A vector-free gene interference system using delaminated Mg–Al-lactate layered double hydroxide nanosheets as molecular carriers to intact plant cells

Background The Mg–Al-lactate layered double hydroxide nanosheet (LDH-NS) has shown great potential as an optimal nanocarrier for extensive use in plants. However, previous studies in plant sciences have not provided a clear description of the application for the LDH-NSs-based double-stranded RNA (dsRNA) delivery (LDH-dsRNA) system in different tissues of both model and non-model species. Results LDH-NSs were synthesized by using the co-precipitation method, while the dsRNAs targeting genes of interest were prepared in vitro using T7 RNA polymerase. The LDH-dsRNA bioconjugates with a neutral charge were produced by incubating with the mass ratio of LDH-NSs to dsRNA at 3:1, which were then introduced into intact plant cells using three different approaches, including injection, spray, and soak. The LDH-dsRNA delivery method was optimized by inhibiting the expression of the Arabidopsis thaliana ACTIN2 gene. As a result, soaking A. thaliana seedlings in a medium containing LDH-dsRNA for 30 min led to the silencing of 80% of the target genes. The stability and effectiveness of the LDH-dsRNA system were further confirmed by the high-efficiency knockdown of plant tissue-specific genes, including that encoding phytoene desaturase (PDS), WUSCHEL (WUS), WUSCHEL-related homeobox 5 (WOX5), and ROOT HAIR DEFECTIVE 6 (RHD6). In addition, the LDH-dsRNA system was employed in cassava, where it was found that the expression of the gene encoding nucleotide-binding site and leucine-rich repeat (NBS-LRR) was significantly reduced. As a result, the resistance of cassava leaves to pathogens was weakened. Noteworthy, the injection of LDH-dsRNA into leaves resulted in a significant downregulation of target genes in both stems and flowers, indicating the successful transport of LDH-dsRNA from leaves to other parts of plants. Conclusions LDH-NSs have proven to be a highly effective molecular tool for delivering dsRNA into intact plant cells, enabling accurate control of target gene expression. Supplementary Information The online version contains supplementary material available at 10.1186/s13007-023-01021-1.


Background
Genetically modified crops have ushered in a new era of agriculture, offering a potential solution to many of the pressing challenges facing humanity today, including global climate change, food scarcity in developing nations, and pollution caused by traditional petroleumbased farming. The process of genetic engineering in crops requires the incorporation of functional nucleotides into plant cells, which cannot penetrate the plasma membrane spontaneously due to their negative charge [2]. Therefore, the delivery of these molecules through the membrane barrier into cells is a crucial step in genetic engineering within living cells.
In contrast to animal cells, plant cells have peripheral cell walls, which mainly consist of cellulose and pectin polysaccharides. This protective barrier acts to prevent foreign pathogens and particles from entering the cell and damaging the plasma membrane. Various methods of delivering biomolecules to plant cells have been developed to overcome the cell wall barrier. Commonly used methods include organism-based methods such as Agrobacterium-mediated delivery [8], chemical methods such as polyethylene glycol-mediated delivery [10,17], and physical methods such as particle bombardment [7]. However, the limitations associated with these methods, including host specificity, low efficiency, potential tissue damage, and high costs have prompted researchers to explore alternative approaches for delivering biomolecules efficiently [1,11,16]. In this regard, nanocarrier-based delivery systems have shown promise for solving these problems by providing a broad-spectrum approach that is non-host specific, low cost, highly efficient, and low in cytotoxicity.
Studies have shown that the intracellular delivery of biomolecules can be achieved through surface spraying, infiltration, or encapsulation using nanoparticles such as gold, starch, mesoporous silica, carbon nanotubes, and LDH-NSs [9,11,24,25]. Nanomaterials have been effectively utilized to deliver various biomolecules into plants, leading to the modification of target gene expression either constitutively or transiently [19]. For instance, single-walled carbon nanotubes (SWNTs) and mesoporous silica nanoparticles (MSNs) have been demonstrated to facilitate the penetration of plasmid DNA into intact plant cells [5,6]. Magnetic nanoparticles (MNPs) were used as nanocarriers to deliver exogenous genes into the seeds of cotton and maize through their pollen tubes [20,27], LDH-NSs have been employed to transport various RNA molecules, including dsRNA [13,22], long noncoding RNA (lncRNA) [15], small interfering RNA (siRNA) [25], and circular RNA (circRNA) [23] into intact cells of diverse plant species. This delivery method has proven effective in regulating the expression of target genes.
Layered double hydroxides (LDHs) belong to a class of pillared clay lamellar materials comprised of positively charged hydroxide layers that are intercalated with anions and water. Lactate intercalated LDHs can be easily delaminated into the water, resulting in LDH-NSs formation. The size of LDH-NSs can be modified within a range of 50-300 nm by using different methods, which affects its efficiency in transporting biomolecules [4,14,21]. Studies showed that non-aqueous precipitation methods were used to prepare Mg 2+ -Al 3+ -based LDH-NSs that successfully delivered dsRNA into intact tobacco leaf cells, providing sustained protection against plant viruses [13]. Aqueous methods were used to synthesize Mg 2+ -Al 3+ -based LDH-NSs that can deliver various types of RNAs into plant cells [15,22,23,26]. Furthermore, Wu et al. demonstrated that the Mg 2+ -Al 3+ -based LDH-NSs prepared through aqueous methods exhibit very low cytotoxicity to plants [21]. These findings suggest that LDH-NSs may be an excellent nanocarrier with versatile applications in plant and agricultural sciences.
In this study, LDH-NSs were synthesized using the chemical co-precipitation method with titration temperatures ranging from 0 to 40 °C. The dsRNAs were prepared in vitro using the T7 RNA polymerase and conjugated with the LDH-NSs. The LDH-dsRNA bioconjugates were then used to modify the expressions of target genes in intact plant cells without transgene integration. Overall, we constructed comprehensive guidelines for a vector-free LDH-dsRNA system that can be used across various tissues in both model and non-model species.

