Effect of a Microemulsion System on Hapten-Peptide Reactivity

Aug 26, 2010 - Alternative Testing Methods for Skin Sensitization: NMR Spectroscopy for ... peptide reactivity assay (PPRA) as an alternative to anima...
0 downloads 0 Views 251KB Size
Download PDF
Chem. Res. Toxicol. 2010, 23, 1433–1441

1433

Effect of a Microemulsion System on Hapten-Peptide Reactivity Studies: Examples of Hydroxycitronellal and Citral, Fragrance Skin Sensitizers, with Glutathione Fabien Merckel,† Guillaume Bernard,† Julien Mutschler,† Elena Gime´nez-Arnau,† G. Frank Gerberick,‡ and Jean-Pierre Lepoittevin*,† Laboratoire de Dermatochimie, Institut de Chimie, CNRS and UniVersite´ de Strasbourg, 4 rue Blaise Pascal, 67070 Strasbourg, France, and Central Product Safety, Miami Valley InnoVation Center, The Procter & Gamble Company, Cincinnati, Ohio 45253 ReceiVed February 5, 2010

In chemico methods, based on the assessment of hapten reactivity toward peptides, have been proposed as alternative methods for the assessment of the skin sensitizing potential of chemicals. However, even if these approaches seem very promising, a major drawback inherent to most in vitro methods is the poor water solubility of many organic molecules in aqueous media. Thus, semiorganic media based on buffer solutions and organic cosolvents such as ethanol or acetonitrile have been proposed, but a narrow equilibrium should be found between the peptide and chemical solubilities. Microemulsions have been shown to be very valuable when reacting a lipophilic organic compound soluble in hydrophobic media with a very hydrophilic organic substance insoluble in most organic solvents. However, the reaction rate between polar and apolar reactants can be influenced, in some cases, by the use of microemulsions. On the basis of NMR experiments, we have compared the reactivity of hydroxycitronellal 1 and citral 2, two weak fragrance sensitizers of major clinical relevance, toward glutathione used as a model nucleophile in a water/acetonitrile 2:1 mixture and in a microemulsion based on chloroform/water/tert-butanol/sodium dodecylsulphate. Hydroxycitronellal and citral were found to react with the thiol group of glutathione to form, in both media, identical adducts, but the observed reaction rates were found to be different. In the case of hydroxycitronellal, the observed reaction rate of glutathione addition on the aldehyde was found to be about three times higher in the microemulsion compared to the classical semiorganic mixture. In the case of citral, the situation was more complex as the Michael addition of glutathione on the conjugated double bond was found to be significantly faster in the classical semiorganic mixture, while the subsequent reaction of a second glutathione molecule on the aldehyde was found to be faster in the microemulsion. This chloroform/water/tert-butanol/sodium dodecylsulphate microemulsion, apparently of the bicontinuous type according to DOSY data, could be of potential interest for the in chemico evaluation of lipophilic chemicals toward peptides to solve solubility problems even if the impact on the chemical rate needs to be further investigated. Introduction Since the introduction of the 3Rs concept (replace, reduce, refine) by Russel and Burch in the late fifties (1), the development of alternative methods for the prediction of toxicological end-points has been a major objective. This need has been reinforced by regulations in the European Union and will be soon within the Globally Harmonized System. The relevant EU legislation includes the Dangerous Substances Directive on dangerous chemicals (2) and the Dangerous Preparation Directive (3). With the recent adoption of the Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH) system (4), further emphasis has been placed on the use of a minimal number of animals for the assessment of toxicological endpoints. The need for alternative methods is even more urgent in the field of cosmetic products as the seventh amendment to the EU Cosmetics Directive intends to ban from the European market cosmetic products containing ingredients tested on * To whom correspondence should be addressed. Tel: +33 368 85 15 01. Fax: +33 368 85 15 27. E-mail: [email protected]. † CNRS and Universite´ de Strasbourg. ‡ The Procter & Gamble Company.

animals (5). Prediction of the skin sensitizing potential of molecules is not an exception to this tendency and is even crucial for the cosmetic industry. For many years, assessment of the sensitizing potential of chemicals has been based on animal methods, mainly the guinea pig maximization test (6), the Buehler occluded patch test (7), and more recently the local lymph node assay (8). Alternative strategies for the predictive identification of chemicals, which possess a sensitizing potential, are based on key component parts of the physiopathological process leading to an induction of sensitization. Thus, in vitro approaches based on cytokine measurement, changes in Langerhans cells or their equivalents, for example, have been proposed (9). To induce skin sensitization, a chemical has to fulfill several steps: penetrate into the epidermis across the stratum corneum, form stable association with proteins in order to create an immunogenic complex, cause dermal trauma, and be inherently immunogenic. In immunological terms, chemical allergens are haptens, low molecular weight molecules unable to induce an immune response by themselves but able to react with epidermal proteins to form antigenic structures. Indeed, the very first step of the sensitization process is not biological but chemical and could be used for the development of alternative in chemico

