RPW8.1 enhances the ethylene‐signaling pathway to feedback‐attenuate its mediated cell death and disease resistance in Arabidopsis

Summary The Arabidopsis RESISTANCE TO POWDERY MILDEW 8.1 (RPW8.1) activates confined cell death and defense against different pathogens. However, the underlying regulatory mechanisms still remain elusive. Here, we show that RPW8.1 activates ethylene signaling that, in turn, negatively regulates RPW8.1 expression. RPW8.1 binds and stabilizes 1‐aminocyclopropane‐1‐carboxylate oxidase 4 (ACO4), which may in part explain increased ethylene production and signaling in RPW8.1‐expressing plants. In return, ACO4 and other key components of ethylene signaling negatively regulate RPW8.1‐mediated cell death and disease resistance via suppressing RPW8.1 expression. Loss of function in ACO4, EIN2, EIN3 EIL1, ERF6, ERF016 or ORA59 increases RPW8.1‐mediated cell death and defense response. By contrast, overexpression of EIN3 abolishes or significantly compromises RPW8.1‐mediated cell death and disease resistance. Furthermore, ERF6, ERF016 and ORA59 appear to act as trans‐repressors of RPW8.1, with OAR59 being able to directly bind to the RPW8.1 promoter. Taken together, our results have revealed a feedback regulatory circuit connecting RPW8.1 and the ethylene‐signaling pathway, in which RPW8.1 enhances ethylene signaling, and the latter, in return, negatively regulates RPW8.1‐mediated cell death and defense response via suppressing RPW8.1 expression to attenuate its defense activity.

. The primers and probes used in this study.
Methods S1 Yeast two-hybrid (Y2H) assays Methods S2 Bimolecular fluorescence complementation (BiFC) assay Methods S3 Determination of ethylene biosynthesis rates and electrolyte leakage measurements Methods S4 The leaf senescence assay Methods S5 RNA extraction and reverse transcription quantitative PCR (RT-qPCR)

Methods S10 Protein expression and purification
Methods S11 CRISPR/Cas9 plasmids construction and mutant screening In vivo Co-IP assay. HA-tagged RPW8.1 was co-expressed with Flag-tagged ACO438-175 or RPW8.1 in N. benthamiana leaves. The pair of Flag empty vector (EV-Flag) and HA-tagged RPW8.1 was used as a negative control. Total proteins were extracted and subjected to immunoprecipitation of ACO438-175 or RPW8.1 protein by the Flag antibody (α-Flag), followed by immunoblot analysis with the HA antibody (α-HA). The input proteins were analyzed with α-HA and α-Flag. The experiments were repeated for three times with similar results. (c, d) Y2H analysis shows the interaction between RPW8.1 and ACO438-175 (c), but not between RPW8.1 and ACO1 or ACO2 (d). Full-length RPW8.1 was translationally fused to the Gal4 DNAbinding domain (BD) of the pGBKT7 destination vector and the fusion protein serves as a bait.
ACO1, ACO2 and ACO4 truncated mutants were fused to the Gal4 activation domain (AD) of the pGADT7 destination vector and the fusion proteins serves as preys. Yeast cells grown on SD/-Trp/-Leu medium (-TL) indicated correct co-transformation and co-expression of different combinations of empty vectors and recombinant plasmids. Interactions were demonstrated by the growth of yeast cells on selective media SD/-Ade/-His/-Leu/-Trp (-TLAH), and -TLAH media supplemented with X-α-gal (40 µg ml -1 ) and AbA (200 ng ml -1 ). The pair of pGBKT7-53 (53) and pGADT7-T (T) was used as a positive control, pGBKT7-Lam (Lam) and pGADT7-T (T) was used as a negative control. Plates were photographed after 3 days.

