Low oleic acid-derived repression of jasmonic acid-inducible defense responses requires the WRKY50 and WRKY51 proteins.

Signaling induced upon a reduction in oleic acid (18:1) levels simultaneously up-regulates salicylic acid (SA)-mediated responses and inhibits jasmonic acid (JA)-inducible defenses, resulting in enhanced resistance to biotrophs but increased susceptibility to necrotrophs. SA and the signaling component Enhanced Disease Susceptibility1 function redundantly in this low-18:1-derived pathway to induce SA signaling but do not function in the repression of JA responses. We show that repression of JA-mediated signaling under low-18:1 conditions is mediated via the WRKY50 and WRKY51 proteins. Knockout mutations in WRKY50 and WRKY51 lowered SA levels but did not restore pathogenesis-related gene expression or pathogen resistance to basal levels in the low-18:1-containing Arabidopsis (Arabidopsis thaliana) mutant, suppressor of SA insensitivity2 (ssi2). In contrast, both JA-inducible PDF1.2 (defensin) expression and basal resistance to Botrytis cinerea were restored. Simultaneous mutations in both WRKY genes (ssi2 wrky50 wrky51) did not further enhance the JA or Botrytis-related responses. The ssi2 wrky50 and ssi2 wrky51 plants contained high levels of reactive oxygen species and exhibited enhanced cell death, the same as ssi2 plants. This suggested that high reactive oxygen species levels or increased cell death were not responsible for the enhanced susceptibility of ssi2 plants to B. cinerea. Exogenous SA inhibited JA-inducible PDF1.2 expression in the wild type but not in wrky50 or wrky51 mutant plants. These results show that the WRKY50 and WRKY51 proteins mediate both SA- and low-18:1-dependent repression of JA signaling.

Plants, like animals, have evolved to develop immunity against a wide variety of microbial pathogens, including basal immunity against virulent pathogens, resistance (R) protein-mediated immunity against species-specific pathogens, and systemic immunity against secondary pathogens. R-mediated signaling is well known to induce a very rapid and efficient immune response and is often associated with the development of a hypersensitive reaction (HR), a form of programmed cell death, at the site of pathogen entry (Dangl et al., 1996). The resulting necrotic lesions are one of the first visible manifestations of pathogeninduced defense responses and are thought to aid the confinement of the pathogen to the dead cells.
Downstream signaling induced in response to R gene activation is commonly mediated by one or more phytohormones. Of these, defense signaling mediated by salicylic acid (SA) and jasmonic acid (JA) have been widely studied. These two phytohormones frequently act antagonistically to mediate defense against specific types of pathogens (Kunkel and Brooks, 2002;Glazebrook, 2005;Koornneef and Pieterse, 2008;Spoel and Dong, 2008). For example, accumulation of SA antagonizes JA-mediated responses (Doherty et al., 1988;Peñ a-Cortés et al., 1993;Gupta et al., 2000;Spoel et al., 2003). Infection with virulent Pseudomonas syringae induces SA-derived signaling and enhances susceptibility to Alternaria brassicicola by inhibiting JA-mediated defense responses in Arabidopsis (Arabidopsis thaliana; Spoel et al., 2007). Conversely, JA-derived signaling antagonizes SA-mediated responses, such as the suppression of host SA-derived responses by the bacterial phytotoxin coronatin, a structural analog of JA (Zhao et al., 2003;Brooks et al., 2005;Cui et al., 2005). Characterization of mutants affected simultaneously in both pathways has led to the identification of several molecular components that mediate cross talk between SA-and JA-derived signaling pathways (Petersen et al., 2000;Spoel et al., 2003, Li et al., 2004. The Arabidopsis mutant suppressor of SA insensitiv-ity2 (ssi2) is one such mutant that is affected in both SA-and JA-derived signaling (Kachroo et al., 2001). SSI2 encodes a plastid-localized stearoyl-acyl carrier protein desaturase (SACPD) that desaturates stearic acid to oleic acid (18:1) in the plant chloroplast. The ssi2 mutation causes a truncation in the SSI2 protein, which results in the loss of 90% of SACPD activity as compared with the wild-type protein (Kachroo et al., 2001). The ssi2 mutant plants are stunted in size, exhibit HR-like cell death lesions on their leaves, accumulate high levels of SA, and overexpress pathogenesis-related (PR) genes. Consequently, these plants exhibit enhanced resistance to bacterial and oomycete pathogens (Kachroo et al., 2001(Kachroo et al., , 2003. In contrast to the up-regulation of the SA pathway, ssi2 mutant plants are defective in JA-mediated defense responses. Although ssi2 plants are not altered in the perception or biosynthesis of JA, these plants are unable to induce defensin (PDF1.2) expression in response to JA. Consequently, these plants are hypersusceptible to necrotrophic pathogens such as Botrytis cinerea (Kachroo et al., 2001(Kachroo et al., , 2003Nandi et al., 2005). Lowering the levels of SA via the expression of a bacterial SA hydroxylase does not restore JA-derived responses in ssi2 plants, indicating that high SA alone is not responsible for the noninduction of JA-responsive defenses in these plants (Kachroo et al., 2001). Characterization of ssi2 suppressors has shown that the altered defense-related phenotypes of ssi2 are the result of reduction in 18:1 levels (Kachroo et al., 2003Chandra-Shekara et al., 2007;Xia et al., 2009). Furthermore, the ability to induce altered defense responses upon reduction in 18:1 levels is conserved among diverse plants, including soybean (Glycine max) and rice (Oryza sativa; Kachroo et al., 2008;Jiang et al., 2009). A large majority of the ssi2 suppressors restore 18:1 levels in ssi2 plants, resulting in the normalization of both SA-and JA-mediated signaling (Kachroo et al., 2003Xia et al., 2009).
