Increased Resistance to Biotrophic Pathogens in the Arabidopsis Constitutive Induced Resistance 1 Mutant Is EDS1 and PAD4-Dependent and Modulated by Environmental Temperature

The Arabidopsis constitutive induced resistance 1 (cir1) mutant displays salicylic acid (SA)-dependent constitutive expression of defence genes and enhanced resistance to biotrophic pathogens. To further characterise the role of CIR1 in plant immunity we conducted epistasis analyses with two key components of the SA-signalling branch of the defence network, ENHANCED DISEASE SUSCEPTIBILITY1 (EDS1) and PHYTOALEXIN DEFICIENT4 (PAD4). We demonstrate that the constitutive defence phenotypes of cir1 require both EDS1 and PAD4, indicating that CIR1 lies upstream of the EDS1-PAD4 regulatory node in the immune signalling network. In light of this finding we examined EDS1 expression in cir1 and observed increased protein, but not mRNA levels in this mutant, suggesting that CIR1 might act as a negative regulator of EDS1 via a post-transcriptional mechanism. Finally, as environmental temperature is known to influence the outcome of plant-pathogen interactions, we analysed cir1 plants grown at 18, 22 or 25°C. We found that susceptibility to Pseudomonas syringae pv. tomato (Pst) DC3000 is modulated by temperature in cir1. Greatest resistance to this pathogen (relative to PR-1:LUC control plants) was observed at 18°C, while at 25°C no difference in susceptibility between cir1 and control plants was apparent. The increase in resistance to Pst DC3000 at 18°C correlated with a stunted growth phenotype, suggesting that activation of defence responses may be enhanced at lower temperatures in the cir1 mutant.


Introduction
Plants have a robust innate immune system that affords protection against attack by potential pathogens in their local environment. Detection of pathogen associated molecular patterns (PAMPs) such as flagellin by pattern recognition receptors at the plasma membrane leads to activation of PAMP-triggered immunity (PTI) [1,2]. Successful phytopathogens have evolved mechanisms, including effectors, to subvert or suppress PTI, allowing them to successfully colonise the plant host [1,3]. This in turn led to the evolution of effector-triggered immunity (ETI) in plants, which relies on the direct or indirect detection of pathogen effectors by cognate host resistance (R) proteins [1,4]. While there is a significant overlap between these two branches of the innate immune system, ETI is generally regarded as a stronger and more rapid response, and is associated with the hypersensitive (HR) response [1,3]. The final layer of innate immunity is systemic acquired resistance (SAR), whereby infection of one part of a plant leads to increased resistance of uninfected tissues to subsequent pathogen challenge [4]. SAR is thought to be established by coordinated expression of an array of anti-microbial pathogenesis related (PR) genes [4,5]. All three branches of innate immunity rely on large scale transcriptional re-programming of the host plant, activated via a complex network that is influenced by the crosstalk between salicylic acid (SA), jasmonic acid (JA) and ethylene (Et) signalling [6,7].
Two key signalling components in the SA-signalling branch of the defence network are ENHANCED DISEASE SUSCEPTI-BILITY1 (EDS1) and PHYTOALEXIN DEFICIENT4 (PAD4) [8,9], both of which show homology to eukaryotic acyl lipases. The EDS1-PAD4 node has long been recognised as a central regulator of PTI and of Toll-interleukin-1 receptor-nucleotide bindingleucine-rich repeat (TIR-NB-LRR) R protein mediated ETI against biotrophic and hemi-biotrophic pathogens [9][10][11]. Recent evidence suggests that EDS1 may also play a role in ETI mediated via coiled coil-NB-LRR R proteins, as EDS1 and SA accumulation have been shown to function redundantly in RPS2 and RPP8mediated resistance against avirulent pathogens [12]. Both EDS1 and PAD4 are required for SA accumulation in response to Pseudomonas syringe pv. tomato (Pst) DC3000 or Pst DC3000 avrRps4 [13], and EDS1 and PAD4 gene expression is SAinducible, suggesting the existing of a positive feedback loop [11,13].