10.1021/tx100043b  2010 American Chemical Society Published on Web 08/26/2010

1434

Chem. Res. Toxicol., Vol. 23, No. 9, 2010

Chart 1. Chemical Structures of Hydroxycitronellal 1, 1-[13C]-Hydroxycitronellal 1a, Citral 2, 1-[13C]-Citral 2a and 3-[13C]-Citral 2b

methods based on the assessment of hapten reactivity (10). Investigators have thus been interested in pursuing whether measuring a chemical’s reactivity could be used to develop an in chemico quantitative peptide-based reactivity assay that would have utility for screening a chemical’s skin sensitization potency as defined in the LLNA1 (11-16). Thus, it has been shown that glutathione (GSH) or synthetic peptides containing nucleophilic residues could be very valuable tools for the detection of sensitizing molecules and even their classification according to potency categories (weak, moderate, strong, and extreme sensitizers) as defined by the European Centre for Ecotoxicology and Toxicology of Chemicals (ECETOC) Task Force (17). Even if these approaches seem very promising, a major drawback inherent to most in vitro methods is the poor water solubility of many organic molecules in aqueous media. If this point is very critical in the case of cell based assays for which the addition of an organic cosolvent such as ethanol or dimethylsulfoxide is very limited due to cytotoxicity, more flexibility is foreseen for in chemico approaches. Thus, semiorganic media based on buffer solutions and organic cosolvents such as ethanol or acetonitrile have been proposed, but a narrow equilibrium should be found between the peptide and chemical solubilities. Another option could be the use of microemulsions that are thermodynamically stable, fluid, optically clear dispersions of two immiscible liquids (18). They are based on the use of a mixture of surfactants and cosurfactants to lower the interfacial tension of an organic/water mixture, and depending on the composition, the size of microemulsions can range from 100 to 1000 Å. Many combinations have thus been proposed on the basis of halogenated, aromatic, or hydrocarbon solvents in association with ionic or nonionic surfactants together with alcohol based cosurfactants. In addition, microemulsions have been shown to be very valuable when reacting a lipophilic organic compound soluble in hydrophobic media with a very hydrophilic organic substance insoluble in most organic solvents (19, 20). However, the reaction rate between polar and apolar reactants can be influenced (21). In this article, we report our preliminary investigations on the reaction mechanisms of hydroxycitronellal 1 and citral 2 (Chart 1), two weak fragrance skin sensitizers (22), with GSH used as a classical model nucleophile in a semiorganic medium and in a microemulsion.

Experimental Procedures Chemistry. [13C]-NaCN, deuterated sodium dodecyl sulfate (SDS), and deuterated solvents were purchased from Euriso-Top (Saint Aubin, France). All other chemicals were purchased from 1 Abbreviations: ACD, allergic contact dermatitis; DOSY, diffusion ordered spectroscopy; HMBC, heteronuclear multiple-bond correlation; HSQC, heteronuclear single-quantum correlation; LLNA, local lymph node assay; SDS, sodium dodecyl sulfate.