The vector-free LDH-NS-based gene interference method
The method for LDH-dsRNA gene interference presented here involves six sequential steps ( Fig. 1A): (1) Preparation of LDH through co-precipitation; (2) Delamination of LDH particles to form LDH-NSs; (3) Synthesis of dsRNA using T7 RNA polymerase in vitro; (4) Bioconjugation of LDH-NSs and dsRNA at a 3:1 mass ratio; (5) Delivery of the LDH-dsRNA bioconjugates into plant cells via different approaches such as soaking, injection or spraying; (6) Penetration of the plant cell by the LDH-dsRNA bioconjugates, leading to silencing of the target gene within the cell. Please refer to the Methods section for more comprehensive details regarding these steps.

Optimization of LDH-NSs synthesis
The synthesis of LDH-NSs is temperature-dependent, thus changes in temperature can affect its production. To determine the ideal parameters for LDH-NSs synthesis and dsRNA loading, we synthesized LDH-NSs at different titration temperatures. The efficacy of dsRNA loading on LDH-NSs was tested at different dsRNA: LDH-NSs mass ratios and reaction durations. The temperature of the titration reaction was gradually increased from 0 to 40 °C during the co-precipitation of the LDH-NSs. To examine the structures of the resulting LDH-NSs, we used atomic force microscopy (AFM). Image analysis revealed significant differences in vertical distances between the nanosheets and stripping ratios at different titration temperatures (Fig. 1B). The monolayer ratio of LDH-NSs synthesized at 30 °C is 71.94%, which is significantly higher than the ratios obtained at the other five temperatures. The dsRNA NbPDS was loaded onto LDH-NSs synthesized at different temperatures by utilizing varying mass ratios of dsRNA to LDH-NSs. The loading efficiency was determined through agarose gel electrophoresis, where light and dark bands indicated low and high loading ratios, respectively. The dsRNA was completely loaded onto LDH-NSs synthesized at 25 °C with a dsRNA: LDH-NSs mass ratio of 1:7, onto LDH-NSs synthesized at 30 °C with a dsRNA: LDH mass ratio of 1:3, and onto LDH-NSs synthesized at 35 °C with a dsRNA: LDH mass ratio of 1:7 (Fig. 1C). However, at synthesis temperatures of 0 °C, 15 °C, and 40 °C, the dsRNA failed to load completely onto LDH-NSs even at a mass ratio of 1:10 ( Fig. 1C).
To explore the effects of increased reaction duration on dsRNA adsorption onto the LDH-NSs synthesized at various temperatures, we extended the reaction time incrementally from 0 to 60 min, where 0 min was taken as the control. Despite the synthesis temperature, the negatively charged dsRNA was rapidly adsorbed onto the positively charged LDH-NSs. The total percentage of dsRNA adsorbed onto the LDH-NSs reached a steady state roughly 5 min post-incubation of dsRNA and LDH-NSs, and extending the reaction beyond this point did not increase dsRNA binding to LDH-NSs (Fig. 1D). However, the temperature of LDH-NSs synthesis had a significant influence on dsRNA adsorption, the maximum load of dsRNA on LDH-NSs synthesized at 0 °C 35 °C, and 40 °C being 46.91%, 99.44%, and 99.66%, respectively (Fig. 1D). Overall, our optimization experiments revealed that LDH-NSs synthesized at 30 °C achieved a relatively high maximum dsRNA load (94.59%) at a low dsRNA: LDH mass ratio (1:3). Hence, LDH-NSs synthesized at 30 °C were selected for subsequent experimentation.

