Lipid Peroxidation during the Hypersensitive Response in Potato in the Absence of 9-Lipoxygenases*

Hypersensitive cell death is an important defense reaction of plants to pathogen infection and is accompanied by lipid peroxidation processes. These may occur non-enzymatically by the action of reactive oxygen species or may be catalyzed by enzymes such as α-dioxygenases, lipoxygenases, or peroxidases. Correlative data showing increases in 9-lipoxygenase products in hyper-sensitively reacting cells have so far suggested that a large part of lipid peroxidation is mediated by a specific set of 9-lipoxygenases. To address the significance of 9-lipoxygenases for this type of pathogen response in potato, RNA interference constructs of a specific pathogen-induced potato 9-lipoxygenase were transferred to potato plants. Significantly reduced 9-lipoxygenase transcript levels were observed in transgenic plants after pathogen treatment. In addition, 9-lipoxygenase activity was hardly detectable, and levels of 9-lipoxygenase-derived oxylipins were reduced up to 12-fold after pathogen infection. In contrast to wild type plants, high levels of non-enzymatically as well as 13-lipoxygenase-derived oxylipins were present in 9-lipoxygenase-deficient plants. From this we conclude that during the normal hypersensitive response in potato, lipid peroxidation may occur as a controlled and directed process that is facilitated by the action of a specific 9-lipoxygenase. If 9-lipoxygenase-mediated formation of hydroperoxides is repressed, autoxidative lipid peroxidation processes and 13-lipoxygenase-mediated oxylipins synthesis become prominent. The unaltered timing and extent of necrosis formation suggests that the origin of lipid hydroperoxides does not influence pathogen-induced cell death in potato.

Lipid peroxidation in plants is an important feature of the hypersensitive cell death, a typical defense reaction displayed during incompatible or non-host interactions of plants with pathogens (1). Membrane damage by peroxidation of polyunsaturated fatty acids can be initiated by reactive oxygen species (ROS), 1 lipid radicals, or enzymatically by the action of lipoxy-genases (LOXs) (2). In plants, LOXs introduce molecular oxygen into linoleic (LA) and linolenic acid (LnA) at either the C-9 or the C-13, leading to the formation of the corresponding hydroperoxides. Apart from free fatty acid derivatives, LOXs can also catalyze the oxygenation of membrane lipids (3,4), leading to altered fluidity and permeability of biomembranes.
In plants, correlative data suggest that 9-LOXs are crucial for lipid peroxidation during the hypersensitive response. Thus, 9-LOX-derived oxylipins accumulate in tobacco leaves undergoing a hypersensitive reaction in response to treatment with the elicitor cryptogein (5) and in cotton leaves infected with Xanthomonas campestris pv. malvacearum (6).
In addition to their proposed role in lipid peroxidation, LOXs may also contribute to defense responses in plants by synthesizing antimicrobial compounds and signal molecules. The products of the LOX reaction in plants, 9-and 13-hydroperoxylinole(n)ic acid (9-and 13-HPOD/HPOT), are substrates for at least seven LOX pathway enzymes (7) that catalyze the synthesis of hydroxyoctadecadienoic acid (HOD) or hydroxyoctadecatrienoic acid (HOT), divinyl ether containing fatty acids such as colnele(n)ic and etherole(n)ic acid, 12-oxo-phytodienoic acid (OPDA), and jasmonic acid (JA), -oxo fatty acids, aldehydes, and trihydroxy fatty acids.
For several of these oxylipins, antimicrobial activity has been demonstrated (8 -11), suggesting that LOXs function in plant defense by contributing to pathogen growth inhibition. On the other hand, oxylipins of both the 9-and the 13-LOX pathway have been identified as signal molecules that activate defense responses. JA, its biosynthetic precursor OPDA, and their derivatives are well studied signaling compounds in the response of the plant to wounding (12,13), whereas 13-HOT is able to induce PR gene expression in barley (14). Moreover, the products of the LOX reaction, 9-and 13-HPOD, as well as the corresponding hydroxides, 9-and 13-HOD, induce cell death in protoplasts (15,16).
