Synthesis and Stability Studies of the Glutathione and N

propane-1,2-diol (PAP) as the biomarkers of the toxic oil batches that caused toxic ... (TOS), an intoxication episode that occurred in Spain in 1981,...
0 downloads 0 Views 131KB Size
Chem. Res. Toxicol. 2005, 18, 1721-1728

1721

Synthesis and Stability Studies of the Glutathione and N-Acetylcysteine Adducts of an Iminoquinone Reactive Intermediate Generated in the Biotransformation of 3-(N-Phenylamino)propane-1,2-diol: Implications for Toxic Oil Syndrome Anna Martı´nez-Cabot, Anna Morato´, and Angel Messeguer* Department of Biological Organic Chemistry, Institut d’Investigacions Quı´miques i Ambientals de Barcelona (CSIC), Barcelona, Spain Received June 23, 2005

Epidemiological studies have pointed to fatty acid mono- and diesters of 3-(N-phenylamino)propane-1,2-diol (PAP) as the biomarkers of the toxic oil batches that caused toxic oil syndrome (TOS), an intoxication episode that occurred in Spain in 1981, causing over 400 deaths and affecting more than 20000 people. The biotransformation of PAP administered intraperitoneally to two mouse strains produced potentially toxic metabolites. The identification of 3-(4′hydroxyphenylamino)propane-1,2-diol among those metabolites was important because the compound can generate the quinoneimine intermediate 2. The potential toxicity of quinoneimines has been attributed primarily to their electrophilic character. Accordingly, the reactions of 2 with N-acetylcysteine, N-acetylcysteine methyl ester, and GSH were investigated. Quinoneimine 2 reacts with the N-acetylcysteine methyl ester to give the expected conjugate as a major product, accompanied by the corresponding bis and tris adducts. The monoadduct, when isolated in pure form, undergoes spontaneous oxidation to generate a new quinoneimine intermediate, which in turn rearranges and undergoes hydrolysis to afford the thiol adduct formally derived from the quinoneimine generated from p-aminophenol. The same overall pathway was observed for the reaction of 2 with N-acetylcysteine and GSH. Both thiol reagents reacted with the quinoneimine to give the corresponding adducts in which the addition took place at the ortho position with respect to the amino group. These conjugates were also unstable and ultimately afforded the corresponding adduct derived from p-aminophenol. The relevancy of these results to TOS, as well as their potential generalization for quinoneimines derived from other xenobiotics, is discussed herein.

Introduction Toxic oil syndrome (TOS)1 was a food intoxication episode of serious proportions that occurred in 1981 in central and northern areas of Spain, causing over 400 deaths and affecting more than 20000 people (1, 2). TOS was characterized as a multisystemic disease, which resembled an allergic-toxic syndrome in the acute phase and an autoimmune condition in the chronic phase. It is currently estimated that 30-40% of the affected population continues to suffer mild symptoms or severe sequelae. TOS research has been summarized in recent reviews (3, 4). Epidemiological studies have shown that the intoxication was due to the ingestion of adulterated rapeseed oil purchased from street vendors. Rapeseed oil denatured with 2% aniline was imported for industrial purposes and illegally refined and distributed for human consumption (5, 6). Fatty acid anilides (2, 7) and fatty acid esters of 3-(N-phenylamino)propane-1,2-diol (PAP) * To whom correspondence should be addressed. Tel: + 34-93-400 61 21. Fax: +34-93-204 59 04. E-mail: [email protected]. 1 Abbreviations: N-AcCys, N-acetylcysteine; N-AcCysOMe, N-acetylcysteine methyl ester; ESI-MS, electrospray ionization mass spectrometry; HRMS, high-resolution mass spectrometry; HRP, horseradish peroxidase; PAA, 3-(phenylamino)-L-alanine, PAP, 3-(N-phenylamino)propane-1,2-diol; TOS, toxic oil syndrome.

(8, 9) have been identified as the major contaminants of the imputed oil batches. The differences in concentration among these contaminants in fraudulent and toxic oil batches led the TOS Management Committee to single out PAP esters as biomarkers of toxic oils. Moreover, these esters, in particular, the dioleyl esters of PAP, have been considered the putative toxic substances generated during the refining process at the plant that produced the toxic oil (9-11). Nevertheless, the toxicity mechanisms of PAP esters in biological systems remain unknown. Although the PAP contents of case oils and of the model deodorized oils used for in vitro and in vivo studies are very low (9), PAP can be generated in the gastrointestinal tract by the action of human pancreatic lipase on PAP mono- and diesters (12). Therefore, independently of the low PAP content in oil batches, said lipase activity could yield racemic PAP or an excess of a single enantiomer prone to be absorbed by the intestinal cell. Hypothetically, PAP might then be transported and further metabolized in the liver. On the basis of this aforementioned premise, the in vivo biotransformation of PAP administered intraperitoneally to two mouse strains was recently investigated. PAP, a highly polar substance, undergoes alkyl chain and

10.1021/tx050171n CCC: $30.25 © 2005 American Chemical Society Published on Web 10/01/2005

