Spray‐induced gene silencing for disease control is dependent on the efficiency of pathogen RNA uptake

Summary Recent discoveries show that fungi can take up environmental RNA, which can then silence fungal genes through environmental RNA interference. This discovery prompted the development of Spray‐Induced Gene Silencing (SIGS) for plant disease management. In this study, we aimed to determine the efficacy of SIGS across a variety of eukaryotic microbes. We first examined the efficiency of RNA uptake in multiple pathogenic and non‐pathogenic fungi, and an oomycete pathogen. We observed efficient double‐stranded RNA (dsRNA) uptake in the fungal plant pathogens Botrytis cinerea, Sclerotinia sclerotiorum, Rhizoctonia solani, Aspergillus niger and Verticillium dahliae, but no uptake in Colletotrichum gloeosporioides, and weak uptake in a beneficial fungus, Trichoderma virens. For the oomycete plant pathogen, Phytophthora infestans, RNA uptake was limited and varied across different cell types and developmental stages. Topical application of dsRNA targeting virulence‐related genes in pathogens with high RNA uptake efficiency significantly inhibited plant disease symptoms, whereas the application of dsRNA in pathogens with low RNA uptake efficiency did not suppress infection. Our results have revealed that dsRNA uptake efficiencies vary across eukaryotic microbe species and cell types. The success of SIGS for plant disease management can largely be determined by the pathogen’s RNA uptake efficiency.


Introduction
Crop protection at both the pre-and post-harvest levels is critical for long-term food security. Fungal and oomycete diseases pose a serious threat to crop production worldwide (Bebber and Gurr, 2015). At present, agricultural crops rely almost exclusively on fungicides to control disease, resulting in pesticide residues that often threaten human health and the environment. Further, resistant strains of fungi have been identified against every major fungicide used in agricultural applications (Fisher et al., 2018). Thus, there is an urgent need to develop eco-friendly, effective solutions to control plant diseases. The development of disease control strategies using biomolecules (nucleic acids, lipids, sugars and proteins) that exploit naturally occurring pathways may help circumvent the use of chemical pesticides.
Recent studies have discovered a novel mechanism of communication between plants and their pathogens, termed crosskingdom RNA interference (RNAi), a phenomenon in which fungi send small RNAs (sRNAs) into host plants to silence host immune response genes and, in turn, plants send sRNAs packaged in extracellular vesicles into their fungal pathogens to silence virulence-related genes (Cai et al., 2018b;Huang et al., 2019;Weiberg et al., 2013;Zeng et al., 2019). Cross-kingdom RNAi can be utilized in plant protection strategies by genetically modifying host plants to express sRNAs or double-stranded RNAs (dsRNAs) that target pathogen virulence-related genes, a technology called host-induced gene silencing (HIGS) (Cai et al., 2019;Song and Thomma, 2018;Wang et al., 2017a;Zhang et al., 2016). These transgene-derived artificial RNAs can be delivered from plants to pathogens and pests, including fungi (Cai et al., 2018a;Nowara et al., 2010;Qi et al., 2019;Xu et al., 2018), nematodes (Chen et al., 2015;Shivakumara et al., 2017), oomycetes (Govindarajulu et al., 2015;Jahan et al., 2015) and insects (Abdellatef et al., 2015;Baum et al., 2007;Liu et al., 2019), inducing gene silencing in trans and conferring disease resistance in the host plants. However, the application of HIGS is limited due to the lack of efficient transformation techniques in many crop plants, including many fruit trees, vegetables and flowers (Capriotti et al., 2020;Koch and Kogel, 2014). Furthermore, many consumers still resist genetically modified (GMO) products, which are currently banned in European agricultural production (Kleter et al., 2018;Tsatsakis et al., 2017;Wunderlich and Gatto, 2015). Therefore, innovative non-GMO plant protection strategies are highly desired.
Previous work has found that fungal pathogens Botrytis cinerea and Fusarium graminearum can efficiently take up environmental dsRNAs, which are then processed into sRNAs and induce the silencing of pathogen genes with complementary sequences (Koch et al., 2016;Wang et al., 2016). These discoveries prompted the development of an innovative crop protection strategy, Spray-Induced Gene Silencing (SIGS). As a non-GMO alternative to HIGS, SIGS inhibits pathogen infection by topical application of dsRNA or sRNA molecules onto plants to silence pathogen virulence-related genes Wang and Jin, 2017). SIGS has been successfully utilized to prevent some fungal diseases. Topical application of dsRNAs and sRNAs that target the key components of the RNAi machinery, Dicer-like (DCL)1 and DCL2 genes of the fungal pathogen Botrytis cinerea inhibits grey mould disease on fruits, vegetables and flowers . When applied to barley leaves, dsRNA designed to target fungal cytochrome P450 inhibited the growth of the fungal pathogen Fusarium graminearum (Koch et al., 2016). Additionally, dsRNA sprays can inhibit Botrytis cinerea and Sclerotinia sclerotiorum growth on Brassica napus (McLoughlin et al., 2018). However, it is still not clear what determines SIGS efficacy and how effective SIGS is against a wide variety of eukaryotic pathogens. Establishing the effectiveness of SIGS across a wide range of pathogens is a critical next step in the development of this technology.
