The Oleate-Stimulated Phospholipase D, PLDδ, and Phosphatidic Acid Decrease H₂O₂-Induced Cell Death in Arabidopsis

Abstract

Hydrolysis of common membrane phospholipids occurs in response to various environmental stresses, but the control and cellular function of this hydrolysis are not fully understood. Hydrogen peroxide (H₂O₂) is a pivotal signaling molecule involved in various stress responses. Here, we show that the plasma membrane–bound phospholipase D, PLDδ, is activated in response to H₂O₂ and that the resulting phosphatidic acid (PA) functions to decrease H₂O₂-promoted programmed cell death. The Arabidopsis genome has 12 PLD genes, and knockout of PLDδ abolishes specifically the oleate-stimulated PLD activity. H₂O₂ treatment of Arabidopsis cells activates PLD enzyme activity, and ablation of PLDδ abolishes that activation. PLDδ-null cells display increased sensitivity to H₂O₂-induced cell death. The addition of PA to PLDδ-null cells mitigates the H₂O₂ effect, whereas suppression of the H₂O₂-induced PA formation in wild-type cells increases the effect. PLDδ-ablated plants exhibit increased susceptibility to stress. These results demonstrate that activation of oleate-stimulated PLDδ constitutes an important step in the plant response to H₂O₂ and increasing plant stress tolerance.

INTRODUCTION

Plants generate and accumulate hydrogen peroxide (H₂O₂) in response to a wide variety of biotic and abiotic stresses, including UV irradiation, excess light, wounding, heat, chilling, water deficits, and pathogens (Prasad et al., 1994; Sharma et al., 1996; Bowler and Fluhr, 2000; Breusegem et al., 2001; Mittler, 2002; Neill et al., 2002a; Kwak et al., 2003; Park et al., 2003; Yoshioka et al., 2003). Increasing evidence shows that H₂O₂ triggers various cellular changes and that the responses are dose dependent. Low dosages of H₂O₂ enhance plant tolerance to multiple stresses and pathogens, whereas high dosages promote programmed cell death (Greenberg, 1996; Alvarez et al., 1998; Delledonne et al., 2001). Cell death is a basic biological process that functions in many aspects of animal and plant development and in their responses to hormones and environmental stresses (Greenberg, 1996; Neill et al., 2002a). Although the mechanisms and regulation of plant programmed cell death remain to be defined, H₂O₂ has emerged as an important signal in the promotion of plant cell death (Bethke and Jones, 2001; Delledonne et al., 2001; Ren et al., 2002; Dat et al., 2003). It has been suggested that H₂O₂ induces cell death by initiating a signal transduction pathway rather than by direct killing as a result of phytotoxicity (Levine et al., 1994; Neill et al., 2002a). However, the cellular components involved in the plant response to H₂O₂ are not fully understood (Bowler and Fluhr, 2000; Breusegem et al., 2001; Mittler, 2002; Neill et al., 2002b).

Hydrolysis of membrane phospholipids occurs in plants under various stress conditions, including wounding, freezing, drought, salt, and pathogen elicitation (Chapman, 1998; Laxalt and Munnik, 2002; Wang, 2002). The activation of phospholipase D (PLD) has been thought to constitute an important and early step in stress-induced phospholipid hydrolysis (Ryu and Wang, 1996; Lee et al., 1997; Wang et al., 2000; Austin-Brown and Chapman, 2002; Meijer et al., 2002). PLD, which hydrolyzes phospholipids to phosphatidic acid (PA) and a free head group, forms a predominant family of phospholipases in plants. Arabidopsis has 12 PLD genes that are grouped into five classes (Qin and Wang, 2002). The various types of PLDs are regulated differently by Ca²⁺, polyphosphoinositides, and free fatty acids and display distinguishable substrate selectivity and specificity, suggesting that they are activated differently and may have unique cellular functions (Wang, 2002).

PLDδ is a recently identified PLD with several unique properties (Gardiner et al., 2001; Wang and Wang, 2001). It is activated by oleic acid (Wang and Wang, 2001) and associated with the plasma membrane and the microtubule cytoskeleton (Gardiner et al., 2001; Wang and Wang, 2001). PLDδ mRNA levels are higher in senescent than in young tissues (Wang and Wang, 2001), and the expression of PLDδ increases in response to severe dehydration and high salts (Katagiri et al., 2001). Increases of free unsaturated fatty acids often occur during plant senescence and under adverse conditions. These properties of activation, localization, and expression patterns raise the possibility that PLDδ may have an important function in the cellular response to stress. Here, we show that PLDδ and its derived PA play a critical role in mediating the plant response to H₂O₂ and plant stress tolerance.

RESULTS

Knockout of the PLDδ Gene Causes the Loss of Oleate-Activated PLD Activity in Arabidopsis

A PLDδ mutant was identified by screening T-DNA insertion lines of Arabidopsis. The T-DNA was inserted near the beginning of the 5′ coding region, 63 nucleotides downstream of the initiation codon (Figure 1A). Homozygous knockout plants had no detectable PLDδ protein (Figure 1B). As a control, the same protein extracts were blotted with a PLDα-specific antibody. Both the PLDδ mutants and the wild type showed similar levels of PLDα (Figure 1B). These data also indicate that the knockout of PLDδ did not alter the expression of PLDα.

