Silencing of six susceptibility genes results in potato late blight resistance

Phytophthora infestans, the causal agent of late blight, is a major threat to commercial potato production worldwide. Significant costs are required for crop protection to secure yield. Many dominant genes for resistance (R-genes) to potato late blight have been identified, and some of these R-genes have been applied in potato breeding. However, the P. infestans population rapidly accumulates new virulent strains that render R-genes ineffective. Here we introduce a new class of resistance which is based on the loss-of-function of a susceptibility gene (S-gene) encoding a product exploited by pathogens during infection and colonization. Impaired S-genes primarily result in recessive resistance traits in contrast to recognition-based resistance that is governed by dominant R-genes. In Arabidopsis thaliana, many S-genes have been detected in screens of mutant populations. In the present study, we selected 11 A. thalianaS-genes and silenced orthologous genes in the potato cultivar Desiree, which is highly susceptible to late blight. The silencing of five genes resulted in complete resistance to the P. infestans isolate Pic99189, and the silencing of a sixth S-gene resulted in reduced susceptibility. The application of S-genes to potato breeding for resistance to late blight is further discussed. Electronic supplementary material The online version of this article (doi:10.1007/s11248-016-9964-2) contains supplementary material, which is available to authorized users.


Introduction
The plant immune system comprises an intricate network of active and passive mechanisms that successfully prevent the colonization of a host by a pathogen (Jones and Dangl 2006) (Fig. 1). In many cases, defence is actively triggered upon first contact between a plant and pathogen. Plasma membrane receptors perceive pathogen-associated molecular patterns (PAMPs) or apoplastic effectors (Fig. 1). This perception leads to intracellular signal transduction events, culminating in defence responses that, when effective, induce PAMP-triggered immunity (PTI). A known example is the receptor-like protein ELR (elicitin response). ELR was isolated from the wild potato species Solanum microdontum and can mediate the broad-spectrum recognition of elicitins (referred to as oomycete PAMPs) from several Phytophthora species (Du et al. 2015). The second layer of defence relies on proteins encoded by resistance genes (R-genes) that recognize intracellular avirulence (Avr) effectors. This recognition results in effector-triggered immunity (ETI).
Potato late blight, caused by Phytophthora infestans, is considered to be the most serious potato disease worldwide. An asexual cycle of sporangial proliferation is completed within 5 days (Fry 2008). Depending on environmental conditions, an unprotected potato field with a susceptible cultivar (cv.) can be devastated within 10 days after infection with P. infestans (Fry 2008). The control of potato late blight is dependent on fungicide sprays and the use of cultivars carrying dominant R-genes (Haverkort et al. 2009). All late blight R-genes identified thus far belong to the coiled-coil nucleotide binding leucinerich repeat (CC-NB-LRR or NLR) class and reside inside the plant cell where these genes recognize P. infestans avirulence effectors (Avr) of the RxLR class.
Because the resistance conferred by R-genes is, in general, race-specific, ETI can be broken due to the rapid evolution of pathogen effectors in agricultural practice (Vleeshouwers et al. 2011). For example, Rpi-vnt1, isolated from S. venturii, confers resistance to a broad spectrum of P. infestans lineages and races in European potato growing areas (Pel et al. 2009). However, this resistance can be overcome by the EC1 lineage, which is abundant in Ecuador. Thus, in addition to exploiting dominant R-genes, combinations of resistance traits that are effective against the prevailing P. infestans population are needed for durable late blight resistance. Therefore, pyramiding ETI receptors is expected to enhance resistance durability (Kim and Hwang 2012;Rietman et al. 2012;Zhu et al. 2013).
