Trienoic fatty acids (TAs) are the major polyunsaturated fatty acid species in the membrane lipids in plant cells. TAs are crucial for the adaptation to abiotic stresses, especially low- or high-temperature stress. We show that TAs in chloroplast membrane lipids are involved in defense responses against avirulent bacterial pathogens. Avirulent pathogen invasion of plants induces a transient production of reactive oxygen intermediates (ROI), programmed cell death and subsequent disease resistance. The Arabidopsis fad7fad8 mutation, which prevents the synthesis of TAs in chloroplast lipids, caused the reduction in ROI accumulation in leaves inoculated with Pseudomonas syringae pv. tomato DC3000 (avrRpm1). Linolenic acid, the most abundant TA, activated the NADPH oxidase that is responsible for ROI generation. TAs were transferred from chloroplast lipids to extrachloroplast lipids coincident with ROI accumulation after inoculation with Pst DC3000 (avrRpm1). Furthermore, the fad7fad8 mutant exhibited reduced cell death and was compromised in its resistance to several avirulent P. syringae strains. These results suggest that TAs derived from chloroplast lipids play an important role in the regulation of plant defense responses.
Trienoic fatty acids (TAs), that is 16- and 18-carbon fatty acids having three cis double bonds, are the most abundant fatty acids in membrane lipids of higher plants. In particular, chloroplast membrane lipids commonly contain about 70% TAs. The formation of TAs is catalyzed by ω-3 fatty acid desaturases (FADs). In Arabidopsis, three isozymes, FAD3, FAD7 and FAD8, have been identified (Arondel et al., 1992; Gibson et al., 1994; Iba et al., 1993). FAD3 is localized in microsomes, whereas FAD7 and FAD8 are localized in the chloroplast membranes (Dyer and Mullen, 2001; Froehlich et al., 2003). Chloroplast TAs are thought to function in responses to various environmental stimuli (Iba, 2002). Alteration of TA contents by genetic or transgenic methods has shown that TAs present in chloroplast membrane lipids are crucial constituents for the adaptation to temperature stress (Murakami et al., 2000). The expression of chloroplast ω-3 FAD genes is induced by pathogen infection, wounding, and salt stress (Berberich et al., 1998; Kirsch et al., 1997; Nishiuchi et al., 1997). Therefore, chloroplast TAs could be an active factor in the response to these stresses; however, at present the physiological roles that chloroplast TAs play are unknown.
Arachidonic acid (AA) is a major polyunsaturated fatty acid in human. AA plays a remarkable role in the activation of NADPH oxidase. NADPH oxidase expressed in phagocytic cells consists of two plasma membrane proteins, gp91phox and p22phox, two cytosolic regulatory proteins, p47phox and p67phox, and the small GTPase Rac (Babior et al., 2002). Translocation of cytosolic proteins to the plasma membrane is essential for the formation of the active complex during phagocytosis. Phagocytic cells that are deficient in cytosolic phospholipase A2 fail to activate NADPH oxidase, whereas the addition of AA to the cells restores the activity, demonstrating that liberated AA is required for activating NADPH oxidase (Dana et al., 1998). AA promotes the interaction of p47phox with p22phox by inducing conformational changes in the former (Shiose and Sumimoto, 2000). An additional effect of AA on the activation of phagocyte NADPH oxidase has been suggested to increase the affinity of gp91phox for its substrates (Doussiere et al., 1999; Rubinek and Levy, 1993).
Plants respond to pathogen infection via the Avr/R recognition system (Doke, 1983; Levine et al., 1994). The recognition of pathogen-encoded Avr proteins by the products of plant resistance (R) genes activates a signal transduction pathway leading to the hypersensitive response (HR) and induces systemic acquired resistance (SAR), a mechanism by which plant tissues distant from the infection site become more resistant to subsequent pathogen infection. HRs are characterized by the rapid production of reactive oxygen intermediates (ROI) referred to as the oxidative burst, which prevents further spread of the pathogen, and by programmed cell death. NADPH oxidase, which catalyzes the generation of O, is responsible for the oxidative burst (Doke, 1983; Torres et al., 2002). In several plants, rboh genes, which are homologous to gp91phox encoding the catalytic unit of human NADPH oxidase, have been identified (Keller et al., 1998; Torres et al., 2002; Yoshioka et al., 2003). However, so far no homologs of p47phox and p67phox have been found in plants. It is not yet clear which molecules participate in the activation of NADPH oxidase in plants.