Figure 1.

T-DNA Insertion in PLDδ Abolishes the Production of PLDδ and the Oleate-Stimulated PLD Activity in Arabidopsis.

(A) Gene structure of PLDδ and the site of the T-DNA insertion. Boxes represent exons.

(B) Immunoblot analysis of PLDδ (top gel) and PLDα (bottom gel) proteins in PLDδ mutant and wild-type (WT) plants. The symbols −/− and +/− represent plants that are homozygous and heterozygous, respectively, in the knockout locus.

(C) Oleate-stimulated PLDδ activity in PLDδ mutant and wild-type plants.

Both the activity and immunoblot assays used the microsomal fraction of leaf extracts.

Disruption of PLDδ gene function was confirmed further by assaying the oleate-activated PLD activity. Using plant PLDs expressed in Escherichia coli, it was shown previously that PLDδ, but not PLDα1, PLDβ1, or PLDγ1, was activated by oleate (Wang and Wang, 2001). The homozygous PLDδ-knockout plants exhibited no oleate-stimulated PLD activity (Figure 1C), demonstrating that only PLDδ possesses the oleate-activated PLD activity in Arabidopsis. The PLDδ mutant allele cosegregated with kanamycin resistance and susceptibility in a 3:1 ratio, suggesting that the PLDδ mutant contains a single T-DNA insertion in the genome. Introduction of a wild-type PLDδ gene to the knockout mutant plants restored the expression of PLDδ (Figure 2C).

Figure 2.

Increased H₂O₂-Induced Cell Death in PLDδ-Null Arabidopsis Protoplasts.

(A) Protoplast death as a function of H₂O₂ concentration. WT, wild type.

(B) H₂O₂ (1 mM)-induced cell death as a function of incubation time.

(C) Complementation of the PLDδ-null mutant with the wild-type PLDδ gene and restoration of the mutant response to H₂O₂ to wild type. The inset shows an immunoblot analysis of proteins extracted from wild-type, PLDδ-complemented (PLDδ-C), and PLDδ-null plants with PLDδ-specific antibody. The double asterisks indicate that the mean value is significantly different from that of wild-type protoplasts (P < 0.01).

Dead cells were counted 24 h after H₂O₂ treatments and are expressed as a percentage of total cells.

Ablation of PLDδ Increases the Cell's Sensitivity to H₂O₂-Induced Cell Death

Under normal growth conditions, PLDδ-knockout plants are indistinguishable from wild-type plants in terms of plant height, number of leaves, days required for flowering and seed maturation, and seed yield. In studying the involvement of PLDδ in cellular processes, we observed that ablation of PLDδ rendered Arabidopsis cells more sensitive to H₂O₂-induced cell death (Figure 2). When leaf protoplasts were exposed to 0.2 mM H₂O₂ for 24 h, 35% of PLDδ-null cells died, whereas the death rate was <5% for wild-type cells (Figure 2A). At 2 mM H₂O₂ for 24 h, all PLDδ-null cells were dead, whereas 60% of wild-type cells were still alive. The apparent H₂O₂ concentrations for 50% cell death (DC₅₀) for PLDδ-null and wild-type cells were 0.45 and 2.5 mM, respectively (Figure 2A). PLDδ-null cells died much faster than wild-type cells when the time course was determined using 1 mM H₂O₂. The apparent half-life for PLDδ-null and wild-type cells were 18 and 45 h, respectively (Figure 2B). In the absence of added H₂O₂, PLDδ-null and wild-type cells displayed no difference in the rate of cell death.

When the PLDδ-knockout mutant was complemented with the PLDδ gene with its own promoter, cells from the complemented plants showed reduced sensitivity to H₂O₂ and were indistinguishable from wild-type cells (Figure 2C). This result demonstrates that knockout of the PLDδ gene is responsible for the increased H₂O₂ sensitivity. The H₂O₂-induced cell death was programmed cell death, as indicated by the positive TUNEL (terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling) staining. After treatment with 1 mM H₂O₂ for 3 h, ∼85% of the nuclei of PLDδ-null protoplasts were labeled positive with TUNEL that labels free 3′-OH termini of broken DNA strands (Figure 3A). By contrast, under the same conditions, no clear labeling was found in wild-type cells.

Figure 3.

H₂O₂-Induced Programmed Cell Death and Increased Cell Death in PLDδ-Null Leaves.

(A) TUNEL staining of H₂O₂-induced programmed cell death in wild-type (WT) and PLDδ-null protoplasts. 3′-OH groups of DNA were visualized as bright green spots within the nucleus. Protoplasts were incubated with 1 mM H₂O₂ for 3 h before DNA was labeled using the TUNEL method.

(B) Trypan blue staining of cell death in wild-type and PLDδ-null Arabidopsis leaves. Leaves of 5-week-old plants were infiltrated with 50 mM H₂O₂ or water. The arrow indicates dead cells.