Merckel et al. Sigma-Aldrich (Saint Quentin Fallavier, France). All chemicals and solvents were used as delivered. Air- or moisture-sensitive reactions were conducted in flame-dried glassware under an atmosphere of dry argon. The reactions were followed by TLC, performed on 0.25 mm silica gel plates (60F254; Merck, Darmstadt, Germany). After migration, the TLC plates were inspected under UV light (254 nm) or sprayed with a solution containing phosphomolybdic acid (5 g), cerium(IV) sulfate (2 g), and sulphuric acid (12 mL) in water (188 mL), followed by heating. Column chromatography purifications were performed on silica gel 60 (Merck, Geduran, 40-63 µm). 1H and 13C NMR spectra were recorded on a Bruker Avance 300 spectrometer at 300 and 75 MHz, respectively. The chemical shifts (δ) are reported in ppm and are indirectly referenced to TMS via the solvent signal (acetonitrile-d3 (CD3CN): δ 1H ) 1.94, δ 13C ) 1.32, 118.26. CDCl3: δ 1H ) 7.26, δ 13C ) 77.16). In the 1H NMR spectra, multiplicities are denoted as s (singlet), d (doublet), t (triplet), q (quadruplet), and br (broad). The different types of carbon in the structures were identified by the DEPT-135 technique. 2,6-Dimethylheptan-1,6-diol (4). A solution of dihydromyrcenol 3 (2 mL, 9.93 mmol) in dichloromethane (100 mL) at -78 °C was saturated with ozone until a persistent blue color was obtained. The excess of ozone was then removed by bubbling a stream of oxygen for 15 min, and dimethylsulfide (2.9 mL, 38.73 mmol, 3.9 equiv) was added. The reaction mixture was stirred at room temperature for 12 h, concentrated under reduced pressure, and the residue takenup by diethyl ether (30 mL), washed with water (3 × 30 mL) and brine (3 × 30 mL), dried over MgSO4, and filtered. Organic solvents were removed under reduced pressure and the residue taken up in dry tetrahydrofuran (20 mL). To this solution was slowly added, at -10 °C, a suspension of lithium aluminum hydride (753 mg, 19.84 mmol, 2 equiv) in dry tetrahydrofuran (20 mL). The reaction mixture was stirred at room temperature for 6 h, cooled down at 5 °C, and water (0.8 mL) was slowly added. After 15 min, a solution of NaOH (10% in water, 0.8 mL) was added followed after 15 min by an addition of water (2.4 mL) to give a white suspension. The organic layer was washed with brine (30 mL), dried over MgSO4, filtered, and concentrated under reduced pressure. The crude oil was purified by column chromatography on silica gel (pentane/AcOEt, 4/6) to give 4 (1.44 g, 8.99 mmol, 91% yield) as a yellowish oil. 1H NMR (300 MHz, CDCl3): δ 0.92 (d, 3H, 3JHH ) 6.8 Hz, -CH(CH3)CH2OH), 1.20 (s, 6H, -C(CH3)2OH), 1.42 (m, 6H, -CHCH2CH2CH2-), 1.61 (m, 1H, -CH(CH3)CH2OH), 3.45 (m, 2H, -CH(CH3)CH2OH). 13C NMR (75 MHz, CDCl3): δ 16.6 (-CH(CH3)CH2OH), 21.6 (-CHCH2CH2CH2-), 29.3 (-C(CH3)2OH), 33.8 (-CHCH2CH2CH2-), 35.7 (-CH(CH3)CH2OH), 44.1 (-CHCH2CH2CH2-), 68.2 (-CH2OH), 71.0 (-C(OH)-). 7-Bromo-2,6-dimethylheptan-2-ol (5). To a solution of compound 4 (1 g, 6.28 mmol) in dry dichloromethane (70 mL) were added, at -5 °C, triphenylphosphine (3.3 g, 12.55 mmol, 2 equiv) and tetrabromomethane (2.1 g, 6.33 mmol, 1 equiv). The orange reaction mixture was stirred at -5 °C for 20 min and then for 3 h at room temperature. Solvents were removed under reduced pressure and the residue filtered on silica and washed with diethyl ether (200 mL). After concentration under reduced pressure, the crude product was purified by column chromatography on silica gel (hexane/ AcOEt, 8/2) to give 5 (1.08 g, 4.83 mmol, 77% yield) as a yellow oil. 1H NMR (300 MHz, CDCl3): δ 0.97 (d, 3H, 3JHH ) 6.8 Hz, -CH(CH3)CH2Br), 1,15 (s, 6H, -C(CH3)2OH), 1.37 (m, 6H, -CHCH2CH2CH2-), 1.78 (m, 1H, -CH(CH3)CH2Br), 3.32 (m, 2H, -CH(CH3)CH2Br). 13C NMR (75 MHz, CDCl3): δ 18.7 (-CH(CH3)CH2Br), 21.6 (-CHCH2CH2CH2-), 29.2 (-C(CH3)2OH), 35.2 (-CH(CH3)CH2Br), 35.3 (-CHCH2CH2CH2-), 41.4 (-CH2Br), 43.6 (-CHCH2CH2CH2-), 70.7 (-C(OH)-). 1-[13C]-3,7-Dimethyl-7-hydroxyoctanenitrile (6). To a solution of compound 5 (683 mg, 3.06 mmol) in freshly distilled dimethylformamide (15 mL) was added [13C]-sodium cyanate (153.2 mg, 3.06 mmol, 1 equiv). The reaction mixture was stirred at 80 °C for 1.5 h, cooled down to room temperature, and hydrolyzed with water (50 mL). The aqueous phase was extracted with diethyl ether (4 × 70 mL) and the combined organic layers washed with brine (200 mL), dried over MgSO4, filtered, and concentrated under reduced