Several recent studies demonstrated that fatty acids in chloroplast are involved in programmed cell death and SAR (Kachroo et al., 2001; Nandi et al., 2004). The suppressor of SA insensitivity (ssi2) mutation, which is defective in stearoyl-ACP desaturase FAB2, confers spontaneous cell death and enhanced disease resistance. The suppressors of fatty acid (stearoyl) desaturase deficiency (sfd) mutants suppress the ssi2-conferred cell death and resistance to virulent Pseudomonas syringae pv. maculicola (Psm). All of the sfd ssi2 mutants contain reduced levels of hexadecatrienoic acid (16:3), a TA specifically acylated in chloroplast galactolipids. Therefore, it has been suggested that 16:3 promotes the ssi2-conferred phenotypes. However, it is unclear whether avirulent pathogen-induced responses correspond to ssi2/fab2-conferred phenotypes and whether 16:3 plays a role in inducing the oxidative burst.
We found that TAs in chloroplast membrane lipids are involved in several defense responses against avirulent pathogens. The accumulation of O and H2O2 was reduced in the fad7fad8 double mutant treated with ozone or inoculated with avirulent P. syringae pv. tomato (Pst), respectively. The fad7fad8 mutant displayed a lesser extent of cell death than the wild type after inoculation. In addition, linolenic acid (LA; 18:3), the most abundant TA, was the most effective fatty acid in activating NADPH oxidase. LA was transferred from chloroplast to extrachloroplast membrane lipids concomitant with the accumulation of H2O2. Furthermore, the fad7fad8 mutant exhibited enhanced susceptibility to avirulent Pst strains. These results suggest that TAs in chloroplast lipids play important physiological roles in modulating defense responses.
Reduction in O accumulation in chloroplast ω-3 desaturase-deficient plants
The Arabidopsis fad7fad8 double mutant (SH1) is deficient in the chloroplast-localized ω-3 desaturases, and thus has reduced levels of TA, especially in the chloroplast membrane lipids (McConn et al., 1994). In contrast, the Arabidopsis fad3 mutant (LK70) lacking the endoplasmic reticulum-localized ω-3 desaturase contains reduced amounts of TAs, mainly in extrachloroplast membrane lipids (Browse et al., 1993). We studied the O accumulation in leaves of these plants using an in vivo ROI-generating method (Figure 1). Ozone induces the oxidative burst and mimics a series of plant–pathogen responses, for example the production of O (Rao et al., 2000a). The generation of O was indicated by the precipitation of formazan, which results from the reduction of nitroblue tetrazolium (NBT). Tests were conducted in young leaves (0.8–1 cm long) because older leaves tended to show non-specific precipitates. After ozone treatment, distinctive formazan spots appeared on the leaves of wild-type Col-0. There were no such pronounced spots in the fad7fad8 mutant (Figure 1a) and the tobacco T23 line (Murakami et al., 2000) (data not shown), which has reduced levels of TA in chloroplast lipids. The fad3 mutant, however, reacted similarly to the wild type (Figure 1a). Quantification of the spots demonstrated that the fad7fad8 mutant produced smaller amounts of formazan precipitates than wild-type plants and the fad3 mutant (Figure 1b). Thus, TAs of chloroplast membrane lipids seemed to be required for ozone-induced O accumulation. In fact, complementation of the fad7fad8 mutant by introducing the chloroplast-localized FAD7 ω-3 desaturase gene restored the ozone-inducible accumulation of O (data not shown). Notably, the development of formazan spots in wild-type plants was inhibited by diphenylene iodonium, an inhibitor of NADPH oxidase, suggesting a role for the enzyme in O production (Figure 1a).
Reduction in H2O2 accumulation and attenuated cell death in the fad7fad8 mutant inoculated with Pst DC3000 (avrRpm1)
Plant resistance R proteins that recognize the corresponding pathogen Avr proteins give rise to plant defense responses. Inoculation with avirulent pathogen results in the development of HRs characterized by the oxidative burst and programmed cell death. We examined the accumulation of ROI and cell death in the fad7fad8 mutant inoculated with Pst DC3000 carrying the avrRpm1 gene, which is recognized by the functional RPM1 protein in the Arabidopsis wild-type Col-0 (Boyes et al., 1998) (Figure 2). O produced by NADPH oxidase is dismutated rapidly to H2O2 in the avirulent pathogen-induced oxidative burst (Torres et al., 2002). Therefore, the accumulation of H2O2 was detected in situ by precipitation of polymerized 3,3-diaminobenzidine (DAB). There was no significant DAB precipitation in the fad7fad8 mutant (Col-0 background) 10 h after inoculation with a concentration (106 cfu ml−1) of Pst DC3000 (avrRpm1), although the fad3 mutant (Col-0 background) exhibited a similar accumulation of DAB precipitates as the wild type (Figure 2). Trypan blue staining was performed 12 h after inoculation to visualize dead cells in the mutants under the same conditions as in the DAB experiment. Trypan blue stains were diminished in the fad7fad8 mutant when compared with the wild-type and the fad3 mutant (Figure 2). However, inoculation with a higher concentration (2.5 × 107 cfu ml−1) of Pst DC3000 (avrRpm1) resulted in no obvious differences in DAB precipitates and trypan blue stains between the wild-type and the fad7fad8 mutant (data not shown). These results suggest that TAs in chloroplast membrane lipids play a role in the accumulation of H2O2 and the control of cell death, although the abilities to induce these responses is not completely abolished in the fad7fad8 mutant.
