Ectopic Expression of HbRPW8-a from Hevea brasiliensis Improves Arabidopsis thaliana Resistance to Powdery Mildew Fungi (Erysiphe cichoracearum UCSC1)

The RPW8s (Resistance to Powdery Mildew 8) are atypical broad-spectrum resistance genes that provide resistance to the powdery mildew fungi. Powdery mildew of rubber tree is one of the serious fungal diseases that affect tree growth and latex production. However, the RPW8 homologs in rubber tree and their role of resistance to powdery mildew remain unclear. In this study, four RPW8 genes, HbRPW8-a, b, c, d, were identified in rubber tree, and phylogenetic analysis showed that HbRPW8-a was clustered with AtRPW8.1 and AtRPW8.2 of Arabidopsis. The HbRPW8-a protein was localized on the plasma membrane and its expression in rubber tree was significantly induced upon powdery mildew infection. Transient expression of HbRPW8-a in tobacco leaves induced plant immune responses, including the accumulation of reactive oxygen species and the deposition of callose in plant cells, which was similar to that induced by AtRPW8.2. Consistently, overexpression of HbRPW8-a in Arabidopsis thaliana enhanced plant resistance to Erysiphe cichoracearum UCSC1 and Pseudomonas syringae pv. tomato DC30000 (PstDC3000). Moreover, such HbRPW8-a mediated resistance to powdery mildew was in a salicylic acid (SA) dependent manner. Taken together, we demonstrated a new RPW8 member in rubber tree, HbRPW8-a, which could potentially contribute the resistance to powdery mildew.


Introduction
Plants have developed two unique pathways to response for pathogen attacks, namely pathogen-associated molecular pattern-triggered immunity (PTI) and effector-triggered immunity (ETI), which work together to protect plants [1][2][3]. The plant immune responses usually involve the burst of reactive oxygen species (ROS), activation of mitogen-activated protein kinase (MAPK) signaling and downstream PATHOGENESIS-RELATED (PR) genes, regulation of phytohormone signaling pathways, and induction of hypersensitive response (HR) [4][5][6][7][8][9]. The ROS burst is one of the early cellular responses in plant cells, serving as a toxin barrier against subsequent pathogen infection [10]. Moreover, the ROS burst can also strengthen the plant cell walls through oxidative cross-link of polymers to inhibit pathogen growth [11][12][13]. While HR is a rapid cell death phenomenon induced by pathogen or their metabolites, occurring at the point of pathogen penetration to prevent further pathogen spreading [14][15][16][17][18][19][20]. The plant immune system is also regulated by plant hormones, especially jasmonic acid (JA) and SA, which play an important role in basic resistance against

Isolation and Identification of HbRPW8s
To identify RPW8 genes in rubber tree, AtRPW8.1 and AtRPW8.2 protein full-length sequence was used to blast against the genome of H. brasiliensis. Four HbRPW8 genes were obtained, and designated as HbRPW8-a, HbRPW8-b, HbRPW8-c and HbRPW8-d, respectively. Compared with the other three HbRPW8s (about 800 amino acids), HbRPW8-a was much smaller, with only 189 amino acids (Supplementary Table S1).

Expression of HbRPW8s upon E. Quercicola Infection
To investigate the possible roles of HbRPW8s in powdery mildew disease, the transcriptional expression levels of HbRPW8s gene were measured at different times after E. quercicola infection (Figure 2). The expression patterns of HbRPW8-b and c were similar, with a strong reduction at 48 hours post infection (hpi) (the second penetration process of E. quercicola), while the expression of HbRPW8-d was relatively stable during the whole infection period.
To investigate the possible roles of HbRPW8s in powdery mildew disease, th scriptional expression levels of HbRPW8s gene were measured at different times a quercicola infection (Figure 2). The expression patterns of HbRPW8-b and c were s with a strong reduction at 48 hours post infection (hpi) (the second penetration pro E. quercicola), while the expression of HbRPW8-d was relatively stable during the infection period. Figure 2. Transcriptional expression of HbRPW8s in response to E. quercicola. The relative exp of each gene was calibrated against 0 h uninoculated samples. The experiment was perform three biological replicates, and six plants were used in each independent trial. The data we lyzed by one-way ANOVA and Duncan's multiple comparison tests. Data were showed as m standard error. Bars with different lower-case letters show significant differences at the lev 0.05.
Interestingly, the expression level of HbRPW8-a was significantly induced at (primary infection period of E. quercicola) and 24 hpi (the haustorium developmen during E. quercicola infection) compared to that of control. Combined with the hig larity between HbRPW8-a and AtRPW8.1/8.2, HbRPW8-a was selected for further tigate in plant defense.

