Upon a dark/light shift the conditional flu mutant of Arabidopsis starts to generate singlet oxygen (1O2), a non-radical reactive oxygen species that is restricted to the plastid compartment. Immediately after the shift, plants stop growing and develop necrotic lesions. We have established a protoplast system, which allows detection and characterization of the death response in flu induced by the release of 1O2. Vitamin B6 that quenches 1O2 in fungi was able to protect flu protoplasts from cell death. Blocking ethylene production was sufficient to partially inhibit the death reaction. Similarly, flu mutant seedlings expressing transgenic NahG were partially protected from the death provoked by the release of 1O2, indicating a requirement for salicylic acid (SA) in this process, whereas in cells depleted of both, ethylene and SA, the extent of cell death was reduced to the wild-type level. The flu mutant was also crossed with the jasmonic acid (JA)-depleted mutant opr3, and with the JA, OPDA and dinor OPDA (dnOPDA)-depleted dde2-2 mutant. Analysis of the resulting double mutants revealed that in contrast to the JA-induced suppression of H2O2/superoxide-dependent cell death reported earlier, JA promotes singlet oxygen-mediated cell death in flu, whereas other oxylipins such as OPDA and dnOPDA antagonize this death-inducing activity of JA.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
In plants, reactive oxygen species (ROS) are produced as byproducts of metabolic pathways that are scavenged by antioxidative defense mechanisms (Apel and Hirt, 2004; Foyer and Noctor, 2000). In plants under stress the equilibrium between production and scavenging of ROS may be perturbed resulting in a transient increase in ROS levels (Elstner, 1991) that have been closely associated with the emergence of various disorders such as cell death, disease, and aging (Neill et al., 2002; Overmyer et al., 2003). ROS affect such disorders in two different ways. First, they may react with a large variety of biomolecules, causing irreversible damage that can lead to tissue necrosis and ultimately may kill the plant (Girotti, 2001; Rebeiz et al., 1988). Secondly, ROS may alter the expression of genes and affect signal transduction pathways. These latter observations suggest that cells may have evolved strategies to use ROS as biological stimuli/signals that activate various genetic response programs (Dalton et al., 1999). If ROS act as signals their biological specificities should exhibit a high degree of selectivity and specificity that could be ascribed to their chemical identities and/or the intracellular sites at which they were produced.
To date, only a very few case studies have been performed that suggest a selective signaling effect of a given ROS in plants. This may in part be due to the fact that most of the previous studies have focused on analyzing the signaling role of H2O2 and/or superoxide radicals, whereas other ROS such as hydroxyl radicals and singlet oxygen have been largely ignored (Apel and Hirt, 2004). Recently we have developed an experimental strategy that allows studying the biological activity of singlet oxygen. The conditional flu mutant of Arabidopsis releases singlet oxygen in plastids in a controlled and non-invasive manner (op den Camp et al., 2003). Immediately after the release of singlet oxygen plants stopped growing, initiated a cell death response and activated distinct sets of early stress-response genes that were different from those upregulated by superoxide/hydrogen peroxide. Collectively, the results suggest that the biological activity of singlet oxygen exhibits a high degree of specificity that differs from that of other ROS.
Programmed cell death (PCD) responses of wild-type plants that resemble singlet oxygen-mediated cell death reactions of the flu mutant are attributed to the production of superoxide radical/hydrogen peroxide during incompatible plant–pathogen interactions, and under various abiotic stress conditions (Beers and McDowell, 2001; Hoeberichts and Woltering, 2003; Overmyer et al., 2003). The initiation, spreading, and containment of these cell death reactions was controlled by a complex network of signaling pathways, some of which were activated by the three phytohormones ethylene, salicylic acid (SA) and jasmonic acid (JA). Ethylene and SA are associated with the initiation of PCD, whereas JA has been implicated with the suppression and containment of PCD (Devadas et al., 2002; Overmyer et al., 2000, 2003; Rao et al., 2002). In the present work we have analyzed the role of these phytohormones during singlet oxygen-mediated cell death reactions of the flu mutant in order to assess the extent of commonalities and differences between cell death reactions triggered by singlet oxygen or superoxide radical/hydrogen peroxide. Whereas the effects of SA and ethylene were very similar, the influence of JA and intermediates of its biosynthetic pathway on singlet oxygen-mediated cell death in flu was distinct from that reported earlier for PCD triggered by hydrogen peroxide/superoxide in wild type.
