Artificial miRNA-based silencing of begomovirus gene confers resistance to cotton leaf curl disease in cotton

Cotton leaf curl disease (CLCuD), caused by a complex of begomoviruses and their associated betasatellites, severely reduces cotton productivity. Conventional breeding approaches have shown limited success due to the rapid evolution of viral populations and the involvement of the long duration breeding cycles. To address this challenge, the present study explores an artificial microRNA (amiRNA)-based strategy to confer durable resistance against CLCuD in Gossypium hirsutum cv. HS6. A computationally predicted endogenous miR2949a was engineered to generate the artificial miRNA (amiR2949a), which was designed and computationally validated in silico to simultaneously target key genes (C1, C3, C4, V1, and V2) of Cotton leaf curl Multan virus (CLCuMuV, a begomovirus), as well as the β C1 gene of the betasatellites collectively associated with CLCuD. The engineered pre-amiR2949a construct was transformed into cotton via Agrobacterium tumefaciens -mediated transformation. Putative transgenic plants were subsequently inoculated with viruliferous whiteflies, and resistance was assessed through molecular approaches and phenotypic evaluation. The transgenic lines exhibited strong resistance to CLCuD showing 71-81% disease suppression, and a complete absence of disease symptoms and virus accumulation up to 60 days post-inoculation (dpi). Quantitative expression analysis of CLCuMuV, although limited to the C1 gene, revealed a significant reduction (69%-90%) in its transcript levels. This reduction was accompanied by an inverse correlation between amiR2949a accumulation and C1 expression, indicating efficient amiRNA-mediated silencing of viral gene. These findings demonstrate that targeted suppression of the viral C1 gene effectively restricts CLCuMuV accumulation and symptom development in cotton. Furthermore, the study establishes amiRNA-mediated RNA silencing as a robust and scalable strategy for engineering resistance against the cotton-infecting begomovirus complex, with potential applicability to other economically important crops.

RNA interference (RNAi) is a conserved phenomenon in eukaryotes that regulates gene expression, and acts as a primary antiviral defense response against invading pathogens via small RNA (sRNA) molecules (1). The two major classes of sRNAs i.e. small interfering RNAs (siRNAs) and microRNAs (miRNAs), mediate gene silencing at transcriptional and post-transcriptional levels based on their sequence complementarity with target transcripts. siRNAs are generated following cleavage of double-stranded RNA precursors by the Dicer-like enzymes (DCLs), whereas miRNAs are processed from endogenous MIR gene transcripts through a highly regulated biogenesis pathway involving DCL1 and AGO1, culminating in sequence-specific gene silencing (2, 3).

Cotton (Gossypium hirsutum L.) is a key fiber crop of global agricultural importance. However, its productivity is severely compromised by cotton leaf curl disease (CLCuD), a destructive disease induced by begomoviruses associated with satellite components (4–6). Among these, Cotton leaf curl Multan virus (CLCuMuV), in association with a betasatellite, predominates across the Indian subcontinent.

CLCuMuV possesses a monopartite, circular, and single-stranded DNA genome of ~2.7 kb. The V1 (coat protein) and V2 (pre-coat protein) are encoded by the virion-sense strand of the genome, while the complementary-sense strand is responsible for encoding proteins which participate in replication (C1/Rep, C3/REn), transcriptional regulation (C2/TrAP), and host defense suppression (C4) (7, 8). The betasatellite markedly enhances disease severity through the expression of βC1 protein, a key pathogenicity determinant that interferes with host RNAi pathway and promotes viral accumulation (9–11). The begomovirus-satellite interaction accelerates virus evolution and complicates the development of durable host resistance.

Breeding for CLCuD resistance has shown limited success due to scarce resistance source and the frequent emergence of resistance-breaking viral strains (12, 13). RNAi-based genetic approaches have therefore been explored as promising alternatives to achieve durable resistance against viruses. However, siRNA-mediated strategies often generate heterogeneous siRNA populations that can cause off-target silencing and raise biosafety concerns (14).