In solanaceous plants, oxylipins derived from the 9-LOX pathway are of importance for pathogen defense. Thus, pathogen-induced 9-LOX transcript accumulation was reported in a number of plants, for example in tobacco after infection with Phytophthora parasitica var. nicotianae (17) and in potato infected by Phytophthora infestans (18 -20). The pathogen-induced accumulation of the 9-LOX products in potato cells (21) and plants (11,20) suggests a role of these compounds for defense responses. More importantly, a functional analysis of the elicitor-induced 9-LOX of tobacco revealed its crucial role for the establishment of resistance (17). However, how 9-LOXs exert their effect in the response to pathogens and, in particular, which of the 9-LOX-derived products are of importance have not been clarified yet.
Here we set out to functionally analyze the role of 9-LOXs for the response of potato to pathogen attack. After identification of the pathogen-induced 9-LOX isoform in our system, RNA interference (RNA i ) was used to down-regulate its expression. Transgenic plants have a significant reduction in pathogeninduced 9-LOX activity, and the accumulation of enzymatically produced oxylipins was strongly reduced, but lipid peroxides of autoxidative origin increased. Cell death, however, is not significantly affected suggesting that potato plants can undergo hypersensitive cell death irrespective of the origin of lipid hydroperoxides.

EXPERIMENTAL PROCEDURES
RNA Analyses and Generation of 9-LOX cDNA Fragments by RT-PCR-RNA was isolated from pathogen-infected potato plants and subjected to Northern analyses as described (20). Reverse transcription was carried out using DNase-treated RNA with SuperScript TM II RNase H Ϫ Reverse Transcriptase (Invitrogen). Subsequent PCR was performed using the primers 5Ј-ACAAACAGGCAACTAAGTGTGCT-TCACC-3Ј and 5Ј-GCGATTTGGGAGGTAACCTGCATAAGG-3Ј under standard conditions.
Cloning of RNA i Constructs-A ␤-glucuronidase (GUS) gene cloned in a pUC19 vector was partially deleted at the EcoRV site by cutting the plasmid with EcoRV and SmaI and subsequent re-ligation. The 374-bp FIG. 1. POTLX-3 is the predominant LOX transcript in P. infestans-infected potato leaves. RNA isolated from P. infestans-infected potato leaves at the time points indicated was subjected to RT-PCR using 9-LOX-specific primers. The PCR products were separated on agarose gels either directly (Ϫ) or after restriction digestion with BglII (B) or ClaI (C). dpi, days post-infection. POTLX-3-specific fragment was amplified from RNA of pathogen-infected plants via RT-PCR using the primers 5Ј-TATGTTGCAGTGAAT-GACTCG-3Ј and 5Ј-GGATAGTCTTGAATTAGCAAG-3Ј with sequences for SstI restriction sites at each end. After cloning of the RT-PCR product into the vector pCR2.1 (Invitrogen), the fragment was inserted into an SstI site at the 3Ј-end of the truncated GUS gene upstream of the nopaline synthase polyadenylation signal. The 374-bp POTLX-3-specific fragment was then amplified from the pCR2.1 clone using primers with BamHI and PstI-KpnI sites at either end. This fragment was inserted upstream of the truncated GUS gene via BamHI/PstI in inverse orientation to the SstI fragment. The RNA i construct was inserted into a binary vector carrying the 35STX promoter (22) as a KpnI-EcoRI fragment, and the resulting plasmid was transferred to Agrobacterium tumefaciens GV3101. the duration of the experiment. As controls, water was pipetted onto the leaves.
Cell death was determined by staining whole leaves with trypan blue as described (24). Hydrogen peroxide formation was visualized by staining with diaminobenzidine as described (25).
Determination of LOX Activity and LOX Products-LOX activity was determined using crude protein extracts from plants as described (20). Determination of hydroxy and hydroperoxy fatty acids as well as divinyl ethers was performed by HPLC-based and GC-based analyses. About 0.5 g of frozen leaf tissue was added to 10 ml of extraction medium (isohexane/isopropyl alcohol, 3:2 (v/v) with 0.0025% (w/v) butylated hydroxytoluene) and immediately homogenized with an Ultra Turrax under a stream of argon on ice for 30 s. As an internal standard, (6Z,9Z,11E,13S)-13-hydroxy-6,9,11-octadecatrienoic acid and heptadecanoic acid were added. The extract was centrifuged at 4,500 ϫ g at 4°C for 10 min. The clear upper phase was collected, and the pellet was extracted three times with 3 ml each of extraction medium. To the combined organic phases, a 6.7% (w/v) solution of potassium sulfate was added to a volume of 47 ml. After vigorous shaking, the upper hexanerich layer was removed. The upper organic phase containing the oxylipins and fatty acids was dried under nitrogen and redissolved in 200 l of methanol.
