WRR4B contributes to a broad‐spectrum disease resistance against powdery mildew in Arabidopsis

Abstract Oidium heveae HN1106, a powdery mildew (PM) that infects rubber trees, has been found to trigger disease resistance in Arabidopsis thaliana through ENHANCED DISEASE SUSCEPTIBILITY 1 (EDS1)‐, PHYTOALEXIN DEFICIENT 4 (PAD4)‐ and salicylic acid (SA)‐mediated signalling pathways. In this study, a typical TOLL‐INTERLEUKIN 1 RECEPTOR, NUCLEOTIDE‐BINDING, LEUCINE‐RICH REPEAT (TIR‐NB‐LRR)‐encoding gene, WHITE RUST RESISTANCE 4 (WRR4B), was identified to be required for the resistance against O. heveae in Arabidopsis. The expression of WRR4B was upregulated by O. heveae inoculation, and WRR4B positively regulated the expression of genes involved in SA biosynthesis, such as EDS1, PAD4, ICS1 (ISOCHORISMATE SYNTHASE 1), SARD1 (SYSTEMIC‐ACQUIRED RESISTANCE DEFICIENT 1) and CBP60g (CALMODULIN‐BINDING PROTEIN 60 G). Furthermore, WRR4B triggered self‐amplification, suggesting that WRR4B mediated plant resistance through taking part in the SA‐based positive feedback loop. In addition, WRR4B induced an EDS1‐dependent hypersensitive response in Nicotiana benthamiana and contributed to disease resistance against three other PM species: Podosphaera xanthii, Erysiphe quercicola and Erysiphe neolycopersici, indicating that WRR4B is a broad‐spectrum disease resistance gene against PMs.

mycelium and asexual propagation structures (conidiophores and conidia), which become apparent on the surface of aboveground plant organs at later stages of infection.
PM fungi secrete a large number of effector proteins into plant cells to manipulate the physiological processes of its host to facilitate its invasion (Thordal-Christensen et al., 2018).
Correspondingly, some plants have evolved dominantly or semidominantly inherited resistance (R) genes that detect pathogen effectors inside the plant cell to trigger a hypersensitive response (HR) associated with local host cell death to limit the spread of the fungus (Bourras et al., 2015;Jones & Dangl, 2006;Lu et al., 2016;Praz et al., 2017).These R genes typically encode canonical nucleotidebinding leucine-rich repeat (NB-LRR/NLR) proteins and represent the most variable gene family in plants (Jacob et al., 2013;Jones et al., 2016).Whole-genome sequencing has indicated that higher plant species possess anywhere from 50 (papaya, Carica papaya) to over 1500 (wheat, Triticium aestivum) NLR genes (Gao et al., 2018;Porter et al., 2009), with more than 150 NLR genes in Arabidopsis thaliana (Meyers et al., 2003).The expression levels of NLRs are strictly regulated (Lai & Eulgem, 2018), with low levels observed in healthy plants and highly variable levels in plants challenged by particular pathogens (van Wersch et al., 2020).The NLR protein must reach a certain abundance (dose) to activate defence signalling, but the overexpression of NLRs often results in autoimmunity, with massive fitness costs (Bieri et al., 2004;Holt et al., 2005;Lai & Eulgem, 2018).The N-terminal coiled-coil (CC) or Toll/interleukin-1 receptor (TIR) domains are used to classify plant NLRs into two main groups, termed CC-NLRs (CNLs) and TIR-NLRs (TNLs) (Meyers et al., 2003;Monteiro & Nishimura, 2018;van Wersch et al., 2020).Both the CC and TIR domains play key roles in the formation of oligomers that are essential for the activation of NLRs and the transduction of downstream immune responses (Bernoux et al., 2011;Maekawa, Cheng, et al., 2011).Generally, CNLs are present in monocot and dicot plants, whereas TNLs are only represented in dicots (Jacob et al., 2013).
Several key components involved in effector-triggered immunity (ETI) signalling were isolated through different genetic screens.The first identified gene involved in ETI signalling was NON-RACE-SPECIFIC DISEASE RESISTANCE 1 (NDR1), which was found to be necessary for resistance against bacterial and fungal pathogens carrying a variety of effectors (Century et al., 1995).
In addition, the feedback loop also upregulates the expression of pipecolic acid (N-hydroxyl-pipecolic acid [NHP])-biosynthesis genes, such as FLAVIN-CONTAINING MONOOXYGENASE 1 (Liu et al., 2020), enabling mutual reinforcement and protection against pathogen interference.
Our previous studies have shown that the PM species O. heveae HN1106 triggers the HR in Arabidopsis through EDS1 and PAD4 but not on NDR1, suggesting that TNL-triggered immunity is probably involved in the resistance against O. heveae (Mei et al., 2016).To date, however, none of the canonical NLR genes have been identified to confer resistance to PMs in Arabidopsis.In this study, one typical TNL gene, WHITE RUST RESISTANCE 4 (WRR4B), was found to be involved in the disease resistance against O. heveae.The expression of WRR4B was upregulated, and WRR4B positively controlled the EDS1, PAD4 and SA pathways during O. heveae infection.Furthermore, overexpressing WRR4B induced cell death in Nicotiana benthamiana in an NbEDS1-dependent manner.These findings lay the foundation for further study of the molecular mechanisms by which WRR4B protects against PMs.