1722

Chem. Res. Toxicol., Vol. 18, No. 11, 2005

Scheme 1. Proposed Pathways for Metabolite Formation at the Aromatic Ring after i.p. Administration of Radiolabeled PAP to Two Mice Strains (14)

aromatic oxidations, whereby the majority of the resulting metabolites is excreted in urine. Some of these metabolites are potentially toxic (13, 14), including 3-(4′hydroxyphenylamino)propane-1,2-diol (1) (15). The observation that this aminophenol is prone to oxidation suggested that the corresponding quinoneimine intermediate 2 (Scheme 1)2 might be generated in the process. All efforts to isolate the quinoneimine by independent synthesis failed and 2 could only be detected by NMR. These experiments showed that, among other pathways that could involve oligomer formation, quinoneimine 2 rearranges to give the corresponding Schiff base, which in turn is facilely hydrolyzed in the reaction medium to give p-aminophenol. It is worth noting that p-aminophenol was also identified as a urine metabolite in the aforementioned in vivo assays (14). Thus, the presence of p-aminophenol may be considered as indirect evidence of the generation of the quinoneimine intermediate 2. Irrespective of this particular transformation, the potential toxicity of quinoneimines has been mainly attributed to their electrophilic character and ability to react with nucleophiles present in biological matrices (16). Taking the aforementioned precedent into account, we set about the reaction of 2 with thiols, specifically N-acetylcysteine methyl ester (N-AcCysOMe), N-acetylcysteine (N-AcCys), and GSH, the results of which are reported here. The main goals of the study were to determine the scope of the reactions, the stability of the resulting adducts, and the implications of the overall results for TOS research. In addition, we envisioned the use of these adducts as standards in forthcoming experiments planned by the TOS Management Committee in which deodorized rapeseed oil samples (17) will be administered to animal models. As detailed below, both the mercapturic acid and the GSH conjugates derived from quinoneimine 2 could be isolated and characterized, but they exhibited poor stability, in particular, to oxidation. The oxidative pathway generated new quinoneimine intermediates that evolved as a function of the reagents present in the medium. The relevance of these results concerning the potential biotransformation of xenobiotics capable of generating quinoneimine intermediates similar to compound 2 will also be discussed. 2 The generation of intermediate 2 has also been observed in incubations of PAP with P450 human isoforms (unpublished results).

Martı´nez-Cabot et al.

Materials and Methods Caution: Because PAP derivatives have been implicated in TOS, precautions should be taken when handling these compounds (work should be performed in fumehoods, and gloves should be worn). General. Glycidol (Merck, 98%) was first distilled and then stored at -20 °C. 4-Aminophenol, 4-benzyloxyaniline hydrochloride, and silver oxide were obtained from Aldrich (Germany). N-Acetyl-L-cysteine, L-glutathione (99%), and horseradish peroxidase (HRP) (E.C. 1.11.1.7) were from Sigma (St. Louis, MO). Methanol, acetonitrile, dichloromethane, and other common solvents were of HPLC grade and purchased from Merck (Darmstadt, Germany). 3-(4′-Hydroxyphenylamino)propane-1,2diol (1) was prepared according to a literature procedure (13). All steps involving quinoneimines were carried out under argon atmosphere and protected from light. Purification of metabolites was carried out in an Applied Biosystems 783 semipreparative HPLC and a Waters (Milford, MA) Prep LC 4000 system. HPLC analyses were performed on a Hewlett-Packard 1100 system equipped with a DAD detector, and a Kromasil 100 C18 (25 cm × 0.4 cm, 5 µm, Teknokroma, Spain) column was employed. Elution conditions consisted of gradients of 10 mM NH4OAc in H2O:acetonitrile at a flow rate of 1 mL/min. The detection was set at 220, 245, and 270 nm. The 1H NMR spectra (500 MHz) were recorded on a Varian Unity 500 spectrometer. Spectra were taken in neutralized CDCl3 or D2O as indicated. Chemical shifts are given in ppm related to tetramethylsilane for 1H. Electrospray ionization mass spectrometry (ESI-MS) were obtained with an HP1100 LC/MS system in the positive ionization mode. Synthesis of Quinoneimine 2. The oxidation of phenol 1 was carried out following the procedure described by Bujons et al. (14). Phenol 1 (67 mg, 0.36 mmol) and Ag2O (94 mg, 0.40 mmol) were stirred in a 1:1 methanol:acetonitrile mixture (18 mL) for 20 min at 20 °C, and the reaction was monitored by HPLC. The HPLC profile of the crude reaction mixture contained a major peak that was identified by LC/MS as the expected quinoneimine 2. ESI-MS 181 (M+) (14). Synthesis of Adducts 3-5. The crude reaction mixture containing the quinoneimine 2 was filtered through Celite to remove the metallic species and added dropwise to a solution of N-AcCysOMe (61 mg, 0.35 mmol, prepared from treatment of N-AcCys with thionyl chloride and methanol) in methanol (3 mL). The mixture was stirred for 30 min at 20 °C and monitored by HPLC. The residue obtained after solvent elimination was purified by semipreparative HPLC by using a C-8 Kromasil 100 column (250 cm × 20 cm, 5 µm, Scharlau, Spain) and 10 mM, pH 7.0, NH4OAc buffer (solvent A) and acetonitrile (solvent B) as mobile phases. The gradient used consisted of an isocratic step (from 0 to 10 min) at 5 mL/min with 100% A, followed by a linear ramp to reach 50% A (from 10 to 75 min). The UV detector was set at 220, 245, and 270 nm. The different fractions corresponding to adducts 3-5 were collected and lyophilized. From these compounds, only 5 was stable enough to obtain an NMR spectra. Compound 3 (ESI-MS): m/z 359.127; m/z for (M + H)+ C15H23N2O6S requires 359.120. Compound 4 (ESI-MS): m/z 556.139; m/z for (M + Na)+ C21H31N3NaO9S2 requires 556.150. Compound 5 1H NMR (500 MHz, CD3OD): δ 6.81 (s, 1 H; H-6′), 4.73 (dd, 3 H, J1 ) 14 Hz, J2 ) 5 Hz; CH-N), 3.81 (m, 1 H; CH-O), 3.74 (s, 6 H; COOCH3), 3.67 (s, 3 H; COOCH3), 3.55 (dd, 2 H, J1 ) 5.5 Hz; J2 ) 2.5 Hz, CH2-O), 3.22-2.95 (6 H; CH2S; 2 H; CH2N), 1.97 (s, 6 H; NHCOCH3), 1.95 (s, 3 H; NHCOCH3). ESI-MS: m/z 709.188; m/z for (M + H)+ C27H41N4O12S3 requires 709.181. Synthesis of the N-AcCys Adduct 3-[4′-Amino-3′-(Nacetylcystein-S-yl)hydroxyphenyl]propane-1,2-diol (6). The crude reaction mixture containing the quinoneimine 2 was allowed to react with N-AcCys (60 mg, 0.37 mmol) as described above. The residue obtained after solvent elimination was purified by preparative HPLC using a Perkin-Elmer (Norwalk, CT) semipreparative C18 column (25 cm × 2 cm i.d.) with 10 mM, pH 7.0, NH4OAc buffer containing 20% acetonitrile (solvent