Transient Expression of HbRPW8-a Induces Plant Immune Responses in N. bentham Leaves
AtRPW8.1 and AtRPW8.2 were reported to be localized in the endomembran tem to target the haustorium upon powdery mildew infection. To determine the s lular localization of HbRPW8-a, the HbRPW8-a-mScarlet and RLK-GFP (a plasma brane control) fusion proteins were transiently expressed in N. benthamiana leaves. confocal microscopy, we found that the fluorescence signal of HbRPW8-a-mScarlet protein was co-localized with RLK-GFP at the plasma membrane (Figure 3), ind HbRPW8-a is a membrane localized protein. Transcriptional expression of HbRPW8s in response to E. quercicola. The relative expression of each gene was calibrated against 0 h uninoculated samples. The experiment was performed with three biological replicates, and six plants were used in each independent trial. The data were analyzed by one-way ANOVA and Duncan's multiple comparison tests. Data were showed as means ± standard error. Bars with different lower-case letters show significant differences at the level of p < 0.05.
Interestingly, the expression level of HbRPW8-a was significantly induced at 13 hpi (primary infection period of E. quercicola) and 24 hpi (the haustorium development stage during E. quercicola infection) compared to that of control. Combined with the high similarity between HbRPW8-a and AtRPW8.1/8.2, HbRPW8-a was selected for further investigate in plant defense.

Transient Expression of HbRPW8-a Induces Plant Immune Responses in N. benthamiana Leaves
AtRPW8.1 and AtRPW8.2 were reported to be localized in the endomembrane system to target the haustorium upon powdery mildew infection. To determine the subcellular localization of HbRPW8-a, the HbRPW8-a-mScarlet and RLK-GFP (a plasma membrane control) fusion proteins were transiently expressed in N. benthamiana leaves. Using confocal microscopy, we found that the fluorescence signal of HbRPW8-a-mScarlet fusion protein was co-localized with RLK-GFP at the plasma membrane (Figure 3), indicating HbRPW8-a is a membrane localized protein.
In addition to subcellular localization of HbRPW8-a, we also observed HbRPW8-a induced callose deposition in N. benthamiana leaves, similar to that induced by AtRPW8.2 (Figure 4a,b). Additionally, similar to that of AtRPW8.2-GFP, the HbRPW8-GFP fusion protein can also induce ROS burst in tobacco cells (Figure 4c,d). These results suggest that HbRPW8-a, like AtRPW8.2, could induce plant immune responses.  In addition to subcellular localization of HbRPW8-a, we also observed HbRPW8-a induced callose deposition in N. benthamiana leaves, similar to that induced by AtRPW8.2 (Figure 4a,b). Additionally, similar to that of AtRPW8.2-GFP, the HbRPW8-GFP fusion protein can also induce ROS burst in tobacco cells (Figure 4c,d). These results suggest that HbRPW8-a, like AtRPW8.2, could induce plant immune responses.   In addition to subcellular localization of HbRPW8-a, we also observed HbRPW8-a induced callose deposition in N. benthamiana leaves, similar to that induced by AtRPW8.2 ( Figure 4a,b). Additionally, similar to that of AtRPW8.2-GFP, the HbRPW8-GFP fusion protein can also induce ROS burst in tobacco cells (Figure 4c,d). These results suggest that HbRPW8-a, like AtRPW8.2, could induce plant immune responses.

