A novel class of oxylipins, sn1-O-(12-oxophytodienoyl)-sn2-O-(hexadecatrienoyl)-monogalactosyl Diglyceride, from Arabidopsis thaliana.

The cyclic derivative of 13(S)-hydroperoxolinolenic acid, 12-oxophytodienoic acid, serves as a signal transducer in higher plants, mediating mechanotransductory processes and plant defenses against a variety of pathogens, and also serves as a precursor for the biosynthesis of jasmonic acid, a mediator of plant herbivore defense. Biosynthesis of 12-oxophytodienoic acid from alpha-linolenic acid occurs in plastids, mainly in chloroplasts, and is thought to start with free linolenic acid liberated from membrane lipids by lipase action. In Arabidopsis thaliana, the glycerolipid fraction contains esterified 12-oxophytodienoic acid, which can be released enzymatically by sn1-specific, but not by sn2-specific, lipases. The 12-oxophytodienoyl glycerolipid fraction was isolated, purified, and characterized. Enzymatic, mass spectrometric, and NMR spectroscopic data allowed us to establish the structure of the novel oxylipin as sn1-O-(12-oxophytodienoyl)-sn2-O-(hexadecatrienoyl)-monogalactosyl diglyceride. The novel class of lipids is localized in plastids. Purified monogalactosyl diglyceride was not converted to the sn1-(12-oxophytodienoyl) derivative by the combined action of (soybean) lipoxygenase and (A. thaliana) allene oxide synthase, an enzyme ensemble that converts free alpha-linolenic acid to free 12-oxophytodienoic acid. When leaves were wounded, a significant and transient increase in the level of (12-oxophytodienoyl)-monogalactosyl diglyceride was observed. In A. thaliana, the major fraction of 12-oxophytodienoic acid occurs esterified at the sn1 position of the plastid-specific glycerolipid, monogalactosyl diglyceride.

Oxylipins are a diverse class of acyclic or cyclic oxidation products of fatty acids, many of which have regulatory functions in the cell (1). Among the plant oxylipins, the cyclic octadecanoids, collectively called jasmonates, have received much recent interest because of their diverse biological functions as mediators of mechanotransduction, herbivore and pathogen defense reactions, senescence, and male fertility (2)(3)(4).
Jasmonate biosynthesis is thought to start with induced release of ␣-linolenic acid from membrane glycerolipids (5), and the pathway from ␣-linolenic acid to jasmonic acid (JA) 1 has been worked out in detail (6,7), whereas the processes of ␣-linolenic acid release are not understood. The oxidation of ␣-linolenic acid to 13(S)-hydroperoxolinolenic acid catalyzed by 13-lipoxygenase and the conversion of this intermediate to the first cyclic metabolite of the pathway, 12-oxophytodienoic acid (OPDA), occur in plastids and thus, in the shoot, predominantly in the chloroplast (8 -10). OPDA then leaves the chloroplast and is reduced by the NADPH-dependent flavoprotein oxophytodienoate reductase 3 (11,12), and the reaction product is converted to JA by ␤-oxidation (11).
It is known that plant lipoxygenases may act on glycerolipids as well as on free fatty acids (13 and references therein). This prompted us to survey plant membrane lipids for the presence of cyclic octadecanoids. As a result of this study, we report here that the major fraction of OPDA in Arabidopsis thaliana occurs esterified in the sn1 position of plastid-specific galactolipids. The implications of this novel class of glycerolipids for octadecanoid physiology and chloroplast membrane biology are discussed.

EXPERIMENTAL PROCEDURES
Plant Material-Greenhouse-grown rosettes of 6 -8-week-old A. thaliana C24 were used for all experiments. Plants were raised individually in standard soil at 18°C (night) and 24°C (day) (average temperatures) between 50 and 70% relative humidity and no less than 150 mol of photons m Ϫ2 s Ϫ1 (photosynthetically active radiation; supplementary light, if required, from sodium vapor lamps). Plants were cut above soil and immediately extracted, or they were immersed in liquid nitrogen and stored at Ϫ70°C for minimum amounts of time.