For SIGS technology to be effective, it is crucial that fungi can take up RNA from the environment. Here, we investigated the environmental RNA uptake efficiency of various plant pathogenic fungi. These included S. sclerotiorum, a necrotrophic plant pathogen that infects over 400 plant species worldwide, principally vegetables and ornamental plants (Bolton et al., 2006;Kabbage et al., 2015); Aspergillus niger, a haploid filamentous fungus that can contaminate fruits, cereal grains, and some animal products directly by fungal growth or indirectly through the production of mycotoxins (Hocking et al., 2007;Wilson et al., 2002); Rhizoctonia solani, a well-known soil-borne fungus that causes a wide range of significant crop diseases (Zhou et al., 2016a); Verticillium dahliae, a soil-borne fungus that causes wilt disease in a wide range of plant species (Klosterman et al., 2009;Luo et al., 2014); and Colletotrichum gloeosporioides, which causes anthracnose, a disease characterized by sunken necrotic lesions, which can infect fruits, flowers, leaves, petioles, stolons and crowns (Zhou et al., 2016b). In addition to various pathogenic fungi, we also investigated the RNA uptake efficiency of a non-pathogenic fungus, Trichoderma virens, which is a biocontrol agent of plant diseases (Contreras-Cornejo et al., 2011;Yang et al., 2016). Finally, we investigated the RNA uptake ability of a non-fungal eukaryotic pathogen, the oomycete Phytophthora infestans, which is the causal agent of late blight disease that threatens tomato and potato crops worldwide (Nowicki et al., 2012).
In this study, we found that many eukaryotic microbes can take up RNA from the environment but with different efficiencies. B. cinerea and S. sclerotiorum, as shown previously (McLoughlin et al., 2018;Wang et al., 2016), R. solani, A. niger and V. dahliae could take up environmental RNA with high efficiency, whereas RNA uptake was modest in T. virens, undetectable in C. gloeosporioides and limited in P. infestans. Furthermore, we found that topical application of dsRNAs targeting virulencerelated pathogen genes could suppress disease caused by the pathogens that have high RNA uptake efficiency. In comparison, SIGS did not inhibit infection of C. gloeosporioides due to the lack of environmental RNA uptake. Finally, to examine the longevity of dsRNA-conferred protection on crops, we challenged tomato leaves with B. cinerea over a time-course post-dsRNA treatment.
In summary, we found that most fungi tested can take up dsRNA, and efficient dsRNA uptake in pathogens is essential for the success of SIGS in crop protection.

RNA uptake efficiency varies across fungal species
To determine whether dsRNA uptake by fungi is a common phenomenon, we first tested the efficiency of dsRNA uptake across different fungal species, including pathogens from different classes of Ascomycota: B. cinerea and S. sclerotiorum from the Leotiomycete class, A. niger from the Eurotiomycete class, V. dahliae from the Sordariomycetes class, and a non-pathogenic fungus, T. virens a common soil microbe from the Sordariomycetes class ( Figure S1), that is usually a part of healthy root ecosystems that promote plant growth (Contreras-Cornejo et al., 2009). We also tested a rice pathogen, R. solani, from the Basidiomycota division. To determine RNA uptake efficiency, we treated fungal cells with fluoresceinlabelled YFP-dsRNA. These fungal cells had no obvious autofluorescence ( Figure S2), and therefore, there was no interference with the fluorescence signal from the fluorescein-labelled dsRNA. Micrococcal nuclease (MNase) treatment was performed 30 min before microscopy analysis to remove any labelled dsRNAs that were outside the fungal cells. We found that different fungal species had different rates and efficiencies of dsRNA uptake ( Figure 1 and Figures S3, S4). Fluorescence signals were observed as early as 6 h post-dsRNA treatment (hpt) in B. cinerea, S. sclerotiorum and R. solani (Figures S3 and S4), all of which showed very efficient RNA uptake ( Figure 1a). In A. niger, an intense fluorescence signal was observable 10 hpt, and in V. dahliae 12 hpt ( Figure 1a and Figures S3, S4). However, no fluorescence signal was detectable in C. gloeosporioides even 24 hpt, suggesting that C. gloesporioides cannot take up environmental dsRNA (Figure 1b To further confirm that the dsRNA entered the cytoplasm of the fungal cells, the fluorescein-labelled dsRNA was added to the liquid culture of S. sclerotiorum mycelium, which was subjected to protoplast preparation after culturing for 48 h. Fluorescence signals were clearly observed within S. sclerotiorum protoplasts after MNase treatment ( Figure S5). Overall, we found that B. cinerea, S. sclerotiorum, R. solani, V. dahliae and A. niger took up external dsRNA efficiently, although uptake by V. dahliae and A. niger was at a slower rate (Figure 1 and Figures S3, S4, S6), C. gloeosporioides showed no detectable dsRNA uptake and the non-pathogenic fungus T. virens displayed weak and slow dsRNA uptake ( Figure 1 and Figures S3, S4, S6). Thus, dsRNA uptake exhibited different efficiencies across diverse species of fungi.
Topical application of dsRNA targeting vesicle trafficking pathway genes inhibited the virulence of B. cinerea, S. sclerotiorum and A. niger Our recent work showed that Botrytis single deletion mutant strains of vesicle trafficking pathway genes vacuolar protein sorting 51 (VPS51), dynactin (DCTN1) and suppressor of actin (SAC1) had greatly reduced virulence on Arabidopsis thaliana leaves (Cai et al., 2018b). To validate whether SIGS of vesicle trafficking pathway genes could control grey mould disease, we first generated a Bc-VPS51+DCTN1+SAC1-dsRNA construct by integrating non-conserved fragments of Bc-VPS51 (169 bp), Bc-DCTN1 (179 bp) and BcSAC1 (168 bp) via overlap PCR and transcribing the construct in vitro from both the 5 0 and 3 0 ends  (Table S2). In all dsRNA constructs throughout this paper, non-conserved regions are selected to avoid off-target effects in host plants or beneficial microbes. Bc-VPS51+DCTN1+SAC1 or Bc-DCL1/2 dsRNA (20 µL of 20 ng/µL in RNase-free water) were separately applied to lettuce leaves, tomato fruits, rose petals and grape fruits. After dsRNA treatment, plant materials were then inoculated with B. cinerea at the dsRNA treated area. All plant materials treated with Bc-VPS51+DCTN1+SAC1-dsRNA showed a reduction in disease symptoms as compared to the controls treated with water or YFP-dsRNA (Figure 2a, b). The reduction of the infected areas was comparable to infection phenotypes after Bc-DCL1/2-dsRNA treatment (Figure 2a, b), which was used as a positive control as previous work demonstrated that Bc-DCL1/2-dsRNA was effective in reducing Botrytis virulence . Moreover, the mRNA expression levels of BcVPS51, BcDCTN1 and BcSAC1 were reduced in B. cinerea treated with Bc-VPS51+DCTN1+SAC1-dsRNA ( Figure 2c).