Fig. S3
The aco4 mutant compromised in ethylene production.
Ethylene biosynthesis rate assay. Seedlings of 2-week-old Col-gl and aco4 mutant were used to measure the ethylene biosynthesis rates. Error bars indicated standard deviation (SD) (n=3). The asterisk (*) above the bar indicates significant differences (P<0.05) determined by Student's t-test. Reverse transcription quantitative PCR (RT-qPCR) assay. The relative expression levels of the indicated genes upon ACC (100 μM) treatment were calculated relatively to that of Mock treatment. ACT2 was used as an internal control. Error bars indicated SD (n=3). Asterisks (**) above the bars indicate significant differences (P<0.01) determined by Student's t-test. Reverse transcription quantitative PCR (RT-qPCR) analysis was used to determine the relative expression levels of RPW8.1 in R1Y4 plants grown on the 1/2 MS medium at the two different developmental stages. ACT2 was used as an internal control. Error bars indicate standard deviation (SD) (n=3). The asterisks (**) above the bars indicate significant differences (P<0.01) determined by Student's t-test.  accumulation in the indicated lines, respectively. Note that there is no any cell death in Col-gl and aco4, but some clusters of cell death in R1Y4 and massive clusters of cell death in aco4/R1Y4. Scale bars, 5 mm. (d) Electrolyte leakage analysis showing the conductivity of the leaves from the indicated lines. Leaves from the indicated lines were subjected to electrolyte leakage analysis at 6-week-old with three biological replicates. Error bars indicate standard deviation (SD) (n=3). Different letters above the bars indicate significant differences (P<0.01) determined by One-way ANOVA. (e) Quantitative analysis of the H2O2 accumulation in the leaves shown in (c) using Image J. Four independent leaves from each indicated line were assessed. Error bars indicate SD (n=4). Different letters above the bars indicate significant differences (P<0.01) determined by Oneway ANOVA. (f, g, h) Reverse transcription quantitative PCR (RT-qPCR) analysis shows the relative expressions of FRK1 (f), PR1 (g) and PR2 (h) in the indicated lines. Total RNA was extracted form 6-week-old plants. The expression level of each gene was calculated relative to that of Col-gl. ACT2 was used as an internal control. Error bars indicate SD (n=3). Asterisks (**) above the bars indicate significant differences (P<0.01) determined by Student's t-test.     Reverse transcription quantitative PCR (RT-qPCR) analysis was used to determine the relative expression levels of ERF6, ERF016 and ORA59 in Col-gl upon powdery mildew infection.
Samples were collected at the indicated time points. ACT2 was used as an internal control. Error bars indicated standard deviation (SD) (n=3). Different letters above the bars indicate significant differences (P<0.01) determined by One-way ANOVA. were generated for each gene in R1Y4 (b) and Col-gl (c) backgrounds, i.e., erf016/R1Y4-1 and erf016/R1Y4-2, erf6/R1Y4-1 and erf6/R1Y4-2, ora59/R1Y4-1 and ora59/R1Y4-2 in the R1Y4 background; erf016/Col-gl-1 and erf016/Col-gl-2, erf6/Col-gl-1 and erf6/Col-gl-2, ora59/Col-gl-1 and ora59/Col-gl-2 in the Col-gl background. Numbers indicate the deleted or inserted bases in the mutants. Among them, erf016/R1Y4-1 carried a 61-bp deletion resulting in frame shift at the codon for amino acid residue (aa) 134 that led to an early stop codon after aa 140 (Leucine); erf016/R1Y4-2 carried a 10-bp deletion resulting in frame shift at the codon for aa 133 that led to an early stop codon after aa 138 (Arginine); erf6/R1Y4-1 carried a 1-bp deletion resulting in frame shift at the codon for amino acid residue (aa) 172 that led to an early stop codon after aa 190 (Phenylalanine); erf6/R1Y4-2 carried a 1-bp insertion resulting in frame shift at the codon for aa 172 that led to an early stop codon after aa 174 (Glycine); ora59/R1Y4-1 carried a 1-bp insertion resulting in frame shift at the codon for amino acid residue (aa) 82 that led to an early stop codon after aa 81 (Serine); ora59/R1Y4-2 carried a 2-bp deletion resulting in frame shift at the codon for aa 82 that led to an early stop codon after aa 81 (Serine); erf016/Col-gl-1 carried a 1-bp insertion resulting in frame shift at the codon for amino acid residue (aa) 134 that led to an early stop codon after aa 137 (Arginine); erf016/Col-gl-2 carried a 1-bp insertion resulting in frame shift at the codon for aa 133 that led to an early stop codon after aa 137 (Arginine); erf6/Col-gl-1 carried a 1bp insertion resulting in frame shift at the codon for amino acid residue (aa) 172 that led to an early stop codon after aa 174 (Glycine); erf6/Col-gl-2 carried a 1-bp deletion resulting in frame shift at the codon for aa 172 that led to an early stop codon after aa 190 (Phenylalanine); ora59/Col-gl-1 carried a 1-bp insertion resulting in frame shift at the codon for amino acid residue (aa) 82 that led to an early stop codon after aa 81 (Serine); ora59/Col-gl-2 carried a 1-bp deletion resulting in frame shift at the codon for aa 82 that led to an early stop codon after aa 84 (Glutamic acid).   Reverse transcription quantitative PCR (RT-qPCR) assay was used to determine the expressions levels of RPW8.1 in the indicated plants upon ACC (100 μM) treatment, which were calculated relative to that of Mock treatment. ACT2 was used as an internal control. Error bars indicate SD (n=3). Asterisks (**) above the bars indicate significant differences (P<0.01) determined by Student's t-test. Table S1. The primers and probes used in this study.

