Arabidopsis thaliana seeds were sown on a 2:1 mixture of vermiculite:peat (Finnpeat B2, Kekkilä Oyj, Tuusula, Finland) and stratified at 4°C for 2 days before being transferred to a growth chamber where they were grown at 23°C and a 12 h photoperiod with a light intensity of 280 μmol m^-2 s^-1 and 50–60% relative humidity. The following ecotypes and mutants were used: Columbia-0 (Col-0), C24, gin1-3 (aba2), and per57-1 (GK_325E06) were obtained from the Nottingham A. thaliana Stock Centre. The pyr1pyl1pyl2pyl4pyl5pyl8 (112458) sextuple mutant was obtained from Pedro L. Rodriguez, Instituto de Biología Molecular y Celular de Plantas. A collection of T-DNA activation tagged Arabidopsis lines in the ecotype C24 background was generated by utilizing the vector pSKI105, kindly provided by Detlef Weigel, Max Planck Institute for Developmental Biology.

Escherichia coli clones overexpressing the Pectobacterium wasabiae polygalacturonase gene pehA were generously provided by H. T. Saarilahti, University of Helsinki. The PehA protein was extracted and used to degrade commercially available polygalacturonic acid (Sigma–Aldrich) (Saarilahti et al., 1990). The concentration of the resulting OG-mixture was estimated in comparison to the concentrations of commercially available trimers on aluminum Silica gel 60_F254 TLC-plates and analyzed using mass spectrometry to confirm that the preparation consisted of OGs with a DP between 2 and 19 (Supplementary Figure S5).

Arabidopsis thaliana seeds were placed in individual wells of 12-well plates, with each well containing 1 ml of liquid ½12 MS medium (Gómez-Gómez et al., 1999). The plates were sealed with Parafilm to minimize evaporation. The seeds were stratified for 3 days at 4°C and then allowed to grow in a 12 h light cycle at 20°C. After 8 days, the medium was replaced with fresh liquid, and the seedlings were grown for 2 more days before treatment with OG-containing medium. After treatment, the plants were grown for 8 days before total plant weight was assessed. For all subsequent assays, four plants were combined into each biological replicate, and at least four biological replicates were tested per treatment. All experiments were performed at least three times. The average plant weight per treatment was compared across the treatments and statistically analyzed using Student’s t-test.

Toluidine blue (TB) tests were performed to assess cuticle permeability as previously described by L’Haridon et al. (2011) with the following modifications. Briefly, 6 μL of TB solution (0.05%) in ½14 PDB were applied onto fully expanded leaves, then covered with a transparent cover for 1 h. After 1 h leaves were washed with tap water and photographed. The photos were used to quantify the sizes of staining area in Fiji¹ as described by Cui et al. (2016). For staining with Calcofluor white (Supplementary Figure S4) leaves were decolorized overnight in 96% ethanol, incubated in 0.07 M sodium phosphate buffer (pH 9) for 1 h and incubated for 30 min in 0.05% Calcofluor white in 0.07 M sodium phosphate buffer (pH 9). Then leaves were rinsed in sodium phosphate buffer to remove excess of Calcofluor white, viewed and photographed under UV light on an InGenius LHR system (Syngene). To test cell wall fortification, the supernatant of an overnight culture filtrate of Pectobacterium carotovorum SCC1 was diluted 1:1 with water, and 1 μl of this diluted solution was placed on the adaxial side of the leaf surface of 3-week-old soil plants. The extent of maceration was documented using photography after 6 and 24 h.