Although, artificial miRNA (amiRNA) technology offers a precise and efficient alternative for targeted gene silencing, its application to simultaneously target multiple begomoviral genes and associated betasatellites in cotton remains insignificantly explored. Further, studies demonstrating amiRNA-mediated suppression of critical virus replication genes in transformed cotton are limited.

In the present study, we attempt to address this limitation by engineering transgenic cotton plants expressing a single rationally designed amiRNA directed against the highly conserved C1 gene of CLCuMuV. In addition to experimentally validating suppression of C1, in silico analyses predicted that the engineered amiRNA could potentially target multiple begomoviral genes (C1, C3, C4, V1, and V2) as well as the βC1 gene of the associated betasatellite. Using this strategy, we demonstrate that amiRNA-mediated silencing effectively reduces CLCuMuV accumulation and confers resistance to CLCuD in transgenic cotton (G. hirsutum cv. HS6). These findings highlight the potential of amiRNA-based antiviral engineering as a precise, scalable, and potentially broad-spectrum approach for developing durable virus resistance in cotton.

Designing and construction of amiRNA expression vector

A cotton encoded microRNA, ghr-miR2949a, previously predicted to target overlapping regions of the C1 and C4 genes of Cotton leaf curl Burewala virus (CLCuBuV) (15), was selected as the amiRNA backbone (Supplementary Table 2). To enhance complementarity with conserved regions of viral target genes, mismatched base pairs between the endogenous mature miRNA and the viral targets were manually modified to create an efficient amiRNA. Viral target regions were selected based on sequence conservation and predicted target accessibility to maximize silencing efficiency. The engineered amiRNA demonstrated the ability to in silico target multiple genes from various cotton-infecting begomoviruses, as well as the βC1 gene of associated betasatellites (Supplementary Table 3). In the G. hirsutum pre-miR2949a backbone, the native mature miRNA duplex was manually changed by replacing with the modified amiR2949a-5p (5’-CCUUGUGUACUGGAUGAGCGCG-3’) and amiR2949a-3p (5’-UGUGCUCAUCCAGCACACAAGGUA-3’) sequences (Supplementary Figure 1). The modified pre-amiR2949a construct was synthesized commercially (GenScript, USA), and in silico analysis confirmed the absence of off-targets within the G. hirsutum genome. The secondary structure of the pre-amiRNA was predicted by the MFOLD program (http://www.unafold.org/mfold/applications/rna-folding-form.php).

The pre-amiR2949a (~335 bp) fragment was cloned into the pUC57 cloning vector and subsequently ligated into the pBI121 vector, downstream of the CaMV 35S promoter, generating pBI121-pre-amiR2949a, which was transformed into Agrobacterium tumefaciens strain LBA4404 (16).

Transformation of G. hirsutum cultivar HS6 and plant regeneration

The genetic transformation of G. hirsutum was carried out using an Agrobacterium -mediated method employing shoot tip explants (16). Following co-cultivation, explants were placed on MS selection medium, regenerated plantlets were rooted, acclimatized in Hoagland solution and, hardened under controlled greenhouse conditions, before transfer to soil.

Molecular confirmation and segregation analysis

Genomic DNA was extracted from the T 0, T 1, T 2, and NT control plants employing DNeasy Plant Mini Kit (QIAGEN, Germany). PCR-based detection of the npt II (neomycin phosphotransferase II) transgene was performed using gene-specific primers (Supplementary Table 1) (17).

Two independent transgenic lines (T 0.18 and T 0.24) were self-pollinated to obtain T 1 progeny followed by T 2 generation development. Segregation of the transgene was assessed by PCR-based screening and inheritance patterns were evaluated using the χ² test assuming the 3:1 Mendelian ratio (p ≤ 0.05).