The resistance conferred by R-genes is based on pathogen recognition followed by the induction of defence responses. Another type of resistance, based on the loss-of-function of a susceptibility gene (Sgene), has recently been introduced (Eckardt 2002; Pavan et al. 2011). Plant genes are termed S-genes when a pathogen takes advantage of them for its own Fig. 1 Plant innate immunity: PAMP-triggered immunity (PTI) and effector-triggered immunity (ETI). Apoplastic pathogen-associated molecular patterns (PAMPs/apoplastic effectors), intracellular effectors or modified effector targets are actively perceived by receptors in the plasma membrane or resistance (R) proteins in the cytoplasm, resulting in the activation of PAMP-triggered immunity (PTI) and effectortriggered immunity (ETI). Helper proteins and guard proteins/ decoys are involved in the co-perception of pathogen-derived components (Césari et al. 2014). Pathogens use host proteins (S proteins) encoded by plant susceptibility genes (S-genes) to facilitate entry and growth, resulting in host compatibility (Doehlemann and Hemetsberger 2013). MLO,CESA3 and SWEETs are examples of S proteins in class 1, 2, or 3 according to van Schie and Takken (2014). Class 1 genes provide features that facilitate the entrance of a pathogen, class 2 genes increase innate immunity when the gene is disabled, and class 3 genes encode substrates essential for the pathogen benefit during the colonization of the plant. Pathogens are impeded from colonizing the plant when these genes become dysfunctional as a result of a recessive mutation or non-expression. Thus, when disabled, Sgenes can induce a resistance phenotype resembling that of healthy plants. For example, the loss-offunction mutant in the Powdery Mildew Resistance 6 (PMR6) gene prevents powdery mildew growth (Vogel et al. 2002). Based on the mechanism, S-genes have been divided into three classes (van Schie and Takken 2014). The first class comprises the genes required in early pathogen infection steps. One example is the maize (Zea mays) wax mutant glossy 11, which limits powdery mildew (Blumeria graminis) spore germination (Hansjakob et al. 2011, Fig. 1). The second class of S-genes encodes negative regulators of plant immunity, such as the CesA3 gene, which is involved in cellulose synthesis (Ellis et al. 2002;Ellis and Turner 2001, Fig. 1). The presence of homozygous recessive mutant alleles of the CesA3 gene can confer resistance to multiple pathogens, as a result of an increase of in planta levels of abscisic and jasmonic acid and ethylene. The third class of S-genes includes genes involved in pathogen sustenance, such as metabolite biosynthesis and sugar transport. For example, the Downy Mildew Resistant 1 (DMR1) gene encodes homoserine kinase (HSK). Arabidopsis dmr1 mutants are resistant to the downy mildew fungus Hyaloperonospora parasitica (Huibers et al. 2013;Van Damme et al. 2008;van Damme et al. 2009) and the fungi Fusarium graminearum and F. culmorum, which cause Fusarium Ear Blight (FEB) disease on small grain cereals (Brewer et al. 2014). Other examples include the genes encoding SWEET proteins, identified as factors that are required for susceptibility to Xanthomonas oryzae, which provide a carbon source to the pathogen (Chen et al. 2010;Streubel et al. 2013;Yuan et al. 2010, Fig. 1).
The aim of the present study was to impair S-genes in plants to obtain durable resistance (Pavan et al. 2010). In tomato, we showed that silencing DMR1 and PMR4 orthologues resulted in resistance to the powdery mildew fungus Oidium neolycopersici (Huibers et al. 2013), suggesting that S-genes, such as PMR4 and DMR1, are conserved among plant species and that impaired orthologues confer resistance to the associated pathogens in other plant species. This finding prompted us to identify potato orthologues of S-genes reported in other plant species (particularly A. thaliana) and to generate a proof of the S-gene concept in potato. We silenced 11 S-genes in potato via RNAi and showed that the silencing of five of these genes conferred complete resistance to late blight in potato, while the silencing of a sixth gene only conferred reduced susceptibility to this disease.