GSH Adducts from PAP Quinoneimine Intermediates

Chem. Res. Toxicol., Vol. 18, No. 11, 2005 1723

Figure 1. Reverse phase HPLC profile of the crude reaction mixture from the reaction of quinoneimine 2 with the N-AcCysOMe. Adducts 3-5 were identified by ESI-MS. A) and acetonitrile containing 20% NH4OAc buffer (solvent B) as mobile phases. The solvent gradient consisted of an isocratic step (from 0 to 2 min) at 10 mL/min with 100% A, followed by a linear ramp to reach 95% A (from 2 to 20 min). The UV detector was set at 220 nm. Compound 6 1H NMR (500 MHz, D2O): δ 7.05 (d, 1 H, J ) 2.5 Hz; H-2′), 6.85 (dd, 1 H, J1 ) 2.5 Hz, J2 ) 9.5 Hz; H-6′), 6.81 (d, 1 H, J ) 9 Hz; H-5′), 4.28 (m, 1 H, CH-N), 3.68 (m, 1 H, CH-O), 3.61 (m, 2 H, CH2-O), 3.363.05 (4 H; CH2-N, CH2-S), 1.83 (s, 1.5 H; NHCOCH3), 1.81 (s, 1.5 H; NHCOCH3). ESI-MS: m/z 345.111; m/z for (M + H)+ C14H21N2O6S requires 345.104. Synthesis of 4-Amino-3-(N-acetylcystein-S-yl)phenol (7). p-Aminophenol (25 mg, 0.23 mmol) was treated with 53 mg (0.23 mmols) of Ag2O in a 1:1 methanol:acetonitrile solution for 15 min at 20 °C, and the reaction was monitored by HPLC. The crude reaction mixture was filtered as above and added dropwise to a solution of N-AcCys (34 mg, 0.21 mmol) in methanol (3 mL). The mixture was stirred for 20 min at 20 °C and monitored by HPLC monitoring. The residue obtained after solvent elimination was purified by semipreparative HPLC using the column and eluent mixture described for the purification of adducts 3-5. The solvent gradient consisted of an isocratic step (from 0 to 10 min) at 5 mL/min with 100% A, followed by a linear ramp to reach 50% A (from 10 to 75 min). Compound 7 1H NMR (500 MHz, D2O): δ 7.02 (d, 1 H, J ) 2 Hz; H-2′), 6.90 (d, 1 H, J ) 8.5 Hz; H-5′), 6.79 (dd, 1 H, J1 ) 8.5 Hz, J2 ) 2 Hz; H-6′), 4.31 (m, 1 H, CH-N), 3.39 (dd, 1 H, J1 ) 14 Hz, J2 ) 3.5 Hz; CH2S), 3.13 (dd, 1 H, J1 ) 14 Hz, J2 ) 8 Hz; CH2S), 1.84 (s, 3 H, CH3). ESIMS: m/z 271.07; m/z for (M + H)+ C11H15N2O4S requires 271.075. Synthesis of the Glutathione Adduct 3-[4′-Amino-3′(glutathion-S-yl)phenol]propane-1,2-diol (8). To a solution containing phenol 1 (67 mg, 0.36 mmol), GSH (176 mg, 0.57 mmol), and HRP (6.6 units) in a 0.1 M, pH 4.5, NaOAc buffer solution was added 30 µL of 30% H2O2 (0.29 mmol) dropwise. The mixture was stirred for 30 min at 20 °C and monitored by HPLC monitoring. The crude reaction was lyophilized and then purified by semipreparative HPLC using the column and eluent mixtures described for the purification of adduct 6. The solvent gradient consisted of an isocratic step (from 0 to 2 min) at 10 mL/min with 100% A, followed by a linear ramp to reach 95% A (from 2 to 20 min). Compound 8: (ESI-MS) m/z 489.165; m/z for (M + H)+ C19H29N4O9S requires 489.158. Synthesis of 4-Amino-3-(glutathion-S-yl)phenol (9). This adduct was prepared according to a literature procedure (18). To a solution containing p-aminophenol (25 mg, 0.23 mmol), GSH (140 mg, 0.45 mmol), and HRP (5.2 units) in a 0.1 M, pH 4.5, NaOAc buffer solution was added 30% H2O2 (25 µL, 0.25 mmol) dropwise. The mixture was stirred for 30 min at 20 °C and monitored by HPLC. The crude reaction mixture was lyophilized and purified by semipreparative HPLC as described above for conjugate 8. The solvent gradient consisted of an isocratic step (from 0 to 2 min) at 10 mL/min with 100% A, followed by a linear ramp to reach 95% A (from 2 to 20 min). Compound 9 1H NMR (500 MHz, D2O): δ 6.97 (d, 1 H, J ) 2.5 Hz; H-2′), 6.89 (d, 1 H, J ) 9 Hz; H-5′), 6.77 (dd, 1 H, J1 ) 2.5

