I N T R O D U C T I O N Plants utilize a multi-layered defence system to control diseases. This includes preformed barriers (e.g. cell wall and the cuticle) and inducible defences. Activation of inducible defences occurs on recognition of pathogen-derived factors or factors produced in response to pathogen activity (Jones and Dangl, 2006) . For example, plants have evolved transmembrane pattern recognition receptors to perceive slowly evolving pathogen-associated molecular patterns (PAMPs), resulting in the downstream activation of defences that contribute to PAMP-triggered immunity (PTI). Some pathogens, however, have evolved effectors that are released into the host to suppress PTI (Gimenez-Ibanez et al., 2009; Gohre and Robatzek, 2008) . To counteract the pathogen, resistant plants have evolved intracellular receptors encoded by R (RESISTANCE) genes to recognize pathogen-derived effectors, leading to the activation of a mechanism termed ‘effectortriggered immunity’ (ETI). ETI has been suggested to be an accelerated and amplified PTI response (Jones and Dangl, 2006) .

Salicylic acid (SA) plays a key role in PTI- and ETI-mediated defences and has an important influence on plant defences against numerous pathogens (Tsuda et al., 2008, 2009) . SA levels increase in pathogen-inoculated plants, resulting in the downstream activation of defence genes, including the PR1 (PATHOGENESIS-RELATED1) gene, which has been used as an excellent molecular marker for the activation of SA signalling. Furthermore, the exogenous application of SA and its functional synthetic analogues 2,6-dichloroisonicotinic acid (INA) and benzo(1,2,3)thiadiazole-7-carbothioic acid S-methyl ester (BTH) enhances disease resistance in a variety of plants against pathogens that primarily exhibit a biotrophic life cycle (Glazebrook, 2005) . Resistance is also enhanced in plant genotypes that exhibit constitutively elevated levels of SA (Nandi et al., 2005; Rate et al., 1999; Shah et al., 1999, 2001; Vanacker et al., 2001) . In contrast, the prevention of SA accumulation by the facilitation of its breakdown to catechol in transgenic NahG plants expressing a bacterial salicylate hydroxylase results in plants that exhibit heightened susceptibility to numerous pathogens (Delaney et al., 1994; Gaffney et al., 1993) . Similarly, disease resistance is compromised in the ics1 (isochorismate synthase 1) mutant, which is deficient in SA biosynthesis (Wildermuth et al., 2001) . The NPR1 (NON-EXPRESSOR OF PATHOGENESISRELATED GENES1) gene is a master regulator for SA signalling in Arabidopsis. Mutant npr1 plants are more susceptible to a variety of pathogens. In contrast, plants constitutively overexpressing NPR1 exhibit heightened resistance to pathogens in a variety of plant species (Cao et al., 1998; Chern et al., 2001; Lin et al., 2004; Makandar et al., 2006, 2010; Parkhi et al., 2010) . Studies with several mutants that suppress the SA-insensitive phenotype of npr1 plants have demonstrated that SA also activates PR1 expression and disease resistance through an NPR1-independent mechanism, which is secondary to the NPR1-dependent SA pathway (Bowling et al., 1997; Clarke et al., 1998; Shah et al., 1999) .

Ethylene (ET) and jasmonic acid (JA) are two other important plant hormones involved in plant defence (Bari and Jones, 2009; Pieterse et al., 2009; Spoel and Dong, 2008) . Similar to SA, the application of ET and JA, and the constitutive activation of these signalling pathways in some plant genotypes, results in enhanced resistance primarily against pathogens that exhibit a necrotrophic life style. Co-operative and antagonistic interactions between SA and ET/JA signalling have been reported. These interactions presumably fine tune plant defences for optimal expression (Verhage et al., 2010) .