Identification of potato orthologues of S-genes
Previously, tomato orthologues of 11 A. thaliana Sgenes were identified as described in Huibers et al. (2013). Briefly, A. thaliana protein sequences were used as a query in a TBLASTN programme against the SGN Tomato Combined database (http://solgenomics. net/tools/blast/) to search for homologous sequences. The tomato and Arabidopsis amino acid sequences were aligned, and the tomato sequences that showed a high level of homology with the A. thaliana S-genes were considered orthologues. To identify potato orthologues, A. thaliana protein sequences were used in a BLASTP analysis at the Spud DB Potato Genomics Resource website (http://solanaceae.plantbiology. msu.edu/blast.shtml). Subsequently, protein and mRNA sequences with the lowest E-values were downloaded. Next, phylogenetic analyses were performed by aligning A. thaliana, tomato and potato protein sequences in MEGA5.1 (Tamura et al. 2011) using standard parameters ( Supplementary Fig. 1). The tomato and potato genes showing the highest level of homology (on amino acid level) to A. thaliana Sgenes were considered to be S-gene orthologues and were indicated with the prefix Sl-or St-to represent S. lycopersicum and S. tuberosum, respectively.

Construction of the silencing vectors
The binary vector pHellsgate8 (CSIRO, Australia) has a kanamycin resistance gene as a selectable marker and was used to generate RNAi constructs Waterhouse and Helliwell 2003). This vector contains a CaMV 35S promoter which drives the expression of the inverted repeat. For all genes, except SR4, DND2, and PMR4, primers were designed to amplify fragments of the target genes ranging from 150 to 300 bp from tomato cDNA sequences of the cv. Moneymaker (Supplementary Table 1). For SR4 and DND2, primers were designed to amplify potato cDNA sequences of the cv. Desiree. The forward primer contained CACC at the 5 0 end for directional cloning into the pENTR/D-TOPO vector (Thermo Fisher Scientific). Total RNA was isolated using a RNeasy kit from Qiagen (Germany). RNA was treated with RNAse-free DNase (Qiagen). Subsequently, this RNA was used as a template for cDNA synthesis using an iScript cDNA synthesis kit (Bio-Rad). The PCR products were amplified using Phusion High-Fidelity DNA polymerase (Thermo Fisher Scientific), inserted into the pENTR/D-TOPO cloning vector and transformed into One-Shot TOP10 E. coli cells. The plasmid DNA of the clones was sequenced to verify the insert. To generate the SlPMR4 silencing construct (Huibers et al. 2013), we synthesized a 101-bp DNA fragment that was identical to the first 97 bp of the predicted coding sequence of Soly-c07g053980 (the tomato PMR4 ortholog, Supplementary Fig. 1) and contained CACC at the 5 0 end flanked by attL sites in pUC57 (Genscript, USA). For PMR6, the silencing fragment was based on the coding sequence of tomato Solyc05g014000, but could potentially cross-silence multiple potato PMR6-like homologs (Sotub05g015080, Sotub11g012470, Sotub06g029220 and Sotub03g023350). The primer sequences are provided in Supplementary Table 1. For expression in planta, the RNAi fragments were transferred from the entry clone through an LR clonase reaction to the pHellsgate8 vector.

Growth and development of transgenic potato plants
The tetraploid potato cv. Desiree (susceptible to late blight) was used for transformation according to the protocol of Visser et al. (1991). The transformants were transferred from MS medium (Murashige and Skoog 1962) supplemented with vitamins, 30 g l -1 sucrose and 100 mg l -1 kanamycin to similar fresh MS medium without kanamycin. After 3 weeks of growth at 24°C and a light intensity of 100 W/m 2 , the rooted transformants were transferred to plastic pots containing potting soil in a growth compartment at 21 and 19°C during 16-h days and 8-h nights, respectively. For each RNAi construct, more than eight independent primary transformants were randomly selected and cultured in a greenhouse and subsequently tested for resistance. Three biological replicates were grown for each transformant.

Pathogen inoculations and detached leaflet assay (DLA)
The P. infestans isolate Pic99189 (race 1.2.5.7.10.11) (Flier et al. 2002) was used in the present study. For each experiment, the isolate was grown on rye agar medium supplemented with 2 % sucrose for 10-15 days at 15°C in closed Petri dishes to induce sporangia formation (Caten and Jinks 1968). To release zoospores from sporangia, ice-cold tap water was added to the Petri dishes, followed by incubation for 3 h at 4°C. The zoospore concentration was assessed by bright field microscopy using a Fuchs-Rosenthal counting chamber and adjusted to 5 9 10 4 spores/ml. The resistance of potato RNAi transformants to Pic99189 was examined using a 10-ll droplet inoculation in detached leaflet assays (DLA) (Vleeshouwers et al. 1999). The leaves were harvested from plants after 5-6 weeks of greenhouse growth. The fourth or fifth fully developed leaf (counted from the top) was used. The lesion diameters were measured from 3-6 days post-inoculation using an electronic calliper (Helios DIGI-MET Ò ).

RNA isolation and quantitative real-time (qRT)-PCR
The kanamycin-resistant transformants were confirmed by PCR using Fw-NPTII and Rv-NPTII primers (Supplementary Table 1). The PCR-positive transformants were transferred to the greenhouse. More than eight independent transformants were randomly selected per gene, and the silencing levels of the transformants were evaluated by qRT-PCR using gene-specific primers (Supplementary Table 1, -qPCR), producing products of approximately 200 bp. Plant total RNA was extracted using a MagMAX-96 total RNA Isolation kit (Ambion). The quantity of the isolated RNA was measured using a Nanodrop Spectrophotometer ND-1000 (Isogen), and the cDNA was produced using an iScript cDNA synthesis kit (Bio-Rad). qRT-PCR was performed in triplicate using a C1000TM Thermal Cycler PCR system (Bio-Rad) with iQ SYBR Green supermix (Bio-Rad). The potato EF1a (Sotub06g010680) transcript was used as an internal control to determine the relative transcript levels. The relative level of gene expression was calculated using the 2 ÀDDC t method (Livak and Schmittgen 2001;Nicot et al. 2005). For the qRT-PCR assay, three technical replicates were included for each experiment, and the expression of each gene was investigated in three biological replicates.