ReactiVity of Skin Sensitizers in a Microemulsion pressure. The crude product was purified by column chromatography on silica gel (petroleum ether/AcOEt, 6/4) to give 6 (517.6 mg, 3.04 mmol, 99% yield) as a colorless oil. 1H NMR (300 MHz, CDCl3): δ 1.03 (d, 3H, 3JHH ) 6.7 Hz, -CH(CH3)CH213CN), 1.18 (s, 6H, -C(CH3)2OH), 1.41 (m, 6H, -CHCH2CH2CH2-), 1.83 (m, 1H, -CH(CH3)CH213CN), 2,27 (m, 2H, -CH(CH3)CH213CN). 13C NMR (75 MHz, CDCl3): δ 19.4 (-CH(CH3)CH213CN), 21.5 (-CHCH2CH2CH2-), 24.4 (d, 1JCC ) 55.5 Hz, -CH213CN), 29.2 (CH3), 29.3 (CH3), 30.4 (-CH(CH3)CH213CN), 36.3 (-CHCH2CH2CH2-), 43.6 (-CHCH2CH2CH2-), 70.7 (-C(OH)-), 118.9 (-13CN). 1-[13C]-3,7-Dimethyl-7-hydroxyoctanal (1-[13C]-Hydroxycitronellal) (1a). To a solution of compound 6 (569 mg, 3.34 mmol) in dry dichloromethane (30 mL) at -70 °C was slowly added a solution of diisobutylaluminium hydride (1 M in CH2Cl2, 10.03 mmol, 3 equiv). The reaction mixture was stirred at room temperature for 3 h and then hydrolyzed with a tartrate buffer (30 mL) and kept under stirring for an additional 12 h. The aqueous layer was extracted with diethyl ether (3 × 30 mL) and the combined organic layers washed with brine (75 mL), dried over MgSO4, filtered, and concentrated under reduced pressure. The crude oil was purified by column chromatography on silica gel (pentane/Et2O, 4/6) to give 1a (363 mg, 2.09 mmol, 63% yield) as a colorless oil. 1H NMR (300 MHz, CDCl3): δ 0.95 (d, 3H, 3JHH ) 6.6 Hz, -CH(CH3)CH213CHO), 1.20 (s, 6H, -C(CH3)2OH), 1.42 (m, 6H, -CHCH2CH2CH2-), 2.25 (m, 1H, -CH(CH3)CH213CHO), 2.37 (m, 2H, -CH213CHO), 9.75 (d, 1H, 1JHC ) 169.5 Hz, -13CHO). 13C NMR (75 MHz, CDCl3): δ 19.9 (-CH(CH3)CH213CHO), 21.7 (-CHCH2CH2CH2-), 28.1 (-CH(CH3)CH213 CHO), 29.3 ((-C(CH3)2OH), 37.3 (d, 3JCC ) 3,1 Hz, -CHCH2CH2CH2-), 43.9 (-CHCH2CH2CH2-), 50.8 (d, 1JCC ) 38.8 Hz, -CH213CHO), 70.9 (-C(OH)-), 203.0 (-13CHO). 1-[13C]-3,7-Dimethyloct-2,6-dien-1-nitrile (8a). To a solution of freshly distilled diisopropylamine (3.33 mL, 23.78 mmol, 4 equiv) in dry tetrahydrofuran (20 mL) was slowly added a solution of n-butyllithium (1.44 M in diethylether, 16.5 mL, 23.78 mmol, 4 equiv) at -20 °C. The reaction medium was stirred for 1 h at -20 °C, then cooled down at -70 °C. A solution of 1-[13C]-acetonitrile (500 mg, 11.98 mmol, 2 equiv) in dry tetrahydrofuran (8 mL) was added. After 30 min, the reaction medium was allowed to warm up at 0 °C, and a solution of diethylchlorophosphate (11.89 mmol, 2 equiv) in tetrahydrofuran (8 mL) was added. After 1 h of stirring at 0 °C, a solution of 6-methyl-5-hepten-2-one 7 in dry tetrahydrofuran (8 mL) was added. Then, the mixture was kept under stirring for an additional 3 h and then hydrolyzed with water (30 mL). The aqueous layer was extracted with diethyl ether (3 × 40 mL) and the combined organic layers washed with brine (150 mL), dried over MgSO4, filtered, and concentrated under reduced pressure. The crude oil was purified by column chromatography on silica gel (pentane/Et2O: 9/1) to give 8a (866 mg, 5.76 mmol, 97% yield) as a colorless oil. 1H NMR (300 MHz, CDCl3): E isomer (66%), δ 1.58 (s, 3H, Z-CH3), 1.67 (s, 3H, E-CH3), 2.03 (s, 3H, -CCH3), 2.17 (m, 4H, -CHCH2CH2-), 5.00 (m, 1H, -CHCH2CH2-), 5.08 (d, 1H, 2JHC ) 1.1 Hz, -CH13CN). Z isomer (34%): δ 1.60 (s, 3H, Z-CH3), 1.67 (s, 3H, E-CH3), 1.89 (s, 3H, -CCH3), 2.17 (m, 2H, -CHCH2CH2-), 2.41 (t, JHH ) 7.6 Hz, 2H, -CHCH2CH2-), 5.00 (m, 1H, -CHCH2CH2-), 5.08 (d, 1H, 2 JHC ) 1.1 Hz, -CH13CN). 13C NMR (75 MHz, CDCl3): E isomer: δ 17.6 (Z-CH3), 21.0 (d, 3JCC ) 4.3 Hz, -CCH3), 25.6 (E-CH3), 25.6 (-CHCH2CH2-), 38.5 (d, 3JCC ) 6.8 Hz, -CHCH2CH2-), 95.7 (d, 1JCC ) 78.3 Hz, -CH13CN), 117.2 (-13CN), 122.1 (-CHCH2CH2-), 133.1 (-C(CH3)2), 164.9 (-CH2CCH3). Z isomer: δ 17.6 (Z-CH3), 22.9 (d, 3JCC ) 8.0 Hz, -CCH3), 25.6 (E-CH3), 26.1 (-CHCH2CH2-), 36.2 (d, 3JCC ) 3.1 Hz, -CHCH2CH2-), 96.2 (d, 1JCC ) 78.3 Hz, -CH13CN), 116.9 (-13CN), 122.1 (-CHCH2CH2-), 133.2 (-C(CH3)2), 165.1 (-CH2CCH3). 1-[13C]-3,7-Dimethyloct-2,6-dienal (1-[13C]-Citral) (2a). To a solution of compound 8a (866 mg, 5.76 mmol) in dry dichloromethane (50 mL) at -70 °C was slowly added a solution of diisobutylaluminium hydride (1 M in dichloromethane, 11.53 mL,