Optimization of LDH-dsRNA delivery method by silencing of Arabidopsis AtACTIN2
To determine the most effective LDH-dsRNA delivery methods, we performed interference experiments on the Arabidopsis housekeeping gene AtACTIN2. The dsRNA targeting AtACTIN2 (dsRNA AtACTIN2 ) was produced in vitro by using the T7 RNA polymerase system. The LDH-dsRNA AtACTIN2 bioconjugates were obtained by incubating with the mass ratio of LDH-NSs to dsRNA AtACTIN2 at 3:1. Five-day-old Arabidopsis seedlings were soaked in a solution supplemented with LDH-NSs-dsRNA AtACTIN2 for 24 h. Results showed that AtACTIN2 expression levels did not show a significant difference in comparison with the control group after soaking for 0.5 h. The expression levels of AtACTIN2 significantly decreased after soaking for 2-24 h compared to the control group, with a further significant decrease observed between 12 and 24 h. At the end of 24 h, the expression levels of AtACTIN2 were less than 20% of that in the control group ( Fig. 2A).
The seedlings were soaked in the liquid culture medium supplemented with LDH-dsRNA AtACTIN2 for 0.5 and 6 h, followed by transfer to a dsRNA-free solid culture medium over time (Fig. 2B, C). When seedlings were soaked in the dsRNA-supplemented medium for 0.5 h before being cultured on dsRNA-free solid media, AtACTIN2 expression was significantly downregulated compared to the control at 2-12 h. Notably, the expression level of AtACTIN2 at 24 h was significantly higher than that at 2-12 h, while it was significantly lower than that at 0 h ( Fig. 2B). At 6 h of dsRNA exposure, there was a significant downregulation of AtACTIN2 compared to the control. This downregulation persisted until the end of the experiment at 24 h. (Fig. 2C).
To verify the efficacy of gene silencing using different methods, we administered LDH-dsRNA AtACTIN2 to the flowering A. thaliana Col-0 through both spray and injection methods. Regardless of the methodology employed, a significant decrease in AtACTIN2 transcription abundance was observed in the leaves, stems, and flowers when compared to the control (Fig. 2D). These results suggested that dsRNA delivery methods including soak, spray, and injection are viable options for use under different experimental conditions.