Hz, J2 ) 9 Hz, H-6′), 4.45 (m, 1 H; CH-N), 3.71-3.55 (3 H; CHNH2-COOH (glu), CH2-COOH), 3.32 [dd, 1 H, J1 ) 4.5 Hz, J2 ) 14.5 Hz; CH2-S (cys)], 3.17 [dd, 1 H, J1 ) 4.5 Hz, J2 ) 14.5 Hz; CH2-S (cys)], 2.39 [m, 2 H; CH2-CONH (glu)], 2.04 (m, 2 H; CH2-CHNH). ESI-MS: m/z 437.110; m/z for (M + Na+) C16H22N4NaO7S requires 437.120. Stability Study of the Adducts. Solutions of N-AcCys and the GSH conjugates 6 and 8 (0.5 mg/mL) were independently maintained in 50 mM, pH 7.4, phosphate buffer (200 µL) for 8 h at 37 °C and monitored by HPLC using the methyl ether derived from 1 as internal standard.

Results and Discussion Reaction of Quinoneimine 2 with N-AcCysOMe. The main purpose of this experiment was to establish the reactivity of 2 with physiologically relevant thiols such as N-AcCys and GSH. However, the initial assays were carried out using the N-AcCys methyl ester since the crude reaction mixtures containing adducts from this thiol were more facilely manipulated. Quinoneimine 2 was generated from the reaction of phenol 1 with silver oxide under an inert atmosphere and the absence of light. Reverse phase HPLC monitoring of the crude reaction mixture showed a peak at a higher retention time than 1 that was identified by LC/MS as the quinoneimine intermediate 2 (14). Once the conversion had been completed, the solution containing the crude quinoneimine was added to the test thiol. The LC/MS analysis of the crude reaction mixture showed three new compounds that were identified as the monoadduct 3 (major product), the bis-adduct 4, and the tris-adduct 5 (Figure 1). The crude reaction mixture also contained p-aminophenol and the starting aminophenol 1. The formation of adduct 4 can be explained assuming that adduct 3 is oxidized to generate the corresponding quinoneimine intermediate in an analogous fashion as aminophenol 1. This intermediate would then be captured by the excess of thiol present in the reaction medium to give 4. A similar two-step pathway would account for the formation of the tris adduct 5 from 4 (see Supporting Information). It has been reported that the reactivity of quinones is determined by their substituents. Thus, a quinone with electron-donating groups is less readily oxidized than its corresponding hydroquinone (19, 20). Consequently, the hydroquinone generated from an initial nucleophilic attack can undergo subsequent oxidation and give rise to conjugates from multiple nucleophilic additions (21, 22). Moreover, the fact that the electronegativity of nitrogen is lower than that of oxygen implies a lower redox potential for quinoneimines as compared to quinones. The generation of the corresponding multiple

1724

Chem. Res. Toxicol., Vol. 18, No. 11, 2005

Martı´nez-Cabot et al.

Figure 2. (A) HPLC profile of freshly purified adduct 3. (B) HPLC profile of an aqueous solution of pure adduct 3 after 4 h at 20 °C. The insets show the ESI-MS spectra of the quinoneimine intermediate resulting from the oxidation of 3 and the final product obtained after rearrangement to the Schiff base and subsequent hydrolysis. This last product corresponds to the adduct derived from the capture of the quinoneimine generated from p-aminophenol.

addition products would therefore be favored, as was the case of the reaction of quinoneimine 2 (23). The presence of p-aminophenol in the crude reaction mixture could be rationalized by a pathway comprising the rearrangement of 2 to its corresponding Schiff base followed by hydrolysis (cf. Scheme 1) (14). As aminophenol 1 should have been completely converted to 2 by silver oxide, its presence in the final crude reaction mixture suggested that 2 had been partially reduced. Actually, the crude reaction medium contained reducing species including the thiol itself and adducts 3-5 could also act as reducing agents. Purification of the crude reaction mixture by semipreparative HPLC enabled the isolation of adducts 3-5. All three adducts were chemically unstable. Figure 2 shows the HPLC profile of the freshly purified adduct 3 in aqueous solution (A) and the profile of the same solution after 4 h at 20 °C (B). LC/MS analysis of this mixture led to the identification of two major compounds: the quinoneimine intermediate derived from the oxidation of adduct 3 and the adduct derived from the quinoneimine generated from p-aminophenol oxidation. For those cases in which the purified sample did not contain any nucleophile capable of promoting a Michael addition on the new quinoneimine, the alternative quinoneimine-Schiff base equilibrium occurred, ultimately affording the p-aminophenol adduct by subsequent hydrolysis of the imine (Scheme 2). The instability of adducts 3 and 4 prevented their isolation and storage; hence, only adduct 5 was stable enough to be characterized by 1H NMR. When considered together, these results indicated that quinoneimine 2 was capable of reacting with a thiol to form the expected adduct that results from a Michael type addition. This adduct, however, was unstable and underwent further transformation that was influenced by the presence or absence of nucleophiles in the reaction medium. We then sought to determine whether these reaction pathways could also be observed for the reaction of 2 with physiologically relevant thiols such as N-AcCys or GSH. Thus, we did not attempt to determine the

Scheme 2. Proposed Pathway for the Transformation of Adduct 3 (R ) N-AcCysOMe) to the Adduct Formally Derived from p-Aminophenola

a

The regiochemistry of the addition was not determined.