Protein-protein interaction studies have identified the presence of EDS1 homodimers, as well as EDS1-PAD4, EDS1-SAG101 and EDS1-PAD4-SAG101 protein complexes in plant cells [13][14][15][16]. The formation of the EDS1-PAD4 complex is required for PTI against virulent pathogens, full accumulation of SA and the establishment of SAR, but not for TIR-NB-LRR mediated ETI [15]. While EDS1 homodimers are present predominantly in the cytoplasm, the EDS1-PAD4 complex is found in the cytoplasm and nucleus, and it has been suggested that nuclear EDS1 acts as a transcriptional regulator [14,15]. Enhanced export of EDS1 from the nucleus was found to increase susceptibility to both virulent and avirulent Pst DC3000, as well as Hyaloperonospora arabidopsidis Emwa1 [17], but co-ordination of cytoplasmic and nuclear EDS1 levels may also be important in the plant immune response [17]. In line with its central role in innate immunity in Arabidopsis, EDS1 is targeted by the Pst effectors AvrRps4 and HopA1, and in accordance with the guard hypothesis of Van der Biezen and Jones [18], EDS1 is found in association with the cognate TIR-NB-LRR R proteins RPS4 and RPS6 [19].
EDS1 and PAD4 were identified in Arabidopsis by screening for altered susceptibility to pathogen challenge, and mutant screens have been widely used to dissect the defence signalling network. One class of gain-of-resistance mutants that display SAR-like constitutive disease resistance are the constitutive expressor of PR genes (cpr) mutants [20][21][22]. These mutants display SA signallingdependent constitutive expression of PR genes and enhanced resistance to virulent biotrophic pathogens [21]. The cpr-type mutants can broadly be divided into two groups [23], those that display constitutive HR-like cell death such as cpr5 and lsd1, and those that do not, including cpr1 and dnd1. The cir1 (constitutively induced resistance 1) mutant belongs to the second class of cpr mutants, and was identified in a mutant screen for increased luciferase activity in Col-0 plants carrying a PR-1:LUC reporter [24]. The cir1 mutation is recessive, and homozygous cir1 plants display increased resistance to virulent Pst DC3000 and H. arabidopsidis and constitutive expression of SA-dependent defence genes such as PR-1, PR-5 and WRKY53 (as well as the JA/Etdependent PDF1.2) in the absence of pathogen challenge [24,25]. As reported for other cpr-type mutants, SA accumulation is essential for the increased resistance to virulent biotrophic pathogens displayed by cir1, which appears to be mediated by both NPR1-dependent and independent signalling pathways, since cir1 npr1 double mutants displayed only partial suppression of cir1-mediated resistance [24]. Although cir1 displays increased PDF1.2 expression, it does not display increased resistance to the necrotrophic fungal pathogen Botrytis cinerea [23]. The CIR1 gene maps to the lower arm of chromosome IV and complementation tests have revealed that it is not allelic to previously reported cpr mutations in this region including cpr1 [24].
Epistasis analyses have revealed that EDS1 and PAD4 are required for constitutive PR expression and enhanced disease resistance in several cpr mutants including cpr1 and cpr6 [21,26]. Given these results and the pivotal role of the EDS1-PAD4 regulatory node in SA-mediated defence against biotrophic pathogens we investigated whether cir1-mediated resistance to Pst DC3000 and H. arabidopsidis also requires EDS1 and/or PAD4, and whether CIR1 might in turn regulate EDS1 expression. Our data indicate that CIR1 is a negative regulator of innate immunity that lies upstream of EDS1 and PAD4 in the defence signalling network, and suggest that CIR1 may be involved in the post-transcriptional regulation of EDS1. In addition, we show that the defence and growth phenotypes of the cir1 mutant are modulated by environmental temperature.

