Silencing susceptibility genes in potato hinders primary infection with Phytophthora infestans at different stages

Abstract Most potato cultivars are susceptible to late blight disease caused by the oomycete pathogen Phytophthora infestans. Here we report that the genetic loss of host susceptibility is a new source of resistance to prevent or diminish pathogen infection. Previously, we showed that RNAi-mediated silencing of the potato susceptibility (S) genes StDND1, StDMR1, and StDMR6 leads to increased late blight resistance. The mechanisms underlying this S-gene-mediated resistance have thus far not been identified. In this study, we examined the infection process of P. infestans in StDND1-, StDMR1-, and StDMR6-silenced potato lines. Microscopic analysis showed that penetration of P. infestans spores was hampered in StDND1-silenced plants. In StDMR1- and StDMR6-silenced plants, P. infestans infection was arrested at a primary infection stage by enhanced cell death responses. Histochemical staining revealed that StDMR1- and StDMR6-silenced plants display elevated ROS levels in cells at the infection sites. Resistance in StDND1-silenced plants, however, seems not to rely on a cell death response as ROS accumulation was found to be absent at most inoculated sites. Quantitative analysis of marker gene expression suggests that the increased resistance observed in StDND1- and StDMR6-silenced plants relies on an early onset of salicylic acid- and ethylene-mediated signaling pathways. Resistance mediated by silencing StDMR1 was found to be correlated with the early induction of salicylic acid-mediated signaling. These data provide evidence that different defense mechanisms are involved in late blight resistance mediated by functional impairment of different potato S-genes.


Introduction
Late blight, caused by the oomycete Phytophthora infestans, is one of the most devastating diseases of potato worldwide. The pathogen produces sporangia, asexual spores that spread by wind and rain. Infection starts when a sporangium that has landed on a leaf germinates directly, or develops into a zoosporangium releasing zoospores that encyst and germinate. Thereafter, the emerging germ tube develops into an appressorium that penetrates the leaf cuticle and epidermis. Infection expands throughout the leaf tissue via hyphal growth, resulting in water-soaked lesions that turn black. After a few days, the infected leaf tissue becomes necrotic and starts to sporulate. Depending on environmental conditions, an unprotected potato field can be devastated within 10 days [1].
A strategy to control late blight is to grow resistant potato cultivars obtained by introgressing disease resistance (R) genes that are effective against diverse P. infestans isolates [1]. More than 35 potato R genes conferring late blight resistance have been identified to date, most of which encode nucleotide-binding domain, leucine-rich repeat-containing (NLR) proteins [2][3][4][5][6]. These intracellular proteins mount successive defense responses when they recognize corresponding avirulence (AVR) proteins secreted by the pathogen. These AVR proteins, known as RXLR effectors, display high evolutionary rates, and as a result P. infestans can rapidly escape NLR-mediated resistance, thereby limiting the durability of genetic resistance in cultivars possessing a single R gene [7,8]. Stacking of R genes is a potential strategy to achieve broadspectrum and durable late blight resistance [9], as shown in cultivar 'Sarpo Mira', which has at least five introgressed R genes (i.e. R3a, R3b, R4, Rpi-Smira1, and Rpi-Smira2) from different genetic sources [10]. There is still a risk that such R-gene-rich potato varieties are overcome by the pathogen if no regional resistance gene strategy is developed [10]. Next to R-gene-mediated resistance, infection by P. infestans can be halted by enhancing the overall host defense status, for example through transcriptional activation of defense-related genes, accumulation of defense metabolites and cell wall reinforcement [11][12][13][14].
Another approach to gaining disease resistance is impairment of plant susceptibility genes (S genes), a concept that has been exploited in the last decade to develop novel breeding strategies to control diverse crop diseases [15][16][17]. More than 150 S genes have been described in Arabidopsis, and there is increasing evidence that S-gene orthologs are present in diverse crop species [18][19][20]. In potato, for example, reduced expression via RNAi-mediated silencing of the S gene SYNTAXIN-RELATED1 (StSYR1), an ortholog of AtPEN1 in Arabidopsis, decreased susceptibility to P. infestans [21].
In our previous studies, we selected 11 Arabidopsis thaliana S genes and silenced their orthologs in the late blight-susceptible potato cultivar (cv.) 'Désirée'. RNAi-mediated silencing of six of these genes, which included StDND1, StDMR1, and StDMR6, resulted in increased resistance to P. infestans [15,18]. This showed that loss of function of a putative S gene can be exploited to hamper infection and leaf colonization by P. infestans, and can generate late blight resistance in potato.
Very recently, studies have shown that CRISPRgenerated mutants in these three individual genes led to resistance to various pathogens in different crops [37][38][39][40]. In previous studies, we showed that individual silencing of StDND1, StDMR1, and StDMR6 increases potato resistance against P. infestans. To analyze how these S-gene-silenced plants prime defense responses to arrest colonization we monitored the infection process of P. infestans by microscopic and histological examination, and determined expression profiles of defense marker genes at early infection stages. Our results show that P. infestans infection in diverse Sgene-silenced potato plants is hindered at different stages.

