Anal. Chem. 2005, 77, 7366-7372
Conjugation of Keto Fatty Acids to Glutathione in Plant Tissues. Characterization and Quantification by HPLC-Tandem Mass Spectrometry Ce´line Davoine,† Thierry Douki,‡ Gilles Iacazio,§ Jean-Luc Montillet,† and Christian Triantaphylide`s*,†
Laboratoire de Radiobiologie Ve´ ge´ tale, CEA/Cadarache, DSV-DEVM, 13108 Saint-Paul Lez Durance Cedex, France, Laboratoire “Le´ sions des Acides Nucle´ iques”, Service de Chimie Inorganique et Biologique UMR-E 3 CEA-UJF, CEA/DSM/ De´ partement de Recherche Fondamentale sur la Matie` re Condense´ e, CEA-Grenoble, 38054 Grenoble Cedex 9, France, and Laboratoire de Bioinorganique Structurale, Case 432, UMR CNRS 6517 Chimie, Biologie et Radicaux libres, Faculte´ des Sciences de St Je´ roˆ me, 13397 Marseille Cedex 20, France
Both biotic and abiotic stress activate the oxylipin pathway in plants. As reactive electrophile species (RES), some oxylipins are expected to bind cellular nucleophiles in a Michae1 l-type addition reaction. Using the HPLC-tandem mass spectrometry techniques, we have established the analytical basis for the investigation of oxylipin conjugation to glutathione (GSH) in plant extracts. The GSH adducts to the four keto fatty acid isomers issued from both linoleic and linolenic acids were first produced and their mass spectrometric features analyzed in the positive electrospray ionization mode. In all cases, the main fragmentation (MS2 mode) of the pseudomolecular ion leads to the neutral loss of a glutamyl moiety (-129 Da), affording an ion that gives structural information upon an additional fragmentation (MS3 mode). The glutamyl loss was confirmed by the analysis of other GSH adducts to oxylipin RES and appeared as being characteristic of GSH adducts. It is thus proposed to search GSH adducts in plant extracts by HPLC-MS/MS, using initially the neutral loss mode and then the MS2 mode to further characterize the identified compounds. This methodology was successfully applied to the analysis of GSH adducts upon infiltration into leaves of the four previous keto fatty acids at 5 mM, a concentration inducing cell death. The production of GSH adducts to oxylipin RES was observed for the first time in plant tissues. Furthermore, the levels of adduct production explain in part the observed GSH depletion. These results support the role of RES in altering protein activities and cellular redox balance of plant cells, via addition reactions to cellular nucleophiles. In response to stress situations, plant cells develop oxidative processes including peroxidation of membrane-associated polyunsaturated fatty acids (PUFAs).1,2 Both reactive oxygen species * To whom correspondence should be addressed. Te´l: +33 4 42 25 64 86. Fax: +33 4 42 25 46 56. E-mail:
[email protected]. † CEA/Cadarache. ‡ CEA/DSM/De´partement de Recherche Fondamentale sur la Matie`re Condense´e. § UMR CNRS 6517 Chimie. (1) Howe, G. A.; Schilmiller, A. L. Curr. Opin. Plant Biol. 2002, 5, 230-6.