Results
cir1-mediated resistance to Pst DC3000 and H. arabidopsidis requires EDS1 and PAD4 The increased resistance to Pst DC3000 and H. arabidopsidis Noco2 displayed by the cir1 mutant has previously been shown to be SA-dependent [24]. EDS1 and PAD4 are two key players in  defence against Pst DC3000 and H. arabidopsidis and are essential for SA accumulation in response to infection by these pathogens [13]. To determine whether cir1-mediated resistance to these pathogens is also dependent on EDS1 and PAD4, we generated cir1 eds1 and cir1 pad4 double mutants and examined their disease susceptibility profiles.
As eds1 is in the Ler background, while cir1 is in the Col-0 background, we analysed five independently generated cir1 eds1 double mutants to control for any effects of a mixed Col-0/Ler background on resistance to Pst DC3000. Bacterial titres in these five lines at 48 h post-infection (hpi) were not significantly different from those observed in the eds1-2 mutant, while those observed in cir1 were significantly lower ( Figure 1). There was no significant difference in bacterial titres between the cir1 mutant in the Col-0 background (cir1) and cir1 plants generated by crossing the single mutant to wild-type Ler plants (cir1 Ler), indicating that a mixed Col-0/Ler genetic background has no effect on cir1-mediated resistance to Pst DC3000 ( Figure 1). Similar results were observed for PAD4; cir1 displayed significantly lower bacterial titres 48 hpi than pad4-1, while no statistically significant difference in bacterial titres was observed between pad4-1 and cir1 pad4 plants ( Figure 1). Thus, both EDS1 and PAD4 are required for cir1mediated resistance to Pst DC3000, and are epistatic to CIR1.
We also examined the resistance profile of the double mutants to H. arabidopsidis Noco2. This pathogen is virulent on the Arabidopsis Col-0 ecotype, but avirulent on Ler due to the presence of the RPP5 resistance gene [27]. EDS1 is required for RPP5-mediated resistance to H. arabidopsidis [9], as reflected in the significantly higher conidiospore production and mycelial growth observed in eds1-2 compared to Ler (Figure 2 and S1). Conidiospore production in the five independently derived cir1 eds1 lines was not significantly different from that observed in eds1-2, while that in cir1 was significantly lower (Figure 2 and S1). Again, similar results were observed with the pad4 mutants, with no statistically significant difference in conidiospore production between pad4-1 and cir1 pad4. However, disease symptoms were less severe in pad4-1 in comparison to eds1-2 as previously reported [15]. Together these data suggest that EDS1 and PAD4 are also required for cir1-mediated resistance to H. arabidopsidis, and again function downstream of CIR1. We observed that cir1 in the Ler (cir1 Ler) background displayed the reduced susceptibility to H. arabidopsidis observed in cir1 in the Col-0 backround, rather than the total resistance displayed by wild-type Ler plants ( Figure 2). Intermediate susceptibility to this pathogen has previously been reported in Col-06Ler crosses [27].
EDS1 and PAD4 are required for cir1-modulated defence gene expression As both EDS1 and PAD4 are required for cir1-mediated resistance against virulent biotrophic pathogens, we examined the expression of two downstream defence genes that are SA responsive and require EDS1 and PAD4 for up-regulation in response to Pst DC3000 infection [28,29]. Quantitative PCR analysis confirmed that At2g14160 (PR-1) and At2g31880 (suppressor of BAK1-interacting receptor-like kinase 1, SOBIR1) were up-regulated in Col-0 plants 24 hpi with Pst DC3000, and that mRNA levels were elevated in uninfected cir1 plants compared to uninfected Col-0 plants ( Figure 3). However, the cir1 eds1 and cir1 pad4 double mutants again phenocopied the single eds1-2 or pad4-1 mutants rather than cir1, as mRNA levels of both genes were not elevated in these plants ( Figure 3). These data indicate that EDS1 and PAD4 are required for elevated constitutive expression of these defence genes in cir1, and support the hypothesis that CIR1 is upstream of the EDS1-PAD4 regulatory node.