LA activates NADPH oxidase in vitro
The Avr/R recognition-mediated oxidative burst is caused by the activation of NADPH oxidase (Doke, 1983; Torres et al., 2002). It is possible that the reduced accumulation of O or H2O2 in the fad7fad8 mutant treated with ozone or inoculated with Pst DC3000 (avrRpm1), respectively, is the result of impairment of this activation. To examine the effects of different fatty acid species on NADPH oxidase activity, an in vitro enzymatic assay was performed using plasma membrane prepared from cultures of tobacco bright yellow 2 cells as described by Hamada et al. (1998) (Figure 3). 16:3 (Larodan Fine Chemicals AB, Malmö, Sweden) had minor effects on NADPH oxidase activity (Figure 3a). This is consistent with the ability of plant species lacking this fatty acid species (e.g. rice) to induce the oxidative burst. On the contrary, LA (18:3) significantly stimulated NADPH oxidase activity (Figure 3a). We obtained similar results when cytochrome c was used as an indicator in an in vitro NADPH oxidase enzymatic assay (data not shown). We tested various polyunsaturated fatty acids, but none was as effective as LA (Figure 3b). AA (20:4), which activates NADPH oxidase in animal cells (Henderson and Chappell, 1996), induces the oxidative burst in potato (Yoshioka et al., 2001), but had little effect in our system. The cytosolic fraction did not affect the NADPH oxidase activity (data not shown), indicating that cytosolic components are not required for the activation of the enzyme. Jasmonic acid (JA) is probably synthesized through an octadecanoid pathway starting from LA in the chloroplast membrane lipids. Therefore, the reduction in ROI accumulation in the fad7fad8 mutant might be caused by reduced production of JA. However, JA had a minimal effect on NADPH oxidase activity (Figure 3b).
Transfer of LA from chloroplast to extrachloroplast membranes coincident with H2O2 accumulation
In plant cells, NADPH oxidase is localized in the plasma membrane (Keller et al., 1998; Sagi and Fluhr, 2001). To elucidate the influence of LA present in chloroplast membrane lipids on the activation of enzyme localized in different membrane systems, the effects of ozone or inoculation with Pst DC3000 (avrRpm1) on the leaf fatty acid composition were investigated in the Arabidopsis wild-type Col-0 (Tables 1 and 2). On the basis of the total fatty acids of total lipids, the LA level of monogalactosyldiacylglycerol (MGDG) was reduced by both treatments (P < 0.01, t-test) (Table 1). In contrast, the relative LA levels of phosphatidylethanolamine (PE) and phosphatidylinositol (PI) increased substantially (P < 0.02). Notably, 16:3 was found in PE and PI of extrachloroplast membrane lipids in leaves treated with ozone as well as in leaves inoculated with Pst DC3000 (avrRpm1) (P < 0.01) (Table 2). However, 16:3 was acylated only in chloroplast galactolipids of untreated leaves (Table 2), suggesting that all the acylated 16:3 observed was derived from chloroplast galactolipids. Phospholipids in extrachloroplast membranes contain high amounts of palmitic acid (16:0). The possibility that a conversion of 16:0 to 16:3 occurred in extrachloroplast membranes after the treatments could be excluded because the levels of 16:2 in these membranes increased in the fad7fad8 mutant defective in the desaturation of 16:2 in chloroplast membranes (data not shown). Thus, LA and 16:3 probably were transferred from chloroplast membrane MGDG to PE and PI in extrachloroplast membranes including the plasma membrane. In fact, previous findings suggested a transfer of unsaturated fatty acids from the chloroplast to extrachloroplast membranes (Browse et al., 1986, 1988). Moreover, the intermediates of JA biosynthesis may be transferred from the chloroplast to the peroxisome (Weber, 2002).