| WRR4B contributes to the resistance against O. heveae HN1106
O. heveae infection triggers host disease resistance mechanisms dependent on EDS1 and PAD4, but not on NDR1 (Mei et al., 2016), suggesting that TIR-NBS-LRR proteins are responsible for O. heveae-induced resistance in Arabidopsis.To identify the genes responsible, we performed an RNA-sequencing (RNA-seq) analysis of wild-type (WT) A. thaliana Col-0 plants inoculated with O. heveae HN1106 at 0 and 2 days post-inoculation (dpi) and found that 12 genes encoding TIR-NBS-LRR proteins were differentially expressed in the infected plants, with nine being upregulated and three being downregulated (Table S1).To determine whether these genes are involved in the resistance against O. heveae, homozygous T-DNA insertion lines were obtained for each of the TIR-NBS-LRR genes (Table S1).These mutants, together with the Arabidopsis WT Col-0 and eds1 mutants, were inoculated with O. heveae.At 6 dpi, no obvious changes were observed on the Col-0 leaves.In contrast, the leaf surface of one mutant (SALK_040895C, wrr4b-1) displayed mild white cotton-like symptoms (Figure 1a), but the other 11 T-DNA insertion lines appeared to be unaffected (Table S2).The eds1 mutant developed large and coalesced white patches over the entire leaf surface (Figure 1a).Next, we examined the fungal growth of O. heveae on the WT and mutants.Trypan blue staining showed that O. heveae formed dense hyphal networks and many conidiospores on wrr4b-1 and eds1 (Figure 1b,e), but only limited fungal growth and no conidiospore formation was observed on WT (Figure 1b) and the other 11 T-DNA insertion mutants (Table S2).
The SALK_040895C T-DNA insertion is located within the second exon of AT1G56540 (WRR4B), which was previously reported to be involved in the resistance against the white rust pathogen Albugo candida (Cevik et al., 2019).To further confirm the function of WRR4B in the resistance against O. heveae, another T-DNA insertion line, SALK_072335C (wrr4b-2), was inoculated with O. heveae.
By 6 dpi, wrr4b-2 also displayed a PM-susceptible phenotype with white patches on the leaf surface (Figure 1a); however, fewer conidiospores on the wrr4b-1 and wrr4b-2 mutants than on the eds1 mutant were formed (Figure 1b,e).
In addition, we tested whether WRR4B is required for the early disease resistance during O. heveae infections in Arabidopsis.
Successful host cell penetration and hyphal growth are early events in PM pathogenesis.We calculated the plant cell penetration ratios and measured the length of fungal growth in the early stages of infection.By 1 dpi, there was a host cell entry rate of about 60% on the wrr4b-1 and wrr4b-2 mutants, in contrast to 37% in Col-0 and 78% in the eds1 mutant (Figure 1c).By 2 dpi,