Chem. Res. Toxicol., Vol. 23, No. 9, 2010 1435 11.53 mmol, 2 equiv). The reaction mixture was stirred at room temperature for 2 h and then hydrolyzed with a tartrate buffer (50 mL) and kept under stirring for an additional 12 h. The aqueous layer was extracted with dichloromethane (3 × 50 mL) and combined organic layers washed with brine (150 mL), dried over MgSO4, filtered, and concentrated under reduced pressure. The crude oil was purified by column chromatography on silica gel (pentane/Et2O: 9/1) to give 2a (587 mg, 3.83 mmol, 67% yield) as a yellow oil. 1H NMR (300 MHz, CDCl3): E isomer (66%), δ 1.57 (s, 3H, Z-CH3), 1.64 (s, 3H, E-CH3), 2.13 (m, 5H, -CCH3 and -CHCH2CH2-), 2.19 (m, 2H, -CHCH2CH2-), 5.03 (m, 1H, -CHCH2CH2-), 5.82 (d, 3JHH ) 9.0 Hz, 1H, -CH13CHO), 9.68 (dd, 1H, 1JHC ) 169.3 Hz, 3JHH ) 8.0 Hz, -13CHO). Z isomer (34%): δ 1.55 (s, 3H, Z-CH3), 1.64 (s, 3H, E-CH3), 1.94 (d, 3H, 4 JHH ) 0.9 Hz, -CCH3), 2.13 (m, 2H, -CHCH2CH2-), 2.19 (m, 2H, -CHCHCH2-), 5.05 (m, 1H, -CHCH2CH2-), 5.85 (d, 1H, 3 JHH ) 9.0 Hz, -CH13CHO), 9.58 (dd, 1H, 1JHC ) 169.7 Hz, 3JHH ) 8.3 Hz, -13CHO). 13C NMR (75 MHz, CDCl3): E isomer, δ 17.5 (d, JCC ) 4.9 Hz, -CCH3), 17.6 (-CH3), 25.6 (-CH3), 26.9 (-CHCH2CH2), 40.6 (d, JCC ) 5.5 Hz, -CHCH2CH2), 122.2 (-CHCH2CH2), 127.7 (d, JCC ) 55.5 Hz, -CH13CHO), 132.8 (-C(CH3)2), 163.7 (-CH2CCH3), 191.2 (-13CHO). Z isomer: δ 17.6 (-CH3), 25.0 (d, JCC ) 5.6 Hz, -CCH3), 25.6 (-CH3), 25.6 (-CHCH2CH2), 32.5 (d, JCC ) 3.7 Hz, -CHCH2CH2), 122.5 (-CHCH2CH2), 128.9 (d, JCC ) 55.5 Hz, -CH13CHO), 133.6 (-C(CH3)2), 163.7 (-CH2CCH3), 191.2 (-13CHO). 1-[13C]-5-Methylhex-4-en-1-nitrile (10b). To a suspension of 13 [ C]-sodium cyanide (500 mg, 9.9 mmol, 1 equiv) in freshly distilled dimethylformamide (15 mL) was added a solution of compound 9 (1.6 g, 9.9 mmol, 1 equiv) in dry dimethylformamide (8 mL). The reaction mixture was stirred at room temperature for 12 h and hydrolyzed with water (50 mL). The aqueous phase was extracted with pentane (3 × 40 mL) and the combined organic layers washed with brine (80 mL), dried over MgSO4, filtered, and concentrated under reduced pressure. The crude product was purified by column chromatography on silica gel (pentane/Et2O, 9/1) to give 10b (963 mg, 8.74 mmol, 88% yield) as a colorless oil. 1H NMR (300 MHz, CDCl3): δ 1.60 (s, 3H, E-CH3), 1.68 (s, 3H, Z-CH3), 2.28 (m, 4H, -CHCH2CH2-), 5.09 (m, 1H, -CHCH2CH2-). 13C NMR (75 MHz, CDCl3): δ 17.4 (d, 1JCC ) 55.5 Hz, -CH213CN), 17.6 (Z-CH3), 23.9 (d, 2JCC ) 1.9 Hz, -CHCH2CH2), 25.4 (ECH3), 119.5 (-13CN), 120.1 (d, 3JCC ) 3.1 Hz, -CHCH2CH2), 135.2 (-C(CH3)2). 3-[13C]-3,7-Dimethyloct-2,6-dien-1-nitrile (8b). A solution of compound 10b (963 mg, 8.74 mmol, 1 equiv) in freshly distilled diethylether (15 mL) was slowly added to a solution of methyllithium (1.41 M in diethylether, 9.92 mL, 13.99 mmol, 1.6 equiv) at -10 °C over a period of 40 min. Then the mixture was allowed to warm up at 0 °C and stirred for 3 h. The reaction medium was hydrolyzed with a solution of 3 N chlorhydric acid (20 mL), stirred at 5 °C for an additional 40 min, then neutralized with solid sodium carbonate. The aqueous phase was extracted with diethylether (3 × 60 mL) and the combined organic layers washed with brine (150 mL), dried over MgSO4, filtered, and concentrated under reduced pressure. The crude aldehyde as a yellowish oil was used without further purification. A solution of diethyl cyanomethylphosphonate (2.02 mL, 12.24 mmol, 1.4 equiv) in freshly distilled tetrahydrofuran (9 mL) was slowly added to a solution of sodium hydride (60% in mineral oil, 490 mg, 12.24 mmol, 1.4 equiv) in dry tetrahydrofuran (12 mL) at -10 °C. The reaction medium was stirred for 1 h at room temperature, until it was completely translucent. Then, a solution of the previously prepared aldehyde in dry tetrahydrofuran (9 mL) was added, and the reaction medium was stirred at room temperature for 1 h and hydrolyzed with water (60 mL). The aqueous phase was extracted with pentane (3 × 150 mL) and the combined organic layers washed with brine (150 mL), dried over MgSO4, filtered, and concentrated under reduced pressure. The crude product was purified by column chromatography on silica gel (pentane/Et2O, 95/5) to give 8b (491 mg, 3.24 mmol, 37% yield) as a colorless oil. 1H NMR (300 MHz, CDCl3): E isomer (65%), δ 1.57 (s, 3H, Z-CH3), 1.65 (s, 3H, E-CH3), 2.00 (d, 3H, 2JHC ) 6.2