Sclerotinia sclerotiorum is closely related to B. cinerea. To test whether VPS51, DCTN1 and SAC1 also play an essential role in S. sclerotiorum pathogenicity, we knocked out SsVPS51, SsDCTN1 and SsSAC1 genes by homologous recombination. Ss-Dvps51, Ss-Ddctn1 and Ss-Dsac1 mutant strains showed a reduction in development and mycelium growth after 3 days of culture and reduced sclerotia size on PDA media in comparison with the wild-type (WT) strain (Figure S7a,b). Most importantly, the Ss-Dvps51, Ss-Ddctn1 and Ss-Dsac1 knockout mutants exhibited reduced virulence on lettuce and collard green leaves in comparison with WT ( Figure S7c, d). Thus, S. sclerotiorum VPS51, DCTN1 and SAC1 are important for the pathogenicity of S. sclerotiorum and should be appropriate targets for SIGS.
Next, we generated S. sclerotiorum VPS51+DCTN1+SAC1-dsRNA by integrating SsVPS51 (230 bp), SsDCTN1 (196 bp) and SsSAC1 (194 bp) via overlap PCR and performed in vitro transcription from both the 5 0 and 3 0 ends (Table S2). The Ss-VPS51+DCTN1+SAC1-dsRNA (20 µL of 40 ng/µL) was externally applied to lettuce and collard green leaves. After dsRNA application, plant material was challenged with S. sclerotiorum. All plants treated with Ss-VPS51+DCTN1+SAC1-dsRNA showed reduced disease symptoms and lesion size compared to water and YFP-dsRNA treated controls (Figure 2d, e). Moreover, the Figure 1 Examination of dsRNA uptake efficiencies in multiple fungi using fluorescein-labelled YFP-dsRNA. (a, b) Fluorescein-labelled YFP-dsRNA was added to the spores of Botrytiscinerea, Verticilliumdahliae, Aspergillusniger, C. gloeosporoides and Trichodermavirens, and to the hyphae of Sclerotiniasclerotiorum and Rhizoctoniasolani. Micrococcal nuclease (MNase) treatment was performed 30 min before acquiring images using the confocal microscopy laser scanner (CMLS). Fluorescence signals were detected inside of B. cinerea, S. sclerotiorum, R. solani, A. niger and V. dahliae cells, but not C. gloeosporoides cells. A weak signal was observed in T. virens cells. Pictures were taken at 10 hours post-treatment (hpt) for B. cinerea, S. sclerotiorum, and R. solani, A. niger and V. dahliae, and at 24 hpt for C. gloeosporoides and T. virens. Scale bars = 20 µm except for B. cinerea image (Scale bars = 10 µm). (c) C. gloeosporioides spores are unable to take up dsRNA. Fluorescence signals were only visible on the outer surface of C. gloeosporioides cells before MNase treatment, but disappeared after MNase treatment. Scale bars = 15 µm. mRNA expression levels of SsVPS51, SsDCTN1 and SsSAC1 were also reduced in S. sclerotiorum ( Figure 2f). Furthermore, Ss-DCL1+DCL2-dsRNA application could also inhibit S. sclerotiorum infection, suggesting that DCL proteins are important for the virulence of S. sclerotiorum. Thus, externally applied Ss-VPS51+DCTN1+SAC1-dsRNA or Ss-DCL1+DCL2-dsRNA protected plants from S. sclerotiorum invasion.
Aspergillus niger can cause post-harvest diseases in many fruits and vegetables (Liu et al., 2017). The A. niger genome encodes the vesicle trafficking pathway genes VPS51, DCTN1 and SAC1. A. niger VPS51+DCTN1+SAC1-dsRNA was generated accordingly. Furthermore, previous research suggested that a probable exo-polygalacturonase b (pgxB) is an A. niger virulence gene, which plays an important role in the infection process on the fruit  (Liu et al., 2017). dsRNA targeting a non-conserved region of pgxB (219 bp) was also generated. Topical application of the VPS51+DCTN1+SAC1-dsRNA or the pgxB-dsRNA onto the surface of tomato, apple and grape fruits led to much weaker fungal growth and disease symptoms, resulting in significantly smaller lesions than the control samples treated with YFP-dsRNA or water (Figure 3a, b). In addition, the mRNA expression levels of the target genes were reduced in A. niger after pgxB-and VPS51+DCTN1+SAC1-dsRNA treatment ( Figure 3c). Thus, external application of VPS51+DCTN1+SAC1-dsRNA inhibits infection of Botrytis cinerea, S. sclerotiorum and A. niger.