Over-expressing HbRPW8-a in Arabidopsis Col Enhances Resistance to Powdery Mildew
Different from the Arabidopsis Wa and Ms ecotypes containing AtRPW8.1/AtRPW8.2 with resistance to powdery mildew, the Arabidopsis Col is susceptible to powdery mildew due to the lack of AtRPW8.1/AtRPW8.2. To further investigate the role of HbRPW8-a in resistance to powdery mildew, we overexpressed HbRPW8-a in the Arabidopsis Col plants and obtained T3 transgenic homozygous lines (Supplementary Figure S1A). The high expression level of HbRPW8-a gene in these lines was confirmed by qRT-PCR (Supplementary Figure S1B).
Three T3 HbRPW8-a overexpression (#8, #16 and #27), Col-WT, Wa-WT and the Arabidopsis mutant pad4 at Col background (a susceptible mutant to powdery mildew), were inoculated with powdery mildew (E. cichoracearum UCSC1), respectively. After 7 days of inoculation (dpi), the pad4 displayed severe disease, with the entire leaves covered by powdery mildew fungi hyphae. Col-WT plants also had visible disease symptoms, with large areas of hyphae on the fully expanded leaves. While the HbRPW8-a overexpression lines (#8, #16 and #27) showed limited disease symptoms, with fewer hyphae on the leaves, and the resistant plant Wa-WT was basically disease free ( Figure 5a). Additionally, we measured the basal immune responses in Arabidopsis plants after powdery mildew inoculation. Consistent with their disease symptoms, the callose deposition and ROS accumulation in leaves of HbRPW8-a overexpression lines were significantly higher than those in Col-WT and pad4 mutant (Figure 5b-e). By trypan blue, we observed that blue-stained necrotic cells were abundant in the leaves of HbRPW8-a overexpression lines, but only at the leaf margins in Col-WT ( Figure 5c). We also counted the number of conidiophores per colony from these plants and revealed that the HbRPW8-a overexpression lines had a significantly fewer conidiophores than that of Col-WT ( Figure 5f). Similar to that in AtRPW8.1-expressed Arabidopsis plants [39], the epidermal cells of HbRPW8-a overexpression lines were enriched in hydrogen peroxide surrounding germinating conidia (Figure 5g). These results reveal that overexpression of HbRPW8-a can induce the basal immune response in Arabidopsis Col and prevent powdery mildew colonization effectively.

Over-expressing HbRPW8-a in Arabidopsis Col Enhanced Plant Resistance to Bacteria
Flagellin 22 (flg22), as a conserved molecule pathogen-associated pattern, can trigger innate immune responses in plants [58]. After flg22 treatment, the accumulation of ROS and callose deposition in Col-WT leaves were continuously increased from 24 h to 72 h, and these responses were even stronger in the HbRPW8-a overexpression lines (Figure 6a-d).
To test whether HbRPW8-a could play a role in bacterial infection, we treated the HbRPW8-a overexpression lines with virulent PstDC3000. Col-WT plants displayed severe chlorosis in leaves with enhanced bacterial growth, whereas HbRPW8-a overexpression lines showed only mild disease symptoms with lower bacterial numbers (Figure 6e-f). These results indicate that the expression of HbRPW8-a could also enhance resistance to biotrophic bacteria P. syringae and inhibit the colonization of PstDC3000 in Arabidopsis Col plants.