Primers
Sequence Methods S1 Yeast two-hybrid (Y2H) assays Y2H was performed to screen for RPW8.1-interacting proteins. The coding sequences of RPW8.1 were amplified from the cDNA synthesized from RPW8.1's mRNA in R1Y4 with the genespecific primer pairs listed in Table S1 and were cloned into the vector pGBKT7 (BD) (Clontech), forming the bait construct. An Arabidopsis cDNA library was constructed by cloning the cDNA synthesized from the total RNA extracted from Golovinomyces cichoracearum UCSC1 infected Arabidopsis leaves into the Y2H vector pGADT7 (AD) (Clontech). This library was screened using RPW8.1-BD as a bait in the Saccharomyces cerevisiae strain Y2H Gold (Clontech). Positive clones were selected on the nutrient-deficient (SD/-Leu/-Trp/-His) media, and were further confirmed on the selective media SD/-Ade/-His/-Leu/-Trp. Twenty positive clones were obtained and the cDNA fragments encoding putative RPW8.1-interacting proteins were sequenced to reveal gene identities.
To generate the plasmids for Y2H analysis, we amplified the coding sequences of RPW8.1 and RPW8.2 from the cDNA of R1Y4 and R2Y4 (Huang et al., 2019) with the primer pairs indicated in Table S1. The cDNA fragments were cloned into the yeast bait vector pGBKT7 (BD) (Clontech) at the EcoRI/BamHI sites. The full length and truncated fragment of ACO4 were amplified from the cDNA of R1Y4 with the primer pairs indicated in Table S1 and were cloned into the yeast prey vector pGADT7 (AD) (Clontech) at the EcoRI/BamHI sites. The pairs of Y2H plasmids were cotransformed into the S. cerevisiae strain Y2H Gold (Clontech) according to the user's manual. The pair of pGADT7-T and pGBKT7-53 were used as a positive control, while pGADT7-T and pGBKT7-Lam were used as negative control. Transformants were selected using the nutrientdeficient (SD/-Leu/-Trp) medium at 30 ℃ for 3 days. Then positive transformants were subjected to a 10-fold serial dilution, and 6 μL droplets were placed on selective media SD/-Ade/-His/-Leu/-Trp (-TLAH), and -TLAH media supplemented with X-α-gal (40 µg ml -1 ) and AbA (200 ng ml -1 ) to assess protein and protein interaction. The results were obtained after 3 days at 30 ℃, the blue colonies indicated interactions.