Signaling induced in response to SA and JA is often mediated by defense-related transcription factors, including those belonging to the WRKY family of proteins (Eulgem and Somssich, 2007). For example, the WRKY25 protein negatively regulates SA-responsive PR-1 expression and resistance to P. syringae (Zheng et al., 2007), whereas WRKY33 positively regulates JA-inducible PDF1.2 expression (Zheng et al., 2006). Overexpression of WRKY33 enhances resistance to necrotrophic fungi but increases susceptibility to P. syringae. Several WRKY proteins are also involved in SA-JA cross talk, such as WRKY62, which likely participates in the SA-derived suppression of JA responses (Mao et al., 2007). The WRKY70 protein suppresses the expression of JA-responsive genes. Furthermore, expression of the WRKY70 transcript is up-regulated by SA and down-regulated by JA, indicating that WRKY70 may be involved in integrating SA-and JA-derived signaling pathways (Li et al., 2004). Likewise, WRKY41 may also be involved in cross talk between the SA-and JA-derived pathways, since overexpression of WRKY41 simultaneously induces PR-5 expression and suppresses JA-responsive PDF1.2 expression (Higashi et al., 2008).
The Arabidopsis genome contains 74 WRKY genes, and several of these are induced in response to pathogen infection and/or exogenous application of SA (Dong et al., 2003). Here, we examined the involvement of WRKY proteins in mediating the altered SAand JA-derived responses in ssi2 plants. Genome-wide transcriptional profiling showed that several WRKY genes were induced in the low 18:1-containing ssi2 plants. Mutations in two of these (WRKY50 and WRKY51) restored JA-inducible PDF1.2 expression and basal resistance to B. cinerea in the ssi2 plants, suggesting that WRKY50 and WRKY51 might serve as positive regulators of SA-mediated signaling but negative regulators of JA-mediated signaling.

A Mutation in WRKY70 Does Not Alter ssi2-Related Phenotypes
The well-characterized role of WRKY70 in mediating SA-JA cross talk, together with the fact that ssi2 plants are impaired in JA-derived defense signaling, prompted us to investigate the role of WRKY70 in ssi2mediated signaling. Since the WRKY70 transcript is SA inducible (Li et al., 2004), we first compared the levels of WRKY70 expression between wild-type and ssi2-1 mutant plants. Northern-blot analysis showed that WRKY70 transcription was indeed higher in the ssi2 (high-SA) plants as compared with wild-type plants (Fig. 1A). Interestingly, compared with wild-type plants, WRKY70 was also expressed at higher than basal levels in ssi2 sid2-1 plants, although these levels were lower than in ssi2 plants. This indicated that induction of the WRKY70 transcript in ssi2 plants was only partially regulated by SA, because ssi2 sid2 plants contain basal levels of SA due to a mutation in the SID2-encoded isochorismate synthase (Wildermuth et al., 2001;Kachroo et al., 2005;Venugopal et al., 2009). To test the role of WRKY70 in ssi2-derived defense signaling, we isolated a knockout (KO) mutation in this gene. Lines carrying T-DNA insertion in WRKY70 were screened for homozygous insertion mutants (Supplemental Table S1). Reverse transcription (RT)-PCR analysis of cDNA from the wrky70 line did not detect any WRKY70 transcript, confirming the presence of a KO mutation in this gene (Supplemental Fig. S1). The wrky70 mutant plants, which were morphologically similar to wild-type plants (data not shown), were then crossed with ssi2 plants. The ssi2 wrky70 double mutant plants segregated in a Mendelian double recessive manner. Consistent with their segregation, the ssi2 wrky70 plants showed ssi2-like phenotypes (Fig. 1B); the ssi2 wrky70 plants were stunted in morphology, showed HR-like cell death on their leaves, and constitutively expressed high levels of the PR-1 gene (Fig. 1C). Consistent with their phenotypes, ssi2 wrky70 double mutant plants showed ssi2-like 18:1 levels (Table I).
To determine if the wrky70 mutation restored JAresponsive PDF1.2 expression in ssi2 plants, we applied exogenous JA to wild-type, ssi2, wrky70, and ssi2 wrky70 plants. As expected, PDF1.2 induction was detected in wild-type and wrky70 plants but not in the ssi2 or ssi2 wrky70 plants (Fig. 1D). Together, these results indicate that absence of WRKY70 does not restore the altered SA-/JA-derived defense signaling in ssi2 plants.
Reduction in 18:1 Levels Induces the Expression of Several WRKY Genes Next, we used two parallel approaches to identify other WRKY genes that might regulate the altered SAand/or JA-derived signaling in ssi2 plants. These included genome-wide transcriptional profiling of WRKY genes between wild-type and ssi2 plants and expression analysis of known WRKY transcription factors that participate in cross-talk between SA and   JA pathways. Genome-wide transcriptional profiling was carried out using Affymetrix ATH1 GeneChips arrays. Using this analysis, 17 WRKY genes (including WRKY70) were found to be induced in ssi2 plants, but only three of these were also induced in ssi2 sid2 plants (Supplemental Table S2). Strikingly, several WRKY genes, including WRKY50 and WRKY51, were not detected in our GeneChip arrays or in many of the publicly available arrays. These too were included in the targeted expression profiling. Using these approaches, we identified 19 WRKY genes that were induced in ssi2 plants (data not shown). Northern-blot analysis of the WRKY genes identified in the genomewide and targeted expression profiling showed that only WRKY46, WRKY50, WRKY51, WRKY53, and WRKY60 were induced in a SA-independent manner in ssi2 sid2 plants ( Fig. 2A). To confirm this further, we analyzed the expression of these WRKY genes in glycerol-treated wild-type and sid2-1 plants, since exogenous glycerol application mimics ssi2-like phenotypes in wild-type plants by lowering 18:1 levels . As expected, glycerol application lowered 18:1 levels to induce defense phenotypes in both wild-type and sid2 plants (Supplemental Fig.  S2A). The glycerol-treated sid2 plants did not accu-mulate any SA (Supplemental Fig. S2B) but showed induction of WRKY46, WRKY50, WRKY51, WRKY53, and WRKY60 genes ( Fig. 2A, right panel). These WRKY genes were thus considered candidate participants in the ssi2-mediated induction of defense responses.