Over-Expression of HbRPW8-a Cannot Improve Plant Resistance to Phytophthora Capsici and Botrytis Cinerea
Pathogen-induced cell death is important in plant resistance to biotrophic pathogens [59]. We used B. cinereal and P. capsici to investigate whether HbRPW8-a could also be involved in plant resistance to necrotrophic pathogens. The lesion diameters of B. cinerea spore infected leaves were measured at 3 dpi, however, no significant difference was observed between HbRPW8-a overexpression lines and Col-WT (Figure 7a,b).
P. capsici spores were inoculated on tobacco leaves that transiently expressed GFP, HbRPW8-a-GFP or AtRPW8.2-GFP protein after 24 h. Similarly, the lesions of the infected sites were indistinguishable between control and HbRPW8-a-GFP/ AtRPW8.2-GFP expressed regions at 3 dpi (Figure 7c,d). These results indicate that HbRPW8-a has no contribution to plant resistance to P. capsici and B. cinerea.  Data were shown as the means ± SD from three independent experiments and columns with different letters indicate significant difference (p < 0.05). (e) average numbers of callose deposits per 1 mm 2 are displayed in over-expressing lines, pad4, Col-WT and Wa-WT leaves at 48 hpi. Data were shown as the means ± SD from three independent experiments and columns with different letters indicate significant difference (p < 0.05); (f) quantitative analysis of conidiophore formation on over-expressing lines, pad4, Col-WT and Wa-WT leaves at 7 dpi. Data were shown as the means ± SD from three independent experiments and columns with different letters indicate significant difference (p < 0.05); (g) fungal structures on leaves of Col-WT, pad4 and #16 stained with trypan blue at 3 dpi and 7 dpi; From 3dpi to 7dpi, the number of hyphae on the leaves became more, especially on pad4 leaves, the fungal structures was densely spread on the leaves. The fungal structures of Col-WT was less than that of pad4, but more than that of #16. And the epidermal cells of #16 were enriched in hydrogen peroxide surrounding germinating conidia, which was not find in Col-WT or pad4 leaves; scale bar = 50 µm. indicate significant difference (p < 0.05). (e) average numbers of callose deposits per 1 mm 2 are displayed in over-expressing lines, pad4, Col-WT and Wa-WT leaves at 48 hpi. Data were shown as the means ± SD from three independent experiments and columns with different letters indicate significant difference (p < 0.05); (f) quantitative analysis of conidiophore formation on over-expressing lines, pad4, Col-WT and Wa-WT leaves at 7 dpi. Data were shown as the means ± SD from three independent experiments and columns with different letters indicate significant difference (p < 0.05); (g) fungal structures on leaves of Col-WT, pad4 and #16 stained with trypan blue at 3 dpi and 7 dpi; From 3dpi to 7dpi, the number of hyphae on the leaves became more, especially on pad4 leaves, the fungal structures was densely spread on the leaves. The fungal structures of Col-WT was less than that of pad4, but more than that of #16. And the epidermal cells of #16 were enriched in hydrogen peroxide surrounding germinating conidia, which was not find in Col-WT or pad4 leaves; scale bar = 50 μm.

Over-expressing HbRPW8-a in Arabidopsis Col Enhanced Plant Resistance to Bacteria
Flagellin 22 (flg22), as a conserved molecule pathogen-associated pattern, can trigger innate immune responses in plants [58]. After flg22 treatment, the accumulation of ROS and callose deposition in Col-WT leaves were continuously increased from 24 h to 72 h, and these responses were even stronger in the HbRPW8-a overexpression lines (Figure 6ad).  [59]. We used B. cinereal and P. capsici to investigate whether involved in plant resistance to necrotrophic pathogens. The le spore infected leaves were measured at 3 dpi, however, no sig served between HbRPW8-a overexpression lines and Col-WT (   Figure 7. The responses of HbRPW8-a over-expressing lines and Co inoculation. (a) Macroscopic evaluation of symptoms 3 dpi with B. and transgenic lines; necrotic area was highlighted by the arrow; (b) a size at 3 dpi; Data were shown as the means ± SD from three indepen nificant difference (p < 0.05). (c) macroscopic evaluation of symptoms lation of tobacco leaf expressed with GFP, HbRPW8-a-GFP or AtRPW P. capsici spores were inoculated on tobacco leaves that t HbRPW8-a-GFP or AtRPW8.2-GFP protein after 24 h. Similarly sites were indistinguishable between control and HbRPW8-a pressed regions at 3 dpi (Figure 7c,d). These results indicate th tribution to plant resistance to P. capsici and B. cinerea.