To determine whether the increased H₂O₂-induced cell death also occurred in PLDδ-null plants, leaves on plants were infiltrated with various concentrations (10 to 50 mM) of H₂O₂. No obvious death was noted in wild-type leaves. Plants are able to tolerate high concentrations of H₂O₂ because of their high H₂O₂-scavenging activity (Neill et al., 2002a). However, treatment of PLDδ-abrogated leaves with 50 mM H₂O₂ for 7 days triggered cell death in leaves, which was confirmed by positive trypan blue staining of PLDδ-null, but not wild-type, leaves (Figure 3B).

H₂O₂ Treatment Activates PLDδ

The involvement of PLDδ in H₂O₂-promoted cell death was studied further using protoplasts rather than whole plants because of the amenability of protoplasts for lipid labeling and quantitative analysis. The removal of cell walls also reduces the cell's capacity to scavenge H₂O₂, so protoplasts are more sensitive to H₂O₂ (Neill et al., 2002a). In addition, the protoplast system provides a synchronized cell system for measuring cellular response. Furthermore, freshly isolated protoplasts have proven to be physiological cell systems for studying the various plant signaling mechanisms (reviewed by Sheen, 2001), including H₂O₂ response (Kovtun et al., 2000) and PLD in abscisic acid action (Jacob et al., 1999).

To measure the change of PLD activity in response to H₂O₂ in vivo, protoplasts were prelabeled with fluorescent 1-oleoyl-2-(12-[(7-nitro-2-1,3-benzoxadiazol-4-yl)amino]dodecanoyl)-glycero-3-phosphocholine (NBD-PC) and then challenged with H₂O₂ in the presence of 1-butanol. The alcohol was used because PLD can transfer the phosphatidyl moiety to a primary alcohol at the expense of PA, and this activity is specific to PLD, thus providing a unique indicator for PLD activity (Wang, 2002). Immediately after prelabeling and before H₂O₂ stimulation, no difference was observed in the amounts of NBD-PC– and PLD-derived products—NBD-PA and NBD-phosphatidylbutanol (NBD-PtdBut)—between PLDδ-null and wild-type cells, indicating that the basal metabolism of NBD-PC is not altered by the gene knockout. However, significant increases in PtdBut occurred 30 and 60 min after H₂O₂ treatment in wild-type but not in PLDδ-null cells (Figure 4A). PA, expressed either as a percentage of total NBD lipids or relative NBD-PA (Figure 4B), also increased in H₂O₂-treated wild-type cells but not in PLDδ-null cells. The lack of increase in PA and PtdBut in H₂O₂-treated PLDδ-null cells indicates that, despite the presence of 11 other PLD genes in Arabidopsis, PLDδ is required for the PLD-mediated lipid hydrolysis in response to H₂O₂.

Figure 4.

H₂O₂-Induced Activation of PLD in Wild-Type but Not in PLDδ-Null Protoplasts.

(A) H₂O₂-induced PLD activity measured by the formation of NBD-PtdBut. PtdBut was expressed as percentage of PtdBut fluorescence over the total fluorescence of NBD-(PC+PtdBut+PA). WT, wild type.

(B) H₂O₂-induced PLD activity measured by the formation of PA. PA was expressed as fold increase over the amount of PA before H₂O₂ treatments.

In (A) and (B), protoplasts were prelabeled with NBD-PC and then challenged with 1 mM H₂O₂. Asterisks indicate that the mean value is significantly different from that of the protoplasts without H₂O₂ treatment at the same time at P < 0.05 (one asterisk) or P < 0.01 (two asterisks).

(C) Total PA in protoplasts as measured by mass spectrometry (n = 5).

(D) Immunoblot analysis of PLDδ after exposure of protoplasts to H₂O₂ for 1 and 3 h.

Measurements of total cellular PA in protoplasts revealed that the basal level of PA in PLDδ-null cells was ∼70% of that in wild-type cells (Figure 4C). This difference is similar to that in leaves, in which the amounts of PA in PLDδ-null and wild-type plants were 1.5 ± 0.4 and 2.0 ± 0.3 nmol/mg dry weight (n = 5), respectively. PA levels measured on the basis of milligrams of total protoplast protein (Figure 4C) were approximately twofold greater that those measured on the basis of milligrams of dry leaf weight, because proteins constitute ∼20% of dry leaf weight. The PA level alteration caused by the lack of PLDδ in both protoplasts and leaves indicates that PLDδ activity may contribute to the basal cellular level of PA and that the preparation of protoplasts does not increase PLDδ activity specifically. When cells were exposed to H₂O₂, the total cellular level of PA displayed a small transient increase in wild-type but not in PLDδ-null cells. The extent of the H₂O₂-induced total PA changes after 1 h of exposure to H₂O₂ was not as significant as that measured by fluorescence-labeled PA. This finding is not surprising considering the possibility that total PA comprises intracellular PA from both biosynthetic and hydrolytic pools, the latter of which includes PA derived from the hydrolysis of NBD-PC. Moreover, NBD-PC was localized primarily in the plasma membrane, where PLDδ is localized.