Topical application of dsRNA inhibited the virulence of R. solani Rice sheath blight caused by R. solani is one of the most devastating fungal diseases of rice (Oryza sativa) worldwide. We identified the vesicle trafficking pathway genes DCTN1 and SAC1 in the R. solani genome (Zheng et al., 2013), but VPS51 was not identified. To analyse whether SIGS could be adapted to control sheath blight, we generated Rs-DCTN1+SAC1-dsRNA containing two segments from non-conserved regions of Rs-DCTN1, and Rs-SAC1 and applied it to rice leaves. Compared to control plants treated with either YFP-dsRNA or water, effective protection was observed in plants treated with Rs-DCTN1+SAC1-dsRNA (Figure 4a,b). Polygalacturonase (PG), a pectin degrading enzyme that hydrolyses a-1,4 glycosidic bonds of pectic acid, is a known virulence factor of R. solani (Rao et al., 2019). Stable expression of a PG-RNAi construct in rice suppresses R. solani infection (Rao et al., 2019). We found that application of Rs-PG-dsRNA also resulted in a reduction of disease symptoms and fungal biomass compared to water and YFP-dsRNA treated controls (Figure 4a, b). Further, the mRNA expression levels of the corresponding genes were reduced in R. solani after the dsRNA treatment ( Figure 4c). Therefore, externally applied Rs-DCTN1+SAC1-dsRNA or Rs-PG-dsRNA protected the plants from R. solani infection by targeting and silencing R. solani genes.
RNA is not stable in the soil and pretreatment of roots with dsRNA inhibited the infection of V. dahliae Verticillium dahliae is a root pathogen that causes wilt disease in hundreds of plant species. Previously, HIGS strategies have been successful to suppress V. dahliae infection (Song and Thomma, 2018;Wang et al., 2016;Xu et al., 2018). However, it is unknown whether direct application of dsRNAs can effectively control V. dahliae virulence. The V. dahliae genome encodes DCTN1 and SAC1 genes, but not the VPS51 gene. As with the other pathogens, we generated Vd-DCTN1+SAC1dsRNA and Vd-DCL1+DCL2 dsRNA. We found that dsRNA is not stable in the soil ( Figure 5a) and direct application of dsRNA to the soil did not protect the plant from V. dahliae infection. So, instead, we dipped Arabidopsis roots in Vd-DCL1/2-dsRNA or Vd-DCTN1+SAC1-dsRNA mixed with V. dahliae spores, then planted the seedlings in the soil. These treated plants showed reduced disease symptoms in comparison with control plants treated with either YFP-dsRNA+spores, or only spores (Figure 5b, c). Furthermore, V. dahliae treated with Vd-DCL1/2-dsRNA or Vd-DCTN1+SAC1-dsRNA showed a clear reduction in targeted mRNA expression levels ( Figure 5d). Although pretreatment of roots with dsRNA could protect the plants to a certain degree, it is not practical in the field. Therefore, innovative RNA stabilization strategies, such as using nanoparticles or special formulas, are needed to protect the RNAs in the soil and increase SIGS efficiency against soil-borne pathogens.
Topical application of dsRNA cannot inhibit the virulence of C. gloeosporioides We also assessed the influence of dsRNA on the virulence of C. gloeosporioides, a fungal pathogen that is unable to take up environmental dsRNA (Figure 1c). Previous work found that DCL proteins are important for the development and pathogenesis of C. gloeosporioides (Wang et al., 2017b). Here, we generated Cg-VPS51+DCTN1+SAC1-dsRNA and Cg-DCL1/2-dsRNA from the nonconserved regions of these genes. We then tested the efficacy of these dsRNAs in controlling C. gloeosporioides disease by separately applying the dsRNAs to the surface of various fruits (e.g. cherry, apple and tomato). All the fruits treated with pathogen gene-targeting dsRNA constructs displayed similar severity of disease symptoms to the control fruits treated with water or YFP-dsRNA ( Figure S8a, b). Moreover, C. gloeosporioides treated with dsRNAs showed no reduction in the expression of targeted mRNAs in comparison with the controls (Figure S8c), indicating that SIGS is not effective in controlling C. gloeosporioides.
Phytophthora infestans exhibits low dsRNA uptake efficiency and topical application of dsRNA fails to inhibit the virulence of P. infestans In addition to fungi, oomycetes also cause severe crop diseases worldwide. For example, potato late blight, caused by P. infestans, induces the death of potato stems and leaves and leads to tuber rotting (Leesutthiphonchai et al., 2018;Shattock, 2002). To determine whether SIGS approaches can be effective to control oomycete pathogens, we first measured the dsRNA uptake Figure 2 Topical application of pathogen gene-targeting dsRNAs inhibited the virulence of B. cinerea and S. sclerotiorum. (a) Tomato fruits, lettuce leaves, rose petals and grape fruits were inoculated with B. cinerea spores after treating with controls (water or YFP-dsRNA) or Bc-VPS51+DCTN1+SAC1-dsRNA or Bc-DCL1/2-dsRNA (20 ng/µL). (b) The relative lesion sizes were measured 3 days post-inoculation (dpi) on lettuce leaves, rose petals and 5 dpi on grapes and tomato fruits using ImageJ software. Error bars indicate the SD of 10 samples and three biological repeats were conducted for the relative lesion sizes. Statistical significance (Student's t-test): **, P < 0.01. (c) qRT-PCR analysis of Bc-VPS51, Bc-DCTN1 and Bc-SAC1 expression in Bc-VPS51+DCTN1+SAC1-dsRNA treated B. cinerea cells. The expression levels of target genes were normalized to expression of B. cinereaActin. Error bars represent the SD from three technical replicates. Statistical significance (Student's t-test): *, P < 0.05; **, P < 0.01 between the treatment and the water control. (d) Lettuce and collard green leaves were inoculated with S. sclerotiorum mycelium plugs after treating with controls (water or YFP-dsRNA) or Ss-VPS51+DCTN1+SAC1-and Ss-DCL1/2-dsRNA (40 ng/µL). Pictures were taken at 3 dpi. (e) The relative lesion sizes were measured at 3 dpi for lettuce and collard greens using ImageJ software. Error bars indicate the standard deviations (SD) of 10 samples. Asterisks (**) indicate statistically significant differences (P < 0.01, Student's t-test) between the treatment and the water control. (f) The mRNA expression levels of SsVPS51, SsDCTN1 and SsSAC1 were detected in Ss-VPS51+DCTN1+SAC1-dsRNA treated S. sclerotiorum by qRT-PCR analysis. The expression levels of target genes were normalized to expression of S. sclerotiorumActin. Asterisks (**) indicate statistically significant differences (P < 0.01, Student's t-test) between the treatment and the water control. Similar results were observed from three biological replicates in (c) and (f). efficiency of P. infestans. After incubating hyphae with fluorescein-labelled YFP-dsRNA and subsequently treating with MNase, weak fluorescence signals were detected in some P. infestans hyphae in plug-inoculated rye-sucrose medium cultures, whereas no signal was observed in hyphae derived from the germination of zoospores 12 h post-inoculation in rye medium (Figure 6a and Figure S9). Fluorescence signals were detected in some zoospore cysts and sporangia (Figure 6b, c, and Figure S9). However, fluorescence was only observed in approximately 5-10% of sporangia and zoospore cysts.