In Arabidopsis, several mutant and transgenic plants in which either one or more of these defence signalling pathways are constitutively expressed result in heightened resistance to pathogens. However, in most of these cases, these plants with hyperactive defences also exhibit growth defects, such as dwarfing, spontaneous cell death and lesion development (Bolton, 2009; Heidel et al., 2004) . These defects may result from hormonal imbalances and/or the allocation of resources away from growth and development to sustain the elevated level of these defence mechanisms in these plants. For the future manipulation of these signalling pathways in order to enhance plant disease resistance, it is important to identify the genetic components that de-link the growth and developmental defects from the constitutive expression of defences. In this article, we report the cdd1 (constitutive defence without growth defect1) mutant of Arabidopsis, identified as a suppressor of the npr1-5 mutant phenotype, in which SA accumulation and signalling are constitutively activated without any adverse impacts on growth and development. In addition, the cdd1 mutation also influences the cross-talk of SA and ET/JA signalling mediated by NPR1 protein. R E S U LT S

The cdd1 mutant was identified in a previously described genetic screen (Shah et al., 1999) for mutants that constitutively expressed the SA-inducible PR1 gene in the npr1-5 (alias sai1) mutant background. As shown in Fig. 1, in comparison with the leaves of the wild-type (WT) and the npr1-5 mutant plant, in which PR1 gene expression was undetectable, the PR1 gene was constitutively expressed at elevated levels in leaves of the cdd1 npr1-5 double mutant. SA application resulted in a further increase, albeit minor, in PR1 expression in cdd1 npr1-5 plants (Fig. 1). However, in comparison with the SA-treated WT plants, PR1 expression was lower in the SA-treated cdd1 npr1-5 plants, suggesting that cdd1 does not completely overcome the npr1-5 defect. In support of this hypothesis, we observed that cdd1mediated constitutive PR1 expression was much higher in the cdd1 single mutant plant, which contains the WT NPR1 allele, than in the cdd1 npr1-5 double mutant. The above results suggest that the cdd1 mutation activates both NPR1-dependent and NPR1-independent mechanisms leading to PR1 expression. cdd1 confers enhanced disease resistance To test whether the elevated expression of PR1 is accompanied by enhanced disease resistance, the cdd1 npr1-5 and cdd1 mutant plants were inoculated with a virulent strain of the bacterial pathogen Pseudomonas syringae pv. tomato DC3000 (referred to as Pst DC3000 hereafter) and the bacterial numbers were determined 3 days post-inoculation (dpi). The WT and npr1-5 plants provided the controls for this experiment. As shown in Fig. 2A, as expected, bacterial numbers were significantly higher in the npr1-5 mutant relative to the WT. The presence of the cdd1 allele resulted in a significant reduction in bacterial numbers in the cdd1 npr1-5 double mutant compared with the npr1-5 mutant plant. Similarly, the growth of Pst DC3000 carrying the avrRpt2 avirulence gene was also restricted in the cdd1 npr1-5 double mutant, relative to the npr1-5 plant, confirming that the cdd1 allele enhances resistance against virulent and avirulent pathogens in the absence of the WT NPR1 allele (Fig. 2B). Resistance against virulent and avirulent Pst DC3000 was also higher in the NPR1 allele-containing cdd1 single mutant plant than in the WT plant (Fig. 2A, B). The cdd1 Fig. 2 Growth of pathogens in cdd1 plants.

Four-week-old soil-grown plants were infiltrated with overnight-grown bacterial cultures of Pseudomonas syringae pv. tomato (Pst) DC3000 at 105 colony-forming units (CFU)/mL (A), Pst DC3000 carrying the avirulence gene avrRpt2 at 105 CFU/mL (B), Psm ES4326 at 105 CFU/mL (C,D) and P. syringae pv. phaseolicola at 107 CFU/mL (E), each resuspended in 10 mM MgCl2. (A, C, E) Samples were harvested at 3 days post-inoculation (dpi) and CFUs/disc were determined. Each bar represents the average and standard deviation obtained from four samples, each containing five leaf discs of 5 mm in diameter. Statistical differences among samples (P < 0.01) are labelled with different letters. (D) Photograph of Psm ES4326-infected leaves taken at 3 dpi. WT, wild-type. mutant also exhibited enhanced resistance against P. syringae pv. maculicola ES4326 (Psm ES4326; Fig. 2C). Disease symptom development was also lower in the cdd1 mutant relative to the WT plant (Fig. 2D).