7366 Analytical Chemistry, Vol. 77, No. 22, November 15, 2005
and enzymes such as lipoxygenases (LOXs) and R-dioxygenases are involved in the lipid peroxidation processes, leading to the production of a wide array of bioactive molecules termed oxylipins.3-6 Some oxylipins exhibit R-β unsaturated keto or epoxide structures and may thus act as reactive electrophile species (RES) toward cellular nucleophiles, particularly with thiols and imidazole residues of proteins. In particular, R-β unsaturated keto compounds may undergo Michae¨l addition reactions. Moreover, in cells, glutathione (GSH), the most abundant free nonprotein thiol, is also a potential “target” for these compounds. These reactions occur both spontaneously at physiological pH and via glutathione-S-transferase (GSTs) catalysis.7 Such additions on proteins and GSH may result in the alteration of protein activities and cellular redox status, respectively.7-11 In mammalian cells, various RES, including cyclopentenone prostaglandins, keto fatty acids, and 4-hydroxynonenal, are known to bind GSH in vivo.7,8,10 As a general response, RES can be carcinogens at high doses but induce, at lower doses, the protective “anticarcinogenic” enzymes of “phase 2”, including detoxification enzymes such as GSTs.7,9 Moreover, GSH conjugates derived from arachidonate hydroperoxides, such as LTC4 and FOG7, were shown to possess biological activities, for example, to mediate chemotactic properties of neutrophiles or actin polymerization.12 Oxylipin conjugates to GSH are presently not described in plants. Some of the plant oxylipins issued from the linoleic (18:2) (2) Farmer, E. E.; Almeras, E.; Krishnamurthy, V. Curr. Opin. Plant Biol. 2003, 6, 372-8. (3) Mueller, M. J. Curr. Opin. Plant Biol. 2004, 7, 441-8. (4) Feussner, I.; Wasternack, C. Annu. Rev. Plant Biol. 2002, 53, 275-97. (5) Montillet, J.-L.; Cacas, J.-L.; Garnier, L.; Montane´, M. H.; Douki, T.; Bessoule J. J.; Polkowska-Kowalczyck, L.; Maciejwska, U.; Agnel, J.-P.; Vial, A.; Triantaphylide`s, C. Plant J. 2004, 40, 439-51. (6) Hamberg, M.; Sanz, A.; Rodriguez, M. J.; Calvo, A. P.; Castresana, C. J. Biol. Chem. 2003, 278, 51796-805. (7) Dinkova-Kostova, A. T. Mini Rev. Med. Chem. 2002, 2, 595-610. (8) Hayes, J. D.; McLellan, L. I. Free Radical Res. 1999, 31, 273-300. (9) Xu , C.; Li, C. Y.; Kong, A. N. Arch. Pharm. Res. 2005, 28, 249-68. (10) Ceaser, E. K.; Moellering, D. R.; Shiva, S.; Ramachandran, A.; Landar, A.; Venkartraman, A.; Crawford, J.; Patel, R.; Dickinson, D. A.; Ulasova, E.; Ji, S.; Darley-Usmar, V. M. Biochem. Soc. Trans. 2004, 32, 151-5. (11) Itoh, K.; Tong, K. I.; Yamamoto, M. Free Radical Biol. Med. 2004, 36, 120813. (12) Bowers, R. C.; Hevko, J.; Henson, P. M.; Murphy, R. C. J. Biol. Chem. 2000, 275, 29931-4. 10.1021/ac051155y CCC: $30.25
© 2005 American Chemical Society Published on Web 10/06/2005
Scheme 1. Oxylipin Reactive Electrophile Species Analyzed in This Work, Issued from Both Linoleic and Linoleic Metabolism through 9- and 13-LOXs1-4a
a The hydroperoxide structures obtained from the LOX reactions are also indicated for clarity. Abbreviations: 10-EHDE, 10,11(S,S)epoxy-9(S)-hydroxy-12(Z) octadecenoic acid; 12-EHDE, 12,13(R,S)-epoxy-9(S)-hydroxy-10(E) octadecenoic acid; 10-EHTrE, 10,11(S,S)epoxy-9(S)-hydroxy-12,15(Z,Z) octadecadienoic acid; 12-EHTrE, 12,13(R,S)-epoxy-9(S)-hydroxy-10,15(E,Z) octadecadienoic acid; 9-HPODE, 9(S)-hydroperoxy-10,12(E,Z) octadecadienoic acid; 13-HPODE, 13(S)-hydroperoxy-9,11(Z,E) octadecadienoic acid; 9-HPOTrE, 9(S)-hydroperoxy-10,12,15(E,Z,Z) octadecatrienoic acid; 13-HPOTrE, 13(S)-hydroperoxy-9,11,15(Z,E,Z) octadecatrienoic acid; 9-KODE, 9oxo-10,12(E,Z) octadecadienoic acid; 13-KODE, 13-oxo-9,11(Z,E) octadecadienoic acid; 9-KOTrE, 9oxo-10,12,15(E,Z,Z) octadecatrienoic acid; 13-KOTrE, 13oxo-9,11,15(Z,E,Z) octadecatrienoic acid; 12-oxoHODE, 9-hydroxy-12-oxo-10(E) octadecenoic acid; 12-oxoHOTrE, 9-hydroxy-12-oxo-10,15(E,Z) octadecadienoic acid; 12-OPDA, 12-oxophytodienoic acid.