To examine whether SIGS inhibited P. infestans infection, we selected several target genes that are important for P. infestans virulence. Silencing of the Pihmp1 gene that encodes the haustorium-specific membrane protein PiHMP1 has been shown to lead to a loss of pathogenicity (Avrova et al., 2008). Additionally, the expression of the P. infestans G-protein beta-subunit gene PiGPB1 is required for sporangium formation (Latijnhouwers and Govers, 2003). Oomycetes possess small RNA biosynthesis pathways similar to those of other eukaryotes (Vetukuri et al., 2011), thus, PiDCL1 was also chosen as a target gene. The vesicle trafficking pathway genes DCTN1 and SAC1, but not VPS51 were identified in the P. infestans genome and were also selected as targets. dsRNAs of all target genes, PiDCL1-, PiHMP1-, PiPGB1-and DCTN1+SAC1, were separately applied to potato leaves before inoculation with P. infestans zoospores. These dsRNA treatments did not reduce or slow disease development ( Figure S10a), as the size of disease lesions were indistinguishable from control-treated plants ( Figure S10a, b). Moreover, P. infestans treated with dsRNAs showed no significant reduction in targeted mRNA expression levels ( Figure S10c), demonstrating that traditional SIGS approaches are not effective in controlling P. infestans.

Examining the longevity of dsRNA-mediated plant protection
We next investigated the protection effect on pre-harvest crop material over time by tracking the SIGS-RNAi efficiency against B. cinerea on tomato plants in a greenhouse assay. To simulate a commercial application, 100 ng/µL Bc-DCL1/2-or Bc-VPS51+DCTN1+SAC1-dsRNA was speayed onto the leaves of 5-to 6-week-old tomato plants. This higher concentration of dsRNA was needed here because tomato leaves are more susceptible to B. cinerea and are infected more rapidly than fruits. Subsequently, the leaves were inoculated with B. cinerea spores at 1, 3, 7 and 14 days post-treatment (Figure 7a). Both Bc-DCL1/2-and Bc-VPS51+DCTN1+SAC1-dsRNA showed significant reductions in fungal growth (Student's t-test at P < 0.05) compared to the controls at 1-, 3-and 7-dpt, while at 14-dpt the effect was insignificant (Figure 7b). Taken together, these results show that the dsRNAs could provide protection for up to 7 days under the experimental condition, although the RNAi effect is reduced over time, likely due to dsRNA degradation in the environment.

Discussion
RNA-based disease management products offer an important eco-friendly alternative to standard chemical pesticides for controlling agricultural losses caused by pests and disease (Taning et al., 2020). SIGS has been shown to be effective on controlling Figure 3 Topical application of pathogen gene-targeting dsRNAs inhibited Aspergillus niger virulence. (a) Tomatoes, apples and grapes were inoculated with A. niger spores after treating with controls (water or YFP-dsRNA), AnpgxB-or An-VPS51+DCTN1+SAC1-dsRNA (20 ng/µL). Pictures were taken at 5 dpi. (b) The relative lesion sizes were measured at 5 dpi using ImageJ software. Error bars indicate the SD of 10 samples. (c) The mRNA expression levels of AnVPS51, AnDCTN1, AnSAC1 and AnpgxB were detected in An-VPS51+DCTN1+SAC1-or AnpgxB-dsRNA treated A. niger. The expression levels of target genes were normalized to expression of A. niger ActinA. Error bars indicate the SD of three technical replicates. Statistical significance (Student's t-test) between the treatment and the water control: *, P < 0.05; **, P < 0.01. Similar results were observed from three biological replicates.  (Koch et al., 2016), and white mould caused by S. sclerotiorum (McLoughlin et al., 2018). To explore whether this RNA-based pathogen control strategy can be adapted across a multitude of eukaryotic pathogens, we investigated the dsRNA uptake efficiency of different species of fungi and oomycetes, and their pathogenicity after dsRNA treatment. Here, we provide solid evidence that environmental dsRNA can be taken up by multiple fungal species with different uptake efficiencies, which determines the effectiveness of SIGS.