Methods S2 Bimolecular fluorescence complementation (BiFC) assay
The coding sequences of RPW8.1, ACO4 and ACO438-175 without stop codon were amplified from their respective cDNA derived from the mRNA of R1Y4 with the primer pairs indicated in Table   S1. They were cloned into the pXY104 (carboxyl (C)-half of YFP, contained an HA tag) vector at the BamHI/SalI sites to make constructs expressing RPW8. Confocal microscopy (Nikon A1) was used to capture YFP fluorescence and chloroplast autofluorescence two days after infiltration.

Methods S3 Determination of ethylene biosynthesis rates and electrolyte leakage measurements
Ethylene production of Col-gl and R1Y4 was determined by gas chromatography (GC2014, Shimadzu, Japan) equipped with a flame ionization detector as described previously (Yang et al., 2017;Zhang et al., 2017) after slightly modified. Briefly, 2-week-old seedings without roots were used and placed in a 10 mL penicillin bottle for twenty-four hours with a minimum supply of distilled water, then ethylene in 1 mL gas sample was detected and quantified by the gas chromatography equipped. Ethylene biosynthesis rate was calculated in nL per gram fresh weight per hour [nL (FW.g.h) -1 ].
For electrolyte leakage analysis, the detached leaves were washed in deionized water for 2 hours, the conductivity was measured by pH/Water Quality Analyzer DS-70 (HORIBA). Then both of the leaves and water were autoclaved, the conductivity was measured again after cooling. The ratios of ion leakage before and after autoclave were calculated, the results are used to evaluate the degree of electrolyte leakage.

Methods S4
The leaf senescence assay The methods used for leaf senescence assay were conducted following modification from the previous report (Li et al., 2013). Briefly, the fifth and sixth rosette leaves from 5-week-old Arabidopsis plants were excised and floated on the distilled water with or without 100 µM ACC in the plastic petri dishes and kept at 22 ℃ under the dark condition for 4 days.

Methods S5 RNA extraction and reverse transcription quantitative PCR (RT-qPCR)
RNA extraction and RT-qPCR assay were performed following a previous report . Briefly, total RNA was extracted from the leaves using TRIzol reagent (Invitrogen) and the first strain cDNA was synthesized using ReverTra Ace® qPCR RT Master Mix with gDNA Remover kit (TOYOBO) according to the manufacturer's manual. RT-qPCR was conducted with the gene-specific primers and QuantiTect SYBR Green PCR Kit (QIAGEN) in Bio-Rad CFX96 Real-Time System (Bio-Rad). The primer pairs used for RT-qPCR are listed in the Table S1.
Relative expression levels were calculated with three technique repeats by the 2 -ΔΔCT method (Livak & Schmittgen, 2001). Statistical analysis was performed by one-way ANOVA or Student's t-test. Quantitative data ware processed by GraphPad Prism 7.0.

Methods S6 Pathogen inoculation and microscopy & analysis
Powdery mildew isolate Golovinomyces cichoracearum UCSC1 was maintained on the leaves of pad4/sid2 double mutant plants. The method used for inoculation and quantification of disease susceptibility was conducted following a previous report (Zhao et al., 2015). Briefly, 5-week-old plants were evenly inoculated with dislodged G. cichoracearum UCSC1 conidia that collected from the infected pad4/sid2 mutant leaves using a brush. To analysis the number of the spores, representative inoculated leaves were collected at 10 days post inoculation (dpi) and washed in 3 ml sterile water containing 0.01% Tween 20 by shaking for 20 min at 200 rpm, the blood cell counting plate was used for counting the spores under microscope (Zeiss imager A2). Cell death and H2O2 accumulation in the leaves of 6-week-old plants were examined by trypan blue staining and 3,3'-diaminobenzidine (DAB) staining, respectively (Xiao et al., 2003). Dying cells and the fungal structures in the inoculated leaves were analyzed by trypan blue staining at 10 dpi for powdery mildew according to the previously described methods (Xiao et al., 2003). All the images were processed using Zeiss LSM Image Browser, GraphPad Prism and Adobe Photoshop or Image J.