and linolenic (18:3) fatty acids and expected to act as RES are described in Scheme 1. Among these, keto fatty acids are upstream oxylipins obtained from PUFAs by LOX catalysis through hydroperoxide production, as observed for pea LOX13 and with extracts from potato tubers,14 or via enzymatic dehydrogenation of the corresponding hydroxy fatty acids, as described for Jerusalem artichoke.15 Keto fatty acids were also observed at very low levels during the hypersensitive reaction of Arabidopsis leaves challenged with the avirulent pathogen Pseudomonas syringae pv. tomato.16 Interestingly, their application to leaf tissue, as well as several other RES, can lead to cell death.16 The induction of defense responses including GSTs has also been reported16-18 and (13) Ku ¨ hn, H.; Wiesner, R.; Rathmann, J.; Schewe, T. Eicosanoı¨ds 1991, 4, 9-14. (14) Kimura, H.; Yokota, K. Appl. Biochem. Biotechnol. 2004, 118, 115-32. (15) Chechetkin, I. R.; Medvedeva, N. V.; Grechkin, A. N. Biochim. Biophys. Acta 2004, 1686, 7-14. (16) Vollenweider, S.; Weber, H.; Stolz, S.; Chetelat, A.; Farmer, E. E. Plant J. 2000, 24, 467-76. (17) Almeras, E.; Stolz, S.; Vollenweider, S.; Reymond, P.; Mene-Saffrane, L.; Farmer, E. E. Plant J. 2003, 34, 205-16.
the oxylipin RES response is considered as part of the plant response to insect and pathogen attacks.2 We have evaluated in this work the analytical potential of highperformance liquid chromatography associated with tandem mass spectrometry (HPLC-MS/MS) to investigate the occurrence of oxylipin RES conjugates to GSH in plant tissues. We first chemically synthesized the GSH adducts of the keto fatty acids issued from both 18:2 and 18:3 fatty acids and from both 9- and 13-LOXs. We have characterized their structure and shown that they arise from a 1-4 Michae¨l addition. These four adducts can be separated and analyzed by HPLC-MS/MS, and we have extended our analyses to the other oxylipin RES depicted in Scheme 1. Finally, we have investigated the conjugation in planta upon keto fatty acid infiltration into tobacco leaf tissue. We show that (i) the HPLC-MS/MS methodology is well adapted to the analysis and characterization of such compounds and (ii) in planta, the keto fatty acid adducts to GSH are rapidly produced and their (18) Weber, H.; Chetelat, A.; Reymond, P.; Farmer, E. E. Plant J. 2004, 37, 877-88.
Analytical Chemistry, Vol. 77, No. 22, November 15, 2005
7367
production contributed to cellular GSH depletion, leading thus to leaf tissue collapse. MATERIALS AND METHODS Synthesis of Oxylipin Adducts to Glutathione. Keto fatty acids [KOD(Tr)E; see Scheme 1 for abbreviations] have been produced from linoleic acid and R-linolenic acid using 9- or 13LOXs, as previously described.19 All the other oxylipins (Scheme 1) were purchased from Larodan Fine Chemicals AB (Malmo¨, Sweden). The GSH adducts were nonenzymatically synthesized as described by Blackburn et al.20 The reaction mixture (5 mL), containing 5 mM GSH and 0.5 mM KOD(Tr)E in 20 mM borate buffer pH 10, was incubated under nitrogen at room temperature for 2 h. Completion of the reaction was monitored by TLC analysis on silica plates (TLC aluminum sheets, silica gel 60/Kieselguhr F254 precoated layer thickness 0.2 mm; Merck, Nogent sur Marne, Fr), using acetonitrile/water/acetic acid (80/20/0.1) as solvent. The reaction was stopped by setting the pH to 4 by addition of 0.1 N HCl. The solution was then loaded onto a C18 solid-phase extraction column (Puriflash 60 C18U40/63, Interchim, Montluc¸ on, France) preconditioned with 20 mM sodium borate at pH 4. The column was washed with water, and the conjugates were eluted with 40 or 60% acetonitrile. Fractions containing the conjugates were then pooled and evaporated to dryness under vacuum. Plant Growth and Treatments. Nicotiana tabacum (var. Petit Havana) was grown for 8-9 weeks in greenhouses at 80 µmol‚m-2‚s-1 photosynthetic photon flux density (HQI-BT 400 W-D Osram lamps, Mu¨nchen, Germany) with a light/dark cycle (14/10 h at 25/20 °C) and with 70% relative humidity. Leaves selected in the middle of the stem were detached and infiltrated between secondary leaf