Singlet oxygen-mediated cell death in seedlings and protoplasts of the flu mutant
The conditional fluorescent (flu) mutant of Arabidopsis accumulates the photosensitizer protochlorophyllide in the dark. After a dark-to-light (LDL) shift, the generation of singlet oxygen, a non-radical ROS, starts within the first minute of illumination and is confined to plastids (op den Camp et al., 2003). Immediately after the shift, plants stop growing and develop necrotic lesions (Figure 1a). The impact of phytohormones on the initiation and spreading of singlet oxygen-mediated cell death reactions was studied in protoplasts that facilitated the uptake of active agents and the quantification of cell death reactions. Dead cells were detected by staining them with Evans blue dye (Figure 1b) (Wright et al., 2000). Protoplasts from wild-type and flu seedlings were isolated after a 15-h dark period, and the percentage of dead cells was calculated 1, 2, 4, 8 and 24 h after LDL (Figure 1c). During the first 2 h after shifting protoplasts from the dark back to light there was no significant difference in the percentage of dead cells between flu and wild type, showing that cell death in flu does not merely result from direct physicochemical damage caused by the release of 1O2, which is produced within the first minutes following LDL (op den Camp et al., 2003). Upon further illumination, however, the percentage of dead cells in flu rapidly increased and reached approximately 50% after 8 h and 80% after 24 h, respectively, whereas the number of dead wild-type cells remained almost constant (Figure 1c). The deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) technique has often been used in animals to identify apoptotic cells by detecting in situ nuclei containing digested DNA (Gorczyca et al., 1994). This technique has been adapted to the detection of PCD in plants (Danon and Gallois, 1998; Wang et al., 1996). In experiments performed in parallel to the Evans blue staining, we counted the percentage of TUNEL-positive protoplasts 1, 4, 8 and 24 h after LDL (Figure 1b,d). One hour after the LDL treatment there was little difference between flu and wt, but thereafter the percentage of TUNEL-positive cells rapidly increased in flu, reaching values close to 100% after 24 h (Figure 1d). The fact that the TUNEL assay measures prelytic DNA fragmentation suggests that in flu this degradation happens as a late event in the death process, as is described in animal apoptosis (Cohen, 1997; Green, 1998). Collectively, the above results indicate that 1O2 released by flu, rather than having a direct, cytotoxic effect, acts as a signal and during the cell death response induces DNA fragmentation, one of the hallmarks of PCD (Danon et al., 2000).
Inhibition of singlet oxygen-mediated cell death by vitamin B6
The importance of 1O2 for the induction of cell death in flu protoplasts was tested by using a scavenger that specifically quenched 1O2 in fungal pathogens. Necrotrophic fungi of the group Cercospora produce the cercosporin toxin to kill and collect nutrients from dead plants (Ehrenshaft et al., 1999). Cercosporin kills plant cells primarily via 1O2-mediated peroxidation of membrane lipids (Ehrenshaft et al., 1999). In order to withstand the deleterious effects of their own toxin, Cercospora species produce the protective molecule vitamin B6 (pyridoxine). Vitamin B6 has the capacity to scavenge the toxic effect of cercosporin by quenching 1O2 (Bilski et al., 2000). To study a possible effect of vitamin B6 during the 1O2-mediated cell death response, wt and flu seedlings were grown on standard medium or on medium complemented with 1 mm vitamin B6. Protoplasts from these seedlings were isolated after a 15-h dark period, and the percentage of dead cells was counted 1, 4, 8 and 24 h after LDL. Vitamin B6 had no effect on the percentage of dead cells of wt, but protoplasts isolated from flu grown on medium with vitamin B6 were protected and the percentage of dead cells was reduced, when compared with protoplasts from flu plants grown without added vitamin B6 (Figure 2a). We further analyzed the protective effect of vitamin B6 using several concentrations ranging from 0.002 to 1 mm. Protoplasts from flu grown on medium with 0.1; 0.5 and 1 mm of vitamin B6 were significantly protected from the 1O2-induced cell death in a dose-dependent manner (Figure 2b).
Singlet oxygen-mediated changes in gene expression
The total number of genes activated after the release of singlet oxygen had been assessed earlier in the flu mutant using Affymetrix gene chips that represented more than 95% of the total genome of Arabidopsis (op den Camp et al., 2003). Within the first 30 min of reillumination of flu mutants the expression of approximately 5% of the genome changed (op den Camp et al., 2003). The studies had been carried out with mature Arabidopsis plants that had started to bolt. In the present work 5-day-old seedlings were analyzed that may differ in their response to singlet oxygen. Therefore, changes in gene expression at this developmental stage of the flu mutant were measured by using Affymetrix gene chips that contained approximately 8200 probe sets. Most of the genes affected in seedlings of the flu mutant by the release of singlet oxygen had also been identified previously in mature plants, although in some cases the extent of up- or down-regulation of gene expression was different between the two sets of samples. Among the genes that were upregulated several were identified that encode proteins involved in the biosynthesis or signaling of SA, JA, or ethylene (Table 1). The expression of these genes was used to monitor indirectly the biological activities of the three phytohormones within seedlings and protoplasts of wild type and flu following a dark/light transition.
Table 1. Induction of genes associated with the biosynthesis and the signaling pathways of ethylene, SA, and JA
Induction level: flu/WT
Time after LDL
During oligonucleotide microarray expression analysis, 227 early induced genes were found among a total of 8200 gene-specific fragments to be induced with differences in hybridization signals of threefold or more in flu relative to wt. Among the genes whose expression was increased in flu 15 and/or 30 min after LDL, some are associated with the biosynthesis and the signaling pathways of ethylene, SA, and JA.