| O. heveae HN1106 triggers cell death and H 2 O 2 production on wrr4b mutants in the later stage of infection
A previous study has shown that O. heveae triggers chlorosis symptoms and induces cell death and H 2 O 2 production on WT Col-0 (Mei et al., 2016).To determine whether O. heveae can also induce these defence responses in wrr4b mutants, WT Col-0, wrr4b-1, wrr4b-2 and eds1 mutants were inoculated with O. heveae HN1106.By 12 dpi, chlorosis symptoms occurred on the edge of the WT Col-0 leaves, and the typical white powder appeared on the whole leaf surfaces of the eds1 mutant (Figure 2a), which is consistent with the previous report (Mei et al., 2016).Unlike the WT Col-0 and eds1 mutants, the wrr4b-1 and wrr4b-2 mutants developed large necrosis patches with a few white powders across the leaf surfaces (Figure 2a).Microscopy analysis showed that dense hyphal networks, some conidiospores and mature fungal spores were observed on wrr4b-1, wrr4b-2 and eds1 mutants but not on WT Col-0 and the other 11 TIR-NB-LRR T-DNA insertion lines (Figure 2b, Table S2).Fewer mature fungal spores were present on the wrr4b-1 and wrr4b-2 tissues than were detected for the eds1 mutant (Figure 2b).In addition, the patches of cell death and H 2 O 2 production induced by O. heveae on the wrr4b-1 and wrr4b-2 mutants were both slightly greater than those induced on WT (Figure 2c,d).The eds1 mutant did not display fungal-induced cell death and H 2 O 2 production (Figure 2c,d).These results show that O. heveae still triggers apparent cell death and H 2 O 2 production on the wrr4b mutants, as well as on WT Arabidopsis.

| The expression of WRR4B is upregulated by O. heveae
Plants display large-scale changes in the expression levels of cognate NLR genes when they are challenged by pathogens, with many R genes being upregulated during the infection process (Lai & Eulgem, 2018;Tan et al., 2007;van Wersch et al., 2020).By contrast, in the present study, our RNA-seq analysis revealed that WRR4B was moderately downregulated by O. heveae at 2 dpi.To further explore the mechanisms regulating WRR4B expression following O. heveae infection at varying time points, WT Col-0 was inoculated with O. heveae.Total RNA was isolated from plant samples at 0, 12, 24, 48, 96 and 192 h post-inoculation (hpi), and reverse transcriptionquantitative PCR (RT-qPCR) was used to monitor WRR4B expression.
Compared with its level at 0 dpi, the abundance of WRR4B mRNA was significantly enhanced at 12 and 24 hpi (Figure 3); however, the expression of WRR4B was reduced at 48 hpi and returned to the initial level of 0 hpi (Figure 3) at 96 and 192 hpi, suggesting that WRR4B expression is upregulated in the early stages of O. heveae infection.

| WRR4B positively regulates the expression of genes involved in SA biosynthesis by a self-amplification loop
Activation of TNLs leads to the upregulation of SA biosynthesisrelated genes, such as EDS1, PAD4, ICS1, SARD1 and CBP60g (Ngou et al., 2022;Zeier, 2021), and the EDS1, PAD4 as well as SA pathway are required for O. heveae-triggered resistance in Arabidopsis (Mei et al., 2016).In the present study, the TIR-NB-LRR-type gene WRR4B was shown to contribute to resistance against O. heveae; therefore, we hypothesized that WRR4B may function by regulating the expression of genes involved in SA signalling.To test this hypothesis, WT and wrr4b-1 mutant plants were inoculated with O. heveae.Plant leaves were collected at 0, 2, 4 and 8 dpi, and the transcript levels of EDS1, PAD4, ICS1, SARD1, CBP60g and PR1 were determined using RT-qPCR.Compared with the WT, the expression levels of EDS1, PAD4, ICS1, SARD1, CBP60g and PR1 were significantly downregulated in the mutant at 2 dpi (Figure 4a-e).
At 0, 4 and 8 dpi, the abundances of EDS1, PAD4, SARD1, CBP60g and ICS1 mRNA in the wrr4b-1 mutant were almost the same as those in the WT (Figure 4a-e).The PR1 mRNA abundance in the wrr4b-1 mutant was much lower than in the WT at 0 and 2 dpi, but almost identical to the WT at 4 and 8 dpi (Figure 4f).Compared with their levels at 0 dpi, the expression of EDS1, PAD4, SARD1, CBP60g and ICS1 was more strongly induced at 2 and 4 dpi and mildly upregulated at 8 dpi (Figure 4a-e), whereas PR1 expression was strongly induced at 2, 4 and 8 dpi both in WT and mutant (Figure 4f).In addition, the expression of WRR4B itself was obviously downregulated in the wrr4b-1 mutant at each time point (Figure 4g).Taken together, these results indicate that WRR4B triggers self-amplification and positively regulates the expression of SA biosynthesis-related genes during the early stage of O. heveae infection.