1436

Chem. Res. Toxicol., Vol. 23, No. 9, 2010

Hz, -13CCH3), 2.15 (m, 4H, -CHCH2CH2-), 4.99 (m, 1H, -CHCH2CH2-), 5.07 (m, 1H, -CHCN). Z isomer (35%): δ 1.59 (s, 3H, Z-CH3), 1.65 (s, 3H, E-CH3), 1.86 (d, 3H, 2JHC ) 6.6 Hz, -13CCH3), 2.15 (m, 2H, -CHCH2CH2-), 2.38 (m, 2H, -CHCH2CH2-), 4.99 (m, 1H, -CHCH2CH2-), 5.07 (m, 1H, -CHCN). 13C NMR (CDCl3- 75 MHz): E isomer: δ 17.6 (Z-CH3), 21.0 (d, 1JCC ) 40.7 Hz, -13CCH3), 22.5 (E-CH3), 25.5 (-CHCH2CH2-), 38.2 (d, 1JCC ) 40.0 Hz, -CHCH2CH2-), 94.6 (d, 1JCC ) 74.0 Hz, -CHCN), 117.1 (-CN), 122.1 (-CHCH2CH2-), 133.0 (-C(CH3)2), 164.9 (-CH213CCH3). Z isomer: δ 17.6 (Z-CH3), 21.4 (d, 1JCC ) 41.3 Hz, -13CCH3), 22.5 (E-CH3), 26.0 (-CHCH2CH2-), 35.9 (d, 1JCC ) 40.1 Hz, -CHCH2CH2-), 95.6 (d, 1JCC ) 73.4 Hz, -CHCN), 117.0 (-CN), 122.1 (-CHCH2CH2-), 133.0 (-C(CH3)2), 164.9 (-CH213CCH3). 3-[13C]-3,7-Dimethyloct-2,6-dienal (3-[13C]-Citral) (2b). The same procedure as that used for the synthesis of 2a was used, starting from 8b (491 mg, 3.27 mmol) to give 2b (356 mg, 2.31 mmol, 71% yield) as a yellow oil. 1H NMR (CDCl3 - 300 MHz): E isomer (65%), δ 1.59 (s, 3H, Z-CH3), 1.67 (s, 3H, E-CH3); 2,14 (dd, 3H, 4JHH ) 1.3 Hz, 2JHC ) 6.2 Hz, -13CCH3), 2.21 (m, 4H, -CHCH2CH2-), 5.05 (m, 1H, -CHCH2CH2-), 5.85 (d, 1H, 3JHH ) 8.0 Hz, -CHCHO), 9.96 (d, 1H, 3JHH ) 8.0 Hz, -CHO). Z isomer (35%): δ 1.58 (s, 3H, Z-CH3), 1.67 (s, 3H, E-CH3), 1.96 (dd, 3H, 4JHH ) 1.3 Hz, 2JHC ) 6.4 Hz, -13CCH3), 2.21 (m, 2H, -CHCH2CH2-), 2.56 (m, 2H, -CHCHCH2-), 5.08 (m, 1H, -CHCH2CH2-), 5.85 (d, 1H, 3JHH ) 8.0 Hz, -CHCHO), 9.87 (d, 1H, 3JHH ) 8.2 Hz, -CHO). 13C NMR (CDCl3 - 75 MHz): E isomer, δ 17.4 (d, 1JCC ) 39.7 Hz, -13CCH3), 17.5 (Z-CH3), 25.6 (E-CH3 and -CHCH2CH2-), 40.3 (d, 1JCC ) 39.5 Hz, -CHCH2CH2-), 122.5 (d, 3JCC ) 3.1 Hz, -CHCH2CH2-), 126.9 (d, 1JCC ) 67.2 Hz, -CHCHO), 132.9 (-C(CH3)2), 163.8 (-CH213CCH3), 191.3 (-CHO). Z isomer: δ 17.5 (Z-CH3), 24.8 (d, 1JCC ) 40.1 Hz, -13CCH3), 25.6 (E-CH3), 25.7 (-CHCH2CH2-), 32.8 (d, 1JCC ) 38.8 Hz, -CHCH2CH2-), 122.2 (d, 3JCC ) 2.5 Hz, -CHCH2CH2-), 128.2 (d, 1JCC ) 66.6 Hz, -CHCHO), 133.6 (-C(CH3)2), 163.8 (-CH213CCH3), 190.8 (-CHO). Reaction of 1-[13C]-Hydroxycitronellal 1a, 1-[13C]-Citral 2a, and 3-[13C]-Citral 2b toward an Excess of Glutathione. Semiorganic Conditions. 