SA Is Required for HbRPW8-a-Mediated Powdery Mildew Resistance
SA, as one of the hormone signaling pathways, plays an important role in plant immune responses. Expressing a bacterial salicylate hydroxylase (encoded by NahG) in Arabidopsis plants was shown to convert SA to catechol [60], causing a decreased diseaseresistance response due low accumulated SA [61].
To examine whether HbRPW8-a mediated plant resistance is dependent on SA signaling, we overexpressed HbRPW8-a in the NahG plants at Col background, and selected T3 transgenic homologous lines #1, #5 and #13 for further study (Supplementary Figure S2).
The disease symptom analysis in Figure 8 showed that the overexpression of HbRPW8a in SA-deficient transgenic NahG plants failed to be resistant to powdery mildew, with similar numbers of hyphae distributed on the leaves compared to that of Col-WT (Figure 8a,g). Interestingly, the NahG plants and its HbRPW8-a transgenic lines had significantly higher ROS accumulation but less callose deposition than that of Col-WT, which caused a larger area of cell death at 3 dpi (Figure 8b-f), suggesting that SA signaling may contribute to the HbRPW8-a mediated plant resistance via regulating callose deposition. WT (Figure 8a,g). Interestingly, the NahG plants and its HbRPW8-a transgenic lines had significantly higher ROS accumulation but less callose deposition than that of Col-WT, which caused a larger area of cell death at 3 dpi (Figure 8b-f), suggesting that SA signaling may contribute to the HbRPW8-a mediated plant resistance via regulating callose deposition. (g) quantitative analysis of conidiophore formation on over-expressing lines, NahG and Col-WT leaves at 7 dpi. All the data were shown as the means ± SD from three independent experiments and columns with different letters indicate significant difference (p < 0.05).  (#1, #5, #13), NahG, Col-WT plants were infected with powdery mildew and photographed at 7 dpi; Col-WT as a control, its disease syptom was the same as that of transgenic lines. The NahG disease syptom was serious, with a large number of hyphae covered on leaves. (b) the accumulation of ROS and cell death in over-expressing lines, NahG and Col-WT leaves, photographed at 3 dpi; in NahG background, plant leaves with a large number of ROS clusters, but a few in Col-WT leaves (c) plant cell callose deposition were stained with DAB at 48 hpi in over-expressing lines, NahG and Col-WT leaves, and observed under microscope UV light; in Col-WT leaves, there were a large number of callose deposition, but a few in NahG background plant leaves, scale bar = 100 µm. (d) Severe cell death occured in NahG background plant leaves, photographed at 3 dpi; (e) representation of DAB-positive stained clusters in over-expressing lines, NahG and Col-WT leaves. Number of stained clusters was calculated per cm 2 ; (f) average numbers of callose deposits per 1 mm 2 are displayed in over-expressing lines, NahG and Col-WT leaves at 48 hpi. (g) quantitative analysis of conidiophore formation on over-expressing lines, NahG and Col-WT leaves at 7 dpi. All the data were shown as the means ± SD from three independent experiments and columns with different letters indicate significant difference (p < 0.05).