veins, applying the syringe tip to epidermis of excised leaves. Chemicals (5 mM) were prepared in a 0.5% Tween 80 aqueous solution before infiltration. After infiltration, leaf petioles were dipped into water and then leaves were incubated in the dark for 6 h prior to analysis. Infiltrated zones of leaves were then harvested, pooled, and frozen before different analyses. Analysis of Leaf Glutathione Levels. For the total GSH analysis, frozen leaves were ground in mortar with pestle and extracted in 6.3 mM diethylenetriamine pentaacetic acid (DTPA) containing 1 mM triethylcarboxyphosphine, 0.15% (v/v) trifluoroacetic acid in the presence of 40 µM internal reference, N-acetylL-cysteine. The homogenate was centrifuged at 20000g for 15 min at 4 °C, and the resulting supernatant was filtered (0.22 µm) and used for the determination of total GSH. Filtered extracts (125 µL) were mixed with the reaction buffer (225 µL) consisting of 6.3 mM DTPA, pH 8.2, containing 0.2 M 4-(2-hydroxyethyl)piperazine-1-propanesulfonic acid and monobromobimane (25 mM in acetonitrile; 6 µL). Following 20 min of incubation in the dark at room temperature, the reaction was stopped with methanesulfonic acid (1 M; 150 µL) and stored at 4 °C in the dark until HPLC analysis was carried out according to Sauge-Merle et al.21 Extraction of Adducts from Tobacco Leaves. Frozen infiltrated zones were ground in mortar with pestle, and adducts (19) Iacazio, G. Chem. Phys. Lipids 2003, 125, 115-21. (20) Blackburn, M. L.; Ketterer, B.; Meyer, D. J.; Juett, A. M.; Bull, A. W. Chem. Res. Toxicol. 1997, 10, 1364-71. (21) Sauge-Merle, S.; Cuine, S.; Carrier, P.; Lecomte-Pradines, C.; Luu, D. T.; Peltier, G. Appl. Environ. Microbiol. 2003, 69, 490-4.
7368
Analytical Chemistry, Vol. 77, No. 22, November 15, 2005
were extracted in 20 mM sodium borate buffer at pH 4. The homogenate was centrifuged at 8000g for 15 min, and the supernatant was filtered (0.22 µm) and used for the HPLC-MS/ MS analysis. Mass Spectrometry Characterization and Quantification of Oxylipin Adducts to Glutathione. All analyses were performed in the positive electrospray ionization (ESI) mode. Mass spectra of synthetic 9-KOD(Tr)E- and 13-KOD(Tr)E-GSH adducts were first recorded on an ion trap LCQ spectrometer (Finningan Mat, San Jose, CA). MS1 as well as MS2 and MS3 fragmentation spectra were obtained with the compounds in solution in a 1/1 mixture of 20 mM ammonium formate and methanol introduced at a flow of 10 µL/min. Characterization of the other synthesized adducts and quantitative analyses were performed by HPLC associated with a tandem mass spectrometer. The system consisted in a series 1100 Agilent HPLC system equipped with an Uptisphere ODB (Interchim, Montluc¸ on, France) octadecylsilyl silica gel column (2 × 150 mm i.d., 5-µm particle size). The mobile phase was a gradient of 2 mM ammonium formate and acetonitrile at a flow rate of 200 µL/min. The proportion of acetonitrile rose from 0 to 40% over 36 min. Methanol was also added at the outlet of the column and prior to the inlet of the mass spectrometer at a flow rate of 100 µL/min. The mass spectrometer was a triple quadrupolar API 3000 apparatus (PerkinElmer-Sciex, Toronto, Canada). Characterization studies were first performed in the neutral loss mode that permits the determination of the m/z ratio of pseudomolecular ions undergoing neutral loss of 129 mass units (glutamyl from the GSH moiety) upon fragmentation. In a subsequent step, the same samples were analyzed in the product ion scan mode that provides the full fragmentation spectrum (MS2) of the pseudomolecular ions of interest. Quantitative analyses were performed in the multiple reaction monitoring (MRM) mode. In this configuration, the spectrometer only monitors the signal corresponding to pseudomolecular ions of defined m/z ratio giving rise to a specific fragment determined in the product ion scan analysis. The KOD(Tr)E-GSH mass spectra are characterized by the same neutral loss of glutamyl moiety (-129 Da), which occurs with very similar intensities (less than 10% differences). Thus, for quantification, the response of the mass spectrometer was