Wounding Experiments-Wounding experiments were carried out as previously described (14). Briefly, all leaves of a rosette were crushed on one half of the leaf blade with a hemostat so that ϳ10% of the total leaf area of the plant was wounded. After incubation for the times indicated, 1 g of leaf tissue was collected using the whole leaf blades of wounded leaves and analyzed for OPDA and JA content as described in detail in Ref. 15, using [ 2 H 5 ]OPDA and [ 13 C 2 ]JA (vide infra) as internal standards. Extraction (16) and quantification of lipid-bound OPDA were done as described above.
nate was centrifuged at 10,000 ϫ g, and the supernatant was used to analyze the enzymatic degradation of OPDA. To do this, to 1 ml of the protein extract (1.5 mg of protein), 1 g/ml [ 2 H 5 ]OPDA and [ 13 C 2 ]JA were added, and the assay mixture was then incubated at room temperature. Aliquots were taken at the times indicated and worked up for the analysis of OPDA and JA.
Lipid Extraction and Purification-Initial experiments were carried out using the standard procedure of Bligh and Dyer (16). It was later found that yields and purity of the glycolipid fraction improved when the following preparative procedure was employed. Leaf tissue (500 g) was immersed in 600 ml of boiling methanol containing, as an antioxidant, 200 mg of 2,6-di-tert-butyl-4-methylphenol (butylated hydroxytoluene). After boiling for 15 min, the hot mixture was homogenized with a Polytron mixer, the homogenate was filtered on a Buechner funnel and then over two sequential paper filters, and the methanol was removed by vacuum distillation (25 mbar, 30°C). The aqueous phase was adjusted to 70% (v/v) methanol (crude lipid extract).
The crude lipid extract was passed over a column of 30 g of LiChroprep RP-18 (Merck). The column was eluted with 150 ml isopropyl alcohol: isohexane (80:55, v/v) and further fractionated on Chromabond NH 2 (Macherey-Nagel; aliquots of one-eighth of the LiChroprep RP-18 eluate passed over 3 g of column material that was eluted with 10 ml of the above solvent). The effluents and eluates of the Chromabond NH 2 column were pooled, the lipid extract was concentrated in a rotary evaporator (25 mbar, 30°C), and the crude lipid was finally dried in a stream of N 2 (yield, ϳ11.3 mg from 500 g of leaf tissue). The residue was stored in 4 -6 ml of chloroform:methanol (1:1, v/v) at Ϫ28°C under nitrogen, if not used directly.
The dried crude lipid was redissolved in 4 -5 ml of i-propanol:ihexane (4:3, v/v) and centrifuged to remove particles. Aliquots (1 ml) of the clear solution were injected onto a semipreparative Nucleosil 100 HPLC column (Knauer; 25 mm ϫ 8 mm inner diameter, particle size 15-25 m) and eluted at a flow rate of 5 ml min Ϫ1 as follows: 1 min of eluent A, then a linear gradient in 10 min to 50% eluent B, then 27 min of isocratic flow at 50% solvent B, followed by a linear gradient over 10 min to 100% solvent B (solvent A: isopropyl alcohol: isohexane (4:3, v/v); solvent B: isopropyl alcohol: isohexane:H 2 O (8:6:1.5, v/v/v)) (17). UV detection was at 205 nm. The lipid composition was checked in each fraction by TLC according to (17). The presence of esterified OPDA was checked by enzymatic treatment of each lipid fraction with Rhizopus arrhizus lipase (18). Esterified OPDA-containing fractions (R t Ϸ 15 min) were purified further. To do so, the OPDA-containing lipid from each run on the silica column was pooled, the solvent was removed in a stream of nitrogen, and the residue was redissolved in 1 ml of acetonitrile:aqueous 50 mM ammonium acetate (70:30, v/v), centrifuged, and applied to a Dynamax C18 reversed phase HPLC column (Rainin; 250 mm ϫ 21.4 mm inner diameter, 8 m particle size, 60 Å pore size) at a flow rate of 20 ml min Ϫ1 (isocratic elution for 100 min, UV detection at 221 nm). The OPDA-containing fractions (R t Ϸ 52-57 min) were pooled and rechromatographed on the same column to yield a homogenous peak (R t ϭ 55 min), which was used for all further characterizations. From 1 kg of plant material, ϳ1-2 mg of ester-OPDA lipid was obtained.