ACC synthase 6 (AtACS-6)
Putative protein similarity to EDS1
Allene oxide synthase
Requirement of ethylene for the induction of cell death in flu
Several genes that were induced early after re-illumination in flu (Table 1) are involved in either the biosynthesis or the signaling pathway of ethylene. Ethylene-responsive element-binding factors (ERFs) are members of a family of transcription factors that have a highly conserved DNA-binding domain, which interacts specifically with GCC box sequences (Hao et al., 1998). These proteins represent a plant-specific class of transcription factors that regulate a wide range of developmental processes (Riechmann and Meyerowitz, 1998). As shown in Table 1, four ERF genes were found to be induced in flu 15 and/or 30 min after LDL, AtERF1 (At3g23240), AtERF2 (At5g47220), AtERF5 (At5g47230) and AtERF6 (At4g17490). Changes in transcript levels of these four genes were quantified independently using real-time PCR to test the reliability of the Affymetrix gene chip analysis (Figure 3a). These measurements confirmed the increase in transcript levels of these genes. Transcripts of ERF5 and ERF6 began to increase almost instantaneously following the release of 1O2, whereas the initial accumulation of transcripts of the two other genes began later.
1-Aminocyclopropane-1-carboxylic acid (ACC) synthase (ACS), which converts S-adenosyl-Met to ACC, catalyzes the first committed and rate-limiting step in ethylene biosynthesis. AtACS6 (At4g11280), one of the seven ACS known in Arabidopsis (Wang et al., 2002), was induced in flu 30 min after LDL (Table 1). The induction of AtACS6 in flu after LDL was confirmed by real-time PCR (Figure 3b). The highest induction level of AtACS6 was detected 2 h after LDL. As the level of ACS activity determines the ACC content, which parallels closely the level of ethylene production (Yang and Hoffman, 1984), we also measured ACC in flu and wt seedlings after LDL exposure (Figure 3c). Within 3 h after the beginning of re-illumination ACC in flu accumulated to levels about fourfold higher than in the controls (Figure 3c). Thus, changes in AtACS6 expression correlated closely with ACC production, indicating that enhanced levels of AtACS6 might have contributed to the increase in ACC production in flu during re-illumination (Figure 3b).
L-aminoethoxyvinylglycine (AVG), a potent inhibitor of ACC synthase (Icekson and Apelbaum, 1983), has been widely used to block ethylene biosynthesis (Capitani et al., 2002; Heidstra et al., 1997; Ouaked et al., 2003). To determine whether ethylene-dependent signaling is involved in triggering cell death in flu, seedlings were grown on standard medium or on medium complemented with AVG (5 μm). Protoplasts from flu grown on medium with AVG were partly protected from the light-induced death that occurred in protoplasts from flu seedlings grown on control medium (Figure 3d). To confirm that this effect of AVG was due to its inhibition of ACC accumulation, seedlings were grown on a medium with AVG (5 μm) that was complemented with ACC (5 μm). As shown in Figure 3(d), addition of ACC abolished the protecting effect of AVG. These results suggest that cell death induced in flu is controlled, at least in part, by ethylene.
Requirement of salicylic acid for the induction of cell death in flu
Among the genes that were rapidly upregulated during re-illumination of pre-darkened flu seedlings, three were found that are known to be induced by SA. EDS1 has been implicated in R-gene-mediated resistance against pathogens and is known to be part of an SA-dependent signaling pathway. This gene, together with an EDS1-like gene (Falk et al., 1999), was induced within the first 30 min of re-illumination (C. Ochsenbein, F. Laudgraf and K. Apel, ETH Zürich, Zürich, Switzerland, in preparation, Table 1). Pathogen-related (PR) proteins have also been used as molecular markers for SA-dependent defense reactions such as the hypersensitive reaction and systemically acquired resistance. Upregulation of PR1 (At4g33720) expression was detected in the flu mutant only after 8 h of reillumination (data not shown). However, expression of another PR1-like gene (At2g14610; Table 1) was rapidly induced within the first 15 min of reillumination. These results derived from the microarray analysis can be confirmed by real-time PCR (data not shown). As all three genes require SA for their expression, the possible involvement of SA-dependent signaling pathways in the 1O2-mediated cell death reaction of flu was tested by combining flu with the SA-depleted transgenic NahG plant (Delaney et al., 1994) and the npr1-1 mutant (Cao et al., 1994).