| The transient expression of WRR4B induces Nb-EDS1 dependent hypersensitive responses in N. benthamiana
Elevated R protein expression levels often induce cell death to inhibit the spread of pathogens (Bieri et al., 2004;Holt et al., 2005;Lai & Eulgem, 2018), and only the TIR domain of some NLRs is sufficient to induce cell death (Bernoux et al., 2011;Maekawa, Kufer, et al., 2011;Wan et al., 2019).To gain further insight into the function of WRR4B, the oestrogen-inducible constructs pER8-WRR4B full-length coding sequence and pER8-WRR4B TIR domain were made.
We transiently expressed them in N. benthamiana by infiltrating the plants with Agrobacterium tumefaciens carrying the indicated vectors, and at 1 dpi, the N. benthamiana leaves were sprayed with oestrogen.Only the full-length WRR4B induced typical cell death symptoms in the injection region of the N. benthamiana leaves 2 days after the oestrogen treatment (Figure 5a,c).No obvious changes were observed for the N. benthamiana leaves infiltrated with A. tumefaciens carrying an empty vector or carrying the vector containing the TIR domain of WRR4B alone (Figure 5a,c).In addition, we further expressed WRR4B full-length in WT N. benthamiana and Nb-eds1 mutant leaves.The strong HR was observed on WT N. benthamiana but not on the Nb-eds1 mutant after 2 days of treatment with oestrogen (Figure 5b,d

| WRR4B contributes to a broad-spectrum disease resistance against PMs
WRR4B is involved in the disease resistance against O. heveae, and we hypothesized that WRR4B probably also contributes to the disease resistance against other PM pathogens.To test this hypothesis, we isolated the PMs from the hosts of Vigna unguiculata, Murraya exotica and Solanum lycopersicum and individually identified them as Podosphaera xanthii, Erysiphe quercicola and Erysiphe neolycopersici through the morphological characteristics (Figure S1) and the rRNA internal transcribed spacer (ITS) sequencing analysis (Figure S2).WT Col-0 and wrr4b-1 mutants were inoculated with PM isolates P. xanthii JD 2202, E. quercicola JLX 2201 and E. neolycopersici FQ 2205.P. xanthii JD 2202 is a nonadapted PM on Arabidopsis that is capable of forming initial haustoria but is arrested before sporulation.Until 12 dpi, no obvious symptoms were observed both in the WT Col-0 and wrr4b-1 mutants (Figure 6a).However, microscopy analysis showed that JD 2202 developed a few conidiospores on the wrr4b-1 mutant but not on WT Col-0, which only exhibited sparse hyphal networks (Figure 6b,c).By 10 dpi of JLX 2201, compared with the chlorosis symptoms of Col-0, the wrr4b-1 mutant displayed typical PM white patches (Figure 6a), and there were more conidiospores produced on the wrr4b-1 mutant than on Col-0 (Figure 6b,c).Tomato PM has been reported as a well-adapted PM on Col-0 (Xiao et al., 2005;Zhang et al., 2018).We also saw some FQ 2205 fungal growth on Col-0 at 6 dpi.However, the growth was more pronounced on wrr4b-1, where we could see early PM symptoms, which were not present yet on Col-0 leaves (Figure 6a).Furthermore, the whole wrr4b-1 mutant leaves were covered with dense hyphal networks and lots of conidiospores (Figure 6b,c).However, on Col-0 leaves, dendritic hyphal networks with a few conidiospores were observed and did not fuse together (Figure 6b,c).Overall, these results indicated that WRR4B contributes to a broad-spectrum disease resistance against P. xanthii, E. quercicola and E. neolycopersici.