Analysis of 12-Oxophytodienoic Acid-To analyze the isomeric composition, optical purity, and quantity of OPDA, the methods described in detail previously, based on GC-MS, were employed (15, 19, 20, 13 C, Isotec Inc.) was converted to its dimethyl ester using CH 2 N 2 and dissolved in 50 l of 2-(2Z pentyl)-2-cyclopenten-1-one (Firmenich SA, Geneva, Switzerland), followed by the addition of 170 l of 0.2 M sodium methanolate. The mixture was stirred for 1.5 h under nitrogen at room temperature. The reaction was terminated with concentrated acetic acid (4 l), and the sample was dried with nitrogen, redissolved in doubly distilled water (150 l), and hydrolyzed under nitrogen for 18 h at 225°C. After cooling, the reaction mixture was diluted into 0.1 M NaHCO 3 (2-3 ml) and extracted twice with chloroform, and the reaction product (racemic [ 13 C 2 ]methyljasmonate) recovered from the chloroform phase was treated for about 2 h at 60°C with 0.5 M KOH to release the free acid, which was purified by HPLC as described in Ref. 15 (yield, ϳ30%; purity, Ͼ99%; isotopic abundance, Ͼ99% as analyzed by GC-MS).
Mass Spectrometry of Lipids-Routine analyses of fatty acid and OPDA methyl esters were performed on a Finnigan MAT Magnum ion trap mass spectrometer in chemical ionization (reactant gas, methanol) mode (scan range, 50 -398 amu; scan rate, 1 Hz). GC settings were as follows: DB-35 column (J & W Scientific, 30 m ϫ 0.25 mm ϫ 0.25 m coat), helium carrier gas, splitless injection mode (250°C), and temperature program of 1 min at 80°C, 30°C min Ϫ1 to 200°C, 5°C min Ϫ1 to 250°C, and 10 min at 250°C. Typical retention times (methyl esters) were as follows: cis-OPDA, 16 LC-MS analyses were performed on a triple-quadrupole instrument (TSQ 7000, Finnigan MAT) operated in electrospray (positive or negative ion) mode (4.5 kV; capillary temperature, 200°C; atmospheric pressure ionization-collision-induced dissociation voltage, 15 V). The compounds to be analyzed were introduced (injection volume, 5 l) through a Luna C18 HPLC column (Phenomenex; 150 mm ϫ 1 mm inner diameter, 5 m particle size) using methanol, 20 mM ammonium acetate (90:10, v/v) as eluent (isocratic flow rate of 50 l min Ϫ1 obtained with an Applied Biosystems 140 B dual-piston pump). The retention time of the OPDA-lipid was ϳ30 min. MS/MS measurements were performed using argon collision gas (2 ϫ 10 Ϫ3 torr) and a collision energy at quadrupole 2 of 10 V. MALDI-TOF-MS was performed on a PE Biosystems Voyager instrument (matrix, dihydroxybenzoic acid) with positive ion detection (25 kV acceleration voltage, calibrated with human angiotensin I (Sigma)).