Plant seedlings were grown in vitro as described by Ferrari et al. (2007), except that the growth medium base used in this study was 1/2MS and the photoperiod was 12 h. The concentrations that were used in this study were set primarily to induce significant growth retardation and so that the induction of defense genes could be observed. Briefly, the protocol for treating plants was similar to that used in previously described growth inhibition studies (Gómez-Gómez et al., 1999), except that 12 seeds were placed in each well for germination. The seedlings were treated with OG-mix (200 μM) or a mock mixture (1/2MS) for 3 h prior to RNA extraction. For each biological replicate, the plants were harvested from 3 wells, rinsed in MQ water and blotted dry before they were frozen in liquid nitrogen. The plant material was then crushed using metal beads in a shaker. RNA was extracted using a GeneJET Plant RNA Purification mini kit (Thermo Scientific) according to the manufacturer’s protocol. Sample quality was assessed using gel electrophoresis and measured using a NanoDrop (Thermo Scientific). A 2 μg sample of total RNA was treated with RNAse-free DNaseI (Thermo Scientific) and then used as the template for cDNA synthesis with a Maxima Reverse Transcriptase and Ribolock RNase Inhibitor (Thermo Scientific) according to the manufacturer’s protocol. Random hexamers and oligo dT primers were used for cDNA synthesis. For qPCR, HOT FIREPol EvaGreen qPCR Mix Plus (Solis Biodyne) was used in a reaction containing approximately 8 ng of cDNA in a total reaction volume of 10 μl. The reactions were performed on a BioRad CFX using the following amplification program: an initial activation cycle of 95°C for 15 min, 40 cycles at 95°C for 15 s, 60°C for 20 s and 72°C for 20 s, and a final melt curve analysis. The resulting qRT-PCR data were analyzed using geNorm software (Vandesompele et al., 2002). The housekeeping genes that were used as the reference genes for the qRT-PCR analysis were AT4G05320 (UBQ10), AT5G09810 (ACT7), AT5G12250 (TUB6), AT5G60390 (EF1α) and AT3G13920 (eIF4A). These genes were tested to determine their stability across cDNA samples using the geNorm M-value principle in Genex 6 software (MultiD Analyses AB, Gothenburg, Sweden). The two most stable reference genes (EF1α and UBQ10) had M-values < 0.3 across all samples and were therefore used for subsequent normalizations (Vandesompele et al., 2002; Beekman et al., 2011). All of the genes that were analyzed, along with their respective primers, are listed in Supplementary Table S1.

Total RNA was extracted from 3 biological replicates per treatment (OG-mix and mock) and purified to remove rRNA using an Epicenter Ribo-Zero rRNA removal kit (Plant leaf) according to the manufacturer’s instructions. The quality and the amount of the extracted RNA was assayed using a QuBit fluorometer (Life Technologies) and a NanoDrop before sequencing. Sequencing was performed on a SOLiD 5500XL platform. Each sample was tested as part of a pool and run in two separate lanes to generate approximately 9 million 75 bp-long single-end reads that were subsequently mapped to the genome using the SHRiMP platform (Rumble et al., 2009). A total of 33,597 gene sequences were used to align the reads from the sequencing data. Gene expression was considered to be significantly different at FDR values < 0.05 and a log₂ fold-change of ≥0.5 or ≤-0.5, as calculated using the R package DESeq2 (Love et al., 2014). The raw RNA sequencing data can be found in the NCBI GEO repository under the accession number GSE69538.

Pectobacterium carotovorum ssp. carotovorum SCC1 was cultured overnight in liquid Luria-Bertani (LB) medium at 28°C. The bacteria were pelleted using centrifugation and washed twice with and then resuspended in 10 mM MgSO₄. The amount of bacteria in the infection suspension was adjusted to 1 × 10⁵ cfu/ml. Infection was initiated by wounding the surfaces of 4-week-old plant leaves with a pipette tip syringe and then applying 5 μl of bacterial solution to the wound site. Three leaves were infected per plant, and these three leaves were combined to represent one biological sample. The plants were covered with plastic lids to keep the moisture level high.