EDS1 expression is constitutively higher in the cir1 mutant
Given that CIR1 functions upstream of EDS1 and PAD4, we next examined whether CIR1 might regulate EDS1 expression in Arabidopsis. Western blot analysis revealed that EDS1 proteins levels were constitutively higher in uninfected cir1 plants in comparison to PR-1:LUC control plants (the genetic background for the cir1 mutant), and also in cir1 pad4 plants versus the pad4-1 single mutant ( Figure 4A). However, analysis of EDS1 steady-state transcript levels using qPCR revealed no statistically significant difference between cir1 and PR-1:LUC control plants ( Figure 4B), suggesting that CIR1 may exert its regulatory effect on EDS1 via a post-transcriptional mechanism.
The cir1 growth and disease resistance phenotypes are temperature-sensitive A number of gain-of-resistance mutants display a stunted growth phenotype, which is thought to result from the fitness cost of constitutive activation of immune responses [30]. In several mutants, including cpr1, bonzai1 (bon1) and suppressor of npr1-1, constitutive 1 (snc1), this phenotype is temperature dependent, manifesting only at lower growth temperatures [22,31]. While cir1 does not exhibit the dwarf stature characteristic of these mutants at 22uC (our standard growth temperature for Arabidopsis), it does display moderately reduced stature in comparison to PR-1:LUC control plants ( Figure 5). We thus examined the growth phenotype and resistance to Pst DC3000 of the cir1 mutant when grown at 18, 22 or 25uC. We observed that cir1 plants displayed greatly reduced stature compared to PR-1:LUC control plants at 18uC, while at 25uC no obvious difference in size was evident ( Figure 5). ANOVA of Pst DC3000 titres 48 hpi revealed significant (p, 0.001) effects of both temperature and genotype on resistance to Pst DC3000, and a significant interaction term (genotype*temperature, p = 0.015), indicating that the effect of genotype on resistance to Pst DC3000 is modulated by environmental temperature. While bacterial titres were on average 8.1-fold (6 2.3) lower in cir1 plants compared to PR-1:LUC plants when grown at 22uC, at 18uC bacterial titres in the cir1 mutant were on average 30.1-fold (68.1) lower than those in control plants ( Figure 6). In contrast, when grown at 25uC there was no   (Figure 6), suggesting that enhanced resistance to Pst DC3000 was abolished in the cir1 mutant at this temperature. The auto-immune phenotypes displayed by several cpr-type mutants including cpr1 and bon1 when grown at 22uC have been linked to increased expression of the TIR-NB-LRR protein SNC1, with increased SNC1 mRNA levels reported in cpr1 and bon1 [31,37]. However, qPCR analysis revealed that SNC1 transcript levels were not significantly higher in cir1 versus PR-1:LUC plants when grown at either 18 or 22uC (Figure 7).

Discussion
The EDS1-PAD4 regulatory node plays a critical role in both PTI and ETI against biotrophic pathogens in Arabidopsis. Here we investigated whether the enhanced resistance to Pst DC3000 and H. arabidopsidis and constitutive expression of SA-dependent defence genes displayed by the cir1 mutant requires EDS1 and PAD4 by generating double mutants. We observed that both cir1 eds1 and cir1 pad4 plants displayed the enhanced susceptibility to Pst DC3000 and H. arabidopsidis that is characteristic of the single eds1-2 and pad4-1 null mutants (Figures 1 & 2). Similarly, analysis of steady state levels of two SA-dependent defence genes revealed that their elevated expression in uninfected cir1 plants requires both EDS1 and PAD4 (Figure 3). These results are in line with those previously observed for a number of other gain-ofresistance mutants including cpr1, cpr6, bon1, snc1 and suppressor of rps4-RLD 1 (srfr1) [21,22,[32][33][34], and indicate that the SARlike constitutive disease resistance displayed by cir1 also operates via the EDS1-PAD4 regulatory node.