Mutations in WRKY50 and WRKY51 Restore JA Responsiveness in ssi2 Plants To study the roles of WRKY46, WRKY50, WRKY51, WRKY53, and WRKY60 genes in ssi2-mediated signaling, we first isolated KO mutations in these genes. Lines carrying T-DNA insertions in the target genes were screened for homozygous insertional mutants (Supplemental Table S1). RT-PCR analysis of cDNA from the wrky46, wrky50, wrky51, wrky53, and wrky60 lines did not detect any transcript for the respective genes, confirming the presence of KO mutations in each gene (Fig. 2B). All wrky mutants were morphologically similar to wild-type plants (Supplemental   ). Interestingly, the ssi2 wrky46, ssi2 wrky50, ssi2 wrky51, and ssi2 wrky53 double mutant plants accumulated significantly lower levels of SA and SA glucoside (SAG) compared with the ssi2 single mutant (Fig. 2D). SA and SAG levels in ssi2 wrky60 plants were also lower than those in ssi2 plants, but they were over 4-fold higher than those in any of the other ssi2 wrky double mutants. Regardless of their SA levels, all ssi2 wrky plants expressed high levels of the PR-1 gene, likely because the SA/SAG levels in these were still higher (up to 5-fold) than those in wild-type or ssi2 sid2 plants (Fig. 2E). Consistent with their morphological and defense phenotypes, 18:1 levels of ssi2 wrky46, ssi2 wrky50, ssi2 wrky51, ssi2 wrky53, and ssi2 wrky60 double mutant plants were similar to that of ssi2 (Table I).
In addition to their altered SA-derived signaling, ssi2 plants are also impaired in their ability to induce a subset of JA-responsive defense gene expression; ssi2 plants induce Thionin (THI2.1) but not PDF1.2 expression in response to exogenous JA (Kachroo et al., 2001).
To determine if the wrky46, wrky50, wrky51, wrky53, and wrky60 mutations restored JA-responsive PDF1.2 expression in ssi2 plants, we applied exogenous JA to wild-type, ssi2, ssi2 wrky46, ssi2 wrky50, ssi2 wrky51, ssi2 wrky53, and ssi2 wrky60 plants. As expected, PDF1.2 induction was detected in wild-type but not ssi2 plants. Interestingly, ssi2 wrky50 and ssi2 wrky51 plants showed induction of PDF1.2, although these levels were lower than in wild-type plants (Fig. 3A). The ssi2 wrky46, ssi2 wrky53, and ssi2 wrky60 plants did not induce PDF1.2 expression in response to JA, similar to ssi2 plants. The wrky single mutants did not exhibit basal PDF1.2 expression and did induce PDF1.2 expression in response to exogenous JA, similar to wild-type plants (Fig. 3B). In addition to PDF1.2, we also tested the expression of two other JA-responsive genes, Vegetative Storage Protein1 (VSP1; Berger et al., 1995) and Lipoxygenase2 (LOX2; Berger et al., 1995). As seen for THI2.1 (data not shown), ssi2 plants and all ssi2 wrky double mutants showed the induction of both VSP1 and LOX2 transcripts in response to exogenous JA (Fig. 3A). Together, these results showed that ssi2 plants are defective only in the induction of JA-responsive PDF1.2 expression and that the wrky50 and wrky51 mutations are able to restore PDF1.2 expression in these plants without altering their 18:1 levels.
Our analysis of ssi2 suppressors has shown that restoration of wild-type-like phenotypes in ssi2 plants relies upon the restoration of 18:1 levels in these plants. The only exception is the ssi2 enhanced disease susceptibility1 (eds1-1) sid2-1 triple mutant plants, which show wild-type-like morphology in spite of containing ssi2-like levels of 18:1 (Supplemental Fig.  S3C). The ssi2 eds1 sid2 plants are also restored in all the SA-related phenotypes associated with the ssi2 mutation (Venugopal et al., 2009). To determine if JA responsiveness was associated with 18:1 levels, the SA pathway, and/or morphological phenotype, we next tested JA-responsive PDF1.2 expression in the ssi2 eds1 sid2 plants. Northern-blot analysis showed that unlike ssi2 wrky50 or ssi2 wrky51, the ssi2 eds1 sid2 plants were unable to induce PDF1.2 in response to JA (Fig. 3C). As shown in Figure 3C, ssi2, ssi2 eds1, ssi2, and ssi2 eds1 sid2 plants all induced VSP1 expression in response to JA. This indicates that basal levels of 18:1 are essential for JA-inducible expression of PDF1.2 and that . Response to JA of various wrky mutants. The genotype designations include ssi2 wrky46 (sw46), ssi2 wrky50 (sw50), ssi2 wrky51 (sw51), ssi2 wrky53 (sw53), ssi2 wrky60 (sw60), and wrky50 wrky51 (wrky50 51). A, Northern-blot analysis of PDF1.2 (defensin), VSP1, and LOX2 expression in response to treatment with water or JA. B, Northern-blot analysis of PDF1.2 expression in response to treatment with water or JA. C and D, Northern-blot analysis of PDF1.2 and VSP1 expression in response to treatment with water or JA (C) or in response to treatment with water+JA or SA+JA (D). Ethidium bromide staining of rRNA was used as a loading control.
this phenotype is independent of morphological size. Furthermore, these data suggest that WRKY50 and WRKY51 act as negative regulators of JA-responsive PDF1.2 expression downstream of 18:1 levels.