StDMR6-silenced plants
Previously, we demonstrated that silencing of StDND1, StDMR1, or StDMR6 in potato results in enhanced resistance to P. infestans isolate Pic99189 [15], and additionally to three genetically diverse isolates, Pic99177, USA618, and EC1, upon silencing of StDND1 [18]. In this study, we tested all three S-gene-silenced potato lines with multiple P. infestans isolates (Supplementary Table S1). The reduction in S-gene expression was quantified by qRT-PCR ( Fig. 1a) and resistance was assessed in detached leaf assays by measuring lesion size at 3-7 days postinoculation (dpi) in two independent experiments. All three tested isolates were able to establish infection on the susceptible control cv. 'Désirée' efficiently, resulting in sporulating lesions at 7 dpi (Fig. 1b). In contrast, lesion development on the S-gene silenced plants was largely hampered. StDND1-and StDMR6-silenced plants showed no lesion growth upon inoculation with Pic99177 and, when inoculated with the aggressive isolates USA618 and EC1, the lesions were significantly smaller as compared with those on cv. 'Désirée'. No lesion growth was observed on leaves of StDMR1-silenced plants inoculated with all three isolates ( Fig. 1b and c). Dark necrotic spots surrounding the inoculation site were visible on StDMR1and StDMR6-silenced plants at 6 dpi, whereas StDND1silenced plants did not show such response (Fig. 2a). These results confirm that silencing StDND1 leads to resistance to multiple P. infestans isolates [18] and show that the same holds true for the other two S genes, StDMR1 and StDMR6. Further, as in our previous studies [15,18] dwarfness was observed in StDMR1-silenced plants and autonecrosis in some of the StDND1-silenced plants. Disease symptoms on leaves of cv. 'Désirée' and S-gene-silenced lines after inoculation with the P. infestans isolates Pic99177, USA618, and EC1 at 7 dpi. Scale bars represent 10 mm. c Lesion development on leaves inoculated with P. infestans isolates Pic99177, USA618, and EC1. Three plants were used for each potato genotype, and each was spot-inoculated with 12 droplets of P. infestans inoculum. Asterisks indicate significant differences from the recipient cv. 'Désirée' ( * * P < .01; * * * P < .001). Two independent experiments were performed with similar results.

Hindered P. infestans infection in StDND1-, StDMR1-, and StDMR6-silenced potato plants
To follow the infection process in more detail, leaves were inoculated with zoospores of the transgenic P. infestans reporter strains 14-3-GFP and EY6 (GUS labeling) and microscopically monitored at different time points of infection [0-96 hours post-inoculation (hpi)]. To monitor strain EY6, a histological study was performed at seven time points (Supplementary Fig. S1). At the two earliest time points, i.e. 0 and 6 hpi, no zoospores were found on any of the genotypes, indicating that they were not attached to the leaf surface and thus were likely washed off during slide preparation. At 12 hpi, germinating cysts with primary appressoria were observed ( Fig. 2; Supplementary Fig. S2). Leaf tissue below infection sites on StDMR6-silenced plants started turning yellow at 12 hpi (Fig. 2). From 24 hpi onwards, intracellular hyphae were observed in leaves of cv. 'Désirée', whereas on StDMR1-and StDMR6-silenced plants cysts with a short germination tube and intercellular hyphae were found ( Fig. 2; Supplementary Fig.   S2). Exceptions were StDND1-silenced plants, on which no P. infestans was observed at any time point (Fig. 2).
Growth and proliferation of P. infestans were analyzed in detail by monitoring isolate 14-3-GFP at eight time points (Supplementary Fig. S1). Three hours after inoculation with P. infestans cysts, germination was observed on all plants, except on StDND1-silenced lines ( Supplementary Fig. S3a). On cv. 'Désirée' plants, hyphal elongation started at 16 hpi. Extensively branched hyphae with collapsed cells underneath became evident at 48 hpi ( Supplementary Fig. S3b). Mycelium developed sporangiophores from 72 hpi onwards, releasing numerous sporangia at 96 hpi. On StDMR1-and StDMR6-silenced plants, the development of P. infestans was similar to that observed on cv. 'Désirée' at 6 and 16 hpi (Supplementary Fig. S3a). However, at later time points hyphal elongation was arrested ( Supplementary Fig. S3b). This was associated with local cell death at the infection sites that became apparent at 16 hpi ( Supplementary Fig. S3a). On StDND1-silenced plants, only germinated cysts with short germ tubes were found ( Supplementary Fig. S3).