Given that CIR1 is epistatic to EDS1, we investigated whether CIR1 might regulate EDS1 expression in Arabidopsis. Western blot analysis of the cir1 mutant revealed that EDS1 protein levels in uninfected plants were higher than in PR-1:LUC control plants (Fig. 4A). As cir1 displays constitutively elevated expression of SAdependent genes, and EDS1 gene expression is SA-inducible [11], elevated EDS1 protein levels in cir1 may simply be a consequence of upregulated EDS1 transcription in this mutant. However, qPCR analysis revealed no significant difference in steady state levels of the EDS1 transcript between the cir1 and PR-1:LUC plants ( Figure 4B), suggesting that CIR1 may instead function as a negative regulator of EDS1 via a post-transcriptional regulatory mechanism. One possible mechanism for CIR1 action might be via regulation of EDS1 protein accumulation or stability. Analysis of EDS1 protein levels in the pad4-1 and cir1 pad4 mutants ( Figure 4A) offers limited support to this hypothesis; while EDS1 protein levels were reduced in the pad4-1 single mutant, in line with previous reports that PAD4 stabilises EDS1 [14], the cir1 pad4 double mutant displayed similar EDS1 levels to cir1 plants ( Figure 4A). This suggests that the cir1 mutation may be able to compensate for the lack of PAD4 in the stabilisation of EDS1, and that CIR1 might serve as a negative regulator of EDS1 protein levels in planta. Recent studies have indicated that the balance between nuclear and cytoplasmic EDS1 pools is important in EDS1 function in innate immunity [17]. EDS1 protein levels are higher in both nuclei-enriched and nuclei-depleted fractions in the snc1 mutant [17], and interaction between SNC1 and EDS1 has been detected in both nucleus and cytoplasm [19]. However, the sub-cellular distribution of EDS1 has not been investigated in cir1, and so it is unclear whether EDS1 protein levels are elevated in both sub-cellular compartments, or only in one.
Environmental conditions modulate plant-pathogen interactions, with temperature known to play an important role in determining the strength of the host response to pathogen challenge [30,35]. Higher temperatures within the ambient growth temperature range of the plant have been shown to reduce the effectiveness of both PTI and ETI [35,36]. The gain-of-resistance mutants cpr1, bon1 and srfr1 all display temperature modulated growth and defence phenotypes, with dwarfism and increased resistance to biotrophic pathogens observed at 22uC but not 28uC. Similarly, we observed that growth and resistance to Pst DC3000 infection are modulated by temperature in cir1 ( Figure 5 & 6). At 18uC an obvious reduction in biomass production in cir1 was accompanied by enhanced resistance to Pst DC3000 relative to cir1 plants grown at 22uC. In contrast, at 25uC plant size and resistance to Pst DC3000 were not significantly different between cir1 and control plants. While these phenotypes are similar to those reported for cpr1, bon1 and srfr1, the temperature at which they occur is lower in cir1; while cpr1, bon1 and srfr1 all exhibit a dwarf phenotype at 22uC, cir1 displays only a modest reduction in size at this temperature.
The apparently identical temperature sensitivity of cpr1, bon1 and srfr1 may result from their convergence on the TIR-NB-LRR protein SNC1. Transcript and/or protein levels of SNC1 are elevated in cpr1, srfr1 and bon1 mutants, indicating that all three proteins act as negative regulators of SNC1 [31,37,38]. CPR1 is an F-box protein which interacts in vivo with SNC1 [37] suggesting that it regulates SNC1 protein levels via 26S proteasome-mediated degradation. Arabidopsis mutants expressing a constitutively active version of SNC1 (snc1-1) display dwarfism, constitutive defence gene expression and increased resistance to biotrophic pathogens [34]. As with cir1, cpr1, bon1 and srfr1, these phenotypes are temperature dependent, manifesting at 22uC but not at 28uC [31,36]. Analysis of the progeny from a cross between bon1 and snc1-11 (a null allele) has revealed that SNC1 is essential for the dwarfism, constitutive PR gene expression and enhanced resistance to biotrophic pathogens displayed by bon1 at 22uC [31]. Similarly, dwarfism and constitutive PR-1 gene expression are abolished in the srfr1 snc1-11 and cpr1 snc1-11 double mutants [22,38], suggesting that the auto-immune phenotypes of cpr1, bon1 and srfr1 at 22uC result largely from de-repression of SNC1.