Nuclear Magnetic Resonance Spectroscopy-The purified OPDA-containing lipid (ϳ20 mg) was dissolved in 0.7 ml of CD 3 OD. NMR spectra were recorded with Bruker AC 200 and AMX 500 (5-mm inverse TXI probe head) instruments. NMR conditions were as follows: H-H COSY, 45 0 mixing pulse; HMQC, phase-sensitive mode using time-proportional phase increment, bilinear rotation decoupling sequence, globally optimized alternating-phase rectangular pulses decoupled; HMBC, phase-sensitive using time-proportional phase increment, delay tuned to long range couplings, 71 ms.
Enzymatic Conversion of Monogalactosyl Diglyceride (MGDG)-MGDG was isolated from Brassica napus according to Bligh and Dyer (16). After drying, the crude lipid was purified as described above using solid phase extraction and semipreparative HPLC (Nucleosil 100, 12-14-min fraction). The identity of the isolated MGDG was verified by electrospray ionization-MS. The presence of the 18:3/16:3 molecular species was indicated by the molecular ion [M ϩ NH 4 ] ϩ ; m/z ϭ 764.61. The purified MGDG (10 mg) was treated with soybean lipoxygenase (1 mg of protein) and recombinant allene oxide synthase (2 mg of protein representing total soluble protein from Escherichia coli lysates from cells expressing the A. thaliana enzyme (16,19)) in 50 mM potassium phosphate, pH 8.0, at room temperature for 6 h. Thereafter, the reaction mixture was reextracted (16), and the lipids were hydrolyzed for 1 h in 0.5 N KOH (100°C). The released acyl moieties were analyzed by GC-MS (19). To monitor the functionality of the enzymes used, the reaction mixture contained, in addition to plant MGDG, [ 2 H 5 ]linolenic acid. This allowed us to monitor in parallel by GC-MS the conversion of the esterified ␣-linolenic acid in the MGDG substrate as well as that of free ␣-linolenic acid by the two enzymes.
General and Biochemical Procedures-The presence and quantity of lipid-bound (esterified) OPDA was determined by lipase treatment of the corresponding fractions using an adaptation of the procedure described in Ref. 18, followed by GC-MS analysis (see above) of the released metabolites (OPDA, fatty acids) as their methyl esters. OPDA was released by the sn1-selective lipase from R. arrhizus (triacylglycerol lipase, EC 3.1.1.3, Roche Molecular Biochemicals) as follows. To each fraction to be analyzed was added an internal standard of cis-[ 2 H 5 ]OPDA (250 ng), and the sample was dried in a stream of nitrogen. The residue was then redissolved in 1 ml of 50 mM Tris borate (pH 7.5). To this solution were added 25 units of lipase in 50 l of 100 mM Tris buffer (pH 7.5); controls received only buffer. The samples were then incubated at 37°C (for the times indicated or overnight, if reactions were to go to completion), acidified (20 l of concentrated HCl), and extracted two times with 2 ml of ethyl acetate followed by standard workup (20). In preparative experiments aimed at purification of the OPDA-lipid, no internal standard ([ 2 H 5 ]OPDA) was used.
Additional lipases tested were those from Candida rugosa (Sigma), Mucor javanicus (Sigma), and wheat germ (Sigma), and phospholipases tested were the phospholipases A 2 (EC 3.1.1.4) from Apis mellifica (Sigma) and Naja mossambica (Sigma). All lipases were assayed at pH 7.5; the two phospholipases A 2 were assayed at pH 8.9. Alkaline hydrolysis of lipids was for 1 h at 100°C in 0.5 M KOH, followed by acidification and further workup as above.

Discovery of Membrane Lipid-bound 12-Oxophytodienoic
Acid-During the course of studies on OPDA metabolism in A. thaliana, it was noticed that [ 2 H 5 ]OPDA added to leaf crude extracts was rapidly metabolized, whereas metabolism of endogenous (i.e. unlabeled) free OPDA followed a completely different time course; over at least 1 h, its level actually increased, before it started to decline also (Fig. 1). The pool from which this OPDA was released in crude extracts could be localized to the membrane fraction (100,000 ϫ g pellet), whereas the 100,000 ϫ g supernatant harbored the OPDA-metabolizing activity. When crude lipid from A. thaliana leaves obtained through Bligh and Dyer extraction (16) was subjected to alkaline hydrolysis (saponification conditions), it released OPDA. Control reactions verified that this OPDA fraction was not just dissolved in the lipid phase but covalently bound in an alkalilabile manner, i.e. most likely esterified (Fig. 2).