The stimulation of PLD activity was not related to changes in PLD proteins. The level of PLDδ protein remained unchanged for the first hour and decreased 3 h after H₂O₂ treatment (Figure 4D). Immunoblot analysis with a PLDδ-specific antibody detected PLDδ in wild-type but not in PLDδ-null mutant plants. These data suggest that the H₂O₂-induced formation of PLD reaction products results from the activation of preexisting PLDδ rather than from the synthesis of the enzyme. We then tested whether H₂O₂ itself might activate PLDδ directly. Coincubation of purified PLDδ (Wang and Wang, 2001) with various concentrations of H₂O₂ failed to stimulate the PLDδ activity (data not shown).

PLDδ-Derived PA Attenuates H₂O₂-Induced Cell Death

To examine whether the lipid product PA is responsible for decreasing H₂O₂-promoted cell death, 1-butanol was added together with H₂O₂ to suppress H₂O₂-induced PA production. The percentage of cells that died in the wild type increased nearly twofold with increased 1-butanol concentrations, and the optimal stimulation of cell death occurred at 0.1% 1-butanol (Figure 5A). 2-Butanol, which is not a substrate of PLD, exhibited no significant effect on cell death (Figure 5B). By contrast, PLDδ-null cells showed no significant increase in cell death when incubated with 1-butanol (Figure 5A). The lack of a 1-butanol effect was consistent with other results showing that no H₂O₂-induced activation of PLD occurred in the PLDδ-null cells (Figure 4). These results suggest that the PLDδ-mediated production of PA is involved in reducing H₂O₂-promoted cell death.

Figure 5.

Effects of 1- and 2-Butanol and PA on H₂O₂-Promoted Cell Death.

(A) 1-Butanol treatment increased cell death in wild-type (WT) protoplasts.

(B) Specificity of 1-butanol on H₂O₂-induced cell death. Protoplasts were treated with 0.1% 1- or 2-butanol (But) for 30 min before the addition of 1 mM H₂O₂.

Double asterisks in (A) and (B) indicate significant differences between 1-butanol–treated and untreated protoplasts at P < 0.01.

(C) PA decreased cell death in PLDδ-null protoplasts. PA was added 30 min before the addition of H₂O₂. Cell death was counted 24 h after H₂O₂ treatment. Asterisks indicate significant differences between PA-treated and untreated protoplasts at P < 0.05 (one asterisk) or P < 0.01 (two asterisks).

PA then was supplied to protoplasts to determine whether it might mitigate the H₂O₂ effect. Protoplasts have the advantage of easy incorporation of PA and other lipids into the cellular membranes, as shown with NBD-PC labeling (Figure 4). Incubation of protoplasts with dioleoyl-PA decreased cell death in PLDδ-null cells (Figure 5C), but incubation with dioleoyl-PC at the same concentration produced no effect on cell death in PLDδ-null or wild-type cells (data not shown). The greatest effect occurred at 50 μM, at which level PLDδ-null and wild-type cells were indistinguishable in their response to H₂O₂-promoted cell death (Figure 5C). However, PA had no mitigating effect on H₂O₂-treated wild-type cells. This differential response to the PA treatment between the two genotypes suggests that PA produced in vivo by wild-type cells (Figure 4) is sufficient and likely a better mediator than the added PA. Also, PA lost its mitigating effect on H₂O₂-induced cell death when the concentration was >100 μM (Figure 5C), suggesting that appropriate cellular concentrations of PA are critical to its function.

Knockout of PLDδ Diminishes the H₂O₂-Induced Activation of Mitogen-Activated Protein Kinase Activity

The mitogen-activated protein kinases (MAPKs) have been implicated in various cellular processes, including plant programmed cell death and response to H₂O₂ (Zhang and Klessig, 1997; Kovtun et al., 2000; Neill et al., 2002a). H₂O₂ has been shown to activate MAPK cascades in various tissues, although how such activation is achieved is unknown (Kovtun et al., 2000; Neill et al., 2002b). To determine whether the PLDδ function is linked to MAPK activation, an in-gel kinase assay was conducted to compare MAPK activity between PLDδ-null and wild-type protoplasts in response to H₂O₂ (Figure 6). One major MAPK band at ∼49 kD was observed (Figure 6), and the size was similar to those reported for plant MAPKs (Zhang and Klessig, 1997; Kovtun et al., 2000). PLDδ-null and wild-type cells had similar levels of basal MAP activity but differed in H₂O₂-induced kinase activity. In wild-type cells, the kinase activity increased at 0.2 and 1 mM H₂O₂. However, H₂O₂ failed to increase kinase activity in PLDδ-null cells (Figure 6).

Figure 6.

Ablation of H₂O₂-Promoted MAPK Activity in PLDδ-Null Arabidopsis.

Autoradiogram from an in-gel kinase assay using protein extracts from wild-type (WT) and PLDδ-null Arabidopsis protoplasts. Myelin basic proteins were embedded on a 10% polyacrylamide gel as the substrate. After electrophoresis, proteins were renatured on the gel and assayed for phosphorylation activity using γ-³²P-ATP.