Discussion
RPW8 genes were first reported as an atypical R gene in Arabidopsis in 2001 [30]. They encoded proteins containing an N-terminal transmembrane domain and one or two coiledcoil domains, known as the RPW8 domain [30]. In recent years, proteins with conserved RPW8 domains have also been found in Brassica napus, strawberries (Fragaria × ananassa), V. pseudoreticulata, Cucurbita pepo L., Akebia trifoliata and Lagenaria siceraria (Molina) Standl. (Figure 9); (for sequence information, see Supplemental Table S3) [40][41][42][62][63][64]. Among them, several RPW8s have been proved to be resistant to powdery mildew. In this study, we identified four HbRPW8s with the conserved RPW8 domain in rubber tree (Figure 1). By phylogenetic analysis, we showed that HbRPW8-a together with AtRPW8.1 and AtRPW8.2 were grouped into the same clade, indicating that HbRPW8-a protein was similar to that of AtRPW8.1 and AtRPW8.2 ( Figure 1B). Similar to that of Lsi04g015960 (encoded RPW8-NBS-LRR protein) and Cp4.1LG10g02780 (encoded RPW8-NBS-LRR protein), HbRPW8-a was strongly induced by powdery mildew at 13 and 24 hpi, which corresponds to the early stages of the interaction between powdery mildew and host plants, suggesting the role of HbRPW8-a in rubber trees in response to powdery mildew. Standl ( Figure 9); (for sequence information, see Supplemental Table S3) Among them, several RPW8s have been proved to be resistant to powdery m study, we identified four HbRPW8s with the conserved RPW8 domain in ru ure 1). By phylogenetic analysis, we showed that HbRPW8-a together with AtRPW8.2 were grouped into the same clade, indicating that HbRPW8-a pr ilar to that of AtRPW8.1 and AtRPW8.2 ( Figure 1B). Similar to that of Lsi coded RPW8-NBS-LRR protein) and Cp4.1LG10g02780 (encoded RPW8tein), HbRPW8-a was strongly induced by powdery mildew at 13 and 24 h responds to the early stages of the interaction between powdery mildew a suggesting the role of HbRPW8-a in rubber trees in response to powdery m AtRPW8 was reported as a broad-spectrum resistance gene [39,40], AtRPW8 leads to enhanced resistance to four Arabidopsis powdery milde bacco powdery mildew, Hyaloperonospora parasitica, Cauliflower mosaic vi oryza, Xanthomonas oryzae pv. oryzae and other plant pathogens [30,65]. Ho the VpRPW8s could also enhance plant resistance to P. viticola, P. capsici, a [41]. Additionally, MdRNL2 and MdRNL6 (encode RPW8-NBS-LRR protein ered to be spectrum fungal resistance genes [66]. By transient expression in and overexpression in Col-WT Arabidopsis, we showed that HbRPW8-a is trum resistance gene with resistance to E. cichoracearum UCSC1 and PstDC pression of AtRPW8.1 and AtRPW8.2 can active oxygen burst and callose d induce plant HR response, which promotes neoplastic pathogens Alternar spp. infection. In contrast, overexpressing of HbRPW8-a could not increase AtRPW8 was reported as a broad-spectrum resistance gene [39,40], expression of AtRPW8 leads to enhanced resistance to four Arabidopsis powdery mildew species, tobacco powdery mildew, Hyaloperonospora parasitica, Cauliflower mosaic virus, Pyricularia oryza, Xanthomonas oryzae pv. oryzae and other plant pathogens [30,65]. Homogeneously, the VpRPW8s could also enhance plant resistance to P. viticola, P. capsici, and P. parasitica [41]. Additionally, MdRNL2 and MdRNL6 (encode RPW8-NBS-LRR protein) were considered to be spectrum fungal resistance genes [66]. By transient expression in tobacco leaves and overexpression in Col-WT Arabidopsis, we showed that HbRPW8-a is a broad-spectrum resistance gene with resistance to E. cichoracearum UCSC1 and PstDC3000. Overexpression of AtRPW8.1 and AtRPW8.2 can active oxygen burst and callose deposition and induce plant HR response, which promotes neoplastic pathogens Alternaria and Botrytis spp. infection. In contrast, overexpressing of HbRPW8-a could not increase the sensitivity of plants to neoplastic pathogens B.cinerea and P. capsici. However, further study is required to understand the different roles between HbRPW8-a and AtRPW8.1/AtRPW8.2.
Plant RPW8s are diverse in terms of subcellular localization [41]. Both AtRPW8.1:YFP and AtRPW8.2:YFP appeared to be localized mainly in the endomembrane system [65], especially for AtRPW8.2:YFP, targeted to haustorium when powdery mildew infected, which is key to activation of resistance at the host-pathogen interface. BnHRa YFP and BnHRb-YFP were mainly localized to the extra-haustorial membrane [40]. VpRPW8s mainly located in the cytoplasm [41]. Interestingly, we found that HbRPW8-a was plasma membrane localized protein ( Figure 3). AtADR1(RPW8-NBS-LRR protein), an Arabidopsis HeLo-/RPW8-like domain 'helper' NLR (RNL), was also localized on plasma membrane to induce immune signaling and cell death [67][68][69][70]. It will be interesting to investigate whether the plasma membrane-localization of HbRPW8-a is required for its induction of innate immune signaling.
Previous study showed that AtRPW8-mediated resistance to powdery mildew in Arabidopsis was dependent on SA signaling [71]. Similarly, we found that decreased SA in NahG plants blocked the function of HbRPW8-a in inhibiting fungal growth. Interestingly, ROS was significantly high accumulated in the leaves of NahG plants, which is probably because SA acts as direct or indirect antioxidants to protect cells from oxidative damage, observed in animal systems [72][73][74]. In addition to SA, it is very possible that other signaling pathways are also involved in HbRPW8-a mediated resistance to powdery mildew in Arabidopsis, which requires further study in the future.

Biological Materials and Growth Conditions
The Arabidopsis thaliana ecotype Columbia-0, pad4, Nicotiana benthamiana, and Hevea brasiliensis (Reyan 7-33-97) were all from our laboratory. The Arabidopsis thaliana ecotype Wa was kindly offered by Prof. Wenming Wang's Lab. The Arabidopsis NahG was kindly offered by Chaozu He's Lab. Rubber tree susceptible cultivar CATAS7-33-97 plants [40] were grown in the experimental plantation of Hainan University. N. benthamiana, A. thaliana wild type (Col-0 and Wa ecotype), and the pad4 mutant at Col background were grown in a greenhouse at 24 ± 2 • C under a 16 h light/8 h dark photoperiod. E. cichoracearum UCSC1 was cultured on A. thaliana pad4 mutant plants. PstDC3000 was preserved at −80 • C. B. cinerea and P. capsica were maintained at 28 • C on Potato Glucose Agar (PDA) in the dark.