calibrated by injecting solutions of the synthetic 9-KODE-GSH adduct of well-defined concentration. RESULTS Production, Characterization and HPLC-MS/MS Analysis of Keto Di- and Trienic Fatty Acid Adducts to Glutathione. We first prepared the 13- and 9-KOD(Tr)E-GSH adducts, according to Blackburn et al.20 All fragmentation spectra were recorded in the positive ESI mode, because fragmentation of [M - H]- pseudomolecular ions produced in the negative mode exhibited only the GSH anion (m/z 306), which is poorly informative in terms of structure of the adducts (data not shown). Crude reaction mixtures were analyzed by HPLC-MS/MS, and a single chromatographic peak was obtained with the four oxylipins studied. The corresponding compounds were all found to be 1:1 adducts between the keto fatty acid and GSH, as inferred from their molecular weight. Further characterization of the adducts was obtained by analysis of the fragmentation spectrum of the pseudo-molecular ions of interest, either on the triple
Figure 1. MS2 spectrum of KOD(Tr)E adducts to GSH. Fragmentation of the [M + H]+ pseudomolecular ions generated by ESI in an ion trap mass spectrometer. (A) ESI-MS2 spectrum of 13-KODE-GSH and (B) main fragmentations; (C) ESI-MS2 spectrum of 9-KODE-GSH and (D) main fragmentations. ESI-MS2 spectra for 13-KOTrE-GSH and 9-KOTrE-GSH gave similar fragmentations.
quadrupolar system associated with HPLC or by infusion of the purified compounds in an ion trap mass spectrometer. A typical fragmentation mass spectrum, corresponding to the 13-KODEGSH adduct is shown in Figure 1. The [M + H]+ pseudo molecular ion (m/z ratio of 602) gives rise to a major fragmentation ion of m/z 473, which corresponds to the loss of a glutamyl moiety (see Figure 1A,B). The comparison of the MS2 fragmentation mass spectra of the purified 9- and 13-isomer adducts showed that all adducts followed identical fragmentation pathways (see Figure 1A,B and C,D for 13-KODE-GSH and 9-KODE-GSH, respectively). Thus, the characteristic loss of the glutamyl moiety (-129 Da) can be profitably applied to search for GSH adducts to a various oxylipins. Conversely, the MS3 analyses of the ions issued from the previous neutral loss, i.e., m/z 473 and 471, for KODE and KOTrE adducts, respectively, greatly differed from one compound to the other and provided structural information with respect to the 9- and 13-keto position on the fatty acid backbone (see Figure 2A,B for KODE-GSH). In addition, each of the MS3 analyses strongly suggested that the addition occurs through a 1-4-type rather than a 1-6-type Michae¨l reaction.20 Indeed, as typically observed for the 13-KODE-GSH adduct, the MS3 fragmentation of the m/z 473 ion, issued from the [M + H]+ MS2 fragmentation, showed a transition corresponding to the loss
of the heptanone moiety (m/z 473 f 359; see Figure 2A,B), characterizing both the 13-keto position and the 1-4 Michae¨l addition. Such a transition has been already described in the ESI MS3 mass spectrum of FOG7, but not in the structurally similar eicosanoid, LTC4, and assigned by deuterium labeling experiments to a McLafferty-type rearrangement, issued from the sulfonium ion.22 The MS3 analysis of the 9-KODE-GSH shows also a typical transition (m/z 473 f 287; Figure 2C,D) corresponding again to a McLafferty-type rearrangement, and similar observations were also done on the KOTrE adducts (data not shown). The other ions observed on the MS3 spectra were explained by the loss of the two remaining amino acids and a dehydration reaction. No ions that could arise from a 1-6 adduct could be observed. However, in the single chromatographic peak, which is observed in all our crude reaction mixture analyses, the presence of small amounts of the 1-6 addition products cannot be excluded. Thus, if our results unambiguously demonstrate the spontaneous 1-4 addition of GSH on all the KOD(Tr)E structures, they formally do not exclude the possibility of a 1-6-type Michae¨l reaction. Extension of the HPLC-MS/MS Analysis to Other Upstream Oxylipin RES Adducts to Glutathione. With the aim (22) Hevko, J.; Murphy, R. C. J. Am. Soc. Mass Spectrom. 2001, 12, 763-71.