Knockout of PLDδ Increases Plant Sensitivity to Stress Damage

H₂O₂ is known to accumulate under various stress conditions (Prasad et al., 1994; Sharma et al., 1996; Bowler and Fluhr, 2000; Breusegem et al., 2001; Mittler, 2002; Neill et al., 2002b). We tested how the increased sensitivity to H₂O₂ would affect the plant's ability to produce H₂O₂ and cope with stresses (Figure 7). PLDδ-null and wild-type plants were challenged with UV-B irradiation. This treatment induced H₂O₂ production, as monitored by both 3,3′-diaminobenzidine staining and H₂O₂ measurement (Figures 7A and 7B). The levels of accumulation for H₂O₂ were indistinguishable between PLDδ-null and wild-type plants, indicating that the loss of PLDδ did not compromise the plant's ability to produce H₂O₂ under stress. However, PLDδ-null plants exhibited less tolerance than wild-type plants to UV-B irradiation, and the number of PLDδ-null plants surviving after these treatments was less than half the number of wild-type and PLDδ-complemented plants (Figures 7C and 7D).

Figure 7.

UV-B Light–Induced H₂O₂ Production and Seedling Death in Wild-Type, PLDδ-Null, and PLDδ-Complemented Arabidopsis.

(A) UV-B light–induced H₂O₂ production as measured by 3,3′-diaminobenzidine staining of 10-day-old seedlings. PLDδ-C, PLDδ-complemented; WT, wild type.

(B) UV-B light–induced H₂O₂ production as measured by the Amplex red hydrogen peroxide/peroxidase assay kit. FW, fresh weight.

(C) Increased death of Arabidopsis seedlings after exposure to UV-B light.

(D) Quantitation of UV-B light–induced seedling death. Double asterisks indicate that the mean value is significantly different from that of wild-type plants at P < 0.01.

DISCUSSION

Phospholipid hydrolysis increases under various stress conditions (Chapman, 1998; Laxalt and Munnik, 2002; Wang, 2002). H₂O₂ is a reactive oxygen species that increases under different stresses and has multifaceted functions in plant stress responses (Prasad et al., 1994; Sharma et al., 1996; Bowler and Fluhr, 2000; Breusegem et al., 2001; Mittler, 2002; Neill et al., 2002b; Dat et al., 2003; Kwak et al., 2003; Yoshioka et al., 2003). However, little is known about the cellular function of lipid hydrolysis and the cellular components that mediate the plant response to H₂O₂. Recently, PLDζ1 was found to be involved in both the initiation and maintenance of root hair morphogenesis (Ohashi et al., 2003); on the other hand, NADPH oxidase and reactive oxygen species also were required for root hair development (Foreman et al., 2003). However, the connection between phospholipid signaling and H₂O₂ function has not been elucidated.

The results of this study demonstrate that the PLDδ-mediated hydrolysis of phospholipids plays a positive role in the plant response to H₂O₂, thus providing a link between membrane phospholipid hydrolysis and the signaling of the plant response to oxidative stress. The activation of PLDδ and its derived PA function to decrease H₂O₂-induced programmed cell death and increase plant stress tolerance. The anti-cell death effect of PLDδ has been documented in freshly isolated protoplasts (Figures 2 and 3A) and in leaves on plants (Figure 3B) and also is reflected in stress tolerance in Arabidopsis seedlings (Figure 7). It should be noted that although most of these studies used 1 mM H₂O₂, the pro-cell death effect of H₂O₂ begins at a lower concentration, because the DC₅₀ for applied H₂O₂ in the PLDδ-null cell is 0.45 mM (Figure 2A). In addition, the effective DC₅₀ for H₂O₂ in the cell should be lower than that because cells have a high capacity to metabolize H₂O₂ by several scavenging enzymes (Neill et al., 2002a). These considerations aside, the DC₅₀ for H₂O₂ is still within the physiological range reported for Arabidopsis cells, in which the steady state H₂O₂ levels varied from 60 μM to 7 mM (Karpinski et al., 1999; Veljovic-Jovanovic et al., 2001).

Comparative measurements of the PLD-specific product PtdBut in wild-type and PLDδ-null cells demonstrate that PLDδ is activated specifically in response to H₂O₂, despite the presence of 11 other PLD genes in Arabidopsis. Although the exact mechanism for the H₂O₂-specific activation of PLDδ remains to be elucidated, our results and those of other researchers have provided important insights into the unique activation and function of PLDδ. Immunoblot analysis suggests that the H₂O₂ activation of PLDδ likely results from a modulation of preexisting PLDδ (Figure 4D). Our testing of purified PLDδ with H₂O₂ suggests that H₂O₂ may not interact with and activate PLDδ directly. One possible effector for the H₂O₂-induced activation of PLDδ might be the increase in cytosolic Ca²⁺. H₂O₂ and various stresses have been shown to activate the Ca²⁺ channel, stimulating a transient increase in the cytosolic Ca²⁺ level (Figure 8) (Pei et al., 2000; Murata et al., 2001; Zhang et al., 2001). Ca²⁺ has been shown to stimulate PLDδ activity by increasing its affinity for its lipid substrate (Qin et al., 2002), and it binds to the Ca²⁺/phospholipid binding C2 domain that resides near the N terminus of PLDδ (Zheng et al., 2000; Wang, 2002) (Figure 8). On the other hand, several other PLDs also contain the C2 domain (Qin and Wang, 2002); thus, the specificity of PLDδ activation also should require other properties that are unique to the enzyme.