Protoplasts from wt, flu, flu-NahG, and flu-npr1-1 seedlings were isolated after a 15-h dark period, and the percentage of dead cells was calculated 1, 4, 8 and 24 h after LDL. As shown in Figure 4(a), protoplasts isolated from flu-NahG seedlings were much less susceptible to 1O2-induced cell death than flu controls, suggesting that SA signaling is required for the execution of the cell death program. Protoplasts of flu-NahG seedlings, however, were only partially protected. A similar partial inhibiting effect was also found in protoplasts of flu after their ethylene biosynthesis had been inhibited by the addition of AVG. As both ethylene and SA seem to be required for 1O2-induced cell death, it was of interest to see whether the simultaneous suppression of ethylene and SA could further reduce the susceptibility of flu protoplasts to 1O2. Protoplasts isolated from flu-NahG seedlings that were grown on a medium complemented with AVG (flu-NahG + AVG) were almost fully protected from the death induced in the original flu. The percentage of dead cells in the flu-NahG + AVG population was significantly lower than in flu-NahG-derived protoplasts and was comparable to that in wt. Thus, components of SA and ethylene signaling pathways seem to function synergistically during regulation of cell death induced in flu. In contrast, flu-npr1-1 protoplasts were as susceptible to 1O2 as control flu protoplasts (Figure 4b). NPR1 takes part in the systemic acquired resistance and requires SA. This result suggests that flu-induced cell death depends on SA signaling components that act independently of NPR1. The SA contents of flu and wt seedlings were measured during reillumination. Production of SA in flu did not significantly differ from that in wild type (A. Buchala and A. Danon, unpublished data), suggesting that the basal level of SA of wt plants is sufficient and necessary for the execution of the cell death pathway in flu.
Involvement of oxylipins in the regulation of the singlet oxygen-mediated cell death in flu
JA biosynthesis takes place in chloroplasts and peroxisomes and occurs through the octadecanoid pathway (Turner et al., 2002; Weber, 2002). Its biosynthesis is initiated by the oxidation of linolenic acid (18:3), which is followed by the formation of cyclopentenone 12-oxo-phytodienoic acid (OPDA) catalyzed by allene oxide synthase (AOS) and allene oxide cyclase (AOC). OPDA reductase 3 (OPR3) then reduces OPDA to 3-oxo-2-(2′(Z)-pentenyl)-cyclopentane-1 octanoic acid (OPC-8:0). Finally, OPC:8 is shortened by three cycles of β-oxidation to yield JA (Stintzi and Browse, 2000; Turner et al., 2002; Weber, 2002). Among the genes exhibiting notably higher expression levels in flu after LDL were AOS (At5g42650) and OPR3 (At2g06050; Table 1). Lipoxygenases (LOXs) catalyze the oxygenation of fatty acids to their hydroperoxy derivatives, which serve as substrates for several enzymes, including AOS in JA biosynthesis. As shown in Table 1, LOX2 (At3g45140) and LOX3 (At1g17420) are also induced in flu early after LDL. Lipases are key regulators of stress reactions that control the release of signaling molecules from membrane lipids, and phospholipase A1 is responsible for the release of linolenic acid from membrane lipids (Ishiguro et al., 2001). Among the genes that were induced in flu very early after LDL, we have identified a gene (At1g54000) that encodes a protein closely related to lipases/acylhydrolases (Table 1). We were able to confirm the induction of all the genes mentioned above by real-time PCR (data not shown).
For determining whether the induction of genes involved in JA biosynthesis results in an increase of JA or some of its biosynthetic intermediates, we measured concentration changes in JA, OPDA, and its 16-carbon structural homolog, dinor OPDA (dnOPDA) (Weber et al., 1997), in wt and flu during re-illumination of pre-darkened wt and flu seedlings, respectively. As shown in Figure 5(a,b), concentrations of OPDA, dnOPDA, and JA increased sharply in flu seedlings, starting soon after re-illumination of the seedlings. The possible involvement of JA, OPDA, and dnOPDA in the cell death induced in flu was investigated further by using the following double mutants: flu with opr3, a mutation leading to JA depletion (Stintzi and Browse, 2000), and flu with dde2-2, a mutant defective in the ALLENE OXIDE SYNTHASE gene which is depleted in JA, OPDA, and dnOPDA (von Malek et al., 2002). The extent of cell death reactions in protoplast suspensions prepared from wt, dde2-2, opr3, flu, flu-dde2-2, and flu-opr3 seedlings was analyzed by counting Evans blue-stained cells after different lengths of re-illumination. As shown in Figure 5(c), no differences were found in the basal levels of dead protoplasts between the wt and the two single mutants dde2-2 and opr3. In contrast, protoplasts extracted from flu-opr3 seedlings exhibited a lower susceptibility to death induced by 1O2 release than flu alone. Death in flu-opr3 protoplasts occurred slowly and affected a smaller percentage of cells throughout reillumination indicating a role for JA as an inducer of cell death. To confirm that JA-dependent signaling is involved in triggering cell death in flu, seedlings were grown on standard medium or on medium complemented with JA (0.5 μm). Protoplasts from wt and flu grown on medium with JA were not affected in their viability showing that the low concentration of JA contained in the growth medium was not toxic (Figure 5c,d). In contrast, protoplasts extracted from flu-opr3 seedlings (which are depleted in JA), grown on medium with JA exhibited a higher susceptibility to death induced by 1O2 than protoplasts from seedlings grown on control medium (Figure 5c,d). In protoplasts obtained from flu-opr3 the addition of exogenous JA increased the extent of cell death close to that of the original flu mutant, confirming that the lack of JA in flu-opr3 was responsible for its lower susceptibility to 1O2. In contrast, no obvious difference in the death percentage was detected between the original flu mutant and the double mutant flu-dde2-2, suggesting that apparently the concomitant absence of JA, OPDA, and dnOPDA was not affecting the cell death process induced in flu. If JA was the only oxylipin capable of influencing this cell death induction, we should find the same suppressing phenotype in flu-opr3 and flu-dde2-2. Hence, these results indicate that JA acts as an inducer of cell death, whereas intermediate(s) in its biosynthesis pathway, most likely OPDA and/or dnOPDA, antagonize the induction of the cell death process in flu.