In the present study, we found that the uptake efficiency of RNA was significantly different among various fungi. Moreover, some fungi cannot take up dsRNA from the environment, such as C. gloeosporioides, thus limiting the application of SIGS in the control of these fungal diseases. Taken together, our results demonstrate that externally applied dsRNA can protect plants from fungal pathogens with high RNA uptake efficiency such as B. cinerea, S. sclerotiorum, A. niger and R. solani, but not against pathogens with low or no RNA uptake efficiency such as C. gloeosporioides and oomycete pathogen P. infestans. Thus, RNA uptake efficiency is crucial for controlling pathogens by SIGS. Of course, beyond RNA uptake ability, a pathogen also needs to have Figure 4 Topical application of pathogen gene-targeting dsRNAs inhibited Rhizoctonia solani virulence. (a) Rice leaves were inoculated with R. solani mycelium plugs after treating with controls (water and YFP-dsRNA), Rs-DCTN1+SAC1-dsRNA or Rs-PG-dsRNA (40 ng/µL). Pictures were taken at 3 dpi. (b) Relative biomass of R. solani was calculated by examining the expression of Rs-Actin by qRT-PCR, which is normalized to Os18S rRNA; error bars represent the SD of three replicates. Asterisks (**) indicate statistically significant differences (P < 0.01, Student's t-test) between the treatment and the water control. (c) The mRNA expression levels of RsDCTN1, RsSAC1, and Rs-PG were detected in Rs-DCTN1+SAC1-or Rs-PG-dsRNA treated R. solani by qRT-PCR analysis. The expression levels of target genes were normalized to expression of R. solani Actin1. Asterisks (**) indicate statistically significant differences (P < 0.01, Student's t-test). Similar results were observed from three biological replicates.
functional RNAi machinery for SIGS to be effective. Establishing RNA uptake efficiency of a pathogen is the first key step in developing SIGS disease management strategies. Recently, researchers have found that Zymoseptoria tritici was incapable of exogenous dsRNA uptake, which may also limit RNAi approaches as a control measure of this fungal species (Kettles et al., 2019). Further, we found that different cell types can have distinct dsRNA uptake efficiencies. Specifically, in P. infestans, RNA uptake was observed in only a fraction of tissues. Thus, it was not surprising that treating plants with dsRNAs targeting essential P. infestans genes did not suppress P. infestans virulence. However, previous work has found HIGS strategies to be effective in preventing infection by the oomycete pathogen Bremia lactucae (Govindarajulu et al., 2015), and P. infestans (Jahan et al., 2015). This shows that P. infestans takes up plant-originated RNAs more efficiently, suggesting that mechanisms such as artificial vesicle-mediated delivery may increase RNA uptake efficiency in P. infestans. Notably, P. infestans possesses a cell wall that is chemically dissimilar to those of fungi (Melida et al., 2013). This may also partially account for its low environmental RNA uptake efficiency.
The fungal cell wall is essential for viability, pathogenicity, and also regulates material exchange with the external environment (Cantu et al., 2009). The relative abundance and the chemical structures of the major components in the cell wall differ among Figure 5 RNA is not stable in the soil and pretreatment of roots with pathogen gene-targeting dsRNA reduced the infection of Verticilliumdahliae. (a) Northern blot analysis of Vd-DCL1/2 dsRNAs was performed over a time course after they were mixed with soil. (b) Arabidopsis plants were uprooted and incubated in V. dahliae spore suspensions (10 6 spores/mL) with controls (water and YFP-dsRNA), Vd-DCL1/2-dsRNA, and Vd-DCTN1+SAC1-dsRNA (40 ng/µL) treatments. Pictures were taken at 14 dpi. (c) Relative biomass of V. dahliae was calculated by examining the expression of Vd-actin by qRT-PCR, which was normalized to At-actin2; error bars represent the SD of three replicates. Asterisks (**) indicate statistically significant differences (P < 0.01, Student's t-test) between the treatment and the water control. (d) The mRNA expression levels of VdDCL1 and VdDCL2 were detected in Vd-DCL1/2-dsRNA treated V. dahliae, and the mRNA expression levels of VdDCTN1 and VdSAC1 were detected in Vd-DCTN1+VdSAC1-dsRNA treated V. dahliae by qRT-PCR analysis. The expression levels of target genes were normalized to expression of Vd-Actin. Asterisks (**) indicate statistically significant differences (P < 0.01, Student's t-test) between the treatment and the water control. Similar results were observed from three biological replicates.  (Bowman and Free, 2006;Diaz et al., 1992). Additionally, the number of chitinase genes in fungi display wide variations, from a single gene in Schizosaccharomyces pombe, to 36 genes in T. virens (Yang and Zhang, 2019). All these may impact the viscoelastic properties and dsRNA uptake efficiency of the cell wall. Walker et al. (2018) have found that the cell wall of mammalian fungal pathogens Candida albicans and Cryptococcus neoformans allows for the uptake of liposome vesicles. These findings suggest that the fungal cell wall is deformable and viscoelastic to allow vesicles and large molecules, such as RNAs to pass through.
Finally, to better characterize the commercial potential of SIGS, we determined the longevity of plant protection against fungal infection conferred by dsRNAs. We found that naked dsRNAs can remain active up to a week on plant material, but the efficacy of protection decreased over time. Innovative methods for enhancing RNA stability will be highly desirable for commercial use of SIGS. These methods include RNA modifications as well as the use of protective nanoparticles. An efficient carrier for SIGS-mediated plant protection is BioClay, which has been shown to sustain plant protection against viruses for weeks (Mitter et al., 2017). In applications against the insect pest, rice striped stem borer, researchers have found that carbon quantum dots (CQD) are efficient dsRNA carriers which can induce systemic RNAi in the insect target (Wang et al., 2020). Another class of nanoparticles worth exploring are artificial vesicles or liposomes, which are used in clinical contexts to deliver therapeutic RNAs (Bochicchio et al., 2014) and may mimic the natural pathway of extracellularvesicle mediated sRNA transport from plants to pathogens. They may also have the potential to improve pathogen RNA uptake efficiency, ultimately leading to longer lasting and more robust plant protection against infections.
