RNA products and RNA virus‐based technologies have the potential to transform agriculture by enabling on‐demand crop trait reprogramming and effective pest and disease management (Pasin et al., 2024; Rössner et al., 2022). In virus‐induced gene silencing (VIGS), engineered RNA viruses can redirect the host RNA interference machineries to target gene silencing through the production of gene‐specific small RNAs (sRNAs) (Rössner et al., 2022). Although endogenous sRNAs and those resulting from VIGS are in the 20–30‐nt range, VIGS vectors are engineered to deliver larger inserts of 200–400 nt with homology to a target gene, often located in less conserved regions to ensure specificity (Ahmed et al., 2020). Reducing insert sizes, may enhance the VIGS scalability and applicability to non‐model species.
Nicotiana benthamiana is the most widely used model for optimizing VIGS protocols. However, this host has a complex, allotetraploid genome with functionally redundant homeologous gene pairs, and for which no high‐quality assemblies were available until very recently (Ranawaka et al., 2023). Multi‐gene CRISPR‐Cas9 mutagenesis was applied to tackle functional redundancy in plants (Berman et al., 2025; Ellison et al., 2020). Here, we hypothesized that VIGS insert sizes could be lowered to match those of endogenous sRNAs by combining enhanced genomics and transcriptomics resources to guide the design of virus‐delivered short RNA inserts (vsRNAi) for simultaneous targeting of homeologous gene pairs.
We focused on the magnesium protoporphyrin chelatase subunit I (CHLI) gene, whose downregulation results in leaf yellowing due to a reduction of chlorophyll biosynthesis and levels. Tomato (Solanum lycopersicum) is a diploid relative of N. benthamiana with high‐quality genome assembly and annotation. The tomato CHLI coding sequence (CDS) is distributed across three exons with boundaries supported by transcriptomic expression analysis (Supporting experimental procedures). Analysis of reported N. benthamiana Niben101 and Niben261 annotations, however, revealed CHLI variation including large sequence deletion and insertion, not supported by results obtained by de novo transcriptome assembly (Figure S1). A high‐quality chromosome‐level genome assembly was recently reported for N. benthamiana (Ranawaka et al., 2023). Sequence searches using the tomato CHLI protein identified homologue loci in chromosomes 5 (NbLab360C05) and 10 (NbLab360C10), whose expression and gene structures were validated by mapping sequencing reads from RNA samples. Leveraging this curated annotation, vsRNAi were designed to target CDS regions conserved in N. benthamiana and tomato CHLI (Figure S1).
Custom‐synthesized DNA oligonucleotide pairs spanning vsRNAi sequences were inserted into pLX‐TRV2 of the JoinTRV vector system, which is based on tobacco rattle virus (TRV) (Aragonés et al., 2022), by one‐step digestion‐ligation reactions (Figure 1a). JoinTRV derivatives with 32‐, 28‐, 24‐ and 20‐nt inserts targeting CHLI (vCHLI, vCHLI‐28, vCHLI‐24 and vCHLI‐20, respectively) were inoculated to N. benthamiana. After 10 days, upper uninoculated leaves of plants treated with vCHLI, vCHLI‐28 and vCHLI‐24 showed a yellowing phenotype. Fluorometry revealed a significant reduction of chlorophyll levels in vCHLI (x̄ = 0.11), vCHLI‐28 (x̄ = 0.23) and vCHLI‐24 (x̄ = 0.39) compared with controls (x̄ = 1.00; Figure 1a) and a significant positive correlation with CHLI transcript levels measured by RT‐qPCR (Figure S2).
Figure 1.

Virus‐delivered short RNA inserts (vsRNAi) trigger gene silencing. (a) JoinTRV vector assembly and delivery of 32‐, 28‐, 24‐ and 20‐nt vsRNAi targeting the two N. benthamiana CHLI homeologues (vCHLI, vCHLI‐28, vCHLI‐24 and vCHLI‐20). Leaf phenotypes and chlorophyll levels (mean ± SD, n = 3) are shown; CTRL—control. Different letters indicate significant differences (P < 0.05), by one‐way ANOVA and Tukey's honestly significant difference (HSD) test. (b) Transcriptomics of vCHLI and CTRL samples (n = 3) detects a significant downregulation of CHLI expression (FDR <0.05). (c) In vCHLI samples, small RNA sequencing shows predominance of 21‐ and 22‐nt sRNAs mapping to CHLI, and with a read depth peak (mean, n = 3) localized to the vsRNAi‐targeted region. (d) Leaf phenotypes and chlorophyll levels (mean ± SD, n = 3) are shown for vCHLI‐b and vPDS, targeting CHLI and PDS homeologues, respectively; ***, P < 0.001, Student's t‐test. (e) vsRNAi portability to scarlet eggplant (Solanum aethiopicum). Phylogeny and fruits are shown, alongside leaf phenotypes and chlorophyll levels (mean ± SD, n = 3) of ‘Rossa di Rotonda’ plants treated with vCHLI‐DK, a pTRV1 + pTRV2 derivative targeting CHLI; ***, P < 0.001, Student's t‐test.