StDMR1-and StDMR6-silenced plants display early ROS accumulation upon P. infestans inoculation
To further determine the different interaction stages, we investigated ROS accumulation in leaves inoculated with P. infestans EY6 ( Supplementary Fig. S1 (Fig. 3; Supplementary Fig. S4).

Changes in defense-related gene expression in S-gene-silenced plants
To confirm the microscopic observations and validate our sampling procedure, we quantified pathogen biomass by qRT-PCR at various times after inoculation with strain EY6 (Fig. 4; Supplementary Fig. and Fig S5). In cv. 'Désirée' plants there was a steady increase in P. infestans biomass starting from 3 hpi, reaching a plateau at 24 hpi. In contrast, P. infestans biomass was low in leaves of RNAi plants, confirming increased resistance to P. infestans acquired by silencing of StDND1, StDMR1, and StDMR6 ( Fig. 4; Supplementary Fig. S5).
To determine which signaling pathways play a role in the acquired resistance, we examined the expression of several defense-related marker genes by qRT-PCR (Fig. 4 and; Supplementary Figs S6 and S8). These included StPR1, StPR10a, StNPR1, StWRKY1, and StPP13A2 for the salicylic acid (SA) pathway, StPIN2 for the jasmonate (JA) pathway, and StCHI9 for the ethylene (ET) pathway [41][42][43]. Induction of expression of SA marker genes was observed in cv. 'Désirée' starting from 12 hpi, in contrast to earlier induction in the S-gene-silenced lines. The JA-responsive marker StPIN2 was slightly upregulated in cv. 'Désirée' and StDND1-silenced plants at 48 and 72 hpi while no or hardly any expression was observed in StDMR1-and StDMR6-silenced plants. The ET marker gene StCHI9 showed a clear increase in expression in cv. 'Désirée' from 24 hpi onwards. An earlier, but weaker, induction of StCHI9 expression was detected in StDND1-and StDMR6-silenced plants. These data suggest that in StDMR1-silenced plants only the SA-mediated signaling pathway contributes to increased resistance to P. infestans, while in StDND1-and StDMR6-silenced plants ET-mediated signaling may contribute as well.
To further validate the oxidative burst and cell death responses observed in inoculated StDMR1-and StDMR6silenced plants, we also evaluated the expression of marker genes associated with these events (Fig. 4; Supplementary Figs S9 and S10). We included three genes encoding ROS-scavenging or ROS-generating enzymes, i.e. superoxide dismutase (StSOD2), ascorbate peroxidase (StAPX2), and catalase (StCAT1), and two cell deathrelated genes, StHSR3 and StATG8I. Compared with cv. 'Désirée', slightly higher induction of expression of these genes was observed in StDMR1-and StDMR6-silenced plants from 6 hpi onwards. In contrast, expression of these marker genes was hardly detectable in StDND1silenced plants (Fig. 4), and thus these expression patterns match the ROS accumulation observed in the respective RNAi plants (Figs 2 and 3; Supplementary  Fig. S3).