To determine whether cdd1 also had an impact on resistance against nonhost pathogens, a relatively high dose [107 colonyforming units (CFU)/mL] of P. syringae pv. phaseolicola was infiltrated into the leaves of the cdd1 mutant and WT plants, and the bacterial numbers were monitored at 3 dpi. As shown in Fig. 2E, the bacterial numbers were significantly lower in the cdd1 mutant relative to the WT.

Mutant npr1 plants are highly susceptible to the biotrophic pathogen Hyaloperonospora arabidopsidis (formerly called Peronospora parasitica) (Delaney et al., 1995) . Previously, it has been shown that son1, which is a suppressor of nim1 (allelic to npr1), effectively restores resistance against H. arabidopsidis in the nim1 son1 double mutant (Kim and Delaney, 2002) . To test whether the cdd1 mutation behaves in a similar way in terms of suppression of susceptibility of npr1-5 against this oomycete pathogen, 2-week-old seedlings of WT, cdd1, npr1-5 and cdd1 npr1-5 were challenged with H. arabidopsidis. As shown in Fig. 3A, cdd1 mutant plants were significantly more resistant than WT plants (P < 0.01) and harboured three times fewer numbers of spores than WT. As expected, npr1-5 plants were significantly more susceptible than WT. However, surprisingly, the double mutant cdd1 npr1-5 exhibited only a modest reduction in pathogen growth and sporulation when compared with the npr1-5 mutant (Fig. 3A). The difference between the resistance levels against Pst DC3000 and H. arabidopsidis in the double mutant compared with the npr1 single mutant may be attributed to the difference in the growth stage at which the plants were infected. These results indicate that cdd1-mediated resistance against H. arabidopsidis requires functional NPR1.

In addition to biotrophic and hemibiotrophic pathogens, SA and NPR1 have been reported to affect resistance to necrotrophic pathogens (Ferrari et al., 2003) . Thus, we monitored disease development following inoculation with the necrotrophic fungus Botrytis cinerea in cdd1 and cdd1 npr1-5 double mutant plants, and compared it with that in WT and npr1-5 plants, respectively. The disease severity was comparable between the cdd1 mutant and WT plants (Fig. 3B).We noticed an increase in susceptibility to B. cinerea in the npr1-5 mutant in comparison with WT (Fig. 3B). Similar observations have been made previously by Zimmerli et al. (2001) . However, as shown in Fig. 3B, the presence of the cdd1 allele restored the WT level of resistance against B. cinerea in the cdd1 npr1-5 double mutant, indicating that the cdd1 allele also has an impact on defence against this necrotrophic fungus, which is evident in the absence of NPR1.