Scheme 3. Structure of the Adduct from the Reaction of Quinonimine 2 with N-AcCys

regiochemistry of the thiol addition that leads to 3 or fully characterize the p-aminophenol adduct. Reaction of Quinoneimine 2 with N-AcCys. The reaction of quinoneimine 2 with N-AcCys under the same conditions described above produced the formation of the corresponding adduct 6 as a major product (Scheme 3). In this case, the corresponding bis adduct was detected in a much lower amount. This observation suggests that

GSH Adducts from PAP Quinoneimine Intermediates

Chem. Res. Toxicol., Vol. 18, No. 11, 2005 1725 Scheme 4. Generation of Quinoneimine 2 Using the HRP/H2O2 System Followed by GSH Addition to Give Conjugate 8

Figure 3. (A) HPLC profile of freshly purified adduct 6. (B) HPLC profile of an aqueous solution of pure adduct 6 after 6 h at 20 °C. The inset shows the ESI-MS spectra of the adduct formally derived from the capture of the quinoneimine generated from p-aminophenol 7.

the mercapturic acids may be more protected from oxidation in the presence of N-AcCys. Nevertheless, the starting aminophenol 1 and p-aminophenol were also identified in the crude reaction mixture. To confirm the structure of 6, the crude reaction mixture was purified by semipreparative HPLC. The ESIMS of compound 6 showed an [M + H]+ peak at m/z 345.111, consistent with the exact mass calculated for the target mercapturic acid. The 1H NMR spectrum of 6 in D2O enabled determination of the regiochemistry of the thiol addition to the quinoneimine. Thus, the analysis of the aromatic region (7.06-6.81 ppm) showed a doublet centered at 7.04 ppm (J ) 2.5 Hz) attributed to H-2′, a doublet of doublets centered at 6.85 ppm (J1 ) 2.5 Hz, J2 ) 9.5 Hz) assigned to H-6′, and a doublet centered at 6.81 ppm (J ) 9.5 ppm) attributed to H-5′ (see Supporting Information). These data converge in assuming that the Michael type addition of N-AcCys onto 2 had taken place at the position ortho to the amino group. This assignment is consistent with that previously reported for the corresponding p-aminophenol-GSH conjugate (also vide infra) (24). It was also interesting to note the presence of the two expected diastereomers of 6 through distinct signals for the methyl groups of the N-acetyl moiety (see Supporting Information). The fact that the starting aminophenol 1 was again present in the crude reaction mixture as well as the failure to identify the dimer of N-AcCys in the mixture lent additional support to the hypothesis that aminophenol conjugates such as 6 can act as reductants to regenerate 1 from quinoneimine intermediate 2. It was anticipated that mercapturic acid 6 could be unstable in a fashion similar to the methyl esters of N-AcCys adducts. As shown in Figure 3, after 6 h of incubation at 20 °C, over 50% of the initial adduct 6 had been transformed into other products (see also Supporting Information). The major compound formed was identified by LC/MS to be the mercapturic acid 7. Again, this was the expected adduct from the conjugation of the quinoneimine derived from p-aminophenol. This result indicates that in the case of N-AcCys adduct 6, oxidation to the corresponding quinoneimine followed by isomer-

ization and hydrolysis to generate the p-aminophenol conjugate also occurs in the absence of nucleophiles in the reaction medium. This observation suggests that the mercapturic acid derived from p-aminophenol, a compound with reported nephrotoxicity (18), could be a potential PAP metabolite. To confirm the structure of 7 and to use the compound as a standard for the detection of PAP metabolites in biological samples, the conjugate was independently synthesized. Compound 7 was synthesized as described above, with the exception that commercial p-aminophenol was used instead of aminophenol 1. LC/MS analysis of the crude reaction mixture revealed that in addition to the expected adduct 7, the corresponding bis and the tris adducts were formed as well although in very minor amounts. The ESIMS of pure conjugate 7 showed an [M + H]+ peak at m/z 271.075, consistent with the exact mass expected for this compound. In this case, the analysis of the aromatic region (7.03-6.75 ppm) showed a doublet centered at 7.02 ppm (J ) 2 Hz) attributed to H-2′, a doublet centered at 6.90 ppm (J ) 8.5 Hz) assigned to H-5′, and a doublet of doublets centered at 6.79 ppm (J1 ) 8.5 Hz, J2 ) 2 Hz) attributed to H-6′ (see Supporting Information). Reaction of Quinoneimine 2 with GSH. The poor solubility of GSH in methanol forced the change of conditions for the generation of quinoneimine 2 and its further reaction with the thiol reagent. The aqueous conditions reported by Klos et al. (18) for the case of p-aminophenol in which HRP and H2O2 in a buffer solution act as oxidation agents were used. The HPLC monitoring of the crude reaction mixture showed the formation of conjugate 8 as major product (Scheme 4). The pure conjugate was isolated by semipreparative HPLC. Its structure could be confirmed by the ESI-MS spectrum where the peak [M + H+] at m/z 489.165 was in agreement with the exact mass of this adduct. Unfortunately, 8 showed a strong instability (see below), which prevented the registering of reliable 1H NMR spectra. To evaluate its relative stability, conjugate 8 was incubated in the same conditions as mercapturic acid 6. After 3 h at 37 °C, over 50% of GSH adduct 8 had been transformed to give the corresponding p-aminophenol conjugate. According to this observation, 8 appeared to be more unstable than its corresponding mercapturic acid 6 (see Supporting Information). Once again, in the absence of nucleophilic reagents, GSH conjugate 8 undergoes a transformation pathway similar to that observed for the above adducts 3 and 6 to give ultimately the more stable conjugate 9 (Figure 4). Therefore, the putative detection of this adduct in biological samples coming from treatment of model animals with deodorized rapeseed oil would constitute an indirect evidence of the formation of PAP-related quinoneimines intermediates. As for the case of mercapturic acid 7, the independent synthesis of GSH adduct 9 was carried out, but now using the H2O2/HRP for the oxidation of starting p-aminophenol. In this case, a 1:4 isomer mixture of conjugates