Identification of potential potato S-gene orthologues
To identify potato orthologues of the 11 S-genes listed in Table 1, we used the amino acid sequences of A. thaliana in a BLAST analysis of the potato sequence database. Potato sequences with an amino acid identity higher than 75 % were selected and used in phylogenetic studies ( Supplementary Fig. 1). Based on multiple sequence alignments, sequences showing the highest degree of homology with the S-gene in A. thaliana were considered to be potential orthologues in potato (Table 2, column 2). The closest homolog to AtSR1 in the potato database was Sotub01g012330, and silencing fragments were designed for this gene. However, when a subsequent TBLASTN search was conducted using the NCBI database, the closest homolog of AtSR1 in the potato RefSeq_RNA database was XM_006355276.1, which corresponds to Sotub04g020530. This gene on chromosome 4 was closer to AtSR1 in the phylogenetic tree than to Sotub01g012330 ( Supplementary Fig. 1). Because Sotub01g012330 was closer to AtSR4 and SlSR4 (Yang et al. 2012), we referred to Sotub01g012330 as StSR4 in Table 2.
Silencing of six potato S-genes results in reduced susceptibility to P. infestans To assess the significance of these potential S-genes for susceptibility to P. infestans in potato, RNAi constructs of all of the selected potato orthologues were generated and used to transform the potato cv. Desiree, which is susceptible to late blight, using Agrobacterium tumefaciens-mediated transformation. From each of the 11 S-genes, several independent transformants with highly reduced transcript levels ([60 %) of the targeted S-gene were selected (Table 2, column 3, Supplementary Fig. 2). These transformants were further indicated as well-silenced transformants, and those with no or a low reduction in the transcript level (\60 %) of the targeted S-gene were considered to be negative controls in the following experiments (Table 2, column 4, Supplementary Fig. 2).
To determine whether the silencing of these genes impacted the susceptibility to late blight, the leaves of 4-to-5-week-old silenced plants were inoculated with the P. infestans isolate Pic99189. The leaves of cv. Desiree and A13-013 (a transformant containing the dominant resistance gene Rpi-vnt1.1 in the genetic background of cv. Desiree and conferring resistance to Pic99189, Zhu et al. 2013) were used as susceptible and resistant controls, respectively. The inoculated leaves were visually inspected (Fig. 2a-c). 3-6 days post inoculation (dpi), the lesion size was measured, and lesion growth over time was plotted ( Fig. 2d-f). Large lesions were observed on the inoculated leaves of the cv. Desiree and transformants in which StBIK1, StCPR5, StDND2, StPMR5 or StPMR6 were silenced (Fig. 2a). The lesion size showed a steady increase from 3 to 6 dpi (Fig. 2d).
Smaller lesions were observed on inoculated leaves of StPMR4-silenced potato plants compared with those of Desiree (Fig. 2b). At 6 dpi, the inoculated parts of Desiree leaves were completely blighted with obvious sporulation, while on the leaves of RNAi::StPMR4 potato plants, the sporulation was only visible through a binocular. The lesion growth on the leaves of these transformants was much slower than on Desiree leaves (Fig. 2e). Thus, the silencing of StPMR4 resulted in reduced susceptibility to P. infestans.
In contrast, up to 6 dpi, no lesion growth was visible on the leaves of resistant A13-013 control plants and transformants in which the five individual S-genes (StCESA3, StDMR1, StDMR6, StDND1 and StSR4) were silenced (Fig. 2c, f). For each of these five genes, at least two independent well-silenced transformants (with three plants per transformation event) were tested ( Table 2). All inoculated leaves showed similar results.
Fitness costs associated with S-gene silencing Altered phenotypic characteristics in non-inoculated plants were observed for some RNAi transformants,  (2014) such as reduced growth (dwarfing), necrosis, and lighter green leaves (colour loss, Table 2). StDMR1silenced plants showed reduced growth and light green leaves compared with cv. Desiree (Fig. 3a, b), whereas StDND1-silenced plants displayed auto-necrotic spots only in the leaves of older plants (Fig. 3c). Plants of a few well-silenced StDND1-tranformants also showed dwarfing, but not as severe as that of StDMR1-silenced transformants. The light green leaf colour and dwarfing were observed for plants of all of the well-silenced RNAi::StCPR5 transformants (Fig. 3d), and these plants were susceptible to late blight (Fig. 2a).