To conclude, SIGS disease management strategies show promising results in many, but not all, eukaryotic pathogens. The efficacy of SIGS is directly related to RNA uptake efficiency by the pathogen. Further optimization of delivery vectors and methods could improve the uptake efficiency in pathogens. Finally, before widespread commercial use of RNA-based disease management strategies is practical, it is crucial that environmental RNA stability is enhanced.

Fungi and oomycete culture and infection conditions
Sclerotinia sclerotiorum 1980, R. solani AG1-IA, A. niger CBS 513.88, T. virens Gv29-8, V. dahliae JR2 and C. gloeosporioides Nara gc5 were routinely cultured on PDA medium (24 g of potato dextrose broth and 15 g of agar per litre) at 25°C. B. cinerea B05.10 was cultured on malt extract agar medium (20 g of malt extract, 10 g of Bacto protease peptone, and 15 g of agar per litre). Fungal mycelia used for genomic DNA and total RNA extraction was harvested from cultures grown on medium covered by a sterile cellophane membrane. Vegetables (lettuce and collard greens), rose petals and fruits (tomato, cherry, grape and apple) purchased from local supermarkets were used as experimental host plant materials. For B. cinerea infection, the fungal spores were diluted in 1% sabouraud maltose broth buffer to a final concentration of 10 5 spores/mL for drop inoculation of plants. For S. sclerotiorum infection, mycelium plugs (5 mm in diameter) were obtained from the expanding margins of colonies to inoculate vegetable leaves. Infected leaf tissues were cultured in a light incubator at 25°C for 72 h and fruits for 120 h at constant and high humidity. For R. solani inoculations, R. solani was cultured for 48 h on PDA medium. A mycelium plug (5 mm diameter) of actively growing R. solani was placed onto the surface of rice leaves. The V. dahliae soil-inoculation assay was performed as previously described (Ellendorff et al., 2009). Fungal biomass quantification was performed as described previously (Gachon and Saindrenan, 2004). C gloeosporioides inoculations were based on the standard protocol, with minor modifications (Seo et al., 2014). Fungal inoculation was conducted by applying 10 µL droplets of spore suspension (1 9 10 5 spores per ml in distilled water) onto fruits. The inoculated fruits were placed in high humidity and dark conditions for 1 day to stimulate infection. For A. niger inoculations, the fungal spores were adjusted to 10 6 spores/mL for drop inoculation of fruits. P. infestans 1306 culture and infection conditions were based on standard protocols (Ah-Fong et al., 2017). The strain was maintained in the dark in rye-sucrose agar at 18°C. Developmental stages were obtained as described from rye media cultures. Potato plants used for infections were grown with a 12 h light/dark cycle (25°C day, 350 lmol/m 2 /s fluorescent light; 18°C night) for 4-5 weeks, and then leaves were detached, placed on 0.8% water agar, inoculated with zoospores and incubated at 16°C with a 12 h light/dark cycle with 115 lmol/ m 2 /s illumination in a clear plastic bag to maintain high humidity. The lesion sizes of pathogen-infected plant materials were measured and calculated using ImageJ software.

RNA extraction and qRT-PCR
RNA was extracted using the Trizol extraction method. Purified RNA was treated with DNase I (Thermo Fisher Scientific, Waltham, Ma, USA), then first-strand cDNA was synthesized using the SuperscriptTM III First-Strand Synthesis System (Thermo Fisher Scientific, Waltham, Ma, USA). qRT-PCR was performed using the CFX384 (Bio-Rad, Hercules, CA, USA) Connect system using SYBR Green mix (Bio-Rad) with the following program: 95°C for 15 min, followed by 40 cycles of 94°C for 30 s, 50°C for 30 s and 72°C for 30 s. Melt curves were generated to evaluate the fidelity of amplification. Expression levels were calculated using the DDC t method.

Sclerotinia sclerotiorum protoplast transformation
Sclerotinia sclerotiorum protoplast transformation was based on standard protocols with minor modifications (Rollins, 2003). Protoplasts were suspended at a concentration of 1 9 10 8 protoplasts per mL in four parts STC (0.8 M sorbitol, 50 mM Tris-HCl, pH 8.0, and 50 mM CaCl 2 ) and one-part SPTC media (0.8 M sorbitol, 40% polyethylene glycol 3500, 50 mM Tris-HCl, pH 8.0 and 50 mM CaCl 2 ). A mixture of 5 µg of transforming DNA, 5 µL of spermidine (50 mM stock) and 5 µL of heparin (5 mg/mL in STC) was added to 100 µL of the protoplast suspension. The above mixtures were incubated on ice for 30 min, then gently mixed with 1 mL of SPTC solution and incubated at room temperature for 20 min. The transformation Figure 7 Examining the longevity of dsRNA-mediated plant protection. (a) Tomato leaves were inoculated with Botrytiscinerea spores after spraying with controls (water or YFP-dsRNA) or Bc-VPS51+DCTN1+SAC1-dsRNA or Bc-DCL1/ 2-dsRNA (100 ng/µL) at 1, 3, 7 and 14 dpt. The relative lesion sizes were measured 4 dpi (d). Error bars indicate the SD of at least three independent biological replicates with total of 45 leaflets in each replicate, and the statistical significance (Student's ttest) by *, P < 0.05; ***, P < 0.001 between the treatment and the water control. mixture was added to a 50 mL flask, then 10-20 mL of regeneration medium (RM; 274 g of sucrose, 1 g of yeast extract and 1 g of casamino acid per litre) was added. The mixture was shaken at 100 rpm and left overnight at 25°C. The protoplasts were gently mixed in 200 mL of regeneration medium RMA (RM containing 15 g agar per litre) at 43°C, poured into 9 cm diameter Petri dishes (10 mL per dish) and incubated at 25°C. After 24 h, the plates were overlaid with 10 mL of selective medium containing 100 µg/mL hygromycin B. Transformants were obtained 4-10 days after transformation and transferred to fresh plates of PDA containing 100 µg/mL of hygromycin B. Hygromycin-resistant transformants were hyphal tip transferred a minimum of three times on PDA containing 100 µg of hygromycin per milliliter. Primers used for S. sclerotiorum knock-down mutant construction and detection are listed in Table S1.