1-[13C]-Hydroxycitronellal 1a (1 mg, 5.78 mmol, 1 equiv) was dissolved in a mixture of deuterated acetonitrile (195 µL), water (395 µL), and tert-butanol (10 µL). Glutathione (17.8 mg, 57.8 mmol, 10 equiv) was added to the medium and the whole mixture introduced to a NMR tube to follow the reaction by 13 C NMR. 1-[13C]-Citral 2a or 3-[13C]-citral 2b (1 mg, 6.53 mmol, 1 equiv) was dissolved in a mixture of deuterated acetonitrile (195 µL), water (395 µL), and tert-butanol (10 µL). Glutathione (20.0 mg, 65.3 mmol, 10 equiv) was added to the medium and the whole mixture introduced to a NMR tube to follow the reaction by 13C NMR. Microemulsion Conditions. The microemulsion (600 µL) was prepared by mixing deuterated sodium dodecyl sulfate (60 mg), deuterated water (200 µL), deuterated tert-butanol (200 µL), and deuterated chloroform (200 µL) until only one phase remained. 1-[13C]-Hydroxycitronellal 1a (1 mg, 5.78 mmol, 1 equiv) was dissolved in the microemulsion, and glutathione (17.8 mg, 57.8 mmol, 10 equiv) was added to the medium. 1-[13C]-Citral 2a or 3-[13C]-citral 2b (1 mg, 6.53 mmol, 1 equiv) was dissolved in the microemulsion, and glutathione (20.0 mg, 65.3 mmol, 10 equiv) was added to the medium. Each mixture was introduced to a NMR tube that was sealed and the reaction followed by 13C NMR. NMR Experiments and Structure Assignment. The reactions were followed by monodimensional 13C NMR on a Bruker Avance 300 spectrometer at 75 MHz (NS ) 800; SWH ) 20325.203 Hz; FIDRES ) 0.620 Hz; AQ ) 0.806 s; D1 ) 0.600 s; pulse ) 30°). Chemical shifts (δ) are reported in ppm in comparison to tetramethylsilane, using the residual signal of acetonitrile (1H, δ ) 21.94 ppm; 13C, δ ) 118.26 ppm) or chloroform (1H, δ ) 7.26 ppm; 13 C, δ ) 77.16 ppm) as the internal standard. 13C peak integrations were calculated using the signal of tert-butanol (δ ) 29.23 ppm) or deuterated tert-butanol (δ ) 68.46 ppm) as the internal standard