Since WRKY50 and WRKY51 negatively regulate JA-derived PDF1.2 expression in ssi2 plants, we tested if these proteins did the same in the wild-type background. SA is known to repress JA-inducible gene expression in Arabidopsis (Spoel et al., 2003). Therefore, we tested PDF1.2 and VSP1 expression in wildtype plants, wrky50 and wrky51 single mutants, or the wrky50 wrky51 double mutant treated either with water and JA (water+JA) or SA and JA (SA+JA). As expected, SA inhibited the JA-inducible expression of PDF1.2 and VSP1 in wild-type plants; the SA+JAtreated wild-type plants induced very low levels of these transcripts in comparison with plants treated with water+JA (Fig. 3D). In contrast, SA failed to inhibit the JA-inducible expression of PDF1.2 or VSP1 in the wrky50 and wrky51 single mutants as well as the wrky50 wrky51 double mutant plants. Notably, the levels of PDF1.2 expression in the SA+JA-treated wrky51 plants were much lower than those in the correspondingly treated wrky50 and wrky50 wrky51 plants (Fig. 3D). Together, these results indicate that WRKY50 and WRKY51 negatively regulate the repression of JA-inducible PDF1.2 expression under low-18:1 conditions in ssi2 plants as well as in the presence of SA in wild-type plants.
Mutations in WRKY50 or WRKY51 Restore Basal Resistance to B. cinerea in ssi2 Plants The ssi2 plants exhibit enhanced susceptibility to B. cinerea as compared with wild-type plants, possibly due to their inability to induce JA-responsive defense signaling (Kachroo et al., 2001(Kachroo et al., , 2003. Since the wrky50 and wrky51 mutations both restored JA-responsive PDF1.2 expression in ssi2 plants, we determined whether these mutations also restored basal resistance to B. cinerea in ssi2 plants. Wild-type and mutant plants were inoculated with B. cinerea, and the plants were monitored for PDF1.2 expression and disease progression. As expected, B. cinerea infection induced PDF1.2 expression in wild-type plants but not in ssi2 plants. Consistent with their JA-inducible PDF1.2 expression, ssi2 wrky50 and ssi2 wrky51 plants also induced PDF1.2 in response to B. cinerea (Fig. 4A). In contrast, and consistent with their inability to respond to JA, ssi2 wrky46, ssi2 wrky53, and ssi2 wrky60 plants did not induce PDF1.2 expression in response to B. cinerea (data not shown).
Analysis of disease progression up to 9 d postinoculation (dpi) showed that the ssi2 plants developed profuse necrosis and succumbed to Botrytis infection within 9 dpi (Fig. 4, B and C). In contrast, wild-type plants were more resistant, with nearly 35% of the plants surviving infection at 9 dpi (Fig. 4, B and C). Interestingly, and congruent to their JA/pathogeninducible expression of PDF1.2, ssi2 wrky50 and ssi2 wrky51 plants were as resistant to infection by B. cinerea as wild-type plants; approximately 30% to 35% of these plants survived infection at 9 dpi. In contrast, ssi2 wrky46, ssi2 wrky53, and ssi2 wrky60 plants were as susceptible to B. cinerea as ssi2 plants (Fig. 4, B and C). Although the wrky50 and wrky51 mutations restored resistance to basal levels in the ssi2 background, they did not further enhance resistance in the wild-type background. The single mutant plants (wrky46, wrky50, wrky51, wrky53, and wrky60) exhibited wild-type-like responses to B. cinerea (Supplemental Fig. S4, A and B). Together, these results show that mutations in WRKY50 and WRKY51 restore PDF1.2 expression as well as resistance to B. cinerea in ssi2 plants.
Since WRKY proteins are known to function redundantly (Eulgem and Somssich, 2007), we next tested if WRKY50 and WRKY51 contributed additively to increased PR-1 expression or the repression of JA responses in ssi2 plants. The triple mutant ssi2 wrky50 wrky51 plants were generated and analyzed for PR-1 expression and resistance to B. cinerea. Northern-blot analysis showed that the ssi2 wrky50 wrky51 triple mutant plants continued to express the PR-1 gene constitutively, similar to the ssi2 wrky50 and ssi2 wrky51 double mutant plants (Supplemental Fig. S5). The ssi2 wrky50 wrky51 plants also induced similar levels of PDF1.2 in response to B. cinerea infection and exogenous JA, like the double mutant plants ( Fig. 4A; data not shown). Furthermore, basal resistance to B. cinerea was not further improved in the ssi2 wrky50 wrky51 plants (Fig. 4, B and C). Together, these results showed that WRKY50 and WRKY51 do not function additively in repressing defense to B. cinerea under low-18:1 conditions. Mutations in WRKY50 or WRKY51 Do Not Alter Sensitivity to, or the Production of, Reactive Oxygen Species The ssi2 mutant accumulates high levels of reactive oxygen species (ROS), and increased ROS levels are known to be associated with enhanced susceptibility to necrotrophs, including B. cinerea (Govrin and Levine, 2000). Conversely, tolerance to ROS has been associated with increased resistance to necrotrophic pathogens (Glazebrook, 2005). Therefore, we tested whether the restored basal resistance to B. cinerea in the ssi2 wrky50 and ssi2 wrky51 plants was due to alterations in responses to, or the production of, ROS in these plants. We first tested the possible involvement of the various WRKY genes in sensitivity to ROS production. Changes in gene expression in response to exogenous hydrogen peroxide (H 2 O 2 ) application were analyzed. As reported previously (Miao et al., 2004), the WRKY53 transcript was induced in plants treated with H 2 O 2 . Northern-blot analysis did not detect increased expression of any of the other WRKY genes analyzed (Fig. 5A). Next, we analyzed the wrky single mutant plants for sensitivity to para-WRKY50/51-Derived Repression of JA Signaling quat (1,1#-dimethyl-4,4#-bypiridilium), an agent that promotes ROS formation by inhibiting electron transport during photosynthesis (Farrington et al., 1973;Hiyama et al., 1993). Various concentrations of paraquat (5-50 mM) were spot inoculated on wild-type and wrky mutant leaves, and the lesion size was measured 48 h later. None of the wrky single mutants showed significant differences in paraquat-derived lesion formation as compared with wild-type plants (Fig. 5B, data shown for 15 mM paraquat).