High-performance liquid chromatography (HPLC) analysis indicated that the basal content of SA and its glucoside (SAG) were four and 10 times higher, respectively, in the cdd1 npr1-5 mutant relative to the WT and npr1-5 mutant plants (Fig. 4). In addition, the expression of the ICS1 gene, which is required for SA biosynthesis (Wildermuth et al., 2001) , was also elevated in cdd1 and cdd1 npr1 mutants (Fig. S1, see Supporting Information). We therefore hypothesized that the cdd1-mediated constitutively high expression of PR1 and the enhanced disease resistance phenotypes result from this elevated accumulation of SA and the corresponding hyperactivation of SA signalling. To test this hypothesis, the cdd1 mutant was crossed with a NahG transgenic plant, which expresses an SA-degrading salicylate hydroxylase, and the segregating F2 population from this hemizygous plant was analysed. As NahG expression was dominant, 28 of 37 plants expressed the NahG gene (Fig. 5A) and none of the plants showed cdd1-mediated constitutive PR1 expression. Of the nine plants lacking NahG expression, two (Fig. 5A, plants 1 and 26) showed a constitutively high level of PR1 expression, typical of cdd1 plants, and two other plants (Fig. 5A, plants 11 and 18) showed only modest basal expression, not associated with the cdd1 mutation, as confirmed in the F3 population (data not shown). To further ascertain the negative correlation between the presence of the NahG gene and cdd1-mediated constitutive PR1 expression, several F2 plants were randomly selected and their F3 progeny were analysed to identify plants that were homozygous for the cdd1 allele, but hemizygous for NahG. The F3 population of cdd1 NahG/+ plants showed perfect negative correlation of PR1 and NahG expression (Fig. 5B), confirming that SA was required for cdd1-dependent PR1 expression. The cdd1-mediated enhanced disease resistance was also attenuated in the cdd1 NahG and cdd1 npr1 NahG plants (Fig. 5C). cdd1 and cdd1 npr1-5 double mutations de-link constitutive SA signalling and defects in growth and development Constitutive activation of SA signalling in several previously characterized mutants is accompanied by detrimental phenotypes, including dwarfing, spontaneous lesions and cell death (Bolton, 2009; Heidel et al., 2004) . In contrast, the cdd1 and cdd1 npr1-5 mutants did not exhibit any obvious morphological and growth abnormalities compared with the WT plant under the conditions utilized in this study (Figs 6A, B and S2, see Supporting Information). No significant difference in biomass was observed between the cdd1 and WT plants (Fig. 6C). The cdd1 and cdd1 npr1-5 mutants also did not exhibit any visible lesions or microscopic cell death phenotype, which have been reported in other constitutive SA signalling mutants (Bolton, 2009; Moeder and Yoshioka, 2008) , suggesting that the cdd1 and cdd1 npr1-5 mutants de-link the activation of SA signalling from growth defects. Although, very rarely, we noticed a slight reduction in the size of the cdd1 single mutant, the cdd1 npr1-5 double mutant cultivated in parallel invariably showed a normal morphology that was indistinguishable from WT or npr1-5 plants, despite exhibiting hyperactive SA signalling, as reported in the previous sections. cdd1 mutation suppresses pathogen-induced

Arabidopsis mutants, such as ssi1 (Nandi et al., 2003a; Shah et al., 1999) , dnd1, cpr6 (Jirage et al., 2001) and cpr5 (Clarke et al., 2001; Jirage et al., 2001) , are lesion mimics that accumulate high levels of SA and, at the same time, constitutively activate the ET/JA-inducible gene PDF1.2 (PLANT DEFENSIN1.2). In contrast, mutants such as ssi2 (Nandi et al., 2005) and cpr1 (Jirage et al., 2001) do not show high basal expression of PDF1.2. The triple response of etiolated seedlings to ET after 1-aminocyclopropane-1-carboxylic acid (ACC) treatment and root growth inhibition after methyl jasmonate (MeJA) treatment in cdd1 plants showed WT responses (Fig. S3, see Supporting Information). Moreover, we could not detect any basal expression of PDF1.2 through Northern blot analysis in the cdd1 and cdd1 npr1-5 double mutant plants (data not shown). However, through real-time polymerase chain reaction (PCR), we detected a slight up-regulation of expression of PDF1.2 and the coregulated transcription factor gene ORA59 (Pre et al., 2008) in cdd1 and cdd1 npr1-5 compared with WT and npr1 plants, respectively (Fig. 7 and S4, see Supporting Information). NPR1 has been reported to mediate cross-talk in SA-mediated suppression of ET/JA signalling (Spoel et al., 2003) . Moreover, as discussed in the previous section, although the cdd1 mutation had little effect on the resistance response against B. cinerea, it effectively suppressed the enhanced susceptibility of npr1-5 mutant plants (Fig. 3B). Thus, we tested the B. cinerea infectioninduced expression of PDF1.2 and ORA59 in the cdd1 and cdd1 npr1-5 double mutant plants. Consistent with disease symptom development and the role of NPR1 in the SA-mediated suppression of ET/JA signalling, we noticed elevated PDF1.2 and ORA59 expression in the npr1-5 mutant plant compared with WT (Fig. 7A, B). However, the presence of the cdd1 allele effectively suppressed the npr1-5-determined high-level expression of PDF1.2 and ORA59 in response to B. cinerea infection (Fig. 7A, B). It has been reported previously that ET can bypass the requirement of NPR1 for SA-mediated suppression of JA signalling (Leon-Reyes et al., 2009) . Thus, we estimated the level of ET production in B. cinerea-infected plants. ET production in B. cinerea-infected cdd1 npr1 plants was at an equivalent level to that in WT and cdd1 mutant plants and at a lower level than in npr1 mutants (Fig. S5, see Supporting Information). Thus, cdd1-mediated suppression of JA signalling in cdd1 npr1 double mutants is independent of ET accumulation in these plants.