Pseudomonas syringae pv. tomato DC3000 was grown in 5 ml liquid King’s broth (KB) medium overnight at 28°C, washed twice with 10 mM MgSO4, and then resuspended in 10 mM MgSO4 supplemented with 0.025% silwet. The amount of bacteria was adjusted to 1 × 10⁷ cfu/ml and sprayed onto 4-week-old plants. The plants were covered with plastic lids to keep the moisture level high. At the indicated time points, three leaves were obtained from each plant, and a 0.5 cm² leaf disk was harvested at each site of infection. The samples from one plant were combined to form one biological sample. The number of viable bacteria in each biological sample was then determined by counting cfus.

Botrytis cinerea Pers.:Fr strain B.05.10 was cultured on potato carrot agar (PCA) plates. Spores were harvested in ½14 strength potato dextrose broth (PDB) medium and filtered through Miracloth (Calbiochem) to remove hyphae. Droplets of 5 μL spore suspension (2 × 10⁵ spores mL^-1) were applied to the leaves of 4-week-old plants for quantification of lesion size (mm) after 72 h. Three fully expanded leaves per plant were infected. The plants were covered with plastic lids to keep the moisture level high and transferred to growth chamber 180 μmol m^-2 s^-1, 8 h light/16 h dark at 23°C/18°C (day/night).

The following binary T-DNA destination vectors were used; pMDC32 (Curtis and Grossniklaus, 2003) and pHGWFS7.0 (Karimi et al., 2002). A 1834-bp sequence corresponding to the PER57 promoter was amplified from Col-0 genomic DNA using primers listed in Supplementary Table S1. Agrobacterium tumefaciens carrying pMP90 was used to transform A. thaliana via floral dipping (Clough and Bent, 1998). T₀ seeds were selected on MS medium containing the appropriate antibiotic. Stable transgenic lines showing either an increased expression level by qPCR or reporter gene signal were used for detailed analysis in this study.

DAB and NBT staining for hydrogen peroxide and superoxide, respectively, were performed as described in Piisilä et al. (2015). In situ detection of hydrogen peroxide was performed by vacuum infiltration with 0.1% DAB (Diaminobenzidine tetrahydrochloride, Sigma–Aldrich) in 10 mM MES [2-(N-morpholino) ethanesulfonic acid], pH 6.5, for 30 min. For superoxide staining detached leaves were first vacuum infiltrated with 10 mM K₂PO₄ buffer, pH 7.8, for 30 min and then the buffer was replaced with 0.1% NBT (Nitroblue tetrazolium, Sigma–Aldrich) in 10 mM K₂PO₄ buffer and the leaves were left under table lights for 30 min. After staining the leaves were cleaned by boiling in alcohol-lactophenol (2:1) for 5 min, and then rinsed twice with 50% ethanol and once with MQ water.

In situ detection of callose was performed as described in Daudi et al. (2012). Briefly, three fully expanded leaves from 24- to 28-day-old A. thaliana plants were syringe-infiltrated with ∼100 μL of a diluted microbial elicitor solution (1 μM flg22 or 5 mM OG) and covered with a transparent cover to maintain high humidity for 20 h. At least eight independent plants were used as biological replicates. After 20 h incubation leaves were harvested and placed in sterile 12-well plates. Then washed with acetic acid:ethanol (1:3) for 6 h to destain the chlorophyll from the leaves followed by ethanol (50% v/v) 1 h, ethanol (30% v/v) 1 h and finally MQ water for 2 h. Leaves were stained with a phosphate buffer (pH 7.5) containing 5 mg/mL aniline blue (Sigma) for 20 h in the dark, transferred to 50% glycerol (Sigma) and examined under fluorescence microscopy. For the calculation of callose deposits, the following Fiji² tools were used: ‘Auto Threshold,’ ‘Make Binary,’ and ‘Analyze Particles.’