The notion that two different oxylipin derivatives with opposite effects on cell death exist in flu was also supported by feeding exogenous JA to flu-dde2-2. In contrast to flu-opr3, this mutant was not only devoid of JA but also lacked its antagonist. Addition of exogenous JA accelerated the initiation of the cell death reaction and further increased the number of dead cells to a level that exceeded that of the flu control (Figure 5d).
In contrast to earlier studies that have identified JA as a suppressor of PCD (Devadas et al., 2002; Overmyer et al., 2000, 2003; Rao et al., 2002), in flu it promotes the singlet oxygen-mediated cell death reaction. Previously a similar death-promoting activity of JA had also been found in protoplasts of wild-type Arabidopsis that were incubated with the fungal toxin fumonisin B1 (FB1). FB1-induced PCD requires – like in the flu mutant – SA, JA, and ethylene signaling pathways (Asai et al., 2000). It was of interest to determine whether, similar to flu protoplasts also in FB1-treated wild-type protoplasts, a second oxylipin may contribute to the control of PCD by antagonizing the death-promoting activity of JA.
Protoplasts extracted from opr3 seedlings exhibited a much lower susceptibility to death induced by FB1 than FB1-treated wild-type protoplasts, confirming the role of JA as an inducer of PCD as described by Asai et al. (2000) (Figure 6). In contrast, no difference in the death percentage was detected between FB1-treated protoplasts of the dde2-2 mutant and wt control (Figure 6). These results suggest that as in flu protoplasts, in FB1-treated wild type protoplasts cell death was inhibited only when either JA was absent or its antagonist, most likely OPDA/dnOPDA, was present. This inhibitory effect was erased, when the two oxylipin derivatives were either absent or present both at the same time.
In plants increased levels of ROS have been associated with various stress responses. ROS have been implicated as signals that trigger some of these reactions (Apel and Hirt, 2004). However, it has been difficult to demonstrate this proposed biological activity experimentally, because in stressed plants several chemically distinct ROS are generated simultaneously, thus making it difficult to discern the biological activity of each of these ROS separately (Apel and Hirt, 2004). We have tried to overcome these difficulties by exploiting the conditional flu mutant of Arabidopsis (Meskauskiene et al., 2001). The release of 1O2 in the flu mutant can be triggered non-invasively and in a highly synchronized and time-controlled manner without a concomitant increase in H2O2 or superoxide production (Kochevar, 2004; op den Camp et al., 2003). One of the stress reactions mediated by 1O2 is a cell death response, similar to the PCD during apoptosis in animals (Cohen, 1997; Green, 1998). The 1O2-mediated cell death response in flu is controlled by concurrent signaling pathways that depend on ethylene, SA, and JA to promote the execution of cell death, whereas OPDA and/or dnOPDA seem to suppress this response.
The protoplast assay
To study how oxylipins, salicylate, and ethylene intervene to regulate cell death, we have established a protoplast system, which allows detection and characterization of the death response in flu induced by the release of 1O2. The protoplasts assay constitutes a powerful tool that allows an easy uptake of active agents and a precise quantification of cell death induction in different populations. It is possible that protoplasts are primed for cell death during isolation but our study showed that the basal level of cell death in our controls is low and stable throughout the length of our experiments. In addition, the data obtained from the protoplast assay are applicable to intact seedlings, where it is possible to see differences in the intensity of necrotic lesions even if those differences are not as easily and clearly quantifiable as with the protoplast assay.