SNC1 therefore appears to act as a temperature sensor in Arabidopsis to modulate host immunity in response to changes in the environment. Further evidence for this role comes from the snc1-3 mutant which displays dwarfism and increased resistance to Pst DC3000 at both 22uC and 28uC [36]. It has been suggested that a threshold concentration of SNC1 must be reached in the nucleus to trigger immunity, supported by data showing that in wild-type plants SNC1 nuclear content decreases with increasing temperature, but not in the snc1-3 mutant [36]. Whether the growth and constitutive defence phenotypes of cir1 also require SNC1 is currently unknown, although we did not observe a statistically significant increase in SNC1 transcript levels in cir1 versus PR-1:LUC plants at either 18 or 22uC (Figure 7). Analysis of the cir1 snc1-11 double mutants we are currently generating will address the role of SNC1 in cir1. We are also carrying out genetic mapping to identify the CIR1 gene. First-pass mapping experiments on the F2 progeny of a cross between cir1 and Ler plants indicated that cir1 mapped approximately 9.4 cM below nga111 on the lower arm of chromosome 4 [24]. Subsequent linkage analysis to markers within this region has localised CIR1 to a 46 kb region of chromosome IV, and we are currently analysing candidate genes within this region. Identification of the CIR1 gene will shed light on the exact biochemical roles played by this negative regulator of EDS1 and PAD4-mediated immunity in Arabidopsis.

Plant growth conditions
Arabidopsis seeds were stratified for 48 h at 4uC in the dark prior to sowing on either a 1:1 mix of peat (Jiffy Products, Norway) and vermiculite or on half-strength MS agar plates. Plants were grown under a long-day photoperiod (16 h light, 8 h dark) at 22uC (unless otherwise stated) and 55% relative humidity, and cool white fluorescent light of 80-100 mmol m 22 s 21 .

Pathogen assays
All pathogen assays were carried out on four-week old Arabidopsis plants. Pseudomonas syringae pv. tomato DC3000 infections were carried out as previously described [39]. Hyaloperonospora arabidopsidis Noco2 infections were carried out as described by Parker et al. [27], and the extent of plant cell necrosis and development of H. arabidopsidis mycelium was examined microscopically 7 dpi by lactophenol trypan blue staining [10].

Luciferase assays
Total protein was extracted by homogenizing leaf tissue from four-week-old soil grown plants in 1 mL extraction buffer (100 mM sodium phosphate buffer pH 7.2, 5 mM DTT). The samples were centrifuged for 5 min at 12 0006g to pellet cell debris, and 100 mL of the resulting supernatant added to 100 mL of assay buffer (60 mM Tris-HCl pH 8.0, 20 mM MgCl 2 , 20 mM DTT, 2 mM EDTA, 2 mM ATP). Luciferase activity was measured for 20 s following injection with 100 mL of luciferin buffer (60 mM Tris-HCl pH 8.0, 20 mM MgCl 2 , 20 mM DTT, 2 mM EDTA, 1 mM luciferin) in a Luminoskan TL-Plus luminometer (Labsystems, Finland). Luciferase activity was normalised to total protein concentration as determined by Bradford protein assay.
