/calmodulin signaling pathways and plant drought tolerance—but also to the broad scope of vsiRNAs, which can target host genes, and to the increasingly detailed understanding of the cytobiology of key RNAi components. The latter is exemplified by the functional diversification of homologs linked to spatial distribution and specific non-canonical RNAi pathways (e.g., RDR1-dependent vasiRNAs, RHs, and AGO10). Further studies on antiviral RNAi will not only expand understanding of its complexity and its integration with host processes but, in the long term, also contribute to improving and enhancing the efficiency of topical dsRNA applications to induce and fine-tune RNAi.

As highlighted by the limited scope of current research, key unanswered questions regarding antiviral RNAi in plants include the following: (1) Where does antiviral RNA silencing initiate? Does it occur in any virus-infected cells, or only in specific cells such as phloem companion cells? (Where?) (2) What are the first signals in plants that sense viruses beyond viral dsRNA to activate RNAi? Are there changes in calcium flux, or is hormone signaling triggered by the insect vector feeding on plants? (What/which and how?) (3) In addition to the core RNAi effector components, how many additional critical host factors are required for RNAi, and what are their roles? (What/which and how?) (4) Viral dsRNAs, whether derived from plant DNA or plant RNA viruses, are localized in different cellular compartments. Are DCL proteins strictly compartmentalized, or can they shuttle to compensate for one another’s loss? What determines their localization, and how are DCL proteins accurately translocated to sites for processing nuclear and cytoplasmic viral dsRNAs upon infection? (Where and how?) (5) Although abundant vsiRNAs are present in infected plant cells, viruses can establish effective infection due to suppression of RNAi. What proportion of vsiRNAs are recruited into RISCs and are functionally capable of mediating viral RNA silencing during active infection? (How many?) Could RNAi function to maintain viral RNA homeostasis in plant cells, preventing excessive viral accumulation that would kill the host and ultimately impair viral infection? Is it possible that RNAi may even benefit the virus? (6) Are siRNA-activated RISCs able to enter the viral replication complex or virion, which is protected by virus-induced membrane structures or viral proteins, to target and cleave viral RNA? (Where/how?) (7) Antiviral RNAi inhibits vertical transmission of viruses by seed. Does it also restrict the seed transmission potential of viruses not typically transmitted by seed? (What/how?) (8) How does exogenously applied dsRNA enter plant cells and subsequently trigger intracellular RNAi activity? Do extracellular receptor proteins mediate entry? (How?)

Many important questions remain unanswered and require continued fundamental studies on host defense pathways during plant–virus–host interplay. As this research field continues to advance, these questions will likely be resolved; additional players in antiviral RNAi-based defense will be discovered. Overall, these advances will expand opportunities to maximize the exploitation and efficiency of antiviral RNAi and to develop alternative new strategies for virus-resistant crops.

Funding

We acknowledge support from the National Natural Science Foundation of China (32320103010 and 32172385 to F.L.) and the Agricultural Science and Technology Innovation Program (CAAS-BRC-CB-2025-02 to F.L.).

Acknowledgments

We apologize to those colleagues whose original work and review articles could not be cited in this review due to space constraints. Figures were created using BioRender.com, with permission. No conflict of interest is declared.

Author contributions

F.L. conceived the project. F.L. wrote the initial draft of the manuscript, and F.L., R.K., and X.Z. thoroughly edited the paper. F.L., X.L., and S.Z. generated the figures and Table 1. All authors contributed to the reviewing and proofreading of the manuscript.

Published: August 26, 2025