Constitutive PR1 expression was not observed in F1 plants derived from a cross between the cdd1 mutant and the parental WT plant, suggesting that the cdd1 allele is recessive to the WT CDD1 allele. Indeed, in the F2 generation, 16 of a total of 72 plants exhibited the constitutive PR1 expression phenotype, which is very close (a2 = 0.296, d.f. = 1, N = 72, P = 0.59) to the expected 1 : 3 ratio for a single-locus recessive trait, confirming the recessive nature of the cdd1 allele.

In order to map the cdd1 locus, the mutant plant, which is in the accession Nössen (Nö-0), was crossed with a WT plant of accession Columbia (Col-0). As expected of a recessive cdd1 allele, the F1 hybrids did not constitutively express PR1 at an elevated level (data not shown). The constitutive PR1 expression phenotype segregated in the F2 progeny derived from these F1 hybrids. However, the overall extent of PR1 expression showed a larger variation in this F2 population than that observed when cdd1 was crossed to its parental accession Nö-0, suggesting that other loci from the accession Col-0 modify the cdd1determined PR1 expression. Only the segregants that constitutively expressed PR1 at high levels were used to map cdd1. In any given mapping population, only 16%–18% of the F2 progeny showed a high level of PR1 expression, as opposed to the 25% expected of a single recessive allele. With the help of 180 such high-PR1-expressing F2 plants, the cdd1 locus was mapped to the lower arm of chromosome 5, between the LFY and G2368 markers, 21.6 cM from LFY and 7.1 cM from G2368. This region does not contain any gene that, when mutated, is reported to have phenotypes similar to that of cdd1 (constitutive elevated SA signalling without any adverse impact on plant morphology).

D I S C U S S I O N

We have identified a novel Arabidopsis mutant cdd1, which is hyper-resistant to virulent, avirulent and nonhost pathogens. SA signalling, which was hyperactive in cdd1, was a major contributor to the cdd1-determined hyper-resistance. The NahG-encoded salicylate hydroxylase attenuated the cdd1-determined constitutive expression of PR1 and heightened disease resistance. As the cdd1-determined PR1 expression and resistance in the absence of NPR1 in the cdd1 npr1-5 double mutant also required SA, we suggest that both NPR1-dependent and NPR1-independent SA-stimulated mechanisms are hyperactive in the cdd1 mutant background. JA and ET signalling have been shown previously to contribute to SA- and NPR1-independent resistance in cpr6 and ssi1 plants, both of which also suppress npr1-determined enhanced disease susceptibility (Bowling et al., 1997; Clarke et al., 1998; Nandi et al., 2003a; Shah et al., 1999) . Similarly, JA and ET may contribute to the SA-independent resistance in cdd1containing plants. Indeed, basal expression of JA/ET-regulated PDF1.2 and ORA59 in cdd1 plants was modestly higher than in WT plants. Whether these arms of the JA- and ET-regulated defence mechanisms contribute to the cdd1-determined resistance remains to be resolved. Moreover, we cannot rule out the possibility that the residual amount of SA in the cdd1 NahG and cdd1 npr1-5 NahG plants contributes to the remnant resistance. cdd1 bypasses the NPR1 dependence of SA-mediated suppression of ET/JA signalling Signalling cross-talk between plant hormones, such as SA, ET and JA, fine tunes the plant defence response (Pieterse et al., 2009) . Both synergistic and antagonistic interactions between SA and ET/JA signalling pathways have been reported (LeonReyes et al., 2009; Nandi et al., 2003a) . One of the prominent established interactions is the SA-induced suppression of JA signalling mediated by NPR1 (Spoel et al., 2003) . However, the induction of ET signalling through either exogenous ET application or pathogen challenge can bypass the requirement of the NPR1 protein for SA-mediated suppression of JA signalling (Leon-Reyes et al., 2009) . The cdd1 npr1-5 double mutant produced WT levels of ET in both unchallenged and B. cinereachallenged plants (Fig. S4); however, the cdd1 mutation effectively bypassed the NPR1 dependence of SA-mediated suppression of the ET/JA responsive marker genes PDF1.2 and ORA59 (Fig. 7). These findings tempt us to postulate that the cdd1 mutation is involved in the SA-mediated, NPR1independent suppression of ET/JA signalling, as observed under conditions of high ET (Leon-Reyes et al., 2009) .