Discussion
This study shows that broad-spectrum resistance to P. infestans resistance can be efficiently achieved by silencing of StDND1, StDMR1, or StDMR6 in potato [5,18]. This is in line with several recent studies that realized resistance to a number of pathogens by CRISPR editing these three individual genes in different crops [37][38][39][40], demonstrating the importance of S-genes in breeding crops with improved disease resistance.
Compared with our previous studies and those reported in the literature, this study gathered insight into the infection process of P. infestans on S-gene-silenced potato plants by making use of GUS(β-glucuronidase)and GFP(green fluorescent protein)-labeled P. infestans strains. On StDND1-silenced plants, penetration of P. infestans was hampered and growth was arrested after cyst germination. On StDMR1-and StDMR6-silenced plants, growth of P. infestans was hindered after germ tube emergence, which was associated with cell death and ROS accumulation at the inoculation sites (Fig. 5). For monitoring the activity of defense-related signaling pathways we made use of marker genes. Resistance mediated by silencing StDMR1 was found to be correlated with the early induction of SA-mediated signaling. Although delayed germination and no penetration were observed on StDND1-silenced plants, it is intriguing to notice that, in both StDND1-and StDMR6-silenced plants, the increased resistance was associated with an early onset of SA-and ET-mediated signaling pathways. It thus seems that P. infestans germ tubes release signals that potentially induce defense signaling. The association between the microscopic observations and defense-related gene expression in the S-gene-silenced plants may offer leads for further exploration of the mechanisms underlying enhanced resistance to P. infestans.
The dnd1 mutant of Arabidopsis exhibits enhanced resistance to avirulent isolates of P. syringae, which is not associated with a hypersensitive response [27,28]. In agreement with this, our study demonstrated that a hypersensitive response does not contribute to the acquired resistance to P. infestans by silencing the potato DND1 ortholog. Instead, all (germinated) cysts of the GUS-expressing line EY6 were washed off during slide preparation, suggesting that cyst attachment to the plant surface of StDND1-silenced plants is blocked. This lack of attachment was also observed when the StDND1silenced plants were inoculated with B. cinerea [29]. It might be that the leaf surface of StDND1-silenced plants is chemically and/or physically altered. However, we have no direct evidence to prove the lack of cyst attachment. When no washing steps were included in preparing samples with the GFP-labeled isolate, cysts germinated and produced short germ tubes (Supplementary Fig. S3). Further, similar expression profiles were found for both the StDND1-and StDMR6-silenced plants. These observations may indicate that cysts do attach to the leaf surface of StDND1-silenced plants but are unsuccessful in penetration. A recent study showed that Phytophthora species, including P. infestans, use a slicing mechanism to invade their hosts. The tip of the germ tube transforms into a knife-like structure that cuts through the plant surface at an oblique angle [45]. Using this strategy, Phytophthora is able to penetrate without brute force and with minimal consumption of energy. This study further implied that according to the laws of mechanics Phytophthora is unable to penetrate the plant without first attaching itself tightly to the leaf surface. Taking this into consideration, we speculate that the leaf surface of StDND1-silenced plants is physically altered, and as a result the pathogen cannot attach and penetrate. The DND1 gene encodes a protein that is a member of the CNGC family but as yet there is no indication of direct involvement of CNGC family members in cell wall biogenesis or cuticle formation. CNGCs play roles in conducting Ca 2+ into plant cells and are involved in various physiological processes [27]. The Arabidopsis dnd1 mutant displays elevated SA levels, which are likely required for its resistance to a broad range of pathogens [23]. In the StDND1-silenced potato plants, we observed a constitutively elevated expression of SA marker genes indicative of elevated SA levels ( Fig. 5; Supplementary Fig. S6, [18]). The elevated levels of PR1 expression in StDND1-silenced plants may lead to increased content of PR1 protein, which has been shown to have inhibitory activity against P. infestans [44]. Further, the ET marker gene StCHI9 was induced upon P. infestans infection in the StDND1-silenced potato plants (Fig. 5: Supplementary Fig. S8). This suggests that the increased resistance to P. infestans found in the StDND1-silenced potato plants relies on both SA-and ET-mediated signaling pathways.
In Arabidopsis dmr1 mutants, pathogen resistance without visible hypersensitive response [35] to H. arabidopsidis, O. neolycopersici, Fusarium culmorum, and Fusarium graminearum was found to be associated with homoserine accumulation [32,33,35]. Our results show clear cell death of single epidermal cells of StDMR1-silenced potato plants at the site of P. infestans inoculation. This ensures pathogen arrest at a very early infection stage, when cysts start to germinate. P. infestans is a hemibiotroph that requires an initial biotrophic phase in which nutrients from living host cells are acquired via intracellular haustoria [46]. A rapid host cell death observed in StDMR1-silenced plants thus strongly hinders this crucial biotrophic growth phase (Fig. 3). In contrast to Arabidopsis dmr1 mutants [31], induction of SA marker gene expression and ROS accumulation were detected in the StDMR1silenced potato plants.
Arabidopsis dmr6 mutants were shown to have a reduced susceptibility to H. arabidopsidis, P. syringae (only at adult stage), and P. capsici [34][35][36]. In the dmr6 mutant, haustorium formation of H. arabidopsidis was found to be severely affected; haustoria had aberrant shapes and stayed immature [35]. Here we found that death of multiple cells occurred on StDMR6-silenced plants around 16 hpi after P. infestans inoculation, which was associated with induced expression of SA and ET marker genes. DMR6 encodes a 2-oxoglutarate Fe(II)-dependent oxygenase, suggesting that the observed resistance is caused by the accumulation of a toxic DMR6 substrate or the absence of a DMR6 metabolic product required for pathogen growth [34]. Later, it was shown this was accompanied by elevated SA levels, and that DMR6 functions in a feedback mechanism to tightly control the SA level [36].
Potato breeding for late blight resistance has so far relied on the introgression of dominant resistance (R) genes from crossable relatives [1]. However, R-genemediated resistance is not always durable; P. infestans can easily avoid recognition by mutating or deleting RXLR effector genes so that NLRs encoded by dominant R-genes are not activated to mount effective defense [2]. Thus, we previously proposed an alternative/additive strategy to enhance late blight resistance by impairing plant S-genes that are required by P. infestans for successful host colonization [47]. In potato, several of these S-genes have been identified, including genes encoding a KRBP1 protein [K-homology (KH) RNAbinding protein], three isoforms of the PP1c protein (phosphatase PP1 catalytic subunits) and an NRL1 protein [non-phototrophic hypocotyl 3/root phototropism 2 (NPH3/RPT2)-like protein]. These three host proteins are targeted by the P. infestans RXLR effectors Pi04314, Pi04089, and Pi02860 to promote infection [48][49][50]. Some S-genes function as negative regulators of plant defense, which are often upregulated by pathogens to suppress plant defense [19]. Therefore, silencing of Sgene expression by RNAi can avoid the suppression of plant defense by pathogens. RNAi is especially powerful in polyploid crops like potato since it results in a dominant inherited resistance [15]. Recent advances in genome editing, including CRISPR/Cas9 technology, make S-gene editing a promising strategy for resistance breeding, even in tetraploid potato and other polyploid crops [51].
The three S-genes studied here were not identified by screening potato for susceptibility factors towards late blight disease. Instead, they were pinpointed as potential candidates because they are orthologs of Sgenes identified in Arabidopsis. Our results showed that these orthologs can also be functionally conserved across plant species. For example, downregulation of DND1 expression in potato and tomato led to resistance to powdery mildew, late blight, and grey mold disease [18,29]. Downregulation of DMR1 expression in tomato and pepper was shown to enhance resistance to powdery mildew and P. capsici, respectively [33,52]. Similarly, tomato dmr6 mutants generated by CRISPR/Cas9 were found to be resistant to P. capsici, as well as P. syringae pv. tomato [53]. These results, including those presented here, show that impairment of orthologous S-genes in diverse crop species potentially leads to broadspectrum resistance to multiple/diverse pathogens. However, further studies are needed to fully unravel the molecular basis of resistance mediated by lossof-function mutation of S-genes in different plant species.