Analytical Chemistry, Vol. 77, No. 22, November 15, 2005
7369
Figure 2. Product ions (MS3) obtained upon fragmentation of the m/z 473 daughter ion generated from the [M + H]+ 602 pseudomolecular ions of (A) 13-KODE-GSH and (C) 9-KODE-GSH in an ion trap mass spectrometer. The main fragmentations leading to the keto fatty acid release and to the McLafferty-type rearrangement are described in (B) and (D). These fragmentations were observed similarly with the corresponding m/z 471 ions generated from [M + H]+ 600 pseudomolecular ions of 13-KOTrE-GSH and 9-KOTrE-GSH, respectively.
to evaluate the HPLC-MS/MS methodology for the investigation of oxylipin-GSH production and fate in planta, we have analyzed the adducts to GSH of the different oxylipin RES depicted in Scheme 1. In all cases, the spontaneous addition led to the production of a main adduct, as assessed by the observation of a single chromatographic peak corresponding to compounds producing the expected pseudomolecular ion ESI analysis in the positive mode (Table 1). The MS2 spectra issued from the positive ESI all show a main fragmentation corresponding to the neutral loss of the glutamyl moiety from the [M + H]+ ion. Such a facile cleavage of the γ-glutamate from the glutathionyl portion of the adducts can be retained for a more specific and sensitive tandem mass spectrometry analysis (MRM mode) of adduct mixtures and used for extract analyses. However, the MS2 fragmentation spectra do not provide enough structural information to unambiguously identify the adducts on the basis on their mass spectrometry features, with the exception of some compounds, such as 12-OPDA and 12-oxoHOD(T)E, for which the structure of the adduct is obvious. This limitation may be overcome by the accurate analysis of the HPLC retention times. Indeed, structural isomers can be easily separated by HPLC and characterized with reference to standards. One example of separation of isobaric compounds with pseudomolecular ions [M + H]+ at m/z 618 is provided in Figure 7370 Analytical Chemistry, Vol. 77, No. 22, November 15, 2005
3 showing that product identification can be achieved by both MRM detection and analysis of the retention time. In Vivo Conjugation of KOD(Tr)E to GSH. To establish whether keto adducts could be synthesized in planta, we have infiltrated these compounds in tobacco leaves at 5 mM, a concentration inducing cell death.16 After each infiltration and incubation for 6 h, we have extracted the adducts (in 20 mM borate buffer pH 4) for a direct analysis by HPLC-MS/MS in the positive ESI mode. Analyses were first performed in the neutral loss mode in order to identify all compounds undergoing loss of glutamyl moiety upon fragmentation ([M + H]+ - 129 Da). For all keto fatty acid [KOD(Tr)E] infiltrations, a major adduct with a [M + H]+ 602 (or 600; Figure 4) was detected by HPLC-MS/ MS, as expected. The loss of a glutamyl moiety, as the major fragmentation pathway for these adducts, was confirmed by HPLC-MS/MS analyses carried out in the MS2 mode and providing full fragmentation spectra. Additional analyses were performed in the highly sensitive and specific MRM mode and showed that no other compound exhibiting mass spectrometry features similar to those of the KOD(Tr)E adducts was produced upon infiltration. For example, MRM analysis of 9-KOTrE infiltrated leaves (Figure 4) indicates that the main adduct corresponds to a [M + H]+ 600 adduct having the specific transition
Table 1. HPLC-MS/MS Analysis of Oxylipin RES Adducts to GSHa RES (molecular mass; Da)
pseudomolecular ion [M + H]+
main fragmentationsb (MS2)
retention time (min)
12-EHDE (312)
620
17.5
10-EHDE (312)
620
12-oxoHODE (312)
620
12-EHTrE (310)
618
10-EHTrE (310)
618
12-oxoHOTrE (310)
618
12-OPDA (291)
600
f 491 (-glu) f 602 (-H2O) f 584 (-2 H2O) 502 (584 - C6H10) f 473 (-glu -H2O) f 455 (-glu -2H2O) 373 (502 - glu) f 491 (-glu) f 308 (GSH + H+) f 491 (-glu) f 602 (-H2O) f 473 (-glu - H2O) f 308 (GSH + H+) f 295 (312 + H+ - H2O) f 489 (-glu) f 600 (-H2O) f 582 (-2H2O) 502 (582 - C6H8) f 471 (-glu -H2O) f 453 (-glu - 2H2O) 373 (502 - glu) f 489 (-glu) f 308 (GSH + H+) f 489 (-glu) f 600 (-H2O) f 471 (-glu - H2O) f 308 (GSH + H+) f 293 (310 + H+ - H2O) f 471 (-glu) other fragmentations: same as in KOTrE-GSH (Figure 1)
20.6 21.2
21.7
21.3 24.0
19.6
a The different oxylipins RES described in Scheme 1 were conjugated to GSH according to the protocol described in Materials and Methods. The adducts were analyzed after 2 h of incubation by HPLCMS/MS in the positive ESI mode. b glu: neutral loss of a glutamyl moiety (-129 Da) from the glutathionyl part of the adduct.