Silencing of Arabidopsis AtPDS3, AtWUS, AtWOX5, and AtRHD6
To further validate the effectiveness of the LDH-NSbased dsRNA delivery method in various parts of model plant Arabidopsis, we employed LDH-dsRNA bioconjugates to silence the AtPDS3 gene in the leaves, AtWUS gene in the shoot apical meristems (SAMs), AtWOX5 gene in the root tips, and AtRHD6 gene in the roots, respectively. The expression level of AtPDS3 was significantly downregulated after treatment with both dsRNA AtPDS3 and LDH-dsRNA AtPDS3 compared to double distilled water (ddW) and LDH-NS treatment. Moreover, AtPDS3 gene expression level was lower in LDH-dsRNA AtPDS3 treated plants than those treated with And put in the solid medium for a time gradient. AtACTIN2 transcription levels in five-day-old Arabidopsis seedlings after 0.5-24 h exposure to dsRNA AtACTIN2 in liquid medium. Control seedlings were soaked in media containing double distilled water only (ddW), dsRNA AtACTIN2 only (dsRNA), or LDH-NS only (LDH) for 24 h. B, C AtACTIN2 transcription levels in five-day-old Arabidopsis seedlings soaked in dsRNA AtACTIN2 -supplemented liquid medium for B 0.5 h or C 6 h and then cultured on dsRNA-free solid media for 0-24 h. D The lift picture showed the Arabidopsis that had been injected with LDH-dsRNA AtACTIN2 or had all their tissues sprayed with it. LDH-dsRNA AtACTIN2 . The tissue-specific transcription level of AtACTIN2 of mature Arabidopsis (leaf, stem, and flower) spray, injection for 24 h. The letters on the bars represent p < 0.05, which was calculated using Duncan's multiple range test to indicate statistically significant differences among the various groups dsRNA AtPDS3 (Fig. 3A). These results were consistent with the quantitative analysis of PDS protein content (Fig. 3B). As shown in Fig. 3C, AtWUS transcription level in the SAM of 5-day-old seedlings treated with dsRNA AtWUS and LDH-dsRNA AtWUS exhibited a significant decrease compared to the ddW treatment group. Additionally, AtWUS expression was significantly downregulated after LDH-dsRNA AtWUS treatment in comparison to The letters on the bars represent p < 0.05, which was calculated using Duncan's multiple range test to indicate statistically significant differences among the various groups. nd, not detected dsRNA AtWUS treatment (Fig. 3C). Differences in hypocotyl length were consistent with AtWUS gene expression patterns (Fig. 3D). As shown in Fig. 3E, the transcription levels of AtWOX5 were significantly reduced in both dsRNA AtWOX5 and LDH-dsRNA AtWOX5 treatment groups compared to the ddW treatment group. AtWOX5 transcription level in the LDH-dsRNA AtWOX5 group was significantly lower than that in the dsRNA AtWOX5 group. Compared with the ddW treatment group, main root growth was significantly inhibited in both dsRNA AtWOX5 and LDH-dsRNA AtWOX5 treatment groups. The main root in the LDH-NSs group was significantly longer than that in the ddW group, while it was significantly shorter in the LDH-dsRNA AtWOX5 group compared to the dsRNA AtWOX5 group (Fig. 3F). Additionally, the GFP fluorescence intensity was significantly lower in the root tips of the AtWOX5: GFP transgenic line treated with LDH-dsRNA AtWOX5 compared to those treated with ddW, LDH-NS, and dsRNA AtWOX5 (Additional file 1: Fig.  S2A and B). As shown in Fig. 3G, AtRHD6 transcription levels were significantly decreased in 3-day-old seedlings treated with dsRNA AtRHD6 and LDH-dsRNA AtRHD6 compared to the ddW treatment group. The seedlings treated with LDH-dsRNA AtRHD6 had significantly fewer and shorter root hairs than those in all other treatment groups ( Fig. 3H and I), indicating successful inhibition of root hair initiation and elongation. Overall, these results verified that LDH-NSs are capable of effectively delivering dsRNA to different parts of Arabidopsis, leading to a characteristic phenotypic alteration.

Silencing of tobacco NbPDS, NbWUS, and NbWOX5
To investigate the potential of LDH-NSs to incorporate dsRNAs targeting genes in other plant species, we employed LDH-dsRNA bioconjugates to silence NbPDS, NbWUS, and NbWOX5 genes in tobacco. As shown in Fig. 4A, NbPDS transcription levels were significantly decreased in both the dsRNA NbPDS and LDH-dsRNA N-bPDS treatment groups compared to all other treatment groups. Additionally, the expression levels of NbPDS were significantly lower in the LDH-dsRNA NbPDS treatment group than that in the dsRNA NbPDS treatment group. These results were consistent with the quantitative analysis of PDS protein content (Fig. 4B). As shown in Fig. 4C and D, the expression levels of NbWUS in the SAM region of tobacco were notably reduced and resulted in a significant decrease in hypocotyl length within the dsRNA NbWUS and LDH-dsRNA NbWUS treatment groups compared to all other treatment groups. In comparison to the dsRNA NbWUS treatment group, the LDH-dsRNA NbWUS treatment group exhibited a significant reduction in both NbWUS transcription levels and hypocotyl length (Fig. 4C and D). As shown in Fig. 4E, the transcription levels of NbWOX5 in tobacco root tips showed a significant decrease in both dsRNA NbWOX5 and LDH-dsRNA NbWOX5 groups compared to ddW and LDH-NSs groups. Moreover, NbWOX5 exhibited the lowest expression levels in the LDH-dsRNA NbWOX5 treatment groups among all four treatment groups. Notably, the LDH-dsRNA NbWUS group showed significant inhibition in the growth of the main root compared to the dsRNA NbWUS group. (Fig. 4F). Overall, these findings suggest that LDH-NSs were able to effectively deliver dsRNA into different parts of tobacco.