To scrutinize these initial results, the crude lipid was subjected to treatments with various lipases including phospholipases of the A 2 type (Fig. 3). Whereas there was no (or marginal) release of OPDA by phospholipase A 2 as well as by the wheat germ lipase, the two lipases that exhibited a preference for the sn1 position of glycerolipids (the fungal enzymes from R. arrhizus and M. javanicus (18)) efficiently released OPDA from the crude lipid fraction. It was estimated that ϳ80 -90% of the total OPDA content of A. thaliana shoots occurred covalently bound in the lipid fraction in plants not subjected to any stimulus inducing the jasmonate signaling system. The endogenous activity that released OPDA from the lipid-bound form was associated with the 5,000 ϫ g pellet but was not detected in the 12,000 ϫ g supernatant of crude extracts and, thus, most likely represents a membrane-associated system. This activity was not characterized further in the present study.
Analysis by chiral GC-MS of the isomeric status of OPDA released from the lipid revealed that the lipase-released metabolite was the cis(ϩ) enantiomer, (9S,13S)-12-oxophytodienoic acid, the same enantiomer that is also extracted in free form from plants (19,20). Alkaline hydrolysis, of course, yielded predominantly the trans isomer (9S,13R)-OPDA), which is produced from the cis form by enolization, particularly under alkaline conditions. The method of analysis of OPDA enantiomers has been described in detail elsewhere (19).
Purification and Structural Characterization of the Lipasedegradable, Membrane-associated OPDA-Lipid-The solubility of the OPDA-lipid in aqueous methanol proved optimum around 80% (v/v), whereas free OPDA is most soluble around 50% (v/v) methanol. Thus, although less polar than the free acid, the OPDA-lipid turned out to be a remarkably polar metabolite that could be extracted from the plant material more selectively with aqueous methanol (70% v/v) than with conventional, less polar, lipid-solvent mixtures (e.g. see Ref. 16). Using the procedure detailed under "Experimental Procedures," from 600 mg of crude lipid (corresponding to 26.5 kg of leaf tissue), 20 mg of an apparently homogenous OPDA-lipid fraction was obtained for structure elucidation. During the course of the purification, the compound, in all steps, behaved most similarly to the plastid lipid MGDG. The OPDA-lipid eluted at R t Ϸ 15 min from the first HPLC column (Nucleosil 100, coeluting with MGDG) and at R t Ϸ 52-57 min from the reversed phase column separating the lipid molecular species. The purified material, upon rechromatography, eluted from the same column as a single peak with R t ϭ 55 min (Fig. 4).