Vitamin B6: a 1O2 quencher
Vitamin B6, an efficient quencher of 1O2, is able to protect necrotrophic fungi of the group Cercospora from cercosporin, a toxin they produce to kill plant cells (Bilski et al., 2000). Like Pchlide, cercosporin absorbs light energy and reacts with molecular oxygen to produce 1O2 (Ehrenshaft et al., 1999). We used vitamin B6 to see if we could mimic its 1O2-quenching effect in our system, where cell death is also induced by 1O2. Interestingly, we observed that protoplasts from flu grown on medium with vitamin B6 were largely protected from the 1O2-induced death that occurred in control protoplasts (Figure 2). Other experiments were performed with H2O2 or menadione, a superoxide generator, which showed that the ROS protective effect of vitamin B6 is restricted to 1O2 (data not shown). As 1O2 is released in plants under various stress conditions such as high light (Fryer et al., 2002; Hideg et al., 1998; Macpherson et al., 1993), vitamin B6 could play a role in 1O2 detoxification not only in Cercospora but also in plants. Recently, a broader role has been assigned to vitamin B6 in response to stresses in Phaseolus vulgaris (Graham et al., 2004). Arabidopsis homologues of Cercospora genes implicated in vitamin B6 biosynthesis have been identified, but their possible role during the scavenging of 1O2 has not been analyzed yet.
Role of ethylene and SA signaling in promoting 1O2-induced cell death
When the expression of genes of wt and flu were compared immediately after 1O2 release, some genes implicated in ethylene and SA signaling pathways were rapidly upregulated in response to 1O2 (Table 1). Ethylene responsive factors (ERFs) bind to the GCC box, a DNA motif associated with ethylene and pathogen-induced gene expression. A large number of ERFs have been found in the Arabidopsis genome and other plant species (Riechmann and Meyerowitz, 1998). Nevertheless, only ERF1 takes part directly in the ethylene response pathway (Solano et al., 1998). ERF1 (Lorenzo et al., 2003) and ERF2 (Brown et al., 2003) are inducible by ethylene but the majority of ERFs are not. Instead, SA, JA, salt, drought, and a growing number of other stress factors regulate the expression of these genes (Wang et al., 2002). Thus, multiple signaling pathways may converge on ERFs to regulate gene expression in response to a variety of different biotic and abiotic stresses. Among the ERFs that were tested in Arabidopsis, only ERF5 and ERF6 were induced in flu very early after the release of 1O2 (Figure 3a), which may indicate that they are involved in early steps of transmitting signals released by 1O2 (Figure 7). ERF1 and ERF2 were upregulated only later, and could reflect a more indirect response to 1O2 possibly linked to OPDA/dnOPDA, JA and/or ethylene production (Figure 3a). Aminocyclopropane-1-carboxylic acid (ACC) synthase (ACS), which catalyzes ACC production, is generally the rate-limiting enzyme in ethylene biosynthesis. Among the seven ACS known in Arabidopsis (Wang et al., 2002), only AtACS6 was induced in flu after LDL (Figure 3b). A similar specific induction of AtACS6 has already been described following ozone induction of cell death (Overmyer et al., 2000), which could indicate a more specialized role for this ACS in ethylene-dependent PCD. Upregulation of AtACS6 preceded the increase in ACC production, suggesting that AtACS6 could take part in ethylene production in response to 1O2 (Figure 7). Finally, blocking ACC synthase activities by AVG was sufficient to inhibit partially the death induced in flu. Addition of ACC abolished this inhibiting effect of AVG. Collectively, these results suggest that ACC and ethylene production are necessary for cell death induction in flu as illustrated in Figure 7.
SA has often been described as a positive regulator of pathogen-induced disease resistance, although its function in relation to cell death is still poorly understood (Beers and McDowell, 2001; Hoeberichts and Woltering, 2003; Overmyer et al., 2003). The flu-NahG double mutant was partially protected from the death induced by 1O2 (Figure 4a), indicating a requirement of SA during the cell death response. Reduced susceptibility of the flu-NahG line may be due to the accumulation of catechol rather than the loss of SA (van Wees and Glazebrook, 2003). However, in the flu-NahG line Pchlide accumulated during darkness and the release of 1O2 during reillumination was demonstrated by the rapid upregulation of 1O2-specific marker genes.
Cell death induced in flu depleted of both SA and ethylene was even more reduced, reaching the level of the wt (Figure 4a). Thus, the SA and ethylene signaling pathways may function additionally during the regulation of this cell death response (Figure 7) and resemble the cooperative action of SA and ethylene already described for other types of plant PCD (Beers and McDowell, 2001; Hoeberichts and Woltering, 2003; Overmyer et al., 2003). SA signaling is mediated by at least two signaling pathways, one requiring the NON-EXPRESSOR OF PR1 (NPR1) gene and a second that is independent of NPR1 (Shah, 2003). As the double mutant flu-npr1-1 was not protected from the death induced by 1O2 (Figure 4b) the induced cell death in flu is likely to occur via the NPR1-independent pathway.