Figure 8.

Proposed Model Depicting the Activation and Function of PLDδ and PA in the Cell's Response to H₂O_2.

Stress stimulates the cellular production of H₂O₂ that activates PLDδ associated with the plasma membrane and the microtubule cytoskeleton. Potential mediators in the activation include an increase in cytosolic Ca²⁺ and oleic acid in the membrane. Such interactions increase the affinity of PLDδ to its substrates, stimulating lipid hydrolysis and PA production. PA may bind to target proteins, such as Raf-like MAPK kinase kinase (MAPKK), that contain a PA binding motif (PABM). Such binding may lead to the activation of MAPK cascades, and PA also may be involved directly in the regulation of ANP. In addition, PA may function by modulating membrane trafficking and remodeling. These interactions modulate the cell's ability to respond to oxidative stress and decrease cell death. Dashed lines denote undocumented, hypothetical interactions.

One such unique property is the stimulation of PLDδ by oleic acid (Wang and Wang, 2001). This study has established unequivocally that PLDδ is solely responsible for the oleate-stimulated PLD activity (Figure 1). Such oleate-stimulated PLD activity also has been identified in mammalian cells, but the gene for this activity remains elusive (Liscovitch et al., 2000). Oleate has been shown to reduce apoptosis in animal cells (Hardy et al., 2000). In Arabidopsis, deficiency in oleate synthesis results in an increase in cell death (Kachroo et al., 2001). This raises the intriguing possibility that PLDδ may serve as a direct target of oleate (Figure 8), which might underlie in part oleate's anti-cell death activity. In addition, PLDδ is distinctively different from other Arabidopsis PLDs in its intracellular localization and patterns of expression. Unlike the other PLDs examined, PLDδ is associated with the plasma membrane, binds to tubulin, and is expressed highly in stressed and senescent tissues (Figure 8) (Gardiner et al., 2001; Katagiri et al., 2001; Wang and Wang, 2001). These distinct properties provide the cellular and biochemical bases for the specific activation and unique function of PLDδ in the plant response to H₂O₂. For example, its membrane association indicates that PLDδ-derived PA will be localized on the plasma membrane. This site of PA production may explain why the process can be mimicked by exogenous PA, because it can be incorporated easily into the plasma membrane.

Several lines of evidence indicate that it is PLDδ-derived PA that mediates the anti-cell death effect promoted by H₂O₂. First, wild-type cells, which produced a higher level of PA than PLDδ-null mutant cells in response to H₂O₂, were more tolerant to H₂O₂-induced cell death than were mutant cells (Figure 4). In addition, the estimated rate of oleate-stimulated PLDδ activity in vitro (4 nmol PA·mg⁻¹ microsomal protein·min⁻¹; Figure 1C) is sufficient to account for the actual amount of PA in vivo (∼1 nmol PA·mg⁻¹ total protoplast protein·min⁻¹; Figure 4C). Second, 1-butanol, which decreases PA production by competing with water to be the hydroxyl donor, resulted in more cell death in the presence of H₂O₂ (Figures 5A and 5B). Third, the addition of PA improved cell survival in PLDδ-null cells that failed to produce PA in response to H₂O₂ (Figure 5C). This information is important for determining how the activation of a PLD mediates cellular function, because in addition to PA, the activation of PLD also generates water-soluble head groups and results in a change in membrane lipid composition attributable to the hydrolysis of common membrane lipids (Welti et al., 2002).

PA has been shown to regulate various cellular functions in multiple ways (Ghosh et al., 1996; Liscovitch et al., 2000; Rizzo et al., 2000; Fang et al., 2001; Jones and Huanun, 2002; Wang, 2002). It may serve as a direct activator or inhibitor of enzymes involved in signal transduction and/or function as a membrane anchor for assembling signaling complexes. The direct targets of PA identified in animal cells include protein kinases (Ghosh et al., 1996; Fang et al., 2001) and protein phosphatases (Jones and Huanun, 2002). The molecular target of PA has not been identified in plants (Munnik, 2001; Wang, 2002). Pharmacological data suggest that PA is an upstream regulator of MAPK in plant-wounding signal transduction (Lee et al., 2001). The present study showed that ablation of PLDδ suppressed the H₂O₂ induction of MAPK activity (Figure 6). This finding raises the possibility that the activation of PLDδ and its derived PA are required for the H₂O₂ stimulation of the MAPK activity that is regulated by upstream kinases. In animal cells, PA has been shown to recruit Raf1, a MAPK kinase kinase, to receptors and to help assemble signaling complexes on the plasma membrane (Rizzo et al., 2000). We have observed that PA binds to an Arabidopsis Raf-like MAPK kinase kinase and that the binding occurs through a PA binding motif that is ∼50% identical to the amino acid sequence of the animal Raf1 (Y. Sang and X. Wang, unpublished data). In plants, the non-Raf MAPK kinase kinase, ANP, has been shown to be involved in the plant response to H₂O₂, and this kinase is important for increasing plant stress tolerance (Kovtun et al., 2000). ANP does not contain the PA binding motif found in Raf-like MAPK kinase kinase. However, it is possible that PLDδ/PA may regulate ANP directly through some other mechanism of interaction and/or through the modulation of Raf-like kinases (Figure 8).