Virus inoculation and symptom assessment

Resistance of transgenic plants to CLCuMuV was evaluated following viruliferous whiteflies (B. tabaci)-mediated inoculation (18). Plants were maintained under greenhouse conditions, and symptom development was recorded at 15-days intervals up to 60 days post-inoculation (dpi). Disease severity was assessed using a 0–4 scoring scale, and resistance percentage was calculated using the formula: Resistance (%) = [(total number of inoculated plants − number of symptomatic plants)/total number of inoculated plants] × 100. The assessment was carried out as described previously (13).

Virus detection

Genomic DNA from the inoculated plants and NT controls was used as the template for PCR-mediated CLCuMuV detection using primers specific for the coat protein (CP) gene (Supplementary Table 1) (19).

RNA isolation, cDNA synthesis and expression analysis

Total RNA was isolated from leaf tissues of transgenic (T 2) and NT control plants (20). cDNA synthesis and polyadenylation were performed employing the Mir-X™ miRNA First-Strand Synthesis Kit (Takara Bio, USA), following a 1 hr incubation at 37 °C. qRT-PCR was performed using the TB Green ® Premix Ex Taq™ II (Takara Bio, USA) on a LightCycler ® 96 System (Roche, Switzerland).

Expression of amiR2949a was quantified employing an amiR2949a-specific forward primer and the mRQ 3′ reverse primer, with the U6 SnRNA as the internal reference (Supplementary Table 1) (21). C1 gene expression was analysed using C1 gene-specific primers, normalized to GAPDH (Supplementary Table 1) (22), and relative gene expression was calculated using the delta-delta Ct method (23). All reactions were carried out in three biological and three technical replicates.

Statistical analysis

The data were statistically analysed using one way analysis of variance (ANOVA), followed by Tukey’s test (p ≤ 0.05). Results are presented as mean ± standard deviation (SD), and statistical analyses were performed using OriginPro software (OriginLab, USA; https://www.originlab.com/OriginProLearning.aspx).

Construction of amiRNA expression vector

An amiR2949a was rationally designed to simultaneously target key genes (C1, C3, C4, V1 and V 2) of cotton-infecting begomoviruses, and the βC1 gene of associated betasatellites (Figures 1A, B). In silico analysis revealed the highest sequence complementarity (77-82%) with the C1 and C4 genes across begomoviruses, including CLCuMuV, CLCuBuV, Cotton leaf curl Rajasthan virus (CLCuRV), Cotton leaf curl Kokhran virus (CLCuKoV) and Cotton leaf curl Bangalore virus (CLCuBV) with moderate complementarity (69-73%) to V2, C3, and βC1 genes.

The mature amiR2949a duplex was manually replaced with the modified amiR2949a-5p and amiR2949a-3p duplex sequences within the pre-amiR2949a backbone. Secondary structure prediction confirmed correct folding of the precursor (Supplementary Figure 2). The synthetic pre-amiR2949a was cloned in pBI121 (pBI121-pre-amiR2949a) plasmid and subsequently transformed into A. tumefaciens strain LBA4404 (Figures 1C, D).

Transformation, transgene detection and inheritance analysis

Genetic transformation (Figures 2A–H) yielded six primary transformants (T 0) of G. hirsutum, of which five were PCR positive for the amiRNA transgene (~700 bp; Figure 2I). Two asymptomatic lines (T 0.18 and T 0.24), which showed no detectable virus accumulation at 60 dpi, were advanced for further analysis (Figure 2J).

Stable inheritance of the transgene was confirmed in T 1 (19/25) and T 2 (28/38) progeny (Supplementary Figures 3, 4). Segregation followed a Mendelian 3:1 ratio in both generations (Table 1) indicating predicted single-locus integration (24).

Evaluation of resistance to CLCuD

Non-transgenic plants developed typical CLCuD symptoms following inoculation with viruliferous whiteflies. However, most transgenic (T 1 and T 2) plants remained asymptomatic throughout the 60-days observation period (Table 2; Supplementary Figure 5).