The apoplastic proteins were extracted from leaves using methods from Bindschedler et al. (2006). Briefly, apoplastic proteins were extracted from leaves by grinding them in liquid nitrogen with mortar and pestle. Powder was suspended in 100 mM Tris-HCl buffer pH 7.5 containing 250 mM sucrose and protease inhibitors and centrifuged at 16 100 g for 5 min at 4°C. Pellet was resuspended in 100 mM Tris-HCl pH 7.5 1 M NaCl, 1 mM CaCl₂, 1 mM MgCl₂ and ionically bound proteins from cell walls were extracted by vortexing 30 min at 4°C. After centrifugation at 16 100 g for 5 min at 4°C, and the supernatant was used as the crude extract for peroxidase activity assay. Total protein in the crude extract was estimated by Bradford assay using a Bio-Rad protein assay dye reagent (Biorad).

Peroxidase activity from the crude extract was determined from the rate of oxidation of guaiacol to tetraguaiacol, by monitoring the increase of absorbance in 470 nm (Li et al., 2008). Briefly, 10 μl of the crude extract was added to 150 μl of 5 mM sodium phosphate buffer pH 6.0, in 96 wells microplate. To establish back ground absorbance, 20 μl of 1 % (v/v) guaiacol (Sigma–Aldrich) in sodium phosphate buffer pH 6.0 was added to the reaction mixtures in the microplate and the absorbance at 470 nm was measured after 5 min incubation in room temperature. The oxidation reaction was started by adding 20 μl of 3% hydrogen peroxidase (Sigma–Aldrich) in sodium phosphate buffer pH 6.0 and the increase in absorbance at 470 nm was recorded for 3 min using a plate reader (EnSpire 2300, Perkin Elmer). Linear regression was applied to the absorbance data and peroxidase activity was calculated as a slope of the regression line. The peroxidase activity was normalized using the total protein content in the crude extract.

The effect of ABA and diphenylene iodonium (DPI), an inhibitor of NADPH oxidase-dependent oxidative burst, on cuticle permeability, ROS formation and pathogen infections was determined after leaves were sprayed under humid conditions with 100 μM ABA (Sigma–Aldrich) on developmental days 7 and 14, alternatively 50 μM DPI (Sigma–Aldrich) on developmental day 20. The permeability of the cuticles, ROS formation and symptoms of infection were evaluated 7 days post-ABA treatment and 24 h post-DPI treatment.

To identify and characterize OG-responsive components that are involved in innate immunity, a population of 62000 T-DNA activation tagged A. thaliana ecotype C24 lines, generated previously in our laboratory by using the vector pSKI015 (Weigel et al., 2000), were treated with OG elicitors. The screen was based on OG-triggered growth retardation of A. thaliana seedlings grown in liquid ½12 MS on 96-well plates that allowed high throughput detection of the mutant phenotypes. The screen yielded 46 activation-tagged mutant lines impaired in their growth responses to OGs. Out of the 46 activation-tagged mutant lines that displayed impaired OG responses, one mutant line that was called ohy1 (OG hypersensitive 1) exhibited a strong hypersensitive response to OGs (DP 2–19) (Figures 1B,C) and was therefore chosen for further characterization. The ohy1 mutant was smaller in size (Figure 1A) than the wild-type plants. A co-segregation analysis of progeny obtained from backcrosses of ohy1 and wild-type plants indicated that the mutation linked to the observed phenotypes was dominant and that the phenotype was dependent on the inserted T-DNA. We used next-generation Illumina sequencing to track the location of the T-DNA insertion in the ohy1 mutant within the extragenic DNA region between genes At5g17820 and At5g17830. The CaMV 35S enhancers that were present in the T-DNA were found to be 608 bp upstream of the At5g17820 (PER57) start codon (Figure 1D), and the expression of the PER57 gene encoding a CIII peroxidase was clearly enhanced (Figures 1E,F).