Role of oxylipins in regulating flu-induced cell death
JA is a plant signaling compound, which plays a role during biotic and abiotic stress and wounding responses (Farmer et al., 2003; Turner et al., 2002; Weber, 2002). The precise mechanism by which JA signaling regulates cell death is far from being understood, in part because two opposite effects of JA on cell death had been reported. In FB1-treated protoplasts of Arabidopsis JA enhanced the rate of cell death (Asai et al., 2000), whereas in other studies JA suppressed cell death mediated by superoxide radicals (Devadas et al., 2002; Overmyer et al., 2000; Rao et al., 2002). Early after the release of 1O2, genes involved in the biosynthesis of oxylipins such as AOS, LOX2, and LOX3 exhibited higher expression levels (Table 1), which were followed by the generation of OPDA, dnOPDA, and JA (Figure 5a). The comparison between flu and the flu-opr3 double mutant depleted in JA (Stintzi and Browse, 2000) indicated that JA promotes PCD induced by 1O2 (Figure 5c). This role of JA as a promoter of cell death in flu was confirmed by complementing the mutant with low concentrations of JA (Figure 5d). Unexpectedly, the comparison between flu and the flu-dde2-2 mutant depleted in JA, OPDA, and dnOPDA (von Malek et al., 2002) showed that the concomitant absence of these three oxylipins reestablished the original sensitivity of flu to cell death (Figure 5c). Therefore, death inhibition can occur only when OPDA/dnOPDA is present and JA is absent (flu-opr3), whereas when OPDA/dnOPDA together with JA are present (flu) or absent (flu-dde2-2) at the same time the effect of these oxylipins on cell death is no longer apparent. These results suggest that OPDA and/or dnOPDA are able to antagonize the effect on cell death promoted by JA (Figure 7). Such an inhibitory role of oxylipins derived from the JA biosynthesis pathway has already been demonstrated in defesce signaling (Farmer et al., 2003; Thoma et al., 2003; Turner et al., 2002; Weber, 2002). For instance, in the absence of JA, OPDA and/or dnOPDA inhibit cell death induced by insect and fungal attack in the opr3 mutant (Stintzi et al., 2001).
In FB1-treated wild-type protoplasts of Arabidopsis the activation of PCD is promoted by JA and, as shown in the present work, this JA-dependent stimulation of cell death seems to be suppressed by OPDA/dnOPDA, suggesting that in flu and FB1-treated protoplasts similar signaling pathways are involved in the induction of PCD. FB1 affects PCD by blocking the biosynthesis of sphingolipids. Sphingolipids have been implicated as elicitors of apoptosis in animals (Cuvillier, 2002) and recent evidence indicates that in plants sphingolipids may also play a similar role (Liang et al., 2003). As 1O2 generated in photosensitized cells of animals induces apoptosis via sphingolipids as a second messenger (Wispriyono et al., 2002), it is tempting to assign a similar signaling role to sphingolipids during 1O2-mediated PCD in plants.
In experiments similar to those shown in Figure 5(d), we also tried to verify the proposed inhibiting role of OPDA and dnOPDA by adding them to the protoplast culture. As dnOPDA was not commercially available, only OPDA was tested. The results of these experiments were inconclusive because depending on the concentration used OPDA had either no effect (1 μm) or was toxic for both wt and flu protoplasts (30 μm, data not shown). There are several ways to explain these results. Endogenous production of OPDA may be confined to the plastid whereas exogenous OPDA may also accumulate within the cytoplasm and initiate synthesis of JA that may antagonize the effect of OPDA. Alternatively, OPDA alone may not be sufficient to inhibit PCD but may require the concomitant action of dnOPDA. Finally, it may be dnOPDA, rather than OPDA, that inhibits PCD (Figure 7).
To date, only a very few case studies have been performed that suggest a selective signaling effect of a given oxylipin during PCD in plants. This may in part be due to the fact that most of the previous studies have focused on analyzing the signaling role of JA, whereas other oxylipins such as OPDA and dnOPDA have been largely ignored. Moreover, very often, mutants used in these studies, such as coi1 or jar1, do not allow one to discriminate between different oxylipin signaling pathways (Staswick et al., 2002; Stintzi et al., 2001). The use of mutants such as dde2-2 or opr3 may help to define more precisely the role of the different oxylipins involved in PCD of plants.
Plant materials and growth conditions
Seeds were sterilized and sown under aseptic conditions on Murashige Skoog (MS) agar plates supplemented with 0.5% sucrose. Seedlings were grown for 5 days at 20°C in continuous light (80 μmol m−2 sec−1), transferred to the dark for 15 h and reilluminated for various lengths of time (80 μmol m−2 sec−1). Mutants used in this work were flu (Meskauskiene et al., 2001), npr1-1 (Cao et al., 1994), dde2-2 (von Malek et al., 2002), opr3 (Stintzi and Browse, 2000) and a transgenic NahG line (Delaney et al., 1994). Chemicals were purchased from Sigma (Saint-Louis, MO, USA).l-aminoethoxyvinylglycine (AVG), ACC and vitamin B6 (pyridoxine) stock solutions were prepared in water, JA in ethanol and Fumonisin B1 in methanol.
Within the first minutes of reillumination distinct sets of genes are activated in the flu mutant that can be used as markers of 1O2 production. We used these marker genes to confirm that 1O2 production was not affected in the different mutants or during treatments used in combination with flu. Measurements of Pchlide were used to show that in the flu background its accumulation during darkness was not affected.