Discussion
Impaired potato S-genes: a new source of durable resistance against P. infestans?
Potato is the third largest food crop in the world, with efficient usage of resources and high nutritional quality. For over 180 years, the most devastating disease in potato is late blight disease, caused by P. infestans, which costs billions of Euros annually for crop protection to prevent crop losses. Many potato Rgenes have been cloned and characterized, and some R-genes have been used in breeding. Still, durable resistant potato varieties have not been generated, reflecting new virulent P. infestans races that render resistance genes ineffective. Only a few potato varieties carrying multiple R-genes show a sufficient level of durable resistance, such as the cultivar Sarpo Mira, which possesses multiple major R-genes (Rietman et al. 2012;Tomczyńska et al. 2014). Thus, pyramiding R-genes show the potential to achieve durable resistance.
The results of the present study showed that silencing the potato orthologues of six A. thaliana Sgenes resulted in complete or partial resistance to late blight in potato. These six S-genes were not originally identified as susceptibility factors towards late blight. Thus, these results showed that orthologues of the Sgenes in A. thaliana might be functionally conserved across plant species. Furthermore, the impairment of orthologues of the same S-gene in different plant species could potentially result in resistance to different pathogens.
Because the S-gene concept is relatively new, it is not clear whether the resistance resulting from impaired S-genes is durable. Several natural loss-offunction alleles of plant S-genes have been identified in agriculture as providing durable disease resistance (Pavan et al. 2010). A well-known example is the barley mlo mutant for non-race specific resistance to the powdery mildew B. graminis, which has been successfully used in European agriculture for more than 35 years (Büschges et al. 1997;Lyngkjaer et al. 2000). Natural mutant alleles of the MLO gene have been identified in several plant species, such as tomato (Bai et al. 2008), pea (Pavan et al. 2011) and cucumber (Berg et al. 2015), which confer resistance against different powdery mildew species. Whether impaired potato S-genes are a new source of durable resistance to late blight needs further study. We are currently testing the RNAi transformants with different races of P. infestans to verify whether resistance is broad spectrum. Moreover, an examination of the resistance mechanisms associated with impaired S-genes might shed light on the resistance spectrum and durability (van Schie and Takken 2014).
Fitness costs associated with impaired potato Sgenes Plant S-genes exploited by pathogens during infection might have other intrinsic functions that are important for the plant. Therefore, impairing the function of particular S-genes might have adverse effects. Prior to using impaired S-gene alleles in crop breeding, potential negative side effects should be identified and reduced or even prevented. Mutant alleles of the 11 S-genes, except DMR6 and PMR4, were associated with phenotypic changes in A. thaliana (Table 1).
Remarkably, the silencing of only three of the 11 tested S-gene orthologues generated negative pleiotropic effects in potato ( Table 2), suggesting that fitness costs could be plant species-dependent (Sun et al. 2016). For the RNAi constructs in the present study, we used the cauliflower mosaic virus 35S promoter, which is a strong and constitutively active promoter in most tissues. It is likely that the use of native S-gene promoters could reduce or prevent the negative side effects on plant performance. In the present study, we identified six S-genes whose impairment by RNAi reduced susceptibility to late blight, of which four S-genes were not associated with negative side effects under the conditions tested ( Fig. 2; Table 2). Further studies are needed to determine whether pleiotropic effects are also absent under conditions relevant to agricultural practices.
Strategies for exploiting impaired S-genes in potato breeding As shown in the present study, silenced S-genes are a new source of resistance to late blight. Loss-offunction mutations in S-genes lead to recessively inherited resistance, which is difficult to exploit in an autotetraploid crop, such as potato. In the heterozygous (wild) potato gene pool, only dominant R-genes have been identified. In contrast, the resistance conferred by recessive alleles of S-genes is not readily selected in both the evolution and breeding of heterozygous species, such as potato. As shown in the present study, the RNAi approach can be applied to knockdown the expression of S-genes, which leads to resistance. RNAi has previously been applied in potato to silence the Granule-Bound Starch Synthase I (GBSSI) gene to produce potatoes with amylose-free starch (Visser et al. 1991). Currently, advanced geneediting technologies are available for the targeted modification of all alleles of a gene. For example, TALEN-induced mutation and CRISPR-Cas9 technology have been used to generate mutant alleles of the MLO gene in all three genomes of allohexaploid wheat (Wang et al. 2014). In addition to these advanced techniques, F 1 hybrid potato breeding (Lindhout et al. 2011) will be advantageous for the use of mutated S-genes. With a homozygous diploid line, knockout mutations can be obtained by mutagenesis, such as ethyl methanesulphonate (EMS) treatment. Thus, a realistic outlook towards late blight resistance is within reach by using S-gene mutants with advanced gene-editing technology or EMS in hybrid potato breeding.