Treatment of the purified lipid fraction with the sn1-specific R. arrhizus lipase (Fig. 4, inset) resulted in the immediate release of two acidic metabolites, which were identified by GC-MS as ␣-linolenic acid and OPDA. After a lag phase of ϳ5 min, when the rates of OPDA and ␣-linolenic acid release began to level off and the enzyme began to attack the sn2 position, a third acidic metabolite began to accumulate, which was identified as (7Z,10Z,13Z)-hexadecatrienoic acid (16:3). These results indicated that (i) a mixture of two lipids had been obtained by the purification procedures employed (attempts to resolve this further were unsuccessful), and (ii) compound A of that mixture was a diacyl lipid with ␣-linolenic acid in the  (Table I, Fig. 7). The resonances of sn1, sn2, and sn3 were assigned by inspection of chemical shifts and multiplicities ( Table I). As seen in Fig. 7, the anomeric proton of the sugar moiety at ␤-linkage) couples via HMBC with the sn3 carbon at ␦ 63.9 ppm. The galactopyranosyl nature of this sugar moiety could be proven by analysis of the spin system, showing e.g. the typical broad doublet of the H-4Ј at ␦ 3.82 ppm (J ϭ 3.2 Hz). The signals C-1Љ and C-1ٞ of the unsaturated fatty acids appear at ␦ 174.6 and 175.0 ppm, making it impossible to assign the linkage position at the glycerol skeleton. However, the sn1 linkage of the OPDA has been proven by the enzymatic methods already described above. The chemical shifts of the carbon and proton resonances, as well as the multiplicity, are in agreement with the data reported formerly for OPDA (24) and methyl-(7Z,10Z,13Z)-hexadecatrienoate (25). The overlapping of the signals of the olefinic protons 7Љ, 8Љ, 10Љ, 11Љ, 13Љ, 14Љ, 15ٞ, and 16ٞ at ␦ 5.30 -5.50 ppm did not allow the determination of the E/Z configuration. However, the configuration of the sn2-hexadecatrienoic acid of MGDG from A. thaliana had earlier been shown to be exclusively the 7Z,10Z,13Z isomer (26), and our GC-MS data (see above) proved that the same isomer was present in the sn2 position of the OPDA lipid. The collective evidence thus firmly established the structure of the isolated metabolite to be sn1-O-(9S,13S-12-oxophytodienoyl)-sn2-O-(7Z,10Z,13Z-hexadecatrienoyl)-monogalactosyl diglyceride (MGDG-O, Fig. 8).
Dynamics of MGDG-O Levels in Leaves from Wounded Plants-It is well established that the level of JA, a wound signal produced via OPDA (6), rises rapidly and transiently after wounding of leaves (e.g. see Ref. 27). We therefore analyzed the levels of MGDG-O in parallel with those of free OPDA and JA in wounded A. thaliana leaves (Fig. 9). The results revealed complex dynamics different for each of the three compounds.
Using a combination of soybean lipoxygenase and recombinant allene oxide synthase, it is possible to convert free ␣-linolenic acid via 13(S)hydroperoxylinolenic acid to the allene oxide 12,13-epoxylinolenic acid, which then rearranges spontaneously, with the formation, among other products, of racemic cis-OPDA (19). When purified MGDG (17) was subjected to the same procedures, MGDG-O was not detected, although ␣-linolenic acid used as a control was readily converted (data not shown).

DISCUSSION
During recent years, it has become obvious that plants harbor a rich variety of oxidized lipids, collectively known as oxylipins (1), many of them inducible and of physiological importance as signaling compounds in herbivore defense, pathogen defense, and mechanotransduction or involved in the regulation of developmental processes such as senescence and pollen maturation (2)(3)(4). The majority of oxylipins occur in free form (1), which is believed to originate from fatty acids released from membrane lipids (5,9). Recently, however, it has become clear that esterified or amidated fatty acids can be oxidized by enzymes of plant oxylipin biosynthesis. Thus, lipoxygenase attacks unsaturated fatty acids of triglycerides to convert them to hydroperoxo derivatives, which are then released by lipases and enter ␤-oxidation (28), but lipoxygenase has also been shown to oxidize fatty acids esterified in phospholipids (29). Phospholipases that exhibit a preference for phosphatidylcholine with oxygenated acyl groups have been identified in plant microsomes (30). Moreover, N-acyl(ethanol) amines have been identified in plants (31), and these are thought to play a role in plant pathogen defense (32). It has recently been shown that enzymes of the octadecanoid biosynthetic pathway, namely lipoxygenase and allene oxide synthase, act on N-acyl(ethanol) amines to yield, among other products, the 12-oxo-N-phytodienoylamines (33). Homologs of isoprostanes that are formed nonenzymatically on phospholipids (34) have meanwhile been found in plants, the phytoprostanes (35). Thus, although the biological relevance of these compounds is still not understood, the generation, in situ, of oxylipins in plant membranes probably complements their formation from free fatty acids, and it will be important to elucidate the enzymology of their formation and their physiological role(s).