In addition, PLDδ/PA also may protect cells through their functions in membrane biogenesis, vesicular trafficking, and cytoskeletal rearrangement (Figure 8). PLDδ has been shown to be associated with the microtubule cytoskeleton (Gardiner et al., 2001). A recent study showed that Arabidopsis PLDβ interacts with actin (Kusner et al., 2003). PA also is a known stimulator of phosphatidylinositol-4-phosphate kinase that synthesizes phosphoinositol-4,5-bisphosphate, an important regulator of cytoskeletal movement and membrane trafficking (Liscovitch et al., 2000). Furthermore, PA is fusogenic and also a substrate for glycerolipid synthesis. Whether and how PLDδ and PA affect these cellular processes require further investigations. The present finding that PLDδ is required for an active cellular response to H₂O₂, decreasing cell death and increasing stress tolerance, indicates that such investigation will be important to revealing the networks of signaling cascades in the plant response to reactive oxygen species and stresses.

METHODS

Isolation of a PLDδ-Knockout Mutant and Genetic Complementation

A PLDδ-knockout mutant was isolated from the Wassilewskija ecotype of Arabidopsis thaliana. The mutant was identified by screening T-DNA insertion lines according to the protocol of the Arabidopsis Gene Knockout Research Facility at the University of Wisconsin (Sussman et al., 2000). The loss of PLDδ function was confirmed by the absence of PLDδ's protein, enzymatic activity (Figure 1), and transcript (data not shown).

To make up for the loss of PLDδ, the PLDδ gene with its own 1.5-kb 5′ untranslated region was cloned from wild-type plants by PCR. The gene with its own promoter was ligated to the agrobacterial binary vector pBin19 and introduced into homozygous PLDδ-knockout plants via Agrobacterium tumefaciens–mediated transformation using floral dipping. The T-DNAs used in PLDδ-knockout and pBin19 plants carried the selection marker for kanamycin and hygromycin resistance, respectively. Thus, the PLDδ-complemented plants were selected by their resistance to both antibiotics. The complementation plants were confirmed by PCR for the presence of the original T-DNA insertion allele and the introduced wild-type PLDδ gene as well as by detection for the production of the PLDδ protein (Figure 2C).

Plant Growth and Protoplast Isolation

Seeds of PLDδ-null mutant and wild-type plants were sown in soil and treated at 4°C for 2 days. Plants were grown in a growth chamber under 14-h-day/10-h-night and 23/18°C cycles. The expanded leaves from 4- to 5-week-old plants were used to isolate protoplasts based on a protocol described previously (Kovtun et al., 2000).

Cell Death Assays

Freshly isolated protoplasts (5 × 10⁵) were incubated in the dark with various concentrations of H₂O₂ or other reagents as indicated in Figures 4 and 5. Cells were stained with 0.5 mg/mL fluorescein diacetate for 15 min, and total and live cells were counted with a fluorescence microscope. When the effect of added phosphatidic acid (PA) on cell death was examined, dioleoyl-PA and other phospholipids were emulsified in the incubation solution by sonication before use. Suspended lipids were added to protoplasts, and the lipid-protoplast mixture was incubated with gentle agitation at 22°C for 30 min before H₂O₂ treatment. Incorporation of phospholipids into protoplast membranes was verified using 1-oleoyl-2-NBD (12[(7nitro-2-1,3-benzoxadiazol-4-yl)amino]dodecanoyl)-PA and NBD-phosphatidylcholine (NBD-PC). The presence of the fluorescent PA or PC in membranes was detected by fluorescence microscopy and analyzed as described below.

Programmed cell death was determined by the TUNEL (terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling) assay using an in situ cell death kit (Roche Molecular Biochemicals, Mannheim, Germany). Protoplasts were collected, and the free 3′-OH groups in single- and double-stranded DNA were labeled according to the manufacturer's instructions. Labeled DNA showing bright green fluorescence was visualized by fluorescence microscopy. To test H₂O₂'s effects on cell death in Arabidopsis plants, the first true leaves from 5-week-old plants were infiltrated using a needleless syringe with either 10 to 50 mM H₂O₂ or water. Leaves were excised from plants at 1, 2, and 7 days after infiltration and stained with a lactophenol–trypan blue solution (Rate et al., 1999).