All reported Arabidopsis mutants in which the SA signalling mechanism is hyperactivated, including the reported suppressors of npr1 with hyperactive SA signalling, have growth and/or developmental defects. For example, the ssi1 and ssi2 mutants exhibit dwarfing and spontaneous cell death (Shah et al., 1999, 2001) . The cpr1, cpr5 and cpr6 mutants are dwarfs, with cpr5 also exhibiting cell death (Bowling et al., 1997; Clarke et al., 1998) . Although the son1 (suppressor of nim1-1) mutant, which was identified as a suppressor of the npr1 allele nim1-1, does not exhibit a growth defect, it does not constitutively express PR1 at elevated levels and does not contain constitutively elevated SA content (Kim and Delaney, 2002) . Furthermore, the son1dependent enhanced resistance phenotype was not attenuated by NahG, suggesting that the son1-determined enhanced resistance is caused by an SA-independent mechanism (Kim and Delaney, 2002) . Constitutive plant defence and its association with the development of a lesion mimic phenotype have also been demonstrated recently in the Arabidopsis Atsr1 mutant (Du et al., 2009) . The AtSR1 gene is a negative regulator of plant immunity. The mutation results in the faster activation of SA-dependent defences in response to pathogen infection when plants are cultivated at 25–27 °C. However, when cultivated at 19–21 °C, the Atsr1 mutant plants accumulate elevated levels of SA, constitutively express PR genes and exhibit a short stature, indicating that the constitutive activation of defences in Atsr1 is accompanied by a detrimental growth/developmental phenotype.

To the best of our knowledge, the cdd1 mutant reported here is the only mutant plant of Arabidopsis known to date that uncouples constitutive activation of SA signalling and growth defects. Like all previously reported suppressors of npr1, the cdd1 mutant accumulates high levels of SA, resulting in the constitutive activation of SA signalling and SA-dependent plant defences. Yet, unlike the other npr1 suppressor mutants, cdd1 does not display growth and developmental abnormalities (Figs 6 and S1). The recessive nature of the cdd1 mutant allele suggests that the CDD1 WT allele has a negative impact on SA accumulation. The expression of the SA biosynthesis gene ICS1 increases on infection with Erysiphe and P. syringae and contributes to pathogen-induced SA accumulation in Arabidopsis (Wildermuth et al., 2001) . Indeed, we noticed a two- to five-fold enhanced expression of the ICS1 gene in cdd1 and cdd1 npr1-5 mutant plants in comparison with the corresponding WT and npr1-5 mutant plants (Fig. S1), supporting the role of CDD1 in the regulation of SA biosynthesis. However, it remains to be determined whether SA accumulation in cdd1 is solely mediated through enhanced ICS1 gene expression.