For the analysis of fatty acids, 10 l of an EDAC solution (1 mg of 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide/10 l of methanol) and 370 l of methanol were added to 20 l of the methanol solution and incubated for 2 h. After adding 200 l of 0.1 M Tris-HCl, pH 7.5, the fatty acid methyl esters (FAMEs) were extracted twice each with 1 ml of hexane. The combined organic phases were evaporated to dryness under nitrogen. The FAMEs were redissolved in 10 l of acetonitrile and analyzed using a gas chromatograph/flame ionization detector (GC/FID).
Analysis of oxylipins was carried out by HPLC on an Agilent 1100 HPLC system coupled to a diode array detector. Prior to analysis, the solvent was removed, and the sample was reconstituted in 50 l of methanol/water/acetic acid (85:15:0.1, v/v). At first, oxylipins were purified on reversed phase-HPLC. This was carried out on a ET250/2 Nucleosil 120 -5 C18 column (Macherey-Nagel, 2.1 ϫ 250 mm, 5-m particle size) with a solvent system of methanol/water/acetic acid (85: 15:0.1, v/v), and a flow rate of 0.18 ml min Ϫ1 . Straight phase-HPLC of the hydro(pero)xy derivatives of linoleic and linolenic acid was carried out on a Zorbax Rx-SIL column (Agilent, 2.1 ϫ 150 mm, 5-m particle size) with a solvent system of n-hexane/2-propanol/acetic acid (100:1: 0.1, v/v) and a flow rate of 0.2 ml min Ϫ1 . The absorbance at 234 nm was recorded indicating the conjugated double bond. Chiral phase-HPLC of the hydro(pero)xy fatty acids was carried out on a Chiralcel OD-H column (Daicel, 2.1 ϫ 150 mm, 5-m particle size) with a solvent system of n-hexane/2-propanol/acetic acid (100:5:0.1, v/v) and a flow rate of 0.1 ml min Ϫ1 . Absorbance at 234 nm was monitored. Reversed phase-HPLC of the divinyl ethers was carried out on a ET250/2 Nucleosil 120 -5 C18 column (Macherey-Nagel, 2.1 ϫ 250 mm, 5-m particle size) with a solvent system of methanol/water/acetic acid (90:10:0.1, v/v) and a flow rate of 0.18 ml min Ϫ1 . The absorbance at 250 nm was recorded indicating a conjugated diene system in conjugation with an ether bond of the fatty acids.
For the analysis of esterified oxylipins and fatty acids, triricinoleate and triheptadecanoate were used as internal standards, and the extraction was performed as described above. After removing of the solvent, 333 l of a mixture of toluene and methanol (1:1, v/v) and 167 l of 0.5 mM sodium methoxide were added. After incubation for 20 min, 0.5 ml of 1 M NaCl and 50 l of HCl (37%, v/v) were added, and the FAMEs and oxidized derivatives were extracted twice each with 0.75 ml of hexane. The combined organic phases were evaporated to dryness under nitrogen and dissolved in 200 l of methanol. 20 l of this solution was dried under nitrogen and dissolved in 10 l of acetonitrile for the analysis of the FAMEs by GC/FID-based analysis. The methyl esters of the oxylipins were determined by HPLC according to the analysis of the free oxylipins.
For determination of JA and OPDA, a GC/mass spectrometry-based analysis was used as described (20).