Silencing phenotypes of vCHLI‐treated plants were robust and equivalent to those obtained by using a VIGS vector including a 300‐nt CHLI cDNA fragment (Figure S3). Transcriptomes of plants inoculated with the unmodified JoinTRV (CTRL) or its vCHLI derivative were analysed. Abundance of over 4000 transcripts was significantly altered in the vCHLI/CTRL comparison (FDR <0.05; Figures 1b and S4; Data S1 and S2). Transcriptome‐wide functional analysis revealed an enrichment of gene ontology terms associated with light responses, and biological processes involved in carbohydrate metabolism, cellulose biosynthesis and cell wall biogenesis (Figure S4c; Data S3–S5), consistent with the anticipated reduction of cell photosynthetic capacity caused by the CHLI downregulation.
Among downregulated transcripts, we identified the CHLI homeologues (NbL05g17570.1, and NbL10g22050.1; log2(FC) ≤ −1.92, FDR <0.05) and confirmed a global expression reduction of their genomic loci beyond the 32‐nt region targeted by vCHLI (Figure 1b; Data S6). Viral amplification of large VIGS inserts can lead to overestimation of expression levels of homologous host genes (Figures S5 and S6); by contrast, vsRNAi enable transcriptome‐wide quantification of target gene silencing.
We next sequenced sRNAs to assess if their production is involved in the vCHLI‐induced phenotypic and transcriptomic changes. In vCHLI samples, vsRNAi triggered host‐derived production of sRNAs (Figure S7) and a marked enrichment of sRNAs mapping to CHLI transcripts (Figure S8; Data S7 and S8). 21‐nt sRNAs were predominant, followed by 22‐nt sRNAs (Figure 1c; Data S9), and their levels showed a significant negative correlation with those of CHLI transcripts (Figure S9). The vsRNAi‐targeted region of CHLI transcript sequences exhibited a localized accumulation of 21‐ and 22‐nt sRNAs in vCHLI samples, which was absent in control samples (Figure 1c; Data S10). We concluded that vsRNAi trigger target gene downregulation through region‐specific enrichment of 21‐ and 22‐nt sRNAs, known end products of Dicer‐like 4 (DCL4) and DCL2, respectively (Rössner et al., 2022).
We used vCHLI‐b and vPDS to target a second 32‐nt region conserved in CHLI homeologues or 32‐nt of the PHYTOENE DESATURASE (PDS) gene pair, respectively. In the upper uninoculated leaves of treated plants, we observed a reduction in green pigmentation and chlorophyll levels (Figure 1d), confirming the broad applicability of vsRNAi for N. benthamiana gene functional characterization.
We next assessed vsRNAi portability to crops. The vCHLI insert is conserved in tomato CHLI (Figure S1); ‘Moneymaker’ seedlings inoculated with vCHLI showed leaf yellowing and a significant reduction of chlorophyll levels compared with control plants (Figure S10a). Scarlet eggplant (Solanum aethiopicum; Figure 1e) is an underutilized solanaceous crop (Gramazio et al., 2016). Searches of de novo assembled transcriptomes and of a genomic assembly (Benoit et al., 2025) confirmed that the 32‐nt insert of vCHLI is conserved in scarlet eggplant CHLI (100% identity) (Figure S10b). Seedlings of the ecotype ‘Rossa di Rotonda’ inoculated with vCHLI showed leaf yellowing and a significant reduction of chlorophyll levels compared with control plants (Figure S10b). Using vCHLI‐DK, a derivative of the pTRV1 plus pTRV2 vector system used for VIGS and genome editing (Ellison et al., 2020), we delivered a CHLI‐targeting 32‐nt vsRNAi and obtained equivalent results (Figures 1e and S10c), highlighting the versatility of our approach.
Overall, our results demonstrate that vsRNAi as short as 24 nt can effectively produce phenotypic alterations. Use of 32‐nt vsRNAi results in robust gene silencing phenotypes, informative transcriptome‐wide changes and target transcript downregulation linked to gene‐specific production of 21‐ and 22‐nt sRNAs.
Viral delivery of artificial micro‐RNAs or trans‐acting small interfering RNAs reduces VIGS insert sizes and off‐targets (Cisneros et al., 2025). vsRNAi offer equivalent specificity and greatly simplifies viral vector engineering by eliminating intermediate cloning steps and precursor elements required to activate sRNA production. Simplified cloning of vsRNAi fragments, which are nearly 10‐fold smaller than those of conventional VIGS and can be synthesized at low cost, may enable high‐throughput functional genomics in model plants and crops that contribute food security and on‐demand alteration of crop traits.