1726

Chem. Res. Toxicol., Vol. 18, No. 11, 2005

Martı´nez-Cabot et al. Scheme 5. General Structure of Compounds for Which the Corresponding Quinoneimine Intermediates Could undergo Various Transformations Depending on the Presence or Absence of Nucleophiles in the Reaction Medium

Figure 4. (A) HPLC profile of freshly purified GSH conjugate 8. (B) HPLC profile of an aqueous solution of pure adduct 8 after 24 h at 20 °C. (C) HPLC profile of the purified GSH conjugate 9 prepared independently from p-aminophenol.

resulting from the addition of GSH at C-2 and at C-3, respectively, was identified by HPLC and ESI-MS. It has been previously described that the major component of this mixture was the isomer resulting from the addition to the carbon atom contiguous to the amino residue (18). To confirm this assignation, the crude reaction mixture was purified by semipreparative HPLC and the 1H NMR spectrum of the major peak was registered in D2O. As anticipated, the analysis of the aromatic zone (6.90-6.75 ppm) confirmed the predicted regiochemistry for the GSH addition. Thus, H-2′ appeared as a doublet centered at 7.0 ppm (J ) 2.5 Hz), H-5′ also as a doublet at 6.9 ppm (J ) 9 Hz), and H-6′ as doublet of doublets at 6.7 ppm (J1 ) 2.5 Hz, J2 ) 9 Hz) (see Supporting Information). Although it was not possible to characterize the GSH adduct 8, the coincidence in retention time under the same HPLC conditions between the standard GSH adduct derived from p-aminophenol 9 and that one obtained from the oxidation, isomerization, and hydrolysis of conjugate 8 (Figure 4) suggested that, for the case of quinoneimine 2, the GSH addition took place also at the same relative position of the aromatic ring, that is, at ortho with respect to the amino group. In summary, major findings observed in this study were that the aminophenol 1, a metabolite from PAP, undergoes an easy oxidation to the corresponding quinoneimine 2. This quinoneimine could be trapped by the thiol reagents related to GSH to give the expected adducts. However, these adducts are not stable. They are susceptible of oxidation to give the corresponding quinoneimine intermediate that can be captured by the excess of thiol reagent to give a bis adduct. In the absence of thiol reagent, this quinoneimine establishes an equilibrium mixture with the corresponding Schiff base, which upon hydrolysis gives rise to the thiol adduct formally derived from p-aminophenol. All of these adducts have been characterized and can be used as standards for the putative identification of these conjugates in biological samples obtained from the treatment of sensitive animal species with deodorized rapeseed model oils (17). These experiments have been planned

by the WHO Management Committee and are in due course. The fact that p-aminophenol had been unambiguously identified in urine samples from two mice strains treated intraperitoneally with 14C radiolabel led PAP (14) suggests that conjugates 7 and/or 9 could be also present in biological samples from those in vivo assays. Consequently, their putative detection would constitute a further support to the hypothetical generation of toxic quinoneimine intermediates in the metabolism of PAP derivatives and, ultimately, for implicating these derivatives in the etiology of TOS epidemy. Nevertheless, this implication does not necessarily mean that the putative formation of conjugates 6-7 and or 8-9 in TOS patients would result in toxicity effects of quinoneimine type intermediates as those described for the case of paracetamol (16). Actually, no renal toxicity has been associated with TOS patients or described in the assays carried out with model animal species. It is accepted that paracetamol toxicity is due to the formation of the corresponding iminoquinone intermediate. This species can act as a highly reactive electrophile and promote covalent attachment to nucleophilic cell components. In addition, this iminoquinone can act as an oxidation reagent and induce lipid peroxidation and/or the formation of undesired disulfide bonds between protein thiol residues. In our case, the electrophilicity of iminoquinone 2 has been shown by the adduct formation with different thiol reagents. Moreover, the detection of newly generated p-aminophenol derivative 1 in those crude reaction mixtures could be a consequence of a concomitant oxidative activity of 2 on the thiols present in the reaction media. The generation of a highly reactive species such as quinoneimine 2 opens the possibility of covalent attachment to other macromolecules or cell components that could result in toxic effects. In this context, Li et al. have reported recently the site specific binding of quinones to proteins through thiol addition and addition-elimination reactions suggesting that quinones may be transferred between proteins (25). The potential extension of this reactivity behavior to the case of iminoquinones such as 2 could provide a new insight for elucidating the mechanisms accounting for the toxicity of these activated species. Work along this line is in progress in our laboratory. Finally, in addition to the potential interest of these results in the TOS area, it is possible to imagine that they could be extended to other examples of xenobiotic phase II metabolism. Although the GSH-related conjugates derived from the quinoneimine from p-aminophenol have been subject of previous studies (18, 26, 27), the case of quinoneimines generated from arylamino moieties of the general formula shown in Scheme 5, such as that of