Table 1. 18:3 levels in classes of lipids in Arabidopsis wild-type Col-0 leaves treated with ozone or inoculated with Pst DC3000 (avrRpm1)
18:3 content in lipids per total fatty acids (mol%)
Lipids were extracted from leaves 6 h after treatment with 250 p.p.b. ozone or leaves 6 h after inoculation with Pst DC3000 (avrRpm1) at 5 × 106 cfu ml−1. Values are mean of three independent experiments. 18:3 content in leaves treated with ozone or inoculated with Pst DC3000 (avrRpm1) is significantly different from that of control leaves or mock-inoculated leaves, respectively (*P < 0.01, **P < 0.02).
25.6 ± 2.5
11.6 ± 1.1
2.7 ± 0.1
1.7 ± 0.1
0.9 ± 0.0
4.9 ± 0.0
13.5 ± 1.6*
9.5 ± 1.0
2.0 ± 0.4
3.5 ± 0.3**
1.6 ± 0.1**
3.4 ± 0.5
26.7 ± 0.6
14.7 ± 0.5
2.4 ± 0.1
2.0 ± 0.1
0.9 ± 0.1
5.2 ± 0.1
Pst DC3000 (avrRpm1)
22.7 ± 0.9*
14.0 ± 0.6
2.3 ± 0.1
3.0 ± 0.2**
1.4 ± 0.0**
5.2 ± 0.2
Table 2. 16:3 levels in classes lipids in Arabidopsis wild-type Col-0 leaves treated with ozone or inoculated with Pst DC3000 (avrRpm1)
16:3 content in lipids per total fatty acids (mol%)
Lipids were extracted from leaves 6 h after treatment with 250 p.p.b. ozone or leaves 6 h after inoculation with Pst DC3000 (avrRpm1) at 5 × 106 cfu ml−1. Values are mean of three independent experiments. Dashes (−) indicate trace amounts (<0.05%). 16:3 content in leaves treated with ozone or inoculated with Pst DC3000 (avrRpm1) is significantly different from that of control leaves or mock-inoculated leaves, respectively (*P < 0.04, **P < 0.01).
11.7 ± 0.5
0.3 ± 0.0
5.6 ± 0.6*
0.4 ± 0.0
0.2 ± 0.0**
0.3 ± 0.0**
0.3 ± 0.0**
11.4 ± 0.4
0.3 ± 0.0
Pst DC3000 (avrRpm1)
9.8 ± 0.5*
0.4 ± 0.0
0.1 ± 0.0**
0.1 ± 0.0**
0.2 ± 0.0**
To further elucidate whether the transfer of TAs from chloroplast to extrachloroplast membranes is involved in the Avr/R-mediated responses, we investigated the changes in the TA levels of MGDG and phospholipids in leaves of Arabidopsis Col gl1-1 mutant (control plant) after inoculation with Pst DC3000 (avrRpm1) or with virulent Pst DC3000 (empty vector) and in leaves of the rpm1-1 mutant (gl1-1 background) after inoculation with Pst DC3000 (avrRpm1). The rpm1-1 mutant, a loss-of-function mutant in which a frameshift results in the loss of the RPM1 function, compromises avrRpm1/RPM1-mediated responses (Bisgrove et al., 1994; Grant et al., 1995). The levels of both LA and 16:3 in MGDG were reduced in gl1-1 leaves inoculated with Pst DC3000 (avrRpm1) compared with mock-inoculated leaves (at 2 h, P < 0.02) (Figure 4a). There was no fluctuation in the LA levels of other chloroplast lipids, including digalactosyldiacylglycerol (DGDG), sulfoquinovosyldiacylglycerol and phosphatidylglycerol, and in the 16:3 level of DGDG (data not shown). In addition, the levels of other fatty acids in these lipids showed no obvious changes (data not shown). Unexpectedly, however, both TA levels of MGDG were also reduced in gl1-1 leaves inoculated with Pst DC3000 and in rpm1-1 leaves inoculated with Pst DC3000 (avrRpm1) as well as in gl1-1 leaves inoculated with Pst DC3000 (avrRpm1) (Figure 4a). On the contrary, a significant increase in both TA levels of phospholipids occurred in gl1-1 leaves inoculated with Pst DC3000 (avrRpm1), unlike in mock-inoculated leaves, in gl1-1 leaves inoculated with Pst DC3000 and in rpm1-1 leaves inoculated with Pst DC3000 (avrRpm1) (at 2 h, P < 0.02 for LA; P < 0.01 for 16:3) (Figure 4b).