Protoplast preparation and determination of cell death
Arabidopsis protoplasts were isolated from 5-day-old seedlings grown under continuous light by incubating leaves for 15 h in the dark in the presence of the digestion medium (MS medium, mannitol 0.5 m, cellulase 1.2%, Macérozyme 0.8% pH 5.8). During this dark period cell walls were digested and at the same time protochlorophyllide accumulated in cells of the flu mutant. Protoplasts were then transferred to the light and separated from cellular debris by filtration through a 100-μm mesh filter (Milian SA, Geneva, Switzerland), followed by flotation on a 0.4-m sucrose solution at 60 × g for 10 min. They were then collected and centrifuged at low speed (60 × g) for 5 min and washed in a modified W5 medium (Menczel and Wolfe, 1984), in which the CaCl2 concentration had been changed from 125 to 1 mm and 0.1 m mannitol and 1 mm MgCl2 were added. Finally, protoplasts were resuspended in the culture medium containing MS medium (supplemented with 0.4 m sucrose, 0.4 m mannitol). The percentage of dead cells was determined by staining protoplasts with Evans blue dye (Sigma), which was added to the samples to a final concentration of 0.04%. The number of stained cells was determined using a light microscope.
In situ detection of nuclear DNA fragmentation
For in situ detection of DNA fragmentation 50 μl of a suspension of protoplasts was transferred onto a polylysine slide (Menzel-Glaser GmbH, Braunschweig, Germany), fixed with an equal volume of 10% buffered formalin (Sigma) and dried for 1 h at 42°C. The fixed protoplasts were washed once with PBS buffer (pH 7.4), permeabilized with 10 μg ml−1 proteinase K (Bioprobe, Sumrall, MS, USA) for 10 min at room temperature and rinsed twice with PBS buffer (pH 7.4). The protoplasts were then dried at 42°C for 10 min and labeled for 1 h in the dark at 37°C using a commercially available TUNEL kit (fluorescein, in situ Cell Death Detection Kit; Boehringer, Mannheim, Germany) at a dilution of 1:2 in the reaction buffer. After rinsing three times with PBS buffer (pH 7.4), the slides were viewed with a fluorescence microscope (Zeiss Axiophot; Carl Zeiss AG, Oberkochen, Germany), using a blue wavelength for fluorescein detection (Zeiss, FITC filter Blue-450-490). For each sample photographs of multiple microscopic fields were taken and TUNEL-positive nuclei scored on prints.
RNA isolation and RT-PCR analysis
Total RNA was prepared according to Melzer et al. (1990). RNA was treated with RQ1 RNase-Free DNase (Promega, Madison, WI, USA) and reverse-transcribed using random hexamers and SUPERSCRIPTTM II RNase H− Reverse Transcriptase (Invitrogen, Carlsbad, CA, USA) following the manufacturer's recommendations. Quantitative real-time PCR was performed with an ABI PRISM® 7700 Sequence Detection System (Applied Biosystems, Foster City, CA, USA) using a SYBR Green PCR kit from Applied Biosystems and specific primers for the different genes analyzed. Relative mRNA abundance was calculated using the comparative delta-Ct method and normalized to the corresponding actin2 (At3g18780) gene levels. Equivalent efficiencies between the different probes and our internal standard were observed during the PCRs.
Analysis of gene expression using oligonucleotide microarrays
About 200–300 seedlings from wt and flu were harvested 15 and 30 min after LDL and immediately frozen in liquid nitrogen. RNA was prepared from frozen tissue and equal amounts of RNA were used for the labeling reaction. Preparation of cDNA and biotin-labeled cRNA was performed as recommended by Affymetrix (Santa Clara, CA, USA). The AtGenome1 Array (Affymetrix), which contains more than 8200 probe sets, was used as the GeneChip. Detection of labeled cRNA using streptavidin-phycoerythrin and reading of the arrays using a confocal scanner (Affymetrix) were performed according to the manufacturer's instructions.
Jasmonate and OPDA measurement
About 0.5 g fresh weight of seedlings was homogenized in a mortar and extracted with 10 ml of methanol. Appropriate amounts of (2 H6)JA and (2 H5)OPDA were added as internal standards for GC-MS analysis. Purification, fractionation and quantification was performed as described earlier (Stenzel et al., 2003). The values shown in Figure 5(a,b) represent the mean and standard deviations of three independent experiments.
Frozen seedlings were extracted in 50% methanol for 30 min at 50°C. The content of ACC in the methanol extract was determined by conversion to ethylene and quantification by gas chromatography (Lizada and Yang, 1979).
We are indebted to André Imboden for taking care of the plants, to Dieter Rubli for art-work and to Martha Geier-Bächtold for editorial work. We thank Dr Annick Stintzi for opr3 seeds, Drs Beat Keller and Bernadette von Malek for dde2-2 seeds, and Mrs S. Vorkefeld for GC-MS sample preparation. Financial support through the Swiss Federal Institute of Technology and the Swiss National Science Foundation is gratefully acknowledged.