Figure 3. MRM chromatograms of a mixture GSH adducts of (A) 10-EHTrE, (B) 12-EHTrE, and (C) 12-oxoHOTrE. The transition m/z 618 f 489 is analyzed, corresponding to the neutral loss of the glutamyl moiety (-129 Da) from the [M + H]+ cations (see Table 1).
m/z 600 f 471. Moreover, this adduct coeluted with the chemically synthesized 9-KOTrE-GSH when a mix of infiltrated leaves homogenate and standard was analyzed by HPLC-MS/MS. This result shows that the main GSH adduct, synthesized after 9-KOTrE infiltration, is the 9-KOTrE-GSH conjugate and strongly suggests
that the addition structure is also 1-4 in vivo. Similar observations were also obtained with the other infiltrated compounds leading to the same conclusions. However, in the case of 13-KODE infiltration (Figure 4), if the adduct corresponding to the product of the spontaneous reaction of GSH with 13-KODE (1-4 adduct; retention time: 25.7 min) was characterized both by the MRM and MS2 analyses, a second chromatographic peak was also observed with the same MS2 features (retention time, 25.9 min). Although the MS2 analyses are not informative to characterize the adduct structures (see above), the presence of this second compound in extracts of infiltrated tobacco leaves might be tentatively attributed to a catalyzed 1-6 addition. The leaf areas infiltrated with 5 mM KOD(Tr)E developed the expected necrotic symptoms after 6 h of incubation, as assessed by dehydration (not shown). The analysis of the water-soluble compounds demonstrated in all cases a massive decrease in the GSH levels concomitant with the synthesis of the oxylipin-GSH conjugates in high yield (Table 2). The high concentration of the latter compounds strongly suggests that their formation is a major cause of GSH consumption. However, upon 9- and 13-KOD(Tr)E infiltrations, the levels of GSH adducts only corresponded to 1030% of GSH consumption, suggesting a possible metabolization of the adducts. In addition, the GSH level decrease is more important in the KOTrE-infiltrated leaves (70%) than in the KODE ones (60%), suggesting that the trienoic keto acids (9- and 13KOTrE) are more reactive than the dienoic keto acids (9- and 13-KODE). Moreover, the infiltration of KOD(Tr)E-GSH adducts, at levels (0.1 mM) similar to those previously determined upon KOD(Tr)E infiltration associated with cell death, does not induce the necrotic symptoms (not shown). Thus, these results suggest that the conjugation of oxylipin RES to GSH, which is associated with massive GSH depletion, should be accompanied by conjugation to proteins and other cellular nucleophiles inducing, or leading to, cell death processes. DISCUSSION In plants, a wide array of oxylipins are produced in response to different stress situations.1-4 Some oxylipins are RES and can bind cellular nucleophiles. We show herein that such compounds are able to conjugate with GSH, the most free abundant thiol in cells. This work allowed us to evaluate the basic principles for HPLC-MS/MS analysis of oxylipin RES adducts to GSH. We first examined in depth the fate of KOD(Tr)E, which are upstream oxylipins directly issued from fatty acid hydroperoxides through LOX activities.4,10 We prepared the GSH adducts from the naturally 9- and 13-KOD(Tr)E and demonstrated that a 1-4-type Michae¨l addition occurs with these compounds, both from in vitro and in vivo experiments. As a first result, the MS2 spectra of KOD(Tr)EGSH adducts analyzed in the positive ESI showed a loss of the glutamyl moiety (-129 Da) from the [M + H]+ ion. The other main fragmentation leads to a “reverse Michae¨l” reaction, providing ions corresponding to the keto fatty acid and to GSH. In this way, the MS2 mode showed no specific fragmentation to characterize the keto fatty acid isomers that are involved. Analysis in the MS3 mode of the fragment ion, issued from the glutamyl moiety loss (m/z 473 or 471), is on the contrary very informative to characterize the 9- or 13-keto fatty acid and the regioselectivity of the addition. Indeed, a very specific McLafferty rearrangement Analytical Chemistry, Vol. 77, No. 22, November 15, 2005