E X P E R I M E N TA L P R O C E D U R E S

Arabidopsis plants and pathogens were cultivated as described previously (Nandi et al., 2003b) . In brief, Arabidopsis seeds were germinated on Murashige and Skoog (MS) plates with 1% sucrose, and 9-day-old germinated seedlings were transferred to soil. Plants were grown at 22 °C in a growth room with a 12-h light (80 mE/m/s) and 12-h dark cycle and 65% humidity. Pst DC3000 and Psm ES4326 infections were carried out exactly as described in Nandi et al. (2004) , except for the dose of pathogen, which is indicated in the legend to each figure. Botrytis cinerea infections were carried out as described previously (Nandi et al., 2005) . Hyaloperonospora arabidopsidis isolate Cala2 maintenance and infection were carried out as described previously (Van der Ent et al., 2008) . All pathogen infections were repeated at least three times with similar results.

The cdd1 mutant was identified in a screen for suppressors of npr1-5 (Shah et al., 1999) . Unlike the other npr1-5 suppressors (ssi1, ssi2, ssi4) identified in this screen, the cdd1 npr1-5 mutant did not exhibit any growth defects. To generate the cdd1 single mutant, cdd1 was crossed to the parental line 1/8E/5 (Cao et al., 1997) , which is in the accession Nö. Northern blot analysis was conducted on the F2 progeny to identify plants that constitutively expressed PR1. The presence of the WT NPR1 allele was confirmed by cleaved amplified polymorphic sequence (CAPS) marker, as described previously (Shah et al., 1999) . To generate the cdd1 NahG and cdd1 npr1 NahG plants, the cdd1 npr1 double mutant was crossed with a transgenic NahG plant in the accession Nö. The F2 progeny plants were screened by Northern blotting for expression of the NahG transgene and by PCR for the presence of the NPR1 allele (Shah et al., 1999) .

SA treatment of 4-week-old plants was performed as described previously (Shah, 1997) . Water-treated plants were used as controls. For the MeJA sensitivity assay, seeds were germinated on MS medium supplemented with 10 mM MeJA. After 3 days of cold treatment, plates were placed vertically in the growth room and the root lengths were measured at 10 days postgermination (Nandi et al., 2003b) . For ET treatment, MS plates were supplemented with the ET precursor 10 mM ACC. Surfacesterilized seeds were plated onto ACC-containing agar and the plates were wrapped in aluminium foil and kept in the cold for 3 days, followed by a 7-day incubation at 21 °C.

Leaf tissues were ground in liquid nitrogen and RNA was extracted by guanidinium thiocyanate–phenol–chloroform as described by Chomczynski and Sachhi (1987) . Radiolabelled probes were generated using random primer labelling as described previously (Shah et al., 1999) . cDNA synthesis was performed as described by Beckers et al. (2009). Diluted cDNA was used as template for real-time PCR with an ABI-PRISM 7500 (Applied Biosystem, Singapore) sequence detector system. Realtime PCR experiments were repeated at least twice with two replications for each RNA sample. The following primers were used for real-time PCR: PDF1.2F, TTTGCTGCTTTCGACGCAC; PDF1.2R, CGCAAACCCCTGACCATG; Act2-F, AGTGGTCGTA CAACCGGTATTGT; Act2-R, GATGGCATGAGGAAGAGAGAAAC; ICS1-F, CTAATCTCCGCCGTCTCTGAACT; ICS1-R, TTGGAACCTG TAACCGAACGA.

Leaves were harvested from 4-week-old soil-grown plants. Leaf samples (0.2 g) were extracted once with 3 mL of 90% methanol and once with 3 mL of 100% methanol. The extracts were combined and dried under nitrogen gas. Total SA and SAG determinations were carried out as described previously (Bowling et al., 1994)

Aerial rosettes of 4-week-old plants were incised from the root, weighed and carefully placed in a 35-mL glass vial containing 1 mL of distilled water. Plants were left in open vials for at least 30 min to allow the release of wound-induced ET. Subsequently, four to five leaves of each plant were inoculated with 5 mL of either B. cinerea at 7.5 ¥ 105 spores/mL resuspended in halfstrength potato dextrose agar (PDA) broth or half-strength PDA broth alone as mock treatment. Immediately after infection, the bottles were closed with a rubber cap and tightened with an aluminium closer. After a specific time interval of infection, 1 mL of air was taken with an airtight injection syringe and the ET content was determined by a gas chromatograph (GC955; Synspec, Groningen, the Netherlands).