Next, we examined changes in free TA levels after pathogen inoculation (Figure 5). In gl1-1 leaves inoculated with Pst DC3000 (avrRpm1), the levels of free LA increased significantly within 4 h (P < 0.03), and rose further during the time of observation (Figure 5a). No obvious increase in the levels of free LA was observed in mock-inoculated leaves, in gl1-1 leaves inoculated with Pst DC3000 and in rpm1-1 leaves inoculated with Pst DC3000 (avrRpm1) (Figure 5a). There was no significant difference in the levels of free 16:3 among the examined treatments (Figure 5a, insert).
Furthermore, to confirm that the released LA induced by the inoculation with Pst DC3000 (avrRpm1) is derived from chloroplast lipids, we examined the levels of free LA in the fad7fad8 mutant (Col-0 background) after inoculation with Pst DC3000 (avrRpm1). In contrast to wild-type Col-0 leaves, the accumulation of free LA was not induced in fad7fad8 leaves inoculated with Pst DC3000 (avrRpm1) (Figure 5b). These results demonstrate that the liberated LA after inoculation with Pst DC3000 (avrRpm1) is derived from MGDG because the fad7fad8 mutant compromises the biosynthesis of LA in chloroplast lipids, mostly in MGDG. It is unclear why the reduction in LA levels in MGDG occurred in avrRpm1/RPM1-independent responses but the accumulation of free LA did not (Figures 4a and 5a). The liberated LA from MGDG might be promptly metabolized or converted to some available derivative in Avr/R-independent responses. The increase in the LA levels of phospholipids in extrachloroplast membranes and the accumulation of free LA derived from MGDG were correlated and dependent on an avrRpm1/RPM1-mediated response (Figures 4b and 5a). Taken together, these results suggest that at least the transfer of LA from chloroplast to extrachloroplast is involved in Avr/R-mediated responses, although the step of liberation of LA from MGDG may not be Avr/R-specific.
Figure 6 shows the increase in DAB precipitates following the inoculation of gl1-1 leaves with Pst DC3000 (avrRpm1). The accumulation of DAB precipitates started about 1 h after the inoculation in an avrRpm1/RPM1-dependent manner. These results demonstrate that the transfer of LA from chloroplast to extrachloroplast occurred coincident with H2O2 accumulation, although the transfer did not necessarily precede H2O2 accumulation. Therefore, it seems plausible that LA liberated from chloroplast membranes was utilized for the activation of NADPH oxidase localized in the plasma membrane in leaves inoculated with Pst DC3000 (avrRpm1).
Reduced resistance of the fad7fad8 mutant to avirulent Pst DC3000
Having studied the role of TAs in the induction of Avr/R-mediated HR, we next assessed the resistance to avirulent pathogens in the fad7fad8 mutant (Figure 7). In the wild-type Col-0, proliferation of the avirulent pathogens Pst DC3000 carrying the avrRpm1 or avrRpt2 gene is limited because Col-0 possesses functional RPM1 and RPS2 that recognize the corresponding Avr proteins (Mindrinos et al., 1994), leading to the oxidative burst response. In contrast, suppression of the growth of Pst DC3000 (avrRpm1) and Pst DC3000 (avrRpt2) did not occur in the fad7fad8 mutant (Col-0 background) (P > 0.05, t-test). The fad3 mutant (Col-0 background) exhibited a partial resistance to these avirulent pathogens (P < 0.01). The fad3 mutant also has somewhat reduced level of LA in chloroplast lipids (Browse et al., 1993). All Arabidopsis genotypes tested, however, allowed growth of the virulent Pst DC3000, which lacks avr genes. Taken together, these results support the idea that increased growth of the avirulent pathogens irrespective of avr genes in the fad7fad8 mutant is due to reduced efficiency of the Avr/R recognition system, and that TAs in chloroplast membrane lipids are involved in the resistance response downstream of the Avr/R recognition.