| DISCUSS ION
Natural rubber, an important raw industrial material, is primarily derived from plantations of rubber trees (H.brasiliensis); however, PM disease of rubber trees, caused by O. heveae, seriously affects yields of rubber latex globally.Almost all commercially cultivated rubber trees are prone to infection by this disease because they have a very narrow genetic background cloned from limited germplasm resources.A whole-genome sequencing analysis of H. brasiliensis revealed that typical NLR, EDS1-like and PAD4-like genes are present in rubber trees (Tang et al., 2016), indicating that the ETI pathway functions in this plant species.It is therefore vital to discover resistance genes and breed resistant cultivars to combat PM disease in rubber trees.A previous study showed that O. heveae-triggered disease resistance responses in WT Arabidopsis (Mei et al., 2016).
In the present study, a TNL gene, WRR4B, was found to confer resistance against O. heveae in WT Arabidopsis, which may enable the generation of resistant rubber trees by transforming them with the Arabidopsis WRR4B gene.
The prototypical R genes responsible for resistance against PM fungi have not been found in Arabidopsis.The ecotype Ms-0 is a resistant accession in which the PM-resistance gene RESISTANCE TO POWDERY MILDEW 8 (RPW8) was first characterized (Xiao et al., 2001).RPW8 is not a typical NLR gene; it encodes a predicted N-terminal transmembrane domain, a central CC domain and Cterminal repeats (Barragan et al., 2019;Zhong & Cheng, 2016).In the present study, the loss of WRR4B resulted in significantly higher host cell entry ratios and fungal growth, as well as supporting the formation of conidiospores and mature conidia during O. heveae infection, which indicated that WRR4B is involved in both penetration and post-penetration resistance against O. heveae.However, the susceptible phenotypes of wrr4b did not reach the level of the eds1 mutant, especially the appearance of serious necrosis symptoms, apparent cell death and H 2 O 2 production developed on wrr4b mutants at the later stage of infection, suggesting that WRR4B partially contributed to the disease resistance against O. heveae and that other R gene(s) in response to O. heveae should still be present on Col-0.In this case, we also examined whether WRR4A, a homologue of WRR4B, is required for resistance against O. heveae.No obvious white patches were observed on wrr4a mutants.By 15 dpi, both WT and wrr4a mutants exhibited typical leaf chlorosis symptoms (Figure S3).But with the microscopy assay, the hyphal networks developed by O. heveae HN1106 on wrr4a mutants were slightly denser than on WT Col-0 (Figure S3), suggesting that WRR4A may also partially contribute to the resistance against O. heveae.Nonetheless, it remains unclear whether there is a functional redundancy between WRR4A and WRR4B in the resistance against O. heveae.