Finally, we evaluated the levels of H 2 O 2 (as a measure of their ROS levels) in wild-type, ssi2, and ssi2 wrky double mutant plants before and after inoculation with B. cinerea. The basal levels of H 2 O 2 in ssi2, ssi2 wrky50, and ssi2 wrky51 plants were approximately 2-fold higher than in wild-type plants (Fig. 5C). Inoculation of B. cinerea increased H 2 O 2 levels in wild-type plants by more than 2-fold. In contrast, the B. cinerearesponsive increase in H 2 O 2 levels was only more than 1.2-fold in the ssi2, ssi2 wrky50, and ssi2 wrky51 plants. Furthermore, no appreciable differences were observed between the basal or pathogen-induced H 2 O 2 levels of ssi2, ssi2 wrky50, and ssi2 wrky51 plants. Together, these results showed that restoration of resistance to B. cinerea in the ssi2 wrky50 and ssi2 wrky51 plants was not due to their altered sensitivity to, or endogenous levels of, ROS. Since the ssi2 mutation confers enhanced resistance to bacterial pathogens, we tested the response of the ssi2 wrky50 and ssi2 wrky51 double mutant plants to virulent and avirulent P. syringae. Wild-type plants accumulated approximately 20-fold more virulent bacteria than ssi2 plants (P , 0.005). The ssi2 wrky50 and ssi2 wrky51 double mutant plants accumulated more virulent bacteria than the ssi2 plants, approximately 3-and 5-fold more, respectively (Fig. 6A). However, these levels were still significantly (P , 0.001) lower than in wild-type plants. As in the case of virulent bacteria, wild-type plants accumulated approximately 10-fold more avirulent bacteria (avrRpt2) than ssi2 plants (Fig. 6B). Notably, the ssi2 wrky50 and ssi2 wrky51 double mutant plants accumulated fewer avrRpt2 bacteria than wild-type plants, and these levels were not significantly different from those in ssi2 plants (Fig. 6B). Together, these results suggest that mutations in WRKY50 and WRKY51 partially alter resistance to virulent but not avirulent bacteria in ssi2 plants.
The above results and the fact that WRKY proteins are known to mediate defense responses prompted us to examine the pathogen response of the wrky single mutant plants. The wrky46, wrky50, and wrky60 plants accumulated similar levels of virulent P. syringae as wild-type plants (Fig. 7A). In contrast, wrky51 and wrky53 plants consistently accumulated 3-fold more virulent bacteria than wild-type plants (P , 0.05). The increased susceptibility of the wrky51 and wrky53 plants was also evident when lower bacterial loads were used as inoculum: wrky51 and wrky53 plants were 3-and 5-fold more susceptible than wild-type plants when inoculated with 10 4 or 10 3 virulent bacteria, respectively (Supplemental Fig. S6). Inoculation with avirulent bacteria showed that bacterial proliferation in the wrky46, wrky50, wrky51, and wrky60 mutants was not significantly altered as compared with wild-type plants (Fig. 7B). In contrast, wrky53 plants were significantly more susceptible (P , 0.05), accumulating 3-fold increased avirulent bacteria than wild- Figure 4. Response to B. cinerea in wild-type (SSI2; ecotype Nossen), ssi2, ssi2 wrky46 (sw46), ssi2 wrky50 (sw50), ssi2 wrky51 (sw51), ssi2 wrky53 (sw53), ssi2 wrky60 (sw60), or ssi2 wrky50 wrky51 (sw50w51) plants. A, Northern-blot analysis of PDF1.2 expression in the indicated genotypes at 72 h postinoculation with B. cinerea. Ethidium bromide staining of rRNA was used as a loading control. B, Morphological phenotype of wild-type or mutant plants at 6 dpi with B. cinerea. C, Percentage survival of wild-type (SSI2), ssi2, sw50, sw51, sw60, and sw50w51 plants at 9 dpi with B. cinerea. Results are representative of five separate experiments. Statistical significance was determined using Student's t test. Asterisks denote data significantly different from SSI2 (P , 0.001).
type plants (Fig. 7B). These results show that WRKY51 is required for basal, while WRKY53 participates in both basal and R-mediated, defense to P. syringae.
To determine if enhanced susceptibility was due to an impaired SA pathway, we evaluated SA responsiveness of wrky51 and wrky53 plants; exogenous SA induced wild-type-like levels of the SA-responsive marker PR-1 (Fig. 7C). SA application also restored wild-type-like resistance to virulent bacteria (Fig. 7D). PR-1 expression and resistance to virulent bacteria were also examined in wrky51 and wrky53 plants after treatment with lower concentrations of the SA analog benzo(1,2,3)thiadiazole-7-carbothioic acid (BTH; Supplemental Fig. S7, A-C). Although the results obtained from three independent applications showed variable levels of PR-1 expression, none indicated a defect in SA responsiveness in the wrky51 and wrky53 plants.

DISCUSSION
The Arabidopsis ssi2 mutant plants are constitutively up-regulated in SA-derived signaling and concomitantly defective in their ability to induce JAmediated expression of PFD1.2 and pathogen defense. As a result, these plants exhibit enhanced resistance to biotrophic pathogens but show heightened susceptibility to necrotrophs. This is in agreement with the fact that SA-mediated signaling often contributes to defense against biotrophs, whereas defense to many Figure 5. Role of WRKY genes in sensitivity to and/or production of ROS. A, Northern-blot analysis showing basal expression of the various WRKY genes in wild-type (Col-0) and ssi2 plants or H 2 O 2 -responsive expression in wild-type plants. Induction of glutathione S-transferase1 (GST1) was used as a positive control for the efficacy of H 2 O 2 treatment. Ethidium bromide staining of rRNA was used as a loading control. B, Mean lesion size on wild-type (Col-0) leaves or the various wrky single mutant leaves spot inoculated with 15 mM paraquat. C, H 2 O 2 levels, basal (0 dpi; gray bars) or at 3 dpi with B. cinerea (3 dpi; black bars), in wild-type (Col-0), ssi2, ssi2 wrky50 (sw50), or ssi2 wrky51 (sw51) plants. Figure 6. Response of ssi2 wrky50 (sw50) and ssi2 wrky51 (sw51) double mutant plants to P. syringae. A and B, Bacterial counts in the wild type (Col-0) and the various mutants infiltrated with virulent (DC3000; A) or avirulent (avrRpt2; B) strains of P. syringae. Bacterial numbers were determined at 0 dpi (white bars) or 3 dpi (black bars) and are presented as log 10 values of colony-forming units (cfu) per cm 2 . Error bars indicate SD; n = 4. Statistical significance was determined using Student's t test. Asterisks denote data significantly different from all others (P , 0.005 for A, P , 0.01 for B); a denotes data significantly different from sw50 and sw51 (P , 0.05).