7371
Figure 4. MRM chromatograms of 9- and 13-KOD(Tr)E-GSH adducts produced in tobacco leaves upon keto fatty acid infiltration. The production of the adducts was investigated in the MRM mode by monitoring the transition 602f473 for 9- and 13-KODE-GSH, and m/z 600f471 for 9and 13-KOTrE-GSH. The retention time for each adduct is indicated on the chromatogram. Table 2. Production of Keto Fatty Acid-GSH Adducts in Vivoa
infiltrated molecules
corresponding adduct mass transition (m/z)
GSH decrease (%)
GSH adducts (% of GSH decrease)
9-KOTrE 13-KOTrE 9-KODE 13-KODE
600 f 471 600 f 471 602 f 473 602 f 473
73.6 ( 1.3 73.2 ( 1.4 60.2 ( 1.6 57.9 ( 1.3
30.1 ( 5.8 11.3 ( 1.4 17.5 ( 4.6 16.9 ( 2.4
a Leaves were infiltrated by the different KOD(Tr)E, as indicated in Materials and Methods, and the oxylipin-GSH adducts were analyzed after 6 h by HPLC-MS/MS. Each treatment led to the necrosis of the infiltrated leaf areas with the exception of mock infiltration used as the control. Mock infiltration was 0.5% Tween 80. For each infiltrated compound, the transition corresponding to the oxylipin-GSH adduct is indicated, together with the decrease of the total GSH and the GSH adducts synthesized.
was observed for each 9- or 13-KOD(Tr)E adduct suggesting a 1-4 Michae¨l addition.22 Thus, our data first do not confirm previous report attributing a 1-6-type addition for the noncatalyzed reaction of 13-KODE to GSH.20 They show second that spontaneous and enzymatic catalyzed reactions cannot be distinguished in vivo by this methodology alone. Therefore, further investigations are clearly needed, probably case by case, to address the important question on the origin of the GSH adducts in vivo. We have investigated the MS2 spectra of GSH adducts of oxylipin RES structurally different from keto fatty acids. As a (23) Coleman, J. O. D.; Blake-Kalff, M. M. A.; Davis, E. T. G. Trends Plant Sci. 1997, 2, 144-51.
7372
Analytical Chemistry, Vol. 77, No. 22, November 15, 2005
second result, we observed that all MS2 mass spectra of GSH adducts analyzed in the positive ESI mode led to the loss of the glutamyl moiety (-129 Da) from the [M + H]+ ion. It can thus be proposed to search for GSH adducts in cellular extracts using the HPLC-MS/MS in the neutral loss mode, as preliminary experiments. The adduct identification can then be achieved by HPLC-MS/MS combining the MRM analyses together with retention times provided authentic standards are available. These principles were applied to the analysis of plant extracts following keto fatty acid infiltration. GSH adducts were produced and shown to accumulate at high levels, explaining GSH depletion. Since GSTs are known to be induced by RES in plants,16 the conjugation reaction may occur in vivo both spontaneously and through GSTs catalysis.8,23 Although a 1-6 Michae¨l addition cannot be ruled out, our results demonstrated the production of compounds arising from a 1-4 addition in largest yield. Finally, we have also observed that leaf tissue infiltrated with millimolar levels of 9- or 13-KOD(Tr)E developed cell death symptoms (see also ref 13), which might be explained by the depletion of the cellular GSH, by the failure to completely detoxify the oxylipin RES by GSH conjugation, or by both actions. Our results support the role of RES in altering pools of cellular reductant thiols and by extension of thiols from proteins, modulating their function and activity,10,11 thus, in processes that could be involved in the regulation of plant cell death pathways. Received for review June 29, 2005. Accepted September 9, 2005. AC051155Y