We show, for the first time, that OPDA, a signal in mechanotransduction (36) and the precursor of JA (6), a wound and pathogen defense signal (2-4), occurs in esterified form in the sn1 position of the plastid galactolipid MGDG. This strongly suggests a defined route of its formation. It is likely that this route is a plastidic one, i.e. in green tissue localized in chloroplast. The first indication is the (7Z,10Z,13Z)-hexadecatrienoic acid (16:3) in the sn2 position of MGDG-O. MGDG molecular species carrying sn2-16:3 are synthesized in plastids by a prokaryotic pathway (22,26). A large fraction of sn2-16:3-MGDG carries ␣-linolenic acid in the sn1 position, the precursor of OPDA. It is thus possible that OPDA is generated in situ from the ␣-linolenic acid in the sn1 position of MGDG. However, we have been unable to show conversion of MGDG to MGDG-O in vitro using soybean lipoxygenase and A. thaliana allene oxide synthase, an enzyme combination that converts free ␣-linolenic acid to OPDA. This means that either the enzymes involved in formation of MGDG-O in vivo differ from those involved in synthesizing OPDA or the substrate has to be presented to them in a particular environment (e.g. a bilayer membrane). Alternatively, it is conceivable that esterified OPDA is not formed from MGDG in situ but that free OPDA is synthesized first and then incorporated in the membrane lipid. These issues will be addressed in future studies. It was shown earlier that allene oxide synthase is associated with chloroplast membranes from which it has to be released with detergent (9,37). The recombinant enzyme is soluble and, as such, acts on free fatty acid 13(S)-hydroperoxides (38,39). The chloroplast membrane docking site of allene oxide synthase is unknown, but the membrane association may reflect the fact that the enzyme, in vivo, acts on a membrane substrate rather than on a free fatty acid.
MGDG-O conceivably is a metabolite with a biological significance of its own, and/or it may form a pool from which OPDA could be released immediately when the need arises. Indeed, a membrane-associated enzymatic activity that liberates OPDA from MGDG-O was identified in A. thaliana in this study but awaits its full characterization. The data in Fig. 9, on the other hand, do not show decreases in MGDG-O correlated with the initial accumulation of JA (and OPDA) induced by wounding. This may indicate resynthesis of MGDG-O or simply reflect the fact that only a minor fraction of it is being converted to release OPDA. A decisive answer about precursor-product relationships awaits pulse-chase analyses of metabolic fluxes through these metabolites and experiments with isolated chloroplasts. It is clear, though, that wounding induces a complex, biphasic increase in the MGDG-O level. The transient nature of this increase shows clearly that MGDG-O is not simply a metabolic end product but that it has a significant turnover. MGDG-O levels fluctuate quite independently from those of JA. Thus, distinct physiological roles of MGDG-O and JA should be envisaged. It is interesting to note that the two sn1-specific lipases that effectively cleaved OPDA from MGDG-O are enzymes of fungal origin and that both species, R. arrhizus and M. javanicus, are plant pathogenic saprophytes. A potential function of MGDG-O could thus be as a reporter of lipolytic fungal activity indicative of invading phytopathogenic fungi. OPDA released locally upon fungal attack would (directly or through conversion to JA) signal to the plant the presence of an intruder and help to mount immediate local defense responses. Indeed, octadecanoids act as local, and not as systemic, defense signals in A. thaliana (40). We suggest that subscribers photocopy these corrections and insert the photocopies at the appropriate places where the article to be corrected originally appeared. Authors are urged to introduce these corrections into any reprints they distribute. Secondary (abstract) services are urged to carry notice of these corrections as prominently as they carried the original abstracts.