Protein Extraction and Immunoblot Analysis of PLDs

Total protein from wild-type, PLDδ-null, and PLDδ-complemented leaves was isolated as described previously (Wang and Wang, 2001). Protein contents were determined using a dye binding assay. Equal amounts of protein (20 μg/lane) were separated by SDS-PAGE analysis and then transferred onto polyvinylidene difluoride filters. The filters were blotted with PLDδ- or PLDα-specific antibodies followed by incubation with a second antibody conjugated to alkaline phosphatase. The PLD proteins were made visible by staining the blot showing phosphatase activity (Wang and Wang, 2001).

Phospholipid Labeling and Assays of Oleate-Stimulated PLD Activities

NBD-PC in chloroform was dried under a stream of nitrogen and suspended in the incubation solution by sonication before use. Protoplasts (5 × 10⁶/mL) were incubated with 0.5 mg/mL NBD-PC for 80 min on ice and then transferred to room temperature (22°C) for 10 min (Jacob et al., 1999). In vivo PLD activity was determined according to the production of phosphatidylbutanol (PtdBut). 1-Butanol (0.1%, v/v) was added with 1 mM H₂O₂ (or water) to NBD-PC–labeled protoplasts (∼3 × 10⁵) and incubated in a glass tube at 22°C for the indicated time (Figure 4). Hot isopropanol (75°C) was added and then incubated for 10 min at 75°C to inactivate PLD. Lipids were extracted and separated on thin layer chromatography plates (silica G) with chloroform:methanol:NH₄OH (65:35:5) as described previously (Ryu and Wang, 1996). NBD-PC, NBD-PA, and NBD-PtdBut were well separated and visualized under UV illumination. The spots were scraped, extracted with chloroform:methanol:water (5:5:1), and quantitated with a fluorescence spectrophotometer at 460 nm (excitation) and 534 nm (emission). In vitro PLDδ activity was measured in the presence of oleate as described previously (Wang and Wang, 2001) using proteins extracted from Arabidopsis leaves.

Total PA Measurement

Total lipids were extracted from Arabidopsis leaves and leaf protoplasts and then analyzed using electron spray ionization tandem mass spectrometry (ESI-MS/MS). Hot isopropanol (75°C) was added to protoplasts after the treatments to inactivate PLD activity. The processes of lipid extraction, ESI-MS/MS analysis, and quantification followed the procedure detailed previously (Welti et al., 2002). Total proteins from equivalent amounts of protoplasts were measured using a dye binding assay.

In-Gel Kinase Activity Assay

Protein was extracted from protoplasts with a buffer (100 mM HEPES, pH 7.5, 5 mM EDTA, 5 mM EGTA, 10 mM DTT, 10 mM Na₃VO₄, 10 mM NaF, 1 mM phenylmethylsulfonyl fluoride, 5 μg/mL antipain, 10% glycerol, 7.5% polyvinylpolypyrrolidone) using the method of Zhang and Klessig (1997) with minor modifications. Protoplasts were sonicated twice for 15 s each in a 1.5-mL microcentrifuge tube followed by centrifugation at 18,000g for 20 min. Proteins (16 μg/lane) in the supernatant were subjected to electrophoresis on 10% SDS-polyacrylamide gels embedded with 0.25 mg/mL myelin basic protein as the substrate. Proteins were renatured on the gel and assayed for kinase activity. The reaction buffer for kinase activity contained 25 mM Tris, pH 7.5, 2 mM EGTA, 12 mM MgCl₂, 1 mM DTT, 0.1 mM Na₃VO₄, 200 nM ATP, and 50 μCi of γ-³²P-ATP (300 Ci/mmol). The gel was dried and exposed to x-ray film.

Stress Treatments

Seeds of wild-type, PLDδ-null, and PLDδ-complemented plants were surface-sterilized with 70% ethanol and 1.2% sodium hypochlorite and planted on a plate containing half-strength Murashige and Skoog (1962) mineral salts, pH 5.7, with 0.8% agar. Each plate contained 50 seeds of each of the three genotypes. Plates were kept at 4°C for 2 days, and then seedlings were grown under 36 μmol·m⁻²·s⁻¹ light/dark (12/12 h) at 23°C for 2 weeks. These 2-week-old seedlings were illuminated with 3 μmol·m⁻²·s⁻¹ UV-B light at 23°C for 3 h. Some of the UV-B light–treated seedlings were used to measure H₂O₂ production immediately after irradiation, and others were grown for 7 days before they were scored for survival ratio.

H₂O₂ Measurements

Production of H₂O₂ in plants was measured by staining plants with 3,3′-diaminobenzidine and by assaying H₂O₂ concentrations in the leaves. Briefly, 3 h after UV illumination, whole plants were submerged into a 2.8-mM 3,3′-diaminobenzidine solution, pH 5.5, for 2 h and then boiled in 96% ethanol for 10 min. For quantification, H₂O₂ was extracted from leaves according to the method described previously (Rao et al., 2000). Extract was diluted 10-fold and then measured for H₂O₂ concentration with an Amplex red hydrogen peroxide/peroxidase assay kit (Molecular Probes, Eugene, OR).

Upon request, materials integral to the findings presented in this publication will be made available in a timely manner to all investigators on similar terms for noncommercial research purposes. To obtain materials, please contact Xuemin Wang, wangs@ksu.edu.