Generation of cir1 eds1 and cir1 pad4 double mutants Homozygous cir1 plants were crossed to either eds1-2 (Ler background) or pad4-1 (Col-0 background) mutants, and the resulting F1 progeny allowed to self-fertilise. While the marker used to screen for homozygosity for the cir1 mutation is elevated PR-1:LUC activity, both EDS1 and PAD4 are known to be required for PR-1 expression [13]. Unsurprisingly then, segregation analyses of the F2 progeny revealed a significant deviation from the expected 13:3 ratio of low:high PR-1:LUC activity suggesting that the eds1 and pad4 mutations were affecting PR-1::LUC reporter activity. A two-step process was therefore employed to isolate the double mutants. Homozygous cir1 plants were identified in the F2 generation by screening for high PR-1:LUC activity, comparable to that of the single cir1 mutant. As expected, genotyping revealed that all were heterozygous for either the eds1-2 or pad4-1 mutant allele. The F3 progeny from these plants were then screened by PCR genotyping or RFLP analysis to identify individuals homozygous for the eds1 or pad4 mutation. The PCR primers used for genotyping the EDS1 locus were 59-GTGGAAACCAAATTTGACATTAG-39 and 59-ACACAA-GGGTGATGCGAGACA-39 which generated PCR products of 750 bp (EDS1) or 600 bp (eds1-2). For PAD4 genotyping, the primers used were 59-GCGATGCATCAGAAGAG-39 and 59-TTAGCCCAAAAGCAAGTATC-39 which generated a 391 bp PCR product. The PAD4 amplicon is cleaved by BsmFI to give products of 281 and 110 bp, while the pad4-1 amplicon is not. As eds1-2 is in the Ler background, cir1 was also crossed with Ler to determine whether the mixed Col-0/Ler genetic background affected the penetrance of the cir1 mutation.

Quantitative PCR analysis
Total RNA was extracted using Trizol (Invitrogen), treated with DNase and cDNA synthesised from 1 mg of RNA using Superscript III reverse transcriptase (Invitrogen). Quantitative PCR was performed using a RotorGene RG3000A instrument (Corbett Research, Australia). Reactions consisted of 1 mL template cDNA, 5 mL Kapa SYBR FAST Universal 26qPCR Master Mix (Kapa Biosystems, South Africa), and 200-900 nM of each primer in a final volume of 10 mL. Amplification conditions included an initial step at 95uC for 3 min, followed by 40 cycles of 95uC for 3 s, primer annealing at 60 or 65uC for 20 s and elongation at 72uC for 1 s. Melt curve analysis confirmed that the individual amplified products corresponded to a single, genespecific cDNA fragment. The relative expression level of each gene of interest was calculated with the RotorGene 6000 series software v1.7 using the two standard curve method, with normalisation to the reference gene Actin-2 (At3g18780). Details of the primers used and specific qPCR reaction conditions can be found in Table  S1.

Western Blot analyses
Total protein was isolated from leaf tissue as described by Ingle et al. [40]. Forty mg of total protein was separated on 12% (w/v) SDS PAGE gels, transferred to nitrocellulose membrane and blocked for 2 h at RT in 16TBS-T buffer containing 2% (w/v) skim milk powder. Primary EDS1 antibody [13] was diluted 1:400 in 16TBS-T buffer with 2% (w/v) milk powder, and blots incubated o/n at 4uC. Incubation with primary antibody was followed by 3610 min washes in 16TBST, and incubation with secondary antibody (Rabbit IgG HRP, 1:5000 dilution) for 1 h at RT, prior to band detection by chemiluminescence.

Statistical analyses
Statistical analyses of all data were carried out using Statistica (version 9). Pst DC3000 titre data were log-transformed prior to ANOVA to ensure homogeneity of variance and normality of error. Fisher's LSD post-hoc analysis was used to identify significantly different mean values within an experiment. The raw data obtained from pathogen assays and qPCR experiments that were used in the statistical analyses (and in the generation of Figures 1, 2, 3, 4B, 6 and 7) are provided in Table S2.