Methods S7 Bacterial growth assays
For bacterial growth assay, 5-week-old Arabidopsis plants were used and infiltrated with the virulent strain Pseudomonas syringae pv. tomato DC3000 (Pst DC3000) at the concentration of OD600 = 0.0005. Bacterial propagation was determined by colony counting as previously described (Li et al., 2010) at 0 and 3 dpi, respectively.

Methods S8 Luciferase (LUC) reporter assays in Nicotiana benthamiana
For dual-luciferase reporter assay, the coding sequences of ethylene response factor (ERF) genes were amplified from the cDNA of Col-gl with the gene-specific primer pairs indicated in Table   S1. The DNA fragments of each gene were cloned into the pCAMBIA1300 vector at KpnI/SpeI sites, forming the effecter constructs. The full-length of RPW8.1 promoter was amplified from the genomic DNA of R1Y4 with the primer pairs indicated in Table S1 and was cloned into pCAMBIA1300-LUC vector at the KpnI site, forming reporter construct (named ProRPW8.1-LUC).
RLUC (Renilla luciferase) driven by CaMV 35S was used as an internal control. Agrobacteria strains harboring recombinant plasmids were incubated in LB liquid medium containing kanamycin (50 μg mL -1 ), rifampin (50 μg mL -1 ) and gentamicin (50 μg mL -1 ) at 28°C for overnight on a shaking platform (200 r min -1 ). Subsequently, the bacteria were collected and adjusted to an OD600 value of 1 with MMA buffer (10 mM MgCl2; 10 mM MES; 200 mM acetosyringone). Resuspended Agrobacteria bacterial containing effecter and reporter were mixed in a 1:1 rate and coinfiltrated into N. benthamiana leaves. Total proteins were isolated using the Dual-Luciferase Reporter Assay System (Promega) following the manufacturer's manual two days after infiltration.
LUC and RLUC activity were detected using the Dual-Luciferase Reporter Assay System (Promega) in GLOMAX96 Microplate Luminometer system (Promega) according to the manufacturer's manual. Six independent replicates were performed.
For LUC reporter assay, Agrobacteria strain harboring the individual effecter plasmid was coinfiltrated into N. benthamiana leaves with ProRPW8.1-LUC. All the images were taken using a lowlight cooled CCD imaging apparatus two days after infiltration. The relative LUC fluorescence intensity were determined by Image J and normalized to eYFP control.

Methods S10 Protein expression and purification
The coding sequences of ORA59 were amplified from the cDNA of R1Y4 with the primer pairs indicated in Table S1 and were cloned into the pGEX-6p-1 vector at the BamHI/EcoRI sites to generate Glutathione-S-transferase (GST): ORA59. Escherichia coli BL21 (DE3) harboring the fusion construct was grown in the LB medium containing 100 μg mL -1 ampicillin at 37 ℃ to OD600 = 0.5. The expression of the fusion protein GST-ORA59 was induced by 0.2 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) and incubation at 16 ℃ for overnight. Glutathione-agarose beads (BD biosciences) was used for purifying GST-ORA59 protein according to the manufacturer's manual.
Methods S11 CRISPR/Cas9 plasmids construction and mutant screening To generate ERF6, ERF016 and ORA59 knockout mutants, CRISPR (clustered regularly interspaced short palindromic repeats)/Cas9 plasmids of each gene were constructed as described previously . Briefly, the guide RNA sequences in ERF6, ERF016 and ORA59 were screened and designed by CRISPR-PLANT system (https://www.genome.arizona.edu/crispr/CRISPRsearch.html), the special primers containing gene-specific spacer sequences were used and cloned into the binary vector pHEE401E at BsaI site, resulting in knockout constructs. All the constructs were verified by sequencing and transformed into R1Y4 and Col-gl via Agrobacterium strain GV3101-mediated transformation. In order to confirm the genotype of knockout lines. Hygromycin was used for screening positive T1 plants. Then genomic DNA was extracted from hygromycin-positive lines using CTAB method, PCR amplification was carried out using gene-special primers pairs. PCR products were sequenced to detect mutations. All the primers used in this study were listed in the Table S1.