necrotrophic pathogens requires JA-derived signaling (Glazebrook, 2005). SA is also well known to antagonize JA signaling. Furthermore, SA signaling components such as NPR1 (for nonexpression of PR-1) and WRKY62 have been shown to participate in the SAderived suppression of JA-inducible responses (Spoel et al., 2003;Mao et al., 2007). However, lowering SA levels neither restores the morphological phenotypes nor relieves the inhibition of JA-derived responses in ssi2 plants. Furthermore, both ssi2 npr1 and ssi2 NPR1 plants remain defective in JA-mediated induction of PDF1.2 as well as resistance to B. cinerea (Kachroo et al., 2001). This suggests that the inhibition of JAmediated defenses in ssi2 plants might not be due to antagonism from their increased SA levels or heightened SA-derived signaling. Thus, the ssi2 mutant provides a unique avenue to identify molecular components that mediate cross talk between the SA and JA pathways but is not involved in the SA-mediated antagonism of the JA pathway.
Suppressor analysis has shown that repression of the JA pathway in the ssi2 plants can only be relieved when their 18:1 levels are increased to wild type like or higher. Mutations in ACT1, GLY1, or ACP4 all restore 18:1 levels as well as both SA-and JA-derived signaling in ssi2 plants (Kachroo et al., 2003Xia et al., 2009). On the other hand, although simultaneous mutations in EDS1 and SID2 restore the morphological and constitutive R gene expression phenotypes (Venugopal et al., 2009), the ssi2 eds1 sid2 plants are neither restored for 18:1 levels nor JA-inducible PDF1.2 expression. Thus, EDS1 and SA function redundantly and downstream of 18:1 levels to modulate resistance to biotrophic pathogens, but they do not participate in the low-18:1-regulated repression of JA signaling (Fig. 8). Like EDS1 and SA, WRKY50 and WRKY51 also function downstream of 18:1 levels; the ssi2 wrky50 and ssi2 wrky51 plants contain ssi2-like (low) levels of 18:1. However, unlike eds1 sid2, the wrky50 and wrky51 mutations restore the ability of ssi2 plants to induce JA-responsive PDF1.2 expression as well as basal resistance to the necrotrophic pathogen B. cinerea. Interestingly, although wrky50 and wrky51 mutations lower SA levels in the ssi2 plants, they do not abolish constitutive cell death, PR-1 expression, or enhanced resistance to P. syringae. This may be because the higher-than-basal levels of SA in the ssi2 wrky50 and ssi2 wrky51 plants are sufficient for enhancing PR-1 expression and bacterial resistance. JA and SA are well known to be mutually antagonistic. Thus, it is possible that the restoration of JA responsiveness results in the reduction of SA levels in the ssi2 wrky50 and ssi2 wrky51 plants. However, ssi2 wrky46 and ssi2 wrky53 plants, which are not restored in JAinducible signaling, also contain lower SA than ssi2 plants. Furthermore, while JA is known to antagonize SA-derived signaling, a direct role in SA biosynthesis is not known. Likewise, restoration of JA responses in the ssi2 wrky50 and ssi2 wrky51 plants is likely not associated with their reduced SA levels, because neither ssi2 sid2 nor ssi2 wrky46 and ssi2 wrky53 plants are restored in JA-inducible PDF1.2 expression.
The ssi2 plants exhibit constitutive cell death, and increased cell death has been associated with enhanced pathogenicity of necrotrophic pathogens such as B. cinerea (Govrin and Levine, 2000). Furthermore, ssi2 plants contain high levels of ROS, which are Figure 7. SA-responsive changes in gene expression and defense to P. syringae in the wrky46, wrky50, wrky51, wrky53, and wrky60 plants. A, B, and D, Response to DC3000 (A and D) or avrRpt2 (B) strains of P. syringae in the wild type (Col-0) and wrky mutants. Bacterial counts are presented as log 10 values of colony-forming units (cfu) per cm 2 at 0 dpi (white bars) and 3 dpi (black bars). Error bars indicate SD; n = 4. Statistical significance was determined using Student's t test. Asterisks denote data significantly different from Col-0 (P , 0.05 for A and B, P , 0.01 for D). Black and gray bars in D indicate water-or SA-treated plants at 3 dpi, respectively. C, Northern-blot analysis showing PR-1 expression in water-and SA-treated plants. Ethidium bromide staining of rRNA was used as a loading control.
known to induce cell death and facilitate the spread of necrotrophic pathogens. In fact, many necrotrophs, including Botrytis, are known to induce ROS production in the host to enhance pathogenicity (von Tiedemann, 1997;Govrin and Levine, 2000;Dickman et al., 2001;Govrin et al., 2006). Likewise, increased sensitivity to oxidative stress is also associated with susceptibility to B. cinerea (Tierens et al., 2002;Mengiste et al., 2003;Veronese et al., 2004). The ssi2 plants exhibit spontaneous cell death as well as accumulate increased levels of ROS (Kachroo et al., 2001). However, cell death and ROS likely do not account for the increased susceptibility of ssi2 plants to B. cinerea, since the ssi2 wrky50 and ssi2 wrky51 double mutants also show constitutive cell death and accumulate ssi2-like levels of ROS. This is also consistent with the fact that WRKY50 and WRKY51 transcripts are not inducible by H 2 O 2 , and the wrky50 and wrky51 single mutants show wild-type-like sensitivity to paraquat. Thus, the restoration of basal resistance to B. cinerea in the ssi2 wrky50 and ssi2 wrky51 plants may not be associated with sensitivity to/accumulation of ROS or the cell death phenotype.