Leaves of 4-week-old soil-grown plants were floated in trypan blue staining solution (Rate et al., 1999) and heated in a microwave oven for 1 min, with intermittent pausing, to avoid excessive boiling. After de-staining as described by Rate et al. (1999) , the samples were inspected under a light microscope and photographed.

A C K N O W L E D G E M E N T S This work was supported by a Department of Biotechnology, Government of India grant (BT/PR5777/agr/16/529/2005) and a Capacity Buildup grant from Jawaharlal Nehru University to AKN, a National Science Foundation grant (IOS-0827200) to JS, and Council of Scientific and Industrial Research (CSIR) fellowships to SS and SR.

R E F E R E N C E S Bari, R. and Jones, J.D. (2009) Role of plant hormones in plant defence responses. Plant Mol. Biol. 69, 473–488.

Beckers, G.J., Jaskiewicz, M., Liu, Y., Underwood, W.R., He, S.Y., Zhang, S. and Conrath, U. (2009) Mitogen-activated protein kinases 3 and 6 are required for full priming of stress responses in Arabidopsis thaliana. Plant Cell, 21, 944–953.

Bolton, M.D. (2009) Primary metabolism and plant defense—fuel for the fire.

Mol. Plant–Microbe Interact. 22, 487–497.

S U P P O R T I N G I N F O R M AT I O N Additional Supporting Information may be found in the online version of this article: Fig. S1 ICS1 expression in cdd1 and cdd1 npr1 plants. Total mRNA was isolated from 5-week-old soil-grown plants of the indicated genotype. The level of ICS1 expression was determined by quantitative real-time polymerase chain reaction. Each bar represents the average and standard deviation of two samples. Statistical differences among samples (P < 0.05) are labelled with different letters over the bars. WT, wild-type.

Fig. S2 Morphological phenotypes of cdd1, cdd1 npr1-5 and ssi2 mutants. (A, C) Four-week-old soil-grown plants. (B) Sevenday-old seedlings of the indicated genotypes. WT, wild-type. Fig. S3 Ethylene (ET) and methyl jasmonate (MeJA) responses in wild-type (WT) and cdd1 plants. (A) Seeds were germinated in Murashige and Skoog (MS) medium alone (-) or in the presence of 10 mM 1-aminocyclopropane-1-carboxylic acid (ACC) (+) in the dark, and photographs were taken 4 days after germination. (B) Seedlings were grown in MS medium (-) or in the presence of 10 mM MeJA (+). Photographs were taken 10 days after germination.

Fig. S4 Basal levels of PDF1.2 and ORA59 gene expression. Leaf samples from 5-week-old soil-grown plants were harvested and the basal expression of the indicated genes was determined by real-time polymerase chain reaction. Each bar represents the average and standard deviation of two samples. Statistical differences among samples (P < 0.05) are labelled with different letters over the bars.

Fig. S5 Ethylene (ET) production from wild-type (WT) (W), cdd1 (c), npr1-5 (n) and cdd1npr1-5 (cn) plants after Botrytis cinerea infection. Each bar represents the mean and standard deviation of ethylene produced from ten 4-week-old soil-grown plants. Four leaves from each plant were inoculated with 5 mL of B. cinerea spores having 7.5 ¥ 105 spores/mL suspended in halfstrength potato dextrose agar (PDA) or mock inoculated with M O L E C U L A R P L A N T P A T H O L O G Y ( 2 0 1 1 ) 12( 9 ) , 8 5 5 – 8 6 5 half-strength PDA, and ethylene production was measured by gas chromatography after the indicated time period. Statistical differences among B. cinerea-infected samples (P < 0.05) are labelled with different letters over the bars.

Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.