In vitro synthesis of dsRNA
In vitro synthesis of dsRNA was based on established protocols . Following the MEGAscriptâ RNAi Kit instructions (Life Technologies, Carlsbad, CA), the T7 promoter sequence was introduced at both 5 0 and 3 0 ends of the RNAi fragments by PCR. After purification, the DNA fragments containing T7 promoters at both ends were used for in vitro transcription. Primers used for in vitro synthesis of dsRNAs are listed in Table S1, and the sequences of dsRNAs are listed in Table S2.

Fungal uptake of fluorescein-labelled dsRNA in vitro
In vitro synthesis of fluorescein-labelled dsRNA was based on established protocols . YFP-dsRNA was labelled using the fluorescein RNA Labeling Mix Kit following the manufacturer's instructions (Sigma, St. Louis, MO). For confocal microscopy examination of fluorescent dsRNA uptake by fungal mycelium, 5 µL of 150 ng/µL fluorescent dsRNA was applied to 10 µL of 10 5 spores/mL of B. cinerea, V. dahliae, A. niger, T. virens, P. infestans and C. gloeosporioides. For B. cinerea, A. niger and P. infestans, the dsRNA-spore mixture was grown on PDA medium applied directly to the microscope slides surface up to 12 h. For V. dahliae and T. virens, the incubation period was extended up to 24-30 h since spore germination is slower in these species. S. sclerotiorum and R. solani mycelium plugs (5 mm in diameter) were placed on slides containing PDA medium and 20 µL of 150 ng/µL fluorescent dsRNA was added to the fungi, and incubated for 10-12 h before confocal imaging. All fungal growth slides were cultured in darkness at 25°C. Germinating spores or mycelium was treated with KCl buffer (control) or 75 U micrococcal nuclease enzyme (dissolved in KCl buffer) to degrade the dsRNA in the surface of spores or mycelium before observation at 37°C for 30 min. The fluorescent signal was analysed using a Leica SP5 confocal microscope.

External application of dsRNA on the surface of plant materials
For dsRNAs applied to S. sclerotiorum, C. gloeosporioides and P. infestans, the dsRNAs were adjusted to a concentration of 40 ng/µL with RNase-free water before use. For in vitro synthetic YFP-dsRNA, Ss-DCL1/2 and Ss-VPS51+DCTN1+SAC1-dsRNA, 20 µL of dsRNA (40 ng/µL) was dropped onto the surface of each plant sample. Then, 5 mm diameter S. sclerotiorum mycelium plugs were applied to the same site as the dsRNA treated area. A total of 20 µL dsRNA (40 ng/µL) were dropped onto the surface of each fruit sample and then C. gloeosporioides spores (1 9 10 5 per mL) were applied to the same dsRNA treated area. For in vitro synthetic YFP-, PiDCL1-, PiHMP1-and PiPGB1-dsRNA, 20 µL of dsRNA (40 ng/µL) was dropped onto the surface of potato leaves, then P. infestans zoospores were applied to the same dsRNA treated area. For dsRNAs that were applied in controlling B. cinerea and A. niger, the dsRNAs were adjusted to a concentration of 20 ng/µL for the fruits, vegetables and petals. For applications on tomato leaves, dsRNA (100 ng/µL) was sprayed onto the leaf surface. In all the above experiments, fungal inoculation was performed after dsRNA treatment as described in Wang et al., 2016. For the time-course experiments, fungal inoculation was performed at the days post-RNA treatment indicated in figure legends. For dsRNAs applied to control V. dahliae, Arabidopsis roots were dipped in a mixture of V. dahliae spores and the dsRNA (20 ng/µL) for 3 min (Ellendorff et al., 2009), then seedlings were planted in soil.

Supporting information
Additional supporting information may be found online in the Supporting Information section at the end of the article. Figure S1 Colony morphology of different types of fungi. Figure S2 Mycelium morphology of different types of fungi. Figure S3 Multiple fungal dsRNA uptake efficiencies were observed over a time course. Figure S4 Accumulation of fluorescein signal intensities over a time course in different fungal cells. Figure S5 Fluorescein-labeled YFP-dsRNA was observed in S. sclerotiorum protoplasts after MNase treatment. Figure S6 Accumulation of fluorescein signal intensities in different fungal cells. Figure S7 SsVPS51, SsDCTN1, and SsSAC1 knockout mutants displayed significantly reduced virulence compared to the wildtype strain. Figure S8 Topical application of pathogen gene-targeting dsRNAs had no effect on C. gloeosporioides infection and disease. Figure S9 Accumulation of fluorescein signal intensities in different P. infestans cells. Figure S10 YFP-, PiHMP1-, PiPGB1-, PiDCL1-, and Pi-DCTN1+SAC1-dsRNAs did not inhibit P. infestans infection on potato leaves. Table S1 Primers used in the study.