Author contributions
F.P. conceived the study; J.‐A.D. and F.P. obtained funding and computational resources; A.G. and F.P. designed experiments with input from the other authors; A.G., V.A., P.O.‐G. and F.P. performed the experiments; F.J.H. and J.P. provided materials; A.G., S.G. and F.P. analysed and managed data. F.P. wrote the manuscript with input from A.G. and the other authors; all authors revised and approved the final version.
Conflict of interest
The authors declare no competing interests.
Supporting information
Table S1–S6.
Figure S1–S10.
Data S1–S12.
Acknowledgements
This work was supported by RYC2023‐045411‐I, FPU20/05477 and PID2023‐146418OB‐I00 from Ministerio de Ciencia, Innovación y Universidades (Spain) through the Agencia Estatal de Investigación and PROMETEO CIPROM/2022/21 from Generalitat Valenciana. F.P. gratefully acknowledges the grants MiniVi (ELIXIR‐IIB, Cineca, Italy) and BCV‐2023‐1‐0021 (Red Española de Supercomputación, Spain), and resources provided by Centro de Supercomputación de Galicia (CESGA, Spain). We thank C. Mallor Giménez (BGHZ‐CITA, Spain) for S. aethiopicum seeds.
Data availability statement
Supplemental information accompanies this article. The transcriptomic and small RNA sequencing datasets generated are deposited under NCBI BioProject PRJNA1217923. pLX‐TRV2‐vCHLI is available at Addgene (239842, https://www.addgene.org/239842/).
References
- Ahmed, F. , Senthil‐Kumar, M. , Dai, X. , Ramu, V.S. , Lee, S. , Mysore, K.S. and Zhao, P.X. (2020) pssRNAit: a web server for designing effective and specific plant siRNAs with genome‐wide off‐target assessment. Plant Physiol. 184, 65–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Aragonés, V. , Aliaga, F. , Pasin, F. and Daròs, J.A. (2022) Simplifying plant gene silencing and genome editing logistics by a one‐Agrobacterium system for simultaneous delivery of multipartite virus vectors. Biotechnol. J. 17, e2100504. [DOI] [PubMed] [Google Scholar]
- Benoit, M. , Jenike, K.M. , Satterlee, J.W. , Ramakrishnan, S. , Gentile, I. , Hendelman, A. , Passalacqua, M.J. et al. (2025) Solanum pan‐genetics reveals paralogues as contingencies in crop engineering. Nature 640, 135–145. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Berman, A. , Su, N. , Li, Z. , Landau, U. , Chakraborty, J. , Gerbi, N. , Liu, J. et al. (2025) Construction of multi‐targeted CRISPR libraries in tomato to overcome functional redundancy at genome‐scale level. Nat. Commun. 16, 4111. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Cisneros, A.E. , Alarcia, A. , Llorens‐Gámez, J.J. , Puertes, A. , Juárez‐Molina, M. , Primc, A. and Carbonell, A. (2025) Syn‐tasiR‐VIGS: virus‐based targeted RNAi in plants by synthetic trans‐acting small interfering RNAs derived from minimal precursors. Nucleic Acids Res. 53, gkaf183. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ellison, E.E. , Nagalakshmi, U. , Gamo, M.E. , Huang, P.J. , Dinesh‐Kumar, S. and Voytas, D.F. (2020) Multiplexed heritable gene editing using RNA viruses and mobile single guide RNAs. Nat. Plants 6, 620–624. [DOI] [PubMed] [Google Scholar]
- Gramazio, P. , Blanca, J. , Ziarsolo, P. , Herraiz, F.J. , Plazas, M. , Prohens, J. and Vilanova, S. (2016) Transcriptome analysis and molecular marker discovery in Solanum incanum and S. aethiopicum, two close relatives of the common eggplant (Solanum melongena) with interest for breeding. BMC Genomics 17, 300. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Pasin, F. , Uranga, M. , Charudattan, R. and Kwon, C.T. (2024) Engineering good viruses to improve crop performance. Nat. Rev. Bioeng. 2, 532–534. [Google Scholar]
- Ranawaka, B. , An, J. , Lorenc, M.T. , Jung, H. , Sulli, M. , Aprea, G. , Roden, S. et al. (2023) A multi‐omic Nicotiana benthamiana resource for fundamental research and biotechnology. Nat. Plants 9, 1558–1571. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Rössner, C. , Lotz, D. and Becker, A. (2022) VIGS goes viral: how VIGS transforms our understanding of plant science. Annu. Rev. Plant Biol. 73, 703–728. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1–S6.
Figure S1–S10.
Data S1–S12.
Data Availability Statement
Supplemental information accompanies this article. The transcriptomic and small RNA sequencing datasets generated are deposited under NCBI BioProject PRJNA1217923. pLX‐TRV2‐vCHLI is available at Addgene (239842, https://www.addgene.org/239842/).