Pathogen growth and inoculum preparation
The P. infestans isolates used in this study (Supplementary  Table S1) are genetically diverse isolates, including Pic99177, USA618, EC1, EY6 (88069), and 14-3-GFP (H30P02). They were cultured on rye sucrose agar medium in the dark at 15 • C for 10-14 days. Agar plates fully covered with sporulating mycelium were flooded with cold water (4 • C). Sporangiospore suspensions were harvested and incubated at 4 • C for 1-2 hours to induce zoospore release. Zoospores were isolated by filtration through 15-μm nylon mesh. Experimental procedures are visualized in Supplementary Fig. S1.

Microscopy and histochemical staining
Host penetration and infection by P. infestans was assessed by imaging leaf disks (diameter 10 mm) punched out from spot-inoculated leaf areas. Bright-field images were taken with a Zeiss Stemi 508 stereomicroscope equipped with a digital camera.

Nucleic acid extraction and gene expression analyses
Leaf samples were f lash-frozen, ground in liquid nitrogen, and stored at −80 • C. Total RNA was extracted using a MagMAX-96 Total RNA Isolation Kit (Ambion) and treated with RNAse-free DNAse (Qiagen). RNA concentrations were measured with an Isogen Nanodrop Spectrophotometer ND-1000. Synthesis of cDNA was performed on 1 μg of total RNA using an iScript cDNA Synthesis Kit (Bio-Rad). Quantitative RT-PCR was performed on a C1000™ Thermal Cycler PCR system (Bio-Rad) using iQ SYBR Green Supermix (Bio-Rad). Genespecific primers were designed with the online tool Primer3 (Supplementary Table S2). Expression levels of marker genes were quantified with the 2 -Ct method using StEF1α and Piβ-tubulin expression for normalization [15]. Quantitative RT-PCR assays were performed on three independent biological samples, each consisting of three technical replicates.

Statistical analysis
Data presented are means of at least three biological replicates with error bars indicating standard deviation. Statistical analyses for detached leaf assays and gene expression were conducted using one-tailed t-tests and Duncan's multiple range test performed in SPSS, respectively.