The Arabidopsis genome encodes 74 WRKY genes, and many of the encoded proteins function directly/ indirectly in defense signaling against microbial pathogens (Pandey and Somssich, 2009). This is consistent with our findings that KO mutations in either WRKY51 or WRKY53 lower basal resistance to P. syringae. Furthermore, WRKY53 also participates in R-mediated resistance to avrRpt2-expressing P. syringae. In addition to pathogen or pathogen-derived elicitors, many WRKY genes are induced in response to high SA (Dong et al., 2003). Of the SA-inducible WRKY genes, WRKY46, WRKY50, WRKY51, WRKY53, WRKY60, and WRKY70 are also induced in response to a reduction in 18:1 levels. Furthermore, this low-18:1-inducible expression of WRKY46, WRKY50, WRKY51, WRKY53, WRKY60, and WRKY70 is independent of high SA, since these genes continue to be induced in ssi2 sid2 as well as glycerol-treated sid2 plants, which are unable to accumulate high SA. This raises the possibility that some or all of the WRKY46, WRKY50, WRKY51, WRKY53, WRKY60, and WRKY70 proteins might regulate defense gene expression and thereby the altered signaling in ssi2 plants. Indeed, many defense-related genes, including PR-1, contain W-boxes (WRKY-binding sites) in their promoter regions (Maleck et al., 2000;Yu et al., 2001). In fact, overexpression of WRKY70 was associated with increased expression of PR-1 as well as enhanced resistance to P. syringae (Li et al., 2004). Clearly, though, constitutive PR-1 expression in ssi2 plants is not the result of increased expression of WRKY70 (or WRKY46, WRKY50, WRKY51, WRKY53, and WRKY60), since all of the ssi2 wrky double mutants continue to overexpress PR-1.
WRKY proteins are characterized by the presence of one or two highly conserved domains carrying the WRKYGQK sequence and a zinc-binding motif at the N-terminal end . WRKY proteins bind specific DNA sequences termed W-box elements in the promoters of target genes, and the promoters of several defense-related genes including PR genes contain W-boxes (Eulgem, 2006). Mounting evidence shows that in addition to inducing gene expression, WRKY proteins can also serve as transcriptional repressors (Rushton et al., 2010). This raises the possibility that WRKY50 and/or WRKY51 directly repress PDF1.2 expression, although WRKY proteins have not been reported to regulate the expression of JA signaling components. Absence of these proteins then relieves this repression to restore JA-derived signaling and thereby resistance to B. cinerea in the ssi2 wrky50 and ssi2 wrky51 plants. Analysis of the 5# upstream sequences of PDF1.2 did not detect sequences corresponding to the minimal W-box domain (C/TTGACC/T; Rushton et al., 1996;Eulgem et al., 2000). However, the absence of W-boxes does not rule out the possibility for WRKY50/51 as regulators of PDF1.2 expression, since some WRKY proteins do bind non-W-box sequences as well (Rushton et al., 2010). Interestingly, a preliminary analysis did detect W-box sequences in several other JA-inducible/metabolizing genes (Supplemental Table S3), such as VSP2, Oxophytodienoic Acid Reductase3, and Allene Oxide Synthase. Functional analyses of these W-box-like sequences could reveal an as yet unidentified role for WRKY proteins in modulating the JA signaling pathway.

Plant Growth Conditions and Genetic Analysis
Arabidopsis (Arabidopsis thaliana) plants were grown in MTPS 144 Conviron walk-in-chambers at 22°C, 65% relative humidity, and 14-h photoperiod. For genetic crosses, flowers from the recipient genotype were emasculated and then pollinated with donor pollen. In most cases, single, double, or triple mutant plants were obtained from more than one combination of crosses and showed similar morphological, molecular, and biochemical phenotypes. F2 plants showing the wild-type genotype at the mutant loci were used as controls in all experiments. The wild-type and mutant alleles were identified by PCR, cleaved-amplified polymorphic sequence, or derived cleaved-amplified polymorphic sequence analysis. T-DNA insertion mutants developed by the Salk Institute Genomic Analysis Laboratory (Alonso et al., 2003) were obtained from the Arabidopsis Biological Resource Center. Homozygous T-DNA insertion lines were generated by selfing heterozygous lines and verified by sequencing PCR products obtained with primers specific for the T-DNA left border in combination with gene-specific primers. Mutants used here are ssi2-1 (Kachroo et al., 2001), eds1-1 (Parker et al., 1996), sid2-1 (Wildermuth et al., 2001), and ssi2-1 sid2-1 or ssi2-1 eds1-1 sid2-1 (Venugopal et al., 2009). The SALK identifiers for the T-DNA insertion lines and their mutant designations are as indicated in Supplemental Table S1. Designations for Salk_120706 (wrky60-1; Xu et al., 2006), Salk_025198 (wrky70-1; Knoth et al., 2007), and Salk_034157 (wrky53-1; Miao et al., 2004) were retained from those assigned in the respective previous publications.

Transcriptional Profiling
Total RNA was isolated from 4-week-old plants using TRIzol (Invitrogen) as outlined above. The experiment was carried out in triplicate, and a separate group of plants was used for each set. RNA was processed and hybridized to the Affymetrix Arabidopsis ATH1 genome array GeneChip following the manufacturer's instructions. All probe sets on the GeneChips were assigned hybridization signals above background using Affymetrix Expression Console Software (http://www.affymetrix.com/Auth/support/downloads/manuals/ expression_console_userguide.pdf) version 1.0. A one-way ANOVA test followed by post hoc two-sample t tests were used to analyze the data. The P values were calculated individually and in pair-wise combination for each probe set. The identities of the WRKY genes were obtained from The Arabidopsis Information Resource (www.arabidopsis.org), and sequence identifiers are as indicated in Supplemental Table S1.