GSH Adducts from PAP Quinoneimine Intermediates

PAP, has not been investigated in detail. As shown in this contribution, the presence of hydrogen atoms at the carbon atom adjacent to the anilido moiety makes possible the establishment of an equilibrium mixture between the quinoneimine and the corresponding Schiff base. Then, hydrolysis of the Schiff base shifts the former equilibrium to render two products: the adduct derived from p-aminophenol and a carbonyl derivative. In the present case, glyceraldehyde would be the released carbonyl compound.3 In other cases, this generated R-CHO species could be an additional risk factor. For instance, Mayeno et al. hypothesized that the structural similarity between PAP and 3-(phenylamino)-L-alanine (PAA), one of the contaminants present on the L-tryptophan batches responsible for the eosinophilia-myalgia syndrome that affected 1500 people in the United States in 1989, could link both xenobiotics to a common etiologic agent (28). Adachi et al. reported that the metabolism of PAA in rats produced, among other species, a p-aminophenol derivative of general structure as that depicted in Scheme 5 (29). The further oxidation of this species would afford an iminoquinone intermediate susceptible of reacting similarly as iminoquinone 2. Consequently, the pathway depicted in Scheme 2 should be taken into account because it is possible that for xenobiotics bearing the above general structure, mercapturic acids or GSH conjugates other than those expected from the original quinoneimine could be generated in biological tissues. If this is the case, methodology for their putative detection should be developed and specific studies on their potential toxicity should be carried out.

Acknowledgment. Financial support from the WHO/ TOS Committee is gratefully acknowledged. Dedicated to Prof. Francesc Camps on the occasion of his 70th birthday. Supporting Information Available: Scheme illustrating the proposed pathway for the evolution of adduct 3 in the presence of excess of the N-AcCysOMe, 1H NMR spectra of conjugate 6, time course plots for the evolution of conjugates 6 and 8 in aqueous solution, and comparative 1H NMR profile for the aromatic regions of conjugates 6, 7, and 9. This material is available free of charge via the Internet at http://pubs.acs.org.

References (1) Tabuenca, J. M. (1981) Toxic-allergic syndrome caused by ingestion of rapeseed oil denatured with aniline. Lancet 2, 567. (2) Grandjean, E. C. (1992) Toxic Oil Syndrome. Current Knowledge and Future Perspectives. Introduction, WHO Reg. Publ. Eur. Ser. No. 42, pp 1-4, WHO, Geneva. (3) Yoshida, S. H., German, J. B., Fletcher, M. P., and Gershwin, M. E. (1994) The toxic oil syndrome: A perspective on immunotoxicological mechanisms. Regul. Toxicol. Pharmacol. 19, 60-79. (4) Europe, WHO (2004) Toxic Oil Syndrome. Ten Years of Progress, World Health Organization Regional Office for Europe, Copenhagen, Denmark. (5) Dı´az de Rojas, F., Castro-Garcı´a, M., Abaitua-Borda, I., AlonsoGordo, J. M., Posada de la Paz, M., Kilbourne, E. M., and Tabuenca-Oliver, J. M. (1987) The association of oil ingestion with toxic oil syndrome in two convents. Am. J. Epidemiol. 25, 907911. 3 The presence of glyceraldehyde in the crude reaction mixture resulting from the evolution of an aqueous solution of conjugate 6 (pH 7.4 buffer) for 12 h at 37 °C was confirmed by formation of the corresponding 2,4-dinitrophenylhydrazone and comparison (HPLC, UV) with a standard independently prepared.