We demonstrate that LA derived from chloroplast lipids functions as a lipid molecule signal in promoting efficiency of HR and mounting disease resistance against avirulent Pst. We showed that the accumulation of ROI was reduced in the fad7fad8 mutant after being treated with ozone or inoculated with Pst DC3000 (avrRpm1). These findings suggest that TAs in chloroplast lipids contribute to the production of ROI during HR. NADPH oxidase localized in the plasma membrane is responsible for the production of ROI. In wild-type leaves inoculated with Pst DC3000 (avrRpm1), TAs were liberated from chloroplast MGDG and transferred to extrachloroplast membranes including the plasma membrane. Moreover, LA was particularly effective in activating NADPH oxidase. The liberated LA may diffuse throughout the cytosol, and then be partly incorporated into phospholipids of the plasma membrane and be partly utilized for the activation of NADPH oxidase. It is, however, important to note that the fad7fad8 mutant is not completely lacking in the production of ROI. LA derived from chloroplast may function in the promotion of efficiency of ROI production, but not in the initiation. We also showed that the fad7fad8 mutant displayed attenuated cell death in leaves inoculated with Pst DC3000 (avrRpm1) and the reduced resistance to avirulent pathogens. These results imply that TAs in chloroplast lipids play important roles in the control of cell death and the establishment of resistance. Resistance was moderately suppressed in the fad3 mutant. The level of LA, but not 16:3, is somewhat reduced in chloroplast lipids of fad3 mutant, although to a lesser extent than in that of the fad7fad8 mutant. FAD3 has functional redundancy for the supply of LA to chloroplast lipids. The correlation between this functional redundancy and moderate loss of resistance in the fad3 mutant supports the idea that LA in chloroplast lipids is concerned with the establishment of resistance. Recently, several reports have demonstrated that derivatives of LA, which are metabolized, for example, by lipoxygenase, α-dioxygenase and hydroperoxide lyase, are involved in defense responses including cell death and resistance (Croft et al., 1993; de León et al., 2002; Rustérucci et al., 1999). Therefore, a part of LA liberated from chloroplast lipids may be metabolized under exquisite control of these enzymes during a series of defense responses and each of the metabolites of LA may be utilized for the control of cell death or the establishment of resistance.
The reduction in ROI accumulation in the fad7fad8 mutant suggested that TAs are involved in the mechanism of NADPH oxidase activation. Notably, LA was the most effective fatty acid in the activation of NADPH oxidase. NADPH oxidase was activated to some extent in vitro by 18:2 (Figure 3b). However, 18:2 may not contribute to the activation in vivo because its level is considerably lower than that of LA in chloroplast lipids. Incidentally, contrary to LA, the levels of 18:2 increased in MGDG but not in phospholipids of extrachloroplast membranes after inoculation with Pst DC3000 (avrRpm1) (data not shown). AA activates NADPH oxidase in animal cells by stimulating the association with cytosolic regulatory proteins, for example p47phox (Shiose and Sumimoto, 2000). In plants, cytosolic proteins are not required for activation (data not shown). Although AA activates NADPH oxidase directly in human neutrophiles by increasing the affinity of the oxidase for its substrates (Doussiere et al., 1999; Rubinek and Levy, 1993), AA did activate NADPH oxidase only slightly in our system (Figure 3b). The NADPH oxidase in plant cells has incompletely conserved NADPH-binding motifs (Torres et al., 2002; Yoshioka et al., 2003). We speculate that LA might substitute for AA, which is not present in plant leaves, in the activation of the plant NADPH oxidase and might regulate the affinity of the oxidase for NADPH. Orozco-Cardenas and Ryan (1999) demonstrated that exogenous application of methyl JA induces H2O2 accumulation resulting from NADPH oxidase activity in tomato. However, H2O2-mediated glutathione S-transferase (gst1) gene expression is independent of the JA signaling mediated by COI1, a subunit of the SCF complex, in Arabidopsis leaves inoculated with avirulent Pst (Grant et al., 2000). Furthermore, the JA signaling mediated by JAR1, an adenylate-forming enzyme, is involved in the attenuation of ozone-induced H2O2 accumulation in Arabidopsis. In addition, application of methyl JA suppresses ozone-induced H2O2 accumulation in Arabidopsis ecotype Cvi-0 which has reduced sensitivity to JA (Rao et al., 2000b). This apparent contradiction suggests that JA might activate NADPH oxidase through unidentified signaling pathways that differ from the pathogen-stimulated ones, although the effects of JA on the activation of NADPH oxidase have not yet been demonstrated in vitro (Figure 3b).