Identification of POTLX3 as the Pathogen-induced 9-LOX in
Potato-To reduce the pathogen-induced expression of 9-LOXs in potato by RNA i , we first determined which of the 9-LOX genes are specifically expressed in response to pathogen infection. Phylogenetic analysis of the 9-LOX family in solanaceous plants revealed that 9-LOXs can be divided into two subfami-lies, the group of classical 9-LOX cDNAs first described in potato tubers and another group containing the pathogen-induced 9-LOXs (7). By using cDNAs of both groups as probes, pathogen-induced accumulation of 9-LOX transcripts in potato has been reported (18 -20). However, based on the high sequence similarity of these cDNAs, cross-hybridization might conceal which of these genes are specifically expressed after pathogen attack. Therefore, RNA from plants infected with P. infestans was reverse-transcribed and subjected to PCR using primers designed to amplify a 582-bp fragment from all 9-LOX genes reported in the data base. The mixture of PCR products was digested with either BglII or ClaI, enzymes that cut specifically in fragments derived from POTLX-3 mRNA (19) or from tuber-specific 9-LOX mRNA (26,27), respectively. Subsequently, the restriction fragments were separated on agarose gels (Fig. 1). Starting 1 day after infection with P. infestans, cDNA fragments digestable by BglII but not by ClaI were present, indicating that the majority of 9-LOX-derived cDNA fragments were derived from POTLX-3 transcripts. The portion of cDNA fragments that was not digestable by BglII decreased and was not detectable after 5 days. Thus, POTLX-3 transcripts represent the majority of pathogen-inducible 9-LOX transcripts. Moreover, our results show that transcripts encoding the tuber-specific 9-LOXs, whose cDNA fragments should have been cut by ClaI, cannot be detected in either untreated or pathogen-infected leaves by this method. Interestingly, in cDNA derived from leaves of uninfected plants, neither BglII nor ClaI was able to cut the mixture of PCR fragments, indicating that mRNA encoding a so far unidentified 9-LOX isoform had been amplified in the RT-PCR. This LOX cDNA fragment (StLOX1-4 was subcloned and sequenced (data not shown). Highest sequence similarity was found to a toxin-induced tomato LOX (97%, GenBank TM accession number AY008278) and to the elicitor-induced 9-LOX from tobacco (90%, 28). StLOX1-4 shared less than 85% homology at the nucleotide level with LOX isoforms from tubers. Generation of RNA Interference Constructs and Plant Transformation-The RNA i technique was chosen to decrease the level of pathogen-induced 9-LOX transcripts in potato. RNA i constructs were generated using a 374-bp fragment of POTLX-3 that was cloned in antisense orientation in front of and in sense orientation behind a truncated GUS gene. The fragment was designed to be as specific for POTLX-3 as possible, and it had 87% homology at the nucleotide level to other 9-LOX genes. This 9-LOX-RNA i construct was transferred to a binary vector behind the 35STX promoter. Potato plants were transformed with the 9-LOX-RNA i construct as well as with the empty vector. Regenerated plants were first analyzed for the presence of the transgene by subjecting their DNA to Southern analyses (data not shown). Eleven plants containing the 9-LOX-RNA i construct were chosen for further analyses.
Loss of Pathogen-induced 9-LOX Transcript Accumulation and 9-LOX Activity in Transgenic RNA i Plants-To analyze whether expression of the RNA i construct efficiently interfered with 9-LOX transcript accumulation, potato plants were infiltrated with P. syringae pv. maculicola or infected with P. infestans. RNA analyses revealed that in none of the transgenic plants were 9-LOX transcripts detectable after bacterial infiltration ( Fig. 2A). In contrast, in untransformed potato plants as well as empty vector plants, significant levels of 9-LOX transcripts accumulated in response to pathogen infection. Similarly, transgenic plants were unable to accumulate 9-LOX transcripts in response to infection with P. infestans (Fig. 2B). Because 9-LOX-derived oxylipins accumulate in potato leaves in response to infiltration with P. syringae pv. maculicola earlier and to higher levels than after infection with P. infestans (20), bacterial infiltration was chosen to analyze alterations in LOX enzyme activity and the oxylipin pattern of transgenic RNA i plants. Four of the transgenic RNA i plants (R4, R6, R8, and R9) were compared with transgenic lines containing only the empty vector and untransformed wild type potato plants.
Compared with control plants, 9-LOX activity was significantly reduced in RNA i plants after infiltration of P. syringae pv. maculicola (Fig. 3). Two of the RNA i plants, R4 and R6, had nearly undetectable levels of 9-LOX activity, whereas RNA i plants R8 and R9 showed residual enzyme activity. 13-LOX activity did not differ significantly between control and transgenic plants and, overall, was 20-fold lower than 9-LOX activity (data not shown).