F I G U R E 5
Oestrogen-inducible WRR4B expression triggers a hypersensitive response in Nicotiana benthamiana.Five-week-old wild-type (WT) N. benthamiana and Nbeds1 mutant leaves were infiltrated with Agrobacterium tumefaciens GV3101, and the plants were sprayed with oestradiol at 1 day post-inoculation (dpi).Symptoms were photographed at 2 days post-oestradiol treatment.(a) The symptoms of WT N. benthamiana leaves inoculated with A. tumefaciens GV3101 carrying the pER8 empty vector, pER8-WRR4B full-length coding sequence or pER8-WRR4B TIR domain .(b) The symptoms of WT N. benthamiana and Nbeds1 mutant leaves inoculated with A. tumefaciens GV3101 carrying the pER8-WRR4B full-length coding sequence .(c, d) The protein abundance was determined using an immunoblot assay with an anti-FLAG antibody at 24 h post-oestradiol treatment.These experiments were repeated twice with similar results.
In unchallenged healthy plants, R genes are generally expressed at a low level (Lai & Eulgem, 2018;Tan et al., 2007).Several R genes have been found to be induced by different pathogens; for example, sugar beet (Beta vulgaris) Heterodera schachtii (Hs1 pro-1 ) (Cai et al., 1997), barley (Hordeum vulgare) Mildew resistance locus a (Mla) (Halterman et al., 2003) and rice (Oryza sativa) Xa1 (Yoshimura et al., 1998) and Xa21 (Gu et al., 2005).In this study, RNA-seq analysis showed that WRR4B was downregulated, not induced, at 2 dpi with O. heveae, which appears to be inconsistent with the above notion.Nevertheless, the further RT-qPCR analysis of earlier time points after O. heveae infection displayed that the WRR4B gene was induced by 12 hpi and 24 hpi and later decreased by 48 hpi.Therefore, the expression pattern of WRR4B upon O. heveae infection was upregulated at first and then downregulated to a lower level.The WRR4A gene also displayed a similar expression pattern (Figure S4).It is probably implemented by effector protein(s) secreted from O. heveae to minimize the pathogen recognition by the immune surveillance system.EDS1, PAD4, SARD1 and CBP60g are required for the upregulation of SA and NHP biosynthesis genes such as ICS1 and FMO1 (Ngou et al., 2022;Zeier, 2021), which comprise a positive feedback loop to regulate the signalling pathway.The activation of the TNLs upregulates the expression of these genes and promotes downstream immunity and cell death (Ngou et al., 2022;Zeier, 2021).Here, we explored whether these components are regulated by WRR4B, revealing that the O. heveae-induced EDS1, PAD4, SARD1, CBP60g, ICS1 and PR1 expression levels were significantly lower in the wrr4b mutant than the WT at 2 dpi, but not at 4 and 8 dpi.This indicates that WRR4B takes part in the immune signalling positive feedback loop involving SA production to regulate resistance against O. heveae.In addition, WRR4B can also trigger self-amplification, which is consistent with previous reports that a positive amplification loop exists in TIR domain signalling (Roberts et al., 2013).Overall, these results suggest that WRR4B, being a TNL, has NADase activity that makes AMP, ADP and ATP-related products, which activate EDS1 complexes.
WRR4B was first characterized as an R gene against the white rust pathogen A. candida (Cevik et al., 2019).Four different CCG effector proteins in Ac2V have been found to be recognized by WRR4B, triggering the HR in a WRR4B-dependent manner in Arabidopsis and N. benthamiana (Redkar et al., 2023).The genome and transcriptome of O. heveae strain HO-73 have been sequenced, revealing that 133 candidate effector proteins are present in this strain.
Like O. heveae HN-1106, O. heveae HO-73 also mediated a more susceptible phenotype in the wrr4b mutant (Figure S5), indicating that the effector proteins corresponding to WRR4B should be present in O. heveae HO-73.However, amino acid BLAST searches revealed that none of these 133 effectors display a high similarity to the four CCG effector proteins of Ac2V, suggesting that the effector proteins of O. heveae recognized by WRR4B are probably different from those of A. candida.
In addition, WRR4B contributes to a broad-spectrum disease resistance against P. xanthii, O. heveae, E. quercicola and E. neolycopersici, indicating that there may exist one or more conserved effector protein(s) recognized by WRR4B that are differentially distributed in these four PMs.
It has been shown that many NLRs function with helper NLRs, also termed RNLs due to their N-terminal RPW8-like CC domains, to transduce immune signals (Castel et al., 2019;Chai et al., 2023;Saile et al., 2020).In Arabidopsis, there are five active RNLs: activated disease resistance 1 (ADR1), ADR1-like 1 (ADR1-L1), ADR1-L2, N requirement gene 1.1 (NRG1.1)and NRG1.2, which are partially redundant regulators of immunity and cell death downstream of TNLs (Castel et al., 2019;Saile et al., 2020;Wu et al., 2019).Interestingly, a newly defined TNL, the suppressor of ADR1-L2 1 (SADR1), has been found to be required for the phenotypes driven by ADR1-L2 autoactivity (Jacob et al., 2023), suggesting that TNLs can function downstream of RNL activation.In this scenario, given that WRR4B partially contributes to the disease resistance against O. heveae and mediates a broad-spectrum resistance against many PM species and white rust pathogens, it is postulated that WRR4B may also act as a helper and together with other functional redundant regulators to fine-tune the signalling pathways.In addition, although WRR4B confers race-specific resistance to A. candida, virulent races of Albugo might have a strategy to overcome the WRR4B-mediated resistance rather than lacking cognate avirulence genes.

| Plant materials
The Arabidopsis ecotype Col-0 was used as the WT in this study.

| Powdery mildew infections
The O. heveae strain HN1106 used in this study was maintained in the after which the plants were placed under the settling tower for 1 h to be inoculated before being moved to the growth chamber.

| Host cell entry assay
Three O. heveae-infected leaves were harvested at 1 dpi and stained with Coomassie brilliant blue.The numbers of germinated fungal sporelings were counted, and the proportion of sporelings that developed secondary and more than secondary hyphae in all the germinated sporelings was assessed (a minimum of 50 germinated sporelings/leaves evaluated).The fungal penetration ratio of each plant was quantified in at least three independent experiments.