The fad7fad8 mutant displayed lower accumulation of H2O2 and a lesser extent of cell death than the wild type after inoculation with a low dose of Pst DC3000 (avrRpm1) (Figure 2). In the Arabidopsis variegated mutant (CS3681) inoculated with Pst DC3000 (avrRpt2), the white leaf areas which contain low numbers of functional chloroplasts exhibit a significant reduction in ion leakage, a reaction indicating cell death, compared with green leaf areas (Genoud et al., 2002). Taken together, these findings suggest that chloroplasts are involved in controlling the efficiency of HR. However, inoculation with a higher dose caused similar rates of H2O2 accumulation and cell death in the wild type and the fad7fad8 mutant (data not shown). It is well known that elicitor-induced H2O2 accumulation and cell death are observed in non-photosynthetic suspension-cultured cells obtained from various plants. In the variegated mutant CS3681, cell death does occur in white leaf areas, although to a lesser extent than in green regions (Genoud et al., 2002). Similar to root cells, the pale cells must be expected to contain low amounts of TAs in their undifferentiated plastids. Taken together, we conclude that TAs in chloroplast lipids may be important for the regulation on the strength of HR, but not for initiation. In fact, the transfer of TAs to extrachloroplast membranes did not necessarily precede H2O2 accumulation (Figures 4–6). Moreover, low levels of H2O2 accumulation were detected in the fad7fad8 mutant with a lag of a few hours (data not shown).
Nandi et al. (2003) demonstrated that several sfd mutants that have abnormal fatty acid composition in chloroplast lipids are involved in the ssi2-conferred cell death. sfd mutants were isolated as suppressors of the ssi2 mutation that confers spontaneous cell death and enhanced resistance to virulent Psm. The reduction in 16:3 levels in all sfd ssi2 double mutants had suggested that lipid species containing 16:3 were involved in the expression of ssi2 phenotypes. Interestingly, ssi2 plants have significant amounts of 16:3 in their phospholipids. We also detected the transition of 16:3 from MGDG to phospholipids after inoculation with Pst DC3000 (avrRpm1) (Table 2, Figure 4). Similar to sfd mutants, the fad7fad8 double mutant, which contains no detectable 16:3 in its leaf lipids, showed decreased rates of cell death (Figure 2). Although it is unclear whether avirulent pathogen-induced cell death corresponds to ssi2-conferred cell death, 16:3 itself, its metabolites, or 16:3-containing lipid molecules derived from chloroplast lipids may be involved in the control of cell death execution.
Reduced resistance to avirulent pathogens in the fad7fad8 mutant suggest that TAs in chloroplast lipids are involved in the Avr/R-mediated resistance (Figure 7). ROI have been proposed to be responsible for the establishment of this resistance (Lamb and Dixon, 1997). However, atrboh mutants, which lack the NADPH oxidases that mediate the oxidative burst, are not compromised in their competence to perform the resistance response (Torres et al., 2002). The reduction in ROI accumulation in the fad7fad8 mutant and the effect of LA on the activation of NADPH oxidase may not be the only factors causing weakened resistance. The reduction in 16:3 levels in the sfd1 ssi2 mutant causes the suppression of ssi2-conferred resistance to virulent Psm, suggesting that 16:3 is involved in the mechanism of resistance in the ssi2 mutation (Nandi et al., 2003), although basal resistance to virulent Psm is unaffected in the sfd1 mutant compared with the wild type (Nandi et al., 2004). If 16:3 is necessary to induce resistance to avirulent pathogens, we cannot explain why the resistance was moderately inhibited in the fad3 mutant although its 16:3 level was quite similar to that of the wild type (Browse et al., 1993) (Figure 7). The sfd1 mutant does not compromise the restriction of avirulent pathogen growth, suggesting that 16:3 is not involved in the Avr/R-mediated resistance (Nandi et al., 2004). However, LA levels in chloroplast lipids were somewhat reduced in the fad3 mutant. Incidentally, plants lacking 16:3 fatty acid molecules, for example rice, also show disease resistance. Taken together, LA in chloroplast lipids might be involved in establishing pathogen resistance independently of ROI.
In conclusion, we show that TAs from chloroplast lipids play an important role in the Avr/R-mediated oxidative burst, cell death, and pathogen resistance. In particular, LA activates the NADPH oxidase. However, it is not yet clear how the TAs function in the control of cell death and resistance. To elucidate the mechanisms of these responses, the nature of the TAs and their derivatives, for example hydroperoxides and aldehydes, will have to be examined in detail.
Plant growth and bacterial strains
Arabidopsis thaliana ecotype Col-0 was used as a wild-type for the fad3 and the fad7fad8 mutant plants. Arabidopsis Col gl1-1 mutant was used as a control plant for the rpm1-1 mutant because the rpm1-1 mutant is in a gl1-1 background (Oppenheimer et al., 1991). Plants were grown on MS medium under continuous light at 23°C. Eighteen-day-old plants were used for superoxide detection. For bacterial infection assays, 2-week-old plants were transferred to the soil and then grown under continuous light at 23°C for 1–2 weeks. Pst DC3000 strains and the plasmids containing the avr genes have been described (Bisgrove et al., 1994). Pst DC3000 strains were grown at 28°C on King's B medium supplemented with 100 μg ml−1 rifampicin plus 30 μg ml−1 kanamycin.