PCR analysis confirmed the absence of CLCuMuV in asymptomatic transgenic plants, whereas mildly symptomatic transgenics and all the NT control showed virus accumulation (Figure 3; Supplementary Figure 6). Notably, viral DNA levels in symptomatic transgenic plants were significantly lower than those in the NT controls. On the basis of symptom severity scoring, transgenic lines displayed overall resistance levels of 75-81% (T 1) and 71-78% (T 2).

Overexpression of amiRNA suppress C1 expression

The qRT-PCR analysis revealed variable expression levels of amiR2949a among the tested transgenic (T 2) lines, with the highest expression detected in T 2.18.4.19 (Figure 4A). Correspondingly, C1 transcript levels were markedly reduced in all transgenic lines compared to NT controls, with maximal suppression observed in T 2.18.4.19 (Figure 4B).

A strong inverse correlation between amiR2949a and C1 expression indicates effective amiRNA-mediated silencing of the viral replication gene, thus contributing to reduced virus accumulation and enhanced CLCuD resistance.

Phenotypic assessment

Transgenic lines displayed no detectable developmental abnormalities through the T 2 generation (Figures 4C–E, 5A). A significant increase in plant height (16%-24%) was observed in transgenic T 2 lines relative to NT controls (Figure 5B). Flowering and boll development remained normal in all transgenic plants, even though virus-infected NT plants exhibited reduced boll size (Supplementary Figure 7).

RNAi operates as a homology-dependent, sequence specific pathway that modulates gene silencing, and constitutes a major antiviral defense pathway in plants. Among different sRNA classes, miRNAs are central to plant immunity by regulating host defense genes, and in some cases directly targeting viral transcripts (25). The efficacy of RNAi-mediated resistance is often influenced by the degree of complementarity between sRNA and their targets, as well as by the functional relevance and positional context of the targeted viral gene (26–28).

In this investigation, an amiRNA-mediated approach was implemented to combat CLCuD infection. Transgenic G. hirsutum cv. HS6 plants were engineered to express a single amiRNA designed and in silico predicted to potentially target multiple essential viral genes of the begomovirus and the βC1 gene of associated betasatellites. Unlike conventional siRNA approaches, which often generate heterogenous siRNA populations with variable efficacy and a risk of off-target effects (26, 29), amiRNAs are processed from endogenous miRNA precursors to produce a single, precisely designed 21-nucleotide RNA species (22). This feature ensures highly specific gene silencing while minimizing unintended host gene targeting and recombination risks. Moreover, amiRNA precursors lack extensive double-stranded RNA regions, thus reducing the chances of transgene silencing and enhancing stability of expression (14, 22).

In this study, the developed cotton lines exhibited a high level of resistance to CLCuMuV infection, remained asymptomatic and largely free of detectable virus accumulation up to eight weeks after inoculation. A shortcoming of this study is that T-DNA integration and transgene copy number were not directly validated through Southern blot. Instead, PCR-based segregation analysis of the transgene was performed in T 1 and T 2 progeny. The segregation patterns observed in all independent transgenic lines were consistent with a predicted single-locus insertion. Although this approach is commonly used to infer single-locus integration events, it cannot definitively exclude tandem or closely linked multiple T-DNA insertions (24, 30–32). Furthermore, stable Mendelian inheritance of the amiRNA transgene in the T 1 and T 2 generations supports the genetic stability of this resistance trait, consistent with earlier reports employing RNAi-based antiviral strategies in cotton and other crops (19, 33).

To elucidate the basis of CLCuD resistance, we focused on the viral C1 gene (motif II) encoding the replication associated protein (Rep), and exhibits high sequence complementarity with amiR2949a across multiple begomovirus species. The C1 protein, being highly conserved and multifunctional, is essential for rolling-circle replication of the viral genome (7). Transgenic T 2 plants expressing amiR2949a exhibited a significant reduction in C1 transcript levels, accompanied by a strong inverse correlation between amiRNA presence and C1 expression. These findings provide indirect but compelling evidence of amiRNA-mediated silencing of the viral replication gene and offer a convincing explanation for the observed reduction in virus accumulation and disease severity.