| Analysis of fungal growth
Three independent plants per line were inoculated with a low density of O. heveae, and eight leaves per plant were harvested at 2 dpi and stained with Coomassie brilliant blue.At least 20 images of single colonies per line were taken and analysed using MIE v. 3.1 software (http:// www.mieso ftware.com/ ).
the hyphae had elongated to over 100 μm on the wrr4b-1 and wr-r4b-2 mutants, 180 μm on the eds1 mutant, but only about 35 μm on the WT (Figure 1d).Overall, these results indicated that WRR4B F I G U R E 1 WRR4B contributes to the resistance against Oidium heveae HN1106.Five-week-old plants were inoculated with O. heveae HN1106.(a) Leaf symptoms at 6 days post-inoculation (dpi).(b) Fungal structures stained with trypan blue and imaged using a light microscope at 6 dpi.Bars, 200 μm.(c) Quantitative analysis of host cell entry rates at 1 dpi; at least 50 germinated sporelings/leaves were counted for each experiment.(d) Quantitative analysis of O. heveae hyphal growth at 2 dpi.(e) The numbers of conidiospores per colony were counted at 6 dpi.Twenty fungal colonies were analysed per genotype for (d) and (e).The data represent the mean ± SD of three experiments.Significant difference from Col-0 by Student's t-test, *p < 0.05, **p < 0.01.These experiments were repeated twice with similar results.contributes to resistance against O. heveae in the early infection stage of Arabidopsis.
).Overall, these results indicated that transient expression of WRR4B induced cell death in Nb-EDS1-dependent manner in N. benthamiana.

F
The cell death and H 2 O 2 production in the later stage of Oidium heveae infection.Five-week-old plants were inoculated with O. heveae HN1106.(a) Leaf symptoms at 12 days post-inoculation (dpi).(b) Microscopy analysis of infected leaves with trypan blue staining at 12 dpi.Bars, 400 μm.(c) The whole infected leaves were photographed by a light stereoscope with trypan blue staining at 12 dpi.Bars, 5 mm.(d) The whole infected leaves were photographed by a light stereoscope with 3,3′-diaminobenzidine staining.Bars, 5 mm.These experiments were repeated twice with similar results.

F
The expression of WRR4B is upregulated by Oidium heveae.Five-weekold Arabidopsis wild-type (WT) Col-0 was inoculated with O. heveae HN1106.The abundance of WRR4B mRNA was determined at the indicated time points after O. heveae inoculation using reverse transcription-quantitative PCR.The data represent the mean ± SD of three independent experiments and six RNA replicates for each experiment.Significant difference from 0 h post-inoculation (hpi) of Col-0 by Student's t-test, **p < 0.01.These experiments were repeated twice with similar results.F I G U R E 4 WRR4B positively regulates the expression of EDS1, PAD4, ICS1, SARD1, CBP60g, PR1 and WRR4B.Five-week-old Arabidopsis wild-type Col-0 (WT) and wrr4b-1 plants were inoculated with Oidium heveae HN1106.The abundances of EDS1 (a), PAD4 (b), ICS1 (c), SARD1 (d), CBP60g (e), PR1 (f) and WRR4B (g) mRNA were determined at the indicated time points after O. heveae infection using reverse transcriptionquantitative PCR.The data represent the mean ± SD of three independent experiments and six RNA replicates for each experiment.Significant difference from Col-0 by Student's t-test, *p < 0.05, **p < 0.01.These experiments were repeated twice with similar results.

F
I G U R E 6 WRR4B contributes to resistance against Podosphaera xanthii, Erysiphe quercicola and Erysiphe neolycopersici.Five-week-old Arabidopsis wild-type (WT) Col-0 and wrr4b-1 mutants were infected with P. xanthii JD 2202, E. quercicola JLX 2201 and E. neolycopersici FQ 2205.(a) Leaf symptoms.(b) Fungal structures stained with Coomassie blue were imaged using a light microscope.Bars, 50 μm.(c) The numbers of conidiospores per colony were counted.The data represent the mean ± SD of at least 20 fungal colonies per genotype.These experiments were repeated twice with similar results.Significant difference from Col-0 by Student's t-test, **p < 0.01.
et al. (1998) and was backcrossed multiple times into the Col-0 background.The T-DNA insertion lines SALK_053459, SALK_104727C, leaves of the susceptible H. brasiliensis clone 73397; P. xanthii JD 2202 was maintained in the leaves of V. unguiculata 'ZhiJiang 19'; E. quercicola JLX 2201 was maintained in the leaves of M. exotica; E. neolycopersici FQ 2205 was maintained in the leaves of S. lycopersicum 'YaShu 6'.Mature PM conidia were used as a source of inoculum.Actively growing fungal spores were dusted on top of the modified settling tower (diameter, 40 cm; height, 60 cm) and covered with a nylon mesh,