Silencing of cassava MeLRRs
Previous research has confirmed that the gene encoding cassava leucine-rich repeat (MeLRR) is responsible for conferring cassava resistance to cassava bacterial blight (Xanthomonas axonopodis pv. manihotis, Xam) [26]. To investigate the effectiveness of the LDH-dsRNA system in combating cassava bacterial blight, cassava leaves were infiltrated with both LDH-dsRNA MeLRRs and LDH-NS. As shown in Fig. 5A, the transcription levels of the MeLRR gene in cassava leaves showed a significant decrease following treatment with LDH-dsRNA MeLRRs compared to treatment with LDH-NS. Of these, the expression change of MeLRR1 gene is the largest, while that of MeLRR4 gene is the smallest. The expression change of MeLRR3 gene is smaller than that of MeLRR1 gene, but larger than that of MeLRR2 gene. As shown in Fig. 5B, the cassava leaves that were infiltrated with LDH-dsRNA MeLRRs had significantly higher bacterial abundance than those that were infiltrated with LDH-NS. The larger disease spots indicate weaker disease resistance. Thus, the silencing of the MeLRRs reduces the resistance of cassava leaves to Xam (Fig. 5C).

Discussion
Plant cell walls hinder the efficiency and broad-spectrum applicability of conventional gene delivery methods. New biomolecular delivery technologies for plants are therefore urgently needed. In previous studies, we demonstrated that MgAl-LDH-NSs could be used to deliver single-stranded DNA (ssDNA) and genomic DNA into plants [3,4,21]. Mg-Al-LDH-NSs are cationic nanoparticles that deliver negatively charged dsRNA into the cell via electrostatic interactions, subsequently resulting in gene silencing. We also previously showed that the temperature of the titration reaction during LDH-NSs coprecipitation determined the capacity of the LDH-NSs to adsorb DNA, as well as its zeta potential and particle size [21].
In this study, we first aimed to identify the optimal titration temperature for LDH-NSs co-precipitation. We found that Mg-Al-LDH-NSs synthesized at a titration temperature of 30 °C were optimal for dsRNA adsorption. In addition, AFM images showed that LDH-NSs synthesized at 30 °C had the highest specific surface area which allows more biomacromolecules can be adsorbed and delivered.
Using a housekeeping gene (ACTIN2) in the model plant Arabidopsis, we aimed to optimize LDH-dsRNA treatment duration and conditions. We found that AtACTIN2 was downregulated after 2 h of exposure to the LDH-dsRNA complex and that exposure to the LDH-dsRNA complex for as little as 0.5 h was sufficient to interfere with target gene expression silencing efficiency of about 80%, with effects first noticeable 2 h after exposure. Importantly, target gene expression in the seedlings soaked in the LDH-dsRNA solution for 0.5 h increased after 24 h of growth on an unmodified medium. This suggested that the LDH-dsRNA complex might potentially be useful for transient gene silencing in plant cells, without altering the genetic material of the cell. In mature Arabidopsis plants, target gene expression was reduced significantly compared to controls whether LDH-dsRNA was applied via spray or The letters on the bars represent p < 0.05, which was calculated using Duncan's multiple range test to indicate statistically significant differences among the various groups injection. However, the injection was significantly more effective than spraying.
Because housekeeping gene expression is uniform, stable, and relatively high across all plant tissues, we also tested the efficacy of LDH-NS-mediated dsRNA delivery against several tissue-specific genes in the model plants Arabidopsis and tobacco: PDS, WUS, RHD6, and WOX5, which are particularly strongly expressed in the leaves, SAMs, roots, and root apical meristems, respectively. However, because we did not restrict our gene expression analyses to the tissues specifically expressing each gene, the apparent efficiency of this method seemed lower for the tissue-specific genes than for the housekeeping gene. The study found that treatment with LDH-NSs led to significant suppression of AtWOX5 expression, suggesting that LDH synthesized with metallic elements may have an impact on the development of root tips.
To test whether vector-free LDH-NSs-mediated dsRNA delivery might be useful in agricultural applications, we attempted to silence genes in cassava, an economically important plant. Injection of LDH-dsRNAs targeting MeLRRs significantly reduced the resistance of the cassava-to-cassava bacterial blight (Xam), suggesting that LDH-NS-mediated gene silencing might have wide application prospects in the agricultural industry. The vector-free LDH-NSs-mediated dsRNA delivery method successfully significantly repressed the expression of all of these genes. Our results showed that the LDH-NSs rapidly adsorbed dsRNA and that the dsRNA was efficiently internalized by the mature walled cells of both model plants (Arabidopsis and tobacco) and non-model plants (the economically important cassava).
Sustainable agricultural production practices are facing urgent problems including how to mitigate the reliance on the use of agrochemicals, meanwhile enhancing the quality and yields of agri-foods. The emergence of nanotechnology has shown promise for solving these problems. LDH-NSs-based transient gene  France). The letters on the bars represent p < 0.05, which was calculated using a Student's t-test to indicate statistically significant differences among the various groups silencing system has the advantages of being biodegradable, non-host specificity, non-genome integration, and more simplified operation than conventional transgenic techniques. These advantages demonstrate the great potential of this system for application in agricultural practices, such as pest and disease control by regulating gene expression, while at the same time promoting the quality and yields of agri-foods by loading bioactive ingredients.