The observed overexpression of PER57 in ohy1 suggested that the accumulation of ROS might be enhanced in these plants. To test this, we characterized hydrogen peroxide and superoxide levels in ohy1 and wild-type plants using diaminobenzidine (DAB) and nitroblue tetrazolium (NBT) staining, respectively (Figure 1G). Indeed, the overexpression of PER57 resulted in higher levels of superoxide and hydrogen peroxide accumulation in the leaves of the ohy1 plants than in the wild-type plants, indicating enhanced peroxidase activity (Figure 1G). To confirm that the up-regulation of PER57 in ohy1 plants was responsible for the observed phenotypes, we used the CaMV35S promoter to drive the expression of the PER57 cDNA in wild-type Col-0 plants. Of the three independent transgenic lines that were analyzed, all lines displayed phenotypes that were identical to the ohy1 mutant in that they were smaller than wild-type plants (Figure 1H) and exhibited both hypersensitivity to OG elicitors (Figure 1I) and an increase in the accumulation of superoxide and hydrogen peroxide (Figure 1J). The phenotypes observed in the gain-of-function transgenic lines confirmed that the increased expression of PER57 was indeed responsible for the phenotypes that were observed in the ohy1 mutant plants.

Based on the finding that the overexpression of PER57 conferred enhanced sensitivity to OGs, manifested as increased growth retardation and ROS formation (Figure 1), we hypothesized that ohy1 plants might also exhibit enhanced DAMP-responses, including an increase in the expression of OG-marker genes. To explore this hypothesis, we treated plants with OGs and monitored the expression of known markers of the OG (PAD3, PER4, PGIP1 and GST1), SA (PR1, PR5, and SID2) and JA (VSP2) signaling pathways. The ohy1 line showed no significant changes in markers of SA and JA signaling in response to OGs (Supplementary Figure S1). In contrast, the expression levels of the OG signaling marker genes were much higher in ohy1 plants than those observed in the wild-type controls (Figure 2A). Interestingly, the expression levels of defense genes in non-treated ohy1 plants were similar to those in wild-type plants, but faster and stronger induction was observed in response to OGs in the ohy1 plants. Consequently, we hypothesized that defense responses in general could be primed in the ohy1 mutant plants. To confirm this, we analyzed the level of induction of defense genes in response to the bacterial PAMP flg22 (Figure 2A) and the accumulation of callose in response to both OG- and flg22 elicitors (Figure 2B). We found that flg22 triggered a stronger level of expression of defense genes in ohy1 than those observed in the wild-type controls. In agreement with the observation of an enhanced expression of defense genes in ohy1, we also observed an increase in callose deposits in response to both elicitors (Figures 2B,C), which support the notion that defense responses are indeed primed in ohy1 plants.

Consistent with the stronger induction of defense-related genes and enhanced callose deposition in the ohy1 line, we hypothesized that the response to pathogens could also be enhanced in these plants. To explore the potential effect of the enhanced expression of PER57 in disease resistance, we infected ohy1 and wild-type plants with two necrotrophic pathogens: the soft-rot bacterium P. carotovorum and the gray mold fungus B. cinerea. Interestingly, the ohy1 mutant displayed immunity to B. cinerea and a clearly higher level of resistance to P. carotovorum in comparison to the wild-type controls (Figures 2D–G). To determine whether ohy1 plants demonstrated enhanced resistance to other than necrotrophic pathogens, we infected mutant and wild-type plants with the hemibiotroph P. syringae (Figures 2H,I). Surprisingly, in contrast to the observed increase in resistance to necrotrophic pathogens, the ohy1 line was clearly more susceptible than the wild-type plants to P. syringae. These results indicate that the overexpression of PER57 in the ohy1 line primes plant defenses to necrotrophic pathogens, while sensitizing the plant to pathogens with a hemibiotrophic lifestyle.