Merckel et al. Scheme 1. Synthetic Route for the Preparation of 1-[13C]-Hydroxycitronellal 1a

with NMRTEC/NMR notebook software (version 2.0, NMRtec, Illkirch-Graffenstaden, France). The structure of the products formed during the reactions was assigned by heteronuclear single-quantum correlation (HSQC) and heteronuclear multiple-bond correlation (HMBC) experiments. The bidimensional 1H-13C heteronuclear experiments were carried out on a Bruker Avance 400 spectrometer (1H, 400 MHz; 13C, 100 MHz). Chemical shifts were compared with those calculated using ACD/CNMR and ACD/HNMR Predictor software (version 6.0, ACD/Laboratories, Toronto, Canada). NMR diffusion experiments were carried out on a 500 MHz Avance Bruker NMR spectrometer equipped with a 10A pulsed-gradient unit capable of producing magnetic field pulse gradients in the z direction of about 69 G cm-1. All experiments were carried out using a 2.5 mm microtube inverse probe. The trapezoidal bipolar pulsed gradients (g) were incremented from 0 to 48 G cm-1, and their total duration (δ) in all cases was 3.5 ms. The pulse gradient separation (∆) was 150 ms. The measurements were performed using the stimulated echo diffusion sequence (23, 24). All measurements were performed at 298.0 K. DOSY NMR spectra were processed with the software NMRNotebook-DOSY module (NMRTec trademark), using inverse Laplace transform stabilized by maximum entropy calculation (25).

Results Synthesis. NMR techniques in association with 13C-labeled molecules have been shown to be very efficient tools for the investigation of hapten-protein interaction mechanisms (26-28). Hydroxycitronellal 1 (Chart 1) was prepared 13C labeled at the reactive aldehyde function following a straightforward hemisynthesis starting from dihydromyrcenol 3 (Scheme 1). This monoterpene was subjected to an ozonolysis reaction at -78 °C in dichloromethane followed by dimethyl sulfide reduction to give 2,6-dimethyl-6-hydroxyheptanal together with the corresponding acid. The crude aldehyde was thus directly treated without purification with 