Conclusion
In conclusion, compared to traditional gene-silencing techniques targeting plant cells, such as Agrobacteriummethod transformation and particle bombardment, gene silencing using LDH-NSs as dsRNA nanocarriers has several unique advantages: (i) LDH-NSs mediated gene silencing is vector free with high interfering efficiency up to 80%. The silencing of the target gene(s) can be transient, without causing permanent genetic modification and long-term effects. (ii) LDH-NSs mediated gene silencing is broad spectrum. This method was successful in several kinds of (Arabidopsis, tobacco, and cassava) plant cells tested herein, irrespective of plant age (i.e. seedling or mature), tissue, or species. However, the success of this technique in woody plants and plants with waxy cuticles remains to be confirmed. (iii) LDH-NSsmediated gene silencing is simple, inexpensive, and timesaving. LDH-NSs can be produced in any laboratory without special equipment, while dsRNAs are readily synthesized using the widely available T7 RNA-polymerase system. Therefore, we believe that LDH-NSs-mediated dsRNA delivery will be of great use to researchers in the plant and agricultural sciences.

Preparation of the LDH-NSs
Delaminated LDH-NSs were chemically synthesized from Mg(C 3 H 5 O 3 ) 2 ·3H 2 O, Al(C 3 H 5 O 3 ), and C 3 H 6 O 3 as previously described method [18]. To optimize the preparation of LDH-NSs, the slurry of wet LDH-NSs (with a pH of 10) was stirred while maintaining a constant N 2 atmosphere. The LDH-NSs titration procedure was conducted at various temperatures using the co-precipitation method, including 0 °C, 15 °C, 25 °C, 30 °C, 35 °C, and 40 °C. To avoid the adsorption of CO 2 by reaction products, N 2 shall be continuously injected. The slurry was then centrifuged (5000 rpm 10 min), and the precipitate was rinsed with decarbonated water three times. The dried precipitate was transferred to decarbonated water in a concentration of 1 g L −1 and stirred magnetically at 350 rpm.

Analysis of zeta potential
The zeta potential analysis of delaminated LDH-NSs was conducted using a Malvern 2000 zeta potential analyzer (Malvern Instruments, Malvern, UK) following the previously described methods [3,18].

AFM imaging
The Bruker Dimension AFM system (version 1.80R1, Bruker Company, Billerica, USA) was used for the analysis of AFM images following the previously described methods [3,18].