Ozone treatment and in situ detection of superoxide
Plants were exposed to 250 p.p.b. ozone for 6 h. Ozone was generated with ozone generator (Model M0-5A; Koito Industries Ltd, Tokyo, Japan) and concentrations were monitored with an ozone monitor (Model 1100; Dylec Inc., Ibaraki, Japan). After treatment, young leaves were detached and vacuum-infiltrated with 10 mm potassium phosphate buffer (pH 7.8) containing 10 mm NaN3. After incubation with 0.1% NBT for 30 min, the leaves were washed with lactophenol alcohol and then photographed (Jabs et al., 1996). To quantify formazan generation, washed leaves were boiled in DMSO until formazan precipitates were eluted completely. The amount of formazan was determined as absorbance at 560 nm.
Bacterial infection, DAB staining and trypan blue staining
Pst inoculation of leaves was carried out by syringe infiltration for HR, and by vacuum infiltration for bacterial growth assay. DAB staining and trypan blue staining were performed as described by Torres et al. (2002). To quantify the intensity of the DAB stain, the average index of DAB stain per unit area was estimated using scion image (Scion Corp., Frederick, MD, USA). Four 0.5 cm2 leaf disks were harvested from infected leaves at 0 and 3 days after vacuum infiltration. Leaf samples were ground in 10 mm MgCl2 and plated in serial dilution on King's B agar containing 100 μg ml−1 rifampicin plus 30 μg ml−1 kanamycin. Infection rates were assessed by counting bacterial colonies.
In vitro NADPH oxidase assay
Purification of plasma membranes was performed as described by Hamada et al. (1998). To avoid contamination of chloroplastic constituents with plasma membranes, we used cultures of tobacco bright yellow 2 cells which lack functional chloroplasts. The cultures were pulverized in liquid nitrogen, and then suspended in 50 mm Tris–HCl buffer (pH 7.5) containing 3 mm Na2EDTA, 10 mm ascorbic acid, 5 mm dithiothreitol and 0.25 m sucrose. The homogenate was centrifuged at 5000 g for 10 min. The supernatant was centrifuged at 20 000 g for 20 min, and the remaining supernatant was further centrifuged at 140 000 g for 120 min to obtain the microsomal pellet. The pellet was resuspended in 50 mm Tris–HCl buffer (pH 7.5) containing 3 mm KCl and 0.25 m sucrose, and the plasma membrane was further purified using the aqueous polymer two-phase system. NADPH oxidase activity in the plasma membranes was evaluated by measuring the superoxide dismutase-inhibitable reduction of NBT (Van Gestelen et al., 1997). The reaction mixture consisted of 50 mm Tris–HCl (pH 7.5), 0.25 m sucrose, and 0.1 mm NBT. After 5 min pre-incubation at 25°C, the reaction was initiated by the addition of 0.1 mm NADPH. The reduction of NBT was monitored as the change of absorbance at 560 nm. The activity was expressed as nmol min−1 mg−1 protein using an extinction coefficient of 12.8 mm−1 cm−1. Fatty acids dissolved in DMSO buffer were added (0.1% DMSO final concentration) at 25°C 5 min before the reaction was started. As control, the same volume of DMSO buffer was added.
Leaves were harvested 6 h after ozone treatment or 6 h after inoculation with Pst DC3000 (avrRpm1) (5 × 106 cfu ml−1). Leaf lipids were extracted and separated into classes as described previously (Miquel and Browse, 1992). The fatty acid composition of individual lipids was determined by gas chromatography as described before (Kodama et al., 1994). Free fatty acids were methylated using diazomethane in the presence of 15:0 fatty acid as an internal standard (Kates, 1986).
This work was supported by the Ministry of Agriculture, Forestry and Fisheries of Japan (Rice Genome Project IP-5005) and by Research for the Future from the Japan Society for the Promotion of Science. We thank Miguel Angel Torres for technical advices and critical reading of the manuscript. We thank Jeffery L. Dangl for providing Pst DC3000 strains and the gl1-1 and the rpm1-1 mutant seeds. We also thank Kazuo Shinozaki, Tsutomu Kawasaki, Hideki Sumimoto, and Chris R. Somerville for valuable discussion.