Alterations in the Oxylipin Pattern after Pathogen Treatment-Oxylipin profiling was performed to detect changes in lipid peroxidation during the hypersensitive response in plants with reduced 9-LOX activity. In general, levels of esterified oxylipins in membranes were 5-10-fold higher than those of the free oxylipins. Interestingly, the levels of esterified oxylipins in the 9-LOX-RNA i plants were strongly reduced and increased to lower levels in response to bacterial infiltration compared with wild type plants. However, except for 13-HO(D/T), the esterified oxylipins were derived from autoxidative processes in both wild type and 9-LOX-RNA i plants (data not shown). Because there were no 9-LOX-produced esterified oxylipins, our analysis of the role of 9-LOXs for pathogen defense concentrated on alterations in free oxylipin levels. These were determined 6, 12, 24, 36, and 48 h after infiltration of P. syringae pv. maculicola. Fig. 4 shows combined data for 9-LOX-derived oxylipins measured in plants R4 and R6, plants R8 and R9, as well as empty vector containing and wild type plants. Consistent with the decrease in 9-LOX activity, oxylipins derived from the 9-LOX pathway accumulated to significantly lower levels in RNA i transgenic plants. In general, 9-LOX-derived oxylipins originating from LnA were more affected in RNA i plants than those derived from LA. The prod-ucts of the 9-LOX reaction, 9-HPOD and 9-HPOT, were significantly reduced particularly at early time points. 9-HOD and 9-HOT levels 24 hours post-infiltration were 20 and 10%, respectively, of the wild type levels in the plants R4/R6. Interestingly, 9-HOT and 9-HOD accumulating in wild type plants in response to bacterial infiltration consisted of 98% S-enantiomers (Table I), whereas the lower ratio of S/R-enantiomers in the transgenic plants is indicative of a non-enzymatic origin of 9-HOD and 9-HOT. Divinyl ether levels were also significantly reduced in the plants R4/R6 with the strongest reduction in 9-LOX activity but were similar in plants R8/R9 and the control plants.
The loss of 9-LOX activity in the transgenic plants correlated with increased levels of 13-LOX products and non-enzymatically produced oxylipins (Fig. 5). In both R4/R6 and R8/R9 plants, JA and its biosynthetic precursor, OPDA, are present at higher amounts. 13-HOD and 13-HOT also accumulated to higher levels in the transgenic plants; however, the presence of about 73% S-enantiomers (Table I) indicates an enzymatic as well as a non-enzymatic origin for 13-HOD and 13-HOT for both transgenic and wild type plants. In addition, higher levels of 16and 12-HOT, compounds that are indicative of autoxidative processes (29), are observed in plants with reduced 9-LOX activity. Thus, oxylipin profiling revealed that in transgenic plants with severely reduced 9-LOX activity accumulation of enzymatically produced 9-LOX-derived oxylipins was remarkably reduced, but they contain significantly increased levels of oxylipins derived from 13-LOX activity and from autoxidative lipid peroxidation.
Pathogen-induced Cell Death Is Not Altered in 9-LOX-RNA i Plants-Despite the block in the 9-LOX pathway and despite the decreased levels of 9-LOX-derived oxylipins, there was no difference in the degree, timing, and phenotype of cell death formation upon bacterial infiltration under our experimental conditions (Fig. 6). In addition to the phenotypic evaluation, trypan blue staining did not reveal differences in the extent of cell death after pathogen treatment. Similarly, diaminobenzidine-staining for hydrogen peroxide in 9-LOX-RNA i and in control plants was not significantly altered. These data suggest that lipid peroxide formation during pathogen-induced cell death does not depend on the presence of 9-LOXs and can proceed via autoxidative processes. Moreover, hypersensitive cell death itself occurs independent of the origin of lipid hydroperoxides.

DISCUSSION
Direct peroxidation of membrane lipids as well as the synthesis of cell death-inducing oxylipins are the proposed contributions of LOXs to programmed cell death in animals (30). Animal 15-LOX can initiate programmed organelle disruption by dioxygenating mitochondrial membranes leading to formation of pore-like structures (31). Treatment of mitochondria with the soybean 15-LOX results in disruption of membrane integrity and release of cytochrome c, usually a trigger of apoptosis (30).