Chem. Res. Toxicol., Vol. 18, No. 11, 2005 1727 (6) Rigau-Perez, J. G. (1984) Report of a WHO Meeting: Madrid, 2125 March, 1983, Summary of Case-Control Studies and Case or Cluster Investigations in Toxic Oil Syndrome: Mass Food Poisoning in Spain, World Health Organization Regional Office for Europe, Copenhagen, Denmark. (7) Bernert, J. T., Kilbourne, E. M., Akins, J. R., Posada de la Paz, M., Meredith, N. K., Abaitua-Borda, I., and Wages, S. (1987) Compositional analysis of oil samples implicated in the Spanish toxic oil syndrome. J. Food Sci. 52, 1562-1569. (8) Vazquez Roncero, A., Janer del Valle, C., Maestro Duran, R., and Constante, E. G. (1983) New aniline derivatives in cooking oils associated with toxic oil syndrome. Lancet 2, 1024-1025. (9) Hill, R. H., Schurz., H. H., Posada de la Paz, M., Abaitua Borda, I., Philen, R. M., Kilbourne, E. M., Head, S. L., Driskell, W. J., Barr, J. R., and Needham, L. L. (1995) Possible etiologic agents for toxic oil syndrome: Fatty acid esters of 3-(N-phenylamino)1,2-propanediol. Environ. Contam. Toxicol. 28, 259-264. (10) Posada de la Paz, M., Philen, R. M., Schurz, H. H., Hill, R. H., Gimenez-Ribota, O., Go´mez de la Ca´mara, A., Kilbourne, E. M., and Abaitua-Borda, I. (1999) Epidemiologic evidence for a new class of compounds associated with toxic oil syndrome. Epidemiology 10, 130-134. (11) Posada de la Paz, M., Philen, R. M., Abaitua-Borda, I., Bernert, J. T., Bada-Gancedo, J. C., DuClos, P. J., and Kilbourne, E. M. (1996) Toxic oil syndrome: Traceback of the toxic evidence for a point source epidemic. Food Chem. Toxicol. 34, 51-257. (12) Morato´, A., Martı´nez-Cabot, A., Escabro´s, J., Bujons, J., and Messeguer, A. (2004) Studies on the toxic oil syndrome: Stereoselective hydrolysis of 3-(phenylamino)propane-1,2-diol esters by human pancreatic lipase. Chem. Res. Toxicol. 17, 889-895. (13) Ladona, M. G., Bujons, J., Messeguer, A., Ampurdane´s, C., Morato´, A., and Corbella, J. (1999) Biotransformation and clearance of 3-(phenylamino)propane-1,2-diol, a compound present in samples related to toxic oil syndrome, in C57BL/6 and A/J mice. Chem. Res. Toxicol. 12, 1127-1137. (14) Bujons, J., Ladona, M. G., Messeguer, A., Morato´, A., and Ampurdane´s, C. (2001) Metabolism of (R)- and (S)-3-(phenylamino)propane-1,2-diol in C57BL/6 and A/J-strain mice. Identification of new metabolites with toxicological significance to the toxic oil syndrome. Chem. Res. Toxicol. 14, 1097-1106. (15) Abaitua-Borda, I., and Posada de la Paz, M. (1992) Clinical Findings in Toxic Oil Syndrome: Current Knowledge and Future Perspectives, World Health Organization, Copenhagen, Denmark. (16) Bessems, J. G. M., and Vermeulen, N. P. E. (2001) Paracetamol (Acetaminophen)-induced toxicity: Molecular and biochemical mechanisms, analogues and protective approaches. Crit. Rev. Toxicol. 31, 55-138. (17) Morato´, A., Escabro´s, J., Manich, A., Reig, N., Castan˜o, Y., Abia´n, J., and Messeguer, A. (2005) On the generation and outcome of 3-(phenylamino)propane-1,2-diol derivatives in deodorized model oils related to toxic oil syndrome. Chem. Res. Toxicol. 18, 665674. (18) Klos, C. K., M., Kramer, C., and Dekant, W. (1992) p-Aminophenol nephrotoxicity: Biosynthesis of toxic glutathione conjugates. Toxicol. Appl. Pharmacol. 115, 98-106. (19) Monks, T. J., and Jones, D. C. (2002) The metabolism and toxicity of quinones, quinonimines, quinone methides, and quinonethioethers. Curr. Drug Metab. 3, 425-438. (20) Kalyanaraman, B., Premovic, P. I., and Sealy, R. C. (1987) Semiquinone anion radicals from addition of amino acids, peptides, and proteins to quinones derived from oxidation of catechols and catecholamines. An ESR spin stabilization study. J. Biol. Chem. 262 11080-11087. (21) Lau, S. S., Hill, B. A., Highet, R. J., and Monks, T. J. (1988) Sequential oxidation and glutathione addition to 1,4-benzoquinone: Correlation of toxicity with increased glutathione substitution. Mol. Pharmacol. 34, 829-836. (22) Eckert, K.-G., Eyer, P., Somenbichler, J., and Zetl, I. (1990) Activation and detoxication of aminophenols. II. Synthesis and structural elucidation of various thiol addition products of 1,4benzoquinoneimine and N-acetyl-1,4-benzoquinoneimine. Xenobiotica 20, 351-361. (23) Kutyrev, A. A. (1991) Nucleophilic reactions of quinones. Tetrahedron 47, 8043-8065. (24) Eyanagi, R., Hisanari, Y., and Shigematsu, H. (1991) Studies of paracetamol/phenacetin toxicity: Isolation and characterization of p-aminophenol-glutathione conjugate. Xenobiotica 21, 793803. (25) Li, W., Heinze, J., and Haehnel, W. (2005) Site-specific binding of quinone to proteins through thiol addition and additionelimination reactions. J. Am. Chem Soc. 127, 6140-6141.

1728

Chem. Res. Toxicol., Vol. 18, No. 11, 2005

(26) Fowler, L. M., Foster, J. R., and Lock, E. A. (1994) Nephrotoxicity of 4-amino-3-S-glutathionylphenol and its modulation by metabolism or transport inhibitors. Arch. Toxicol. 68, 15-23. (27) Lock, E. A., Cross, T. J., and Schnellmann, R. G. (1993) Studies on the mechanism of 4-aminophenol-induced toxicity to renal proximal tubules. Hum. Exp. Toxicol. 12, 383-388. (28) Mayeno, A. N., Benson, L. M., Naylor, S., Colberg-Beers, M., Puchalski, J. T., and Gleich, G. J. (1995) Biotransformation of 3-(phenylamino)-1,2-propanediol to 3-(phenylamino)alanine: A

Martı´nez-Cabot et al. chemical link between toxic oil syndrome and eosinophiliamyalgia syndrome. Chem. Res. Toxicol. 8, 911-916. (29) Adachi, J., Mio, T., Ueno, Y., Naito, T., Nishimura, A., Fujiwara, S., Sumino, K., and Tatsuno, Y. (1994) Identification of four metabolites of 3-(phenylamino)alanine, a constituent in L-tryptophan products implicated in eosinophilia-myalgia syndrome, in rats. Arch. Toxicol. 68, 500-505.

TX050171N