Quantitative Real-Time PCR
For qRT-PCR, rubber tree bronze leaves inoculated with powdery mildew spores were sampled at 0, 4, 8,13, 24, 48, and 72 hpi. The total RNA was extracted by a plant total RNA extraction kit (Tiangen Biotech., Beijing, China). The first-strand cDNA was synthesized by the EasyScript ® One-Step gDNA Removal and cDNA Synthesis SuperMix kit (TRAN, Beijing, China). qRT-PCR was then performed with SYBR green (TRAN, Beijing, China) in the LightCycler ® 96 System (Roche, Switzerland). Hb18s was used as the internal control, and all used primers were listed in Supplementary Table S2. Relative expression level was calculated by 2 −∆∆Ct method [75], in triplicates for each gene.
The phylogeny was inferred based on a multiple sequence alignment of the fulllength amino acid sequences of the reported V. pseudoreticulata and A. thaliana RPW8s (AtRPW8.1/8.2, HR1-HR4). The phylogenetic tree was constructed via MEGA 6.0. The numbers at the end of protein names indicate the accession number of each sequence in the NCBI database.

Plant Transformation and Disease Assays
Plant transformation in Arabidopsis Col-WT and NahG was performed by the flower dip method [78]. Three T3 homozygous lines of 35s-HbRPW8-a-mScarlet were selected for subsequent experiments.
P. capsica, B. cinerea, PstDC3000, and E.cichoracearum were used for pathogenicity tests. Four-week-old tobacco leaves were inoculated with A.tumefaciens containing 35s-HbRPW8-a-GFP/35s-AtRPW8.2-GFP recombinant plasmids. After inoculation for 24 hpi, the tobacco leaves were inoculated with P. capsici spore suspension at a concentration of 5 × 10 4 spores mL −1 [79] and were kept under moisture retention in vitro. Five-weekold Arabidopsis leaves were inoculated with B. cinerea spores at the concentration of 1.5 × 10 6 spores mL −1 [80], which were also kept under moisture retention in vitro. The diameter of necrotic spots was measured after three days.
The transgenic Arabidopsis leaves were sprayed by B. cinerea spores at the concentration of 2 × 10 5 spores mL −1 , and the size of lesion was measured after 3 days of treatment. PstDC3000 treatment with OD = 0.6 was used to observe the leave disease, the OD = 0.1 solution was used to count the bacterial concentration at the time of 0, 24, 48 and 72 hpi. For powdery mildew infection, E. cichoracearum grown for about 10 days was inoculated on transgenic lines for disease test [81]. For flg22 treatment, four-week-old Arabidopsis plants of Col-WT and transgenic lines were sprayed with 1 µM flg22 and harvested after 24, 48, 72 hpi.

ROS Measurement and Callose Staining
For ROS burst assay, four-week-old tobacco leaves and Arabidopsis leaves were collected and were stained with 3,3 -diaminobenzidine (DAB, 2 mg mL −1 ) solution overnight. To investigate ROS accumulation, cleaned leaves were boiled in ethanol (95%) until pigments were removed. Callose staining assay was performed as described previously [82]. In brief, treated leaves were collected and socked in ethanol (95%) overnight to remove the pigments, and then incubated in the staining solution (0.1% aniline blue, 67 mmol/L Na 2 HPO 4 , pH 12.0) for 1 h in dark at room temperature. Callose deposition was observed under an epifluorescence microscope (Carl Zeiss, Oberkochen, Germany) and photographed.

Conclusions
In conclusion, we have identified four HbRPW8 genes from rubber tree in this study. By taking advantage of an Arabidopsis Col ecotype that lacks AtRPW8.1/8.2, we show that HbRPW8-a, a protein localized on the plasma membrane, has broad resistance to several plant pathogens including powdery mildew. Our study reveals a new member of the plant RPW8 family, providing a foundation for future studies and potential applications in rubber tree variety improvement.