Optimization of LDH-mediated dsRNA AtACTIN2 delivery system
The LDH-dsRNA targeting AtACTIN2 was co-cultured with 5-day-old Arabidopsis in tissue culture plates for 0.5 and 6 h of soaking. After being soaked in a solution containing LDH-dsRNA AtACTIN2 , Arabidopsis seedlings were transferred to a solid tissue culture plate and kept for 0, 2, 4, 6, 12, and 24 h. The dsRNAs (final concentration 15 mg L −1 ) were incubated with the LDH-NSs (final concentration 45 mg L −1 ) at room temperature for 5 min prior to the experiment. The transcription levels of AtACTIN2 were measured using qRT-PCR after 0.5, 2, 4, 6, 12, and 24 h of soaking.

Comparison of different methods for introducing LDH-dsRNA AtACTIN2 to Arabidopsis
To evaluate the effectiveness of LDH-dsRNA AtACTIN2 in the different parts of Arabidopsis, we applied it to Arabidopsis plants using both spray and injection methods. After 12 h, we measured the transcription levels of AtAC-TIN2 in various tissues using qRT-PCR.

Analyzing the effectiveness of the LDH-dsRNA in A. thaliana
Arabidopsis seeds (Col-0 and AtWOX5: GFP) were surface sterilized and cultured in 6-well tissue culture plates containing ddW, LDH-NS, dsRNA AtRHD6/ AtWOX5/ AtWUX or LDH-dsRNA AtRHD6/ AtWOX5/AtWUX for three days, respectively. The culture solution was replaced every 24 h. Root hairs were examined using an optical microscope (Nikon, Japan). Hypocotyl length, main root length, and root hair length were measured from images of each sample using ImageJ software (version 1.53q). The fluorescence of GFP was captured by confocal microscopy at Ex = 488 nm and Em = 520 nm (SP8, Leica, Germany). The average fluorescence intensity values of GFP in the root tips were measured using ImageJ (version 1.53q). In addition, the injection method was employed to silence AtPDS3 in the leaves of 30-day-old A. thaliana.

Analyzing the effectiveness of the LDH-dsRNA in N. benthamiana
Tobacco seeds were surface-sterilized and then cultured in 6-well tissue culture plates containing ddW, LDH-NS, dsRNA NbPDS/NbWUS/ NbWOX5, and LDH-dsRNA NbPDS/NbWUS/ NbWOX5 for 3 days, respectively. The culture solution was replaced every 24 h. The expression levels of target genes were measured by using qRT-PCR. The contents of PDS proteins were measured using enzyme-linked immunosorbent assay (ELISA) kits (Jiangsu Meimian Industrial Co., China). In addition, the injection method was employed to silence NbPDS in the leaves of 30-day-old N. benthamiana.

Analyzing the effectiveness of the LDH-dsRNA in cassava leaves
The South China 124 (SC124) variety of cassava (M. esculenta) was cultivated in a greenhouse under a 16 h light (28 °C) and 8 h dark (22 °C) cycle. LDH-dsRNA MeLRR bioconjugates were obtained by incubating 90 mg L −1 LDH-NS with 30 mg L −1 dsRNA. Leaves of 4-week-old cassava plants were subjected to syringe infiltration with a solution containing LDH-dsRNA MeLRR1 , LDH-dsR-NA MeLRR2 , LDH -dsRNA MeLRR3 , and LDH-dsRNA MeLRR4 , respectively. To serve as a negative control, leaves were treated with a solution containing 90 mg L −1 LDH-NS using syringe infiltration. The infiltrated plants were then transferred to a greenhouse. The expression level of MeLRR genes was measured after 3 days post-infiltration. The leaves were then syringe-infiltrated with the pathogenic bacterium Xam at a concentration of 4 × 10 6 CFU (colony-forming units) mL −1 . The numbers of Xam colonies were measured using the colony counting method at 0 and 1 days post-Xam infiltration [26]. Cassava leaves infected with Xam were treated with Coomassie brilliant blue stain and subsequently visualized using a Fusion FX7-826 apparatus (Vilber Lourmat, France) [26].

Gene expression analysis
Total RNA was extracted from plant samples using the RNAprep Pure Plant Kit (Tiangen, China). The total RNA was reversely transcribed into cDNA using the Fast-Quant RT Kit with gDNase (Tiangen, China). Quantitative real-time reverse-transcription PCR (qRT-PCR) was performed on an Applied Biosystems QuantStudio 6 Flex Real-Time PCR System using the UltraSYBR mixture (Beijing CoWin Biotech., China) and gene-specific primers (Additional file 1: Table S3). The relative expression levels were determined using the previously described 2 -ΔΔCt method [12].