Whether lipid peroxidation during the pathogen-induced hypersensitive cell death in plants is catalyzed by LOXs or is the result of processes mediated by reactive oxygen species so far has been addressed by the determination of levels and nature of the accumulating lipid peroxides. It is generally assumed that the preponderance of S-enantiomeric hydroperoxides of polyunsaturated fatty acids represents the products of enzymatic reactions and that racemic mixtures result in almost all cases from non-enzymatic reactions, e.g. the autoxidative processes. Thus, the accumulation of large amounts of (9S)-HPOD and (9S)-HPOT in pathogen-infected plant tissue has led to the view that 9-LOXs are the enzymes responsible for lipid peroxidation (5,6). Furthermore, the timing of oxylipin accumulation and the onset of cell death which, in potato, occur concomitantly 6 -12 h after bacterial infiltration (20) support the notion that 9-LOX-mediated lipid oxygenation is the cause for membrane damage during the hypersensitive response. In tobacco treated with the hypersensitive response-inducing elicitor cryptogein, LOX activity starts to increase between 8 and 12 h after elicitor treatment, which correlates with cell death formation between 12 and 24 h (5, 32). However, only functional analyses can prove or disprove a role for LOXs in lipid peroxidation. Gain-of-function experiments using products of the LOX pathway were performed to demonstrate a role for LOXs in inducing cell death. In animals, various lipid hydroperoxides, as well as the substrate of animal LOXs, arachidonic acid, are known to induce apoptosis (30,(33)(34)(35). In plants, both the hydroperoxides and the hydroxides of 9-and 13-LOX products of LA and LnA are able to induce programmed cell death in tomato protoplasts (36) and lentil root protoplasts (30).
Loss-of-function approaches have included inhibitor studies, expression of LOX antibodies in plant cells, and deprivation of oxygen. In lentil root protoplasts, application of different LOX inhibitors resulted in reduction of H 2 O 2 -induced cell death (37). However, LOX inhibitors are generally considered to be problematic because of their unspecific activity on other enzymes. Eicosatetraynoic acid, for example, is considered to be an irreversible inhibitor for LOXs but is also used as a specific inhibitor of phospholipase A 2 (38). In addition to their possible effect on other enzymes, most of the LOX inhibitors can also act as general radical scavengers and may thus interfere with cell death via a LOX-independent mechanism.
Electroporation of anti-LOX monoclonal antibodies into lentil root protoplasts decreased the H 2 O 2 -induced programmed cell death ϳ2-fold, suggesting that the lentil LOX is important for cell death (30). Finally, the failure of cryptogein to induce necrosis formation in tobacco under low oxygen pressure was interpreted that depriving LOXs of oxygen was the reason for lack of cell death formation (5). However, anaerobic conditions have been shown to be above wound stress in the hierarchy of stress responses (26), and this might also be true for pathogen stress.
In the present study, the 9-LOX pathway was specifically blocked by RNA i of the pathogen-induced 9-LOX isoform of potato. Our data show that 9-LOX activity is drastically reduced and that 9-LOX-derived oxylipins are present in lower amounts than in wild type plants. On the other hand, increased amounts of oxylipins derived from autoxidation accumulate in response to pathogen infection. Because there is no phenotypic difference in the hypersensitive response in transgenic and control plants, as judged by trypan blue staining for dead cells and by diaminobenzidine staining for hydrogen peroxide formation, 9-LOX-derived lipid peroxides are apparently not required for the hypersensitive response in potato. Interestingly, the capacity of the plant to perform lipid peroxidation via autoxidative processes becomes apparent in the absence of detectable 9-LOX activity. Lipid peroxide formation is considered a crucial event in the development of hypersensitive cell death, and it appears to be facilitated by direct oxidation of fatty acids by ROS. Presumably, in the presence of 9-LOXs, lipid peroxidation proceeds via enzymatic reactions, possibly by the activation of the specific 9-LOX isoform by ROS. The enzymatic formation of lipid peroxides by LOXs during induction of hypersensitive cell death might be of dual advantage to the response of the plants to pathogens as follows: (i) because LOX-derived lipid peroxides are less toxic substances for the plant cell, because they are rapidly metabolized by enzymes of the LOX pathway; and (ii) several of these metabolites act as specific antimicrobial compounds. Being non-toxic for the plant, they might thus enhance the defense capacity of the plant in addition to the initial hypersensitive cell death reaction.