Trypan Blue Staining of Cell Death
The leaves were vacuum infiltrated with trypan blue stain prepared in 10 mL of acidic phenol, 10 mL of glycerol, and 20 mL of sterile water with 10 mg of trypan blue. The samples were placed in a heated-water bath (90°C) for 2 min and incubated at room temperature for 2 to 12 h. The samples were destained using chloral hydrate (25 g per 10 mL of sterile water; Sigma), mounted on slides, and observed for cell death with a compound microscope. The samples were photographed using an AxioCam camera (Zeiss), and images were analyzed using Openlab 3.5.2 (Improvision) software.

H 2 O 2 Levels and Paraquat Treatment
For determination of H 2 O 2 levels, 50 mg of leaf tissue was homogenized in 1 mL of Tris-HCl (40 mM, pH 7.0). The samples were incubated for 1 h in the dark after addition of 20 mM 2#,7#-dichlorofluorescein and 20 mg mL 21 horseradish peroxidase, followed by measurement of absorption at 488 nm (excitation) and 523 nm (emission). The levels of H 2 O 2 were calculated as mmol mg 21 protein by extrapolating from a standard curve for H 2 O 2 . For paraquat treatments, paraquat was prepared in sterile water and leaves were spot inoculated with 10 mL of 5, 10, 15, 25, or 50 mM solutions. Lesion sizes were measured 48 h after paraquat application using Vernier calipers. Results presented are representative of two or three separate treatments.

Pathogen Infections
The bacterial strain Pseudomonas syringae DC3000 derivatives containing pVSP61 (empty vector) or avrRpt2 were grown overnight in King's B medium containing rifampicin and kanamycin (Sigma). The bacterial cells were harvested, washed, and suspended in 10 mM MgCl 2 . The cells were diluted to a final density of 10 5 colony-forming units mL 21 (A 600 ) and used for infiltration. The bacterial suspension was injected into the abaxial surface of the leaf using a needleless syringe. Three leaf discs from the inoculated leaves were collected at 0 and 3 dpi. The leaf discs were homogenized in 10 mM MgCl 2 , diluted 10 3 -or 10 4 -fold, and plated on King's B medium. P. syringaerelated experiments were repeated three or four times for every genotype analyzed. Botrytis cinerea was grown on V8 plates (20% V8 juice, 0.3% CaCO 3 , and 1.5% agar) at 24°C. Spores were collected from 2-week-old cultures and washed twice with sterile water. Washed spores were suspended in 10 mL of sterile water, and the suspension was filtered through Miracloth to remove mycelia. Four-week-old plants were spray inoculated with spore suspensions (2 3 10 5 spores mL 21 ) and maintained under high humidity until disease assessment. Lesion sizes were measured using Vernier calipers, and approximately 32 lesions were measured for every genotype assessed per experiment. Data presented are representative of three separate experiments. Survival rates were calculated using at least 20 to 30 plants per genotype in three separate inoculations.

Fatty Acid Profiling
Fatty acid analysis was carried out as described previously (Dahmer et al., 1989). For fatty acid profiling, one or few leaves of 4-week-old plants were placed in 2 mL of 3% H 2 SO 4 in methanol containing 0.001% butylated hydroxytoluene. After 30 min of incubation at 80°C, 1 mL of hexane with 0.001% butylated hydroxytoluene was added. The hexane phase was then transferred to vials for gas chromatography. One-microliter samples were analyzed by gas chromatography on a Varian FAME 0.25-mm 3 50-m column and quantified with flame ionization detection. The identities of the peaks were determined by comparing the retention times with known fatty acid standards. Mole values were calculated by dividing the peak area by the M r of the fatty acid.

SA/SAG Quantification and Chemical Treatment of Plants
SA and SAG quantifications were carried out from 300 mg of leaf tissue as described before (Chandra-Shekara et al., 2004). For SA treatments 3-or 4-week-old plants were sprayed or subirrigated with 500 mM SA or 100 mM BTH. For the experiments in Supplemental Figure S7, plants were sprayed or subirrigated with 10 to 75 mM BTH as indicated. For glycerol treatment, plants were sprayed with 50 mM solution prepared in sterile water. For JA treatment, plants were sprayed with 50 mM JA solution prepared in sterile water; plants were covered with a transparent plastic dome to maximize exposure to JA. For the SA/JA/glycerol treatment experiment, results presented are representative of three independent experiments. At least two different RNA samples were analyzed per treatment per experiment.

RNA Extraction, Northern-Blot, and RT-PCR Analysis
Total RNA was isolated from 4-week-old plants using the TRIzol reagent (Invitrogen) as per the manufacturer's instructions. RNA quality and concentration were determined by gel electrophoresis and determination of A 260 . Northern-blot analysis and synthesis of random-primed probes were carried out as described earlier (Kachroo et al., 2003). Full-length cDNAs of PDF1.2 (At5g44420), VSP1 (At5g24780), and LOX2 (At3g45140) were amplified with sequence-specific primers and used as probes in northern-blot analysis. RT and first-strand cDNA synthesis were carried out using SuperScript II (Invitrogen). Two to three independent RNA preparations were analyzed at least twice by RT-PCR (35 amplification cycles). The amplified products were quantified using ImageQuant TL image analysis software (General Electric). Full-length sequence-specific primers were used for RT-PCR analysis.

Supplemental Data
The following materials are available in the online version of this article.
Supplemental Figure S2. Glycerol induces cell death and lowers 18:1 levels in ssi2 plants.
Supplemental Figure S4. Response of wrky mutants to B. cinerea.
Supplemental Figure S6. Response of wrky51 and wrky53 plants to virulent P. syringae.
Supplemental Table S1. T-DNA insertional lines used for analysis of WRKY function.
Supplemental Table S2. Fold change in transcript levels of WRKY genes in wild-type and ssi2 plants.
Supplemental Table S3. JA-responsive/metabolizing genes containing putative W-box elements in their 5# upstream regions.