Our results (Figure 2) showed that the overexpression of PER57 in the ohy1 plants conferred resistance to B. cinerea and P. carotovorum. Interestingly, similar resistance phenotypes have been observed in plants with cuticular defects showing immunity to B. cinerea (Bessire et al., 2007; Voisin et al., 2009; Chassot et al., 2007; L’Haridon et al., 2011). This similarity prompted us to explore whether the permeability of the leaf cuticle was increased in the ohy1 mutant. We examined the permeability of the cuticle of ohy1 leaves by applying the hydrophilic agent TB (Tanaka et al., 2004) to the adaxial sides of the leaves (Figures 3A,B). As expected, no staining was observed in the leaves of wild-type plants, but dark blue staining was clearly visible in the leaves of ohy1 mutant plants as early as 1 min after the TB was applied. These results indicate that the permeability of the leaf cuticle was increased in the ohy1 plants (Figures 3A,B). This increase in cuticular permeability could be associated with a decrease in the fortification of the cell walls. When drops of the P. carotovorum cellular culture filtrates enriched in cell wall-degrading enzymes (CWDE) were applied to the adaxial side of the leaves, increased leaf maceration was observed in ohy1 plants in comparison to the wild-type controls (Figure 3C). This data indicates a clear reduction in the fortification of the cell walls of the ohy1 plants.

To explore the mechanisms by which peroxidase-generated ROS increase permeability in leaf cuticles, we first hypothesized that these ROS might have a negative effect on the expression of genes that affect cutin biosynthesis. Therefore, the expression levels of MYB96, a positive regulator of cutin formation (Seo et al., 2011; Cui et al., 2016), and BDG and LACS2.3, which are both major cutin-biosynthetic genes, were analyzed in the ohy1 line. The expression levels of these genes were strongly down-regulated in the ohy1 line, suggesting that the loss of cuticle integrity was influenced by impaired cutin biosynthesis (Figure 3D). In addition, the enhanced accumulation of cuticular waxes has been observed in bdg, lacs2.3, fdh and lcr mutants that display impaired cuticle integrity and may suggest a mechanism involving a compensatory reaction to the loss of cuticle integrity (Kurdyukov et al., 2006; Bessire et al., 2007; Voisin et al., 2009). To test whether this was the case also with ohy1 plants, we characterized the expression of the cuticular wax-biosynthetic genes CER1, LTP3, LTP4, and KCS2 in wild-type and ohy1 plants. As expected, the expression levels of these genes were much higher in the ohy1 plants than in the wild-type controls (Figure 3E).

In summary, our results indicate that CIII peroxidase-derived ROS in the ohy1 plants modulate the expression of genes known to be associated with cuticle biosynthesis, which leads to permeabilization of the cuticle.

To determine the expression profile of PER57 in vivo, we generated transgenic A. thaliana plants that expressed the GUS reporter gene under the control of the PER57 promoter. The generated transgenic plants were sprayed either with OGs or flg22, or wounded with laboratory forceps. GUS-staining of the transgenic plants indicated that PER57 appears to be specifically expressed in roots and no GUS-activity was detected in leaves even in response to wounding or flg22 and OG elicitors (Figure 4A).

CIII peroxidases exist as large multigene families and often appear to have redundant functions (Cosio and Dunand, 2009), several of which are expressed in leaves, which could explain the effect obtained by ectopic expression of PER57. Consequently, overexpressing other CIII peroxidase genes might result in the same phenotypes that were observed in the PER57 overexpressors, including the ohy1 mutant. To test if the phenotypes observed in PER57 overexpressors is a general effect of elevated levels of CIII peroxidases, we generated transgenic lines that overexpressed six different CIII peroxidases, which, according to our RNA sequence analysis of 10-day-old OG-treated A. thaliana seedlings (Supplementary Table S2; Supplementary Figure S3) were either responsive: PER10, PER28, and PER34, or non-responsive: PER44, PER53, and PER64 to OG elicitors. Overexpression of all six CIII peroxidase genes resulted in phenotypes that were identical to the phenotype of the ohy1 mutant, including an increase in the permeability of the leaf cuticle (Figures 4B,E), a dramatic increase in the formation of superoxide (Figure 4C) and a strong increase in resistance to B. cinerea (Figures 4D,F). The results demonstrated that the effect of PER57 overexpression on cuticular integrity and resistance to pathogens is not specific and the same effect could be achieved by overexpressing other apoplastic CIII peroxidases.