Although amiR2949a was rationally designed and predicted in silico to target multiple viral genes, transcriptional validation in this study was confined to the C1 gene. Moreover, this study was primarily intended as a proof-of-concept demonstration of amiR2949a efficacy rather than an exhaustive analysis of all predicted targets. Expanding expression analyses to all predicted targets (C3, C4, V1, V2, and βC1) would require additional experiments that were beyond the scope and available resources for this initial report. To still support the intended multi-target design, the authors have included in silico amiRNA-target pairing analyses showing the predicted binding interactions with C3, C4, V1, V2 and βC1, which are consistent with the proposed multi-target suppression mechanism. However, the significant reduction in viral accumulation and symptom severity supports its potential for multi-gene silencing.

Additionally, experimental validation of amiRNA-guided cleavage was not performed and remains a limitation of the present study. The targeting of C1 and the overlapping C4 coding region contributes to the broad-spectrum efficacy of this amiRNA, as C1 gene is functionally indispensable and conserved among the begomovirus species. Furthermore, several previous studies have inferred amiRNA specificity indirectly through reductions in target transcript levels and the development of resistance phenotypes, without performing cleavage site analyses (30, 34–39).

While some studies have reported inconsistent correlation between sRNA accumulation and virus resistance (30, 31), our data demonstrate a clear and consistent relationship between amiRNA expression, suppression of an essential viral gene, and enhanced resistance (24, 39–41). Notably, resistant transgenic plants did not exhibit abnormal phenotypic traits in the absence of virus infection, indicating no obvious developmental defects attributable to amiR2949a expression under greenhouse conditions. Following virus inoculation, transgenic lines exhibited significant (16–24%) increase in plant height relative to infected non-transgenic controls, consistent with effective suppression of begomovirus-induced stunting. Together, these observations support the interpretation that improved growth reflects reduced viral accumulation rather than pleiotropic effects of the amiRNA. Importantly, transgenic plants exhibited normal phenotypic development, increased plant height and normal boll development compared to infected non-transgenic controls. This suggests that amiR2949a does not adversely affect endogenous cotton gene expression, and that reduced viral titer contributes to improved growth of the plants. The absence of visible pleiotropic effects underscores the specificity and biosafety of the amiRNA-based approach. Collectively, these findings suggests that an early and efficient suppression of the C1 gene restricts CLCuMuV replication, thereby limiting CLCuD progression. While greenhouse results are promising, multi-location field trials under natural infection pressure are essential to assess the durability, stability and agronomic performance of amiR2949a-expressing cotton lines. Such trials are critical to assess durability and agronomic performance under diverse environmental and epidemiological conditions. However, several challenges limit the immediate implementation of field evaluations, including variable environmental conditions, the presence of diverse and recombinant begomovirus strains and whitefly vector populations, and the occurrence of mixed infections. Further, regulatory approvals, the availability of containment-compliant field sites, and logistical and funding constraints further hinder multi-location trials. Therefore, comprehensive field evaluation remains a prioritized next step for validating the greenhouse proof-of-concept results.

In conclusion, this study establishes that an amiRNA targeting the C1 gene of CLCuMuV confers effective resistance to CLCuD in cotton. Expression of amiR2949a led to a significant suppression of viral C1 transcripts, reduced CLCuMuV accumulation, and improved CLCuD resistance without causing any detectable adverse effect on plant growth or development. The observed inverse relationship between amiRNA expression and viral accumulation signifies the mechanistic basis of the resistance phenotype. Overall, this study establishes amiRNA-mediated silencing as a precise, durable and scalable approach for engineering broad-spectrum virus resistance in cotton. Furthermore, it provides a strong foundation for the future integration of amiRNA-based antiviral approaches into cotton improvement and breeding programs.