Abscisic acid has been shown to influence cuticle formation in several studies (Asselbergh et al., 2007; L’Haridon et al., 2011). To test whether this was the case in CIII peroxidase overexpressors, we analyzed the expression levels of MYB96, BDG and LACS2.3 in ohy1 plants that were treated with ABA. Interestingly, applying ABA completely restored normal expression levels of these genes in the ohy1 line (Figure 3D). Furthermore, similar to the effect of ABA on the expression of cutin biosynthesis genes, the transcription of CER1, LTP3, LTP4, and KCS2 was increased after ABA treatment (Figure 3E). Apparently, ABA plays an important role in activating genes known to be involved in cuticular wax biosynthesis.

Our data suggest that both CIII peroxidase-derived ROS and perturbed ABA levels affected the expression of genes known to be associated with cuticle biosynthesis. To further study the effect of ABA-signaling in cuticle permeability, we next sought to determine the expression levels of BDG and LACS2.3 in ABA-deficient aba2 and ABA-insensitive pyr/pyl 112458 sextuple (Gonzalez-Guzman et al., 2012) mutant plants treated either with mock or ABA. The cuticular genes BDG and LACS2.3 were strongly down-regulated in the mock-treated aba2 and pyr/pyl 112458 mutant plants (Figure 5A). Exposure to ABA restored the expression of these cuticular genes in the aba2 mutant. However, as predicted, the expression of these genes remained strongly down-regulated in the ABA-insensitive pyr/pyl 112458 mutant (Figure 5A).

Because we observed a strong positive correlation between higher levels of peroxidase-derived ROS and increased cuticle permeability (Figures 5B,D), we next sought to explore the effect of ABA on ROS formation in the ohy1 mutant, ABA biosynthesis mutant aba2, and ABA signaling mutant pyr/pyl 112458. The mutant and wild-type plants were sprayed with 100 μM ABA, which resulted in the removal of ROS in all the lines except for the ABA insensitive pyr/pyl 112458 plants. Furthermore, the ABA-induced decrease in the accumulation of ROS was accompanied by the restoration of the leaf cuticle (Figure 5B).

To determine whether NADPH oxidase contributed to the accumulation of ROS in the plants with increased leaf cuticle permeability, we sprayed these plants with 50 μM DPI, which is an inhibitor of NADPH oxidase-dependent oxidative bursts (Figure 5B). However, DPI had no significant effect on ROS formation nor on cuticle permeability (Figure 5D) in any of the tested lines, indicating that these ROS are likely to be produced primarily by peroxidases and not by NADPH oxidases.

We further performed peroxidase activity assays to determine whether ABA had an effect on peroxidase activities in ohy1, aba2 and pyr/pyl 112458 mutant plants. The plants were sprayed either with ABA or mock (Figure 5C). In the absence of ABA, all of the tested mutant plants exhibited higher levels of peroxidase activity than were observed in the wild-type plants. Interestingly, the addition of ABA reduced peroxidase activity to wild-type levels only in aba2-1 mutant. While no significant changes were observed in peroxidase activity in the ABA-insensitive pyr/pyl 112458 mutants, the ohy1 mutant plants continued to demonstrate relatively high peroxidase activity despite their exposure to ABA. These data suggest that while the application of ABA outpaces the constitutive production of ROS by the peroxidases, the ABA-induced removal of ROS has little effect on the actual activity of the peroxidases.

Based on the finding that the application of ABA had a significant impact on the integrity of the cuticle, we suspected that ABA might also influence the enhanced pathogen resistance that was observed in the ohy1 line and in the mutants with impaired ability to respond to or biosynthesize ABA. Indeed, pre-treatment of the ohy1 and aba mutants with 100 μM ABA, led to wild-type level susceptibility to B. cinerea, whereas no alterations in the susceptibility was detected in the ABA-insensitive pyr/pyl 112458 sextuple mutant (Figure 5E).

In summary, our results show a negative effect of ABA on ROS formation, and point to the involvement of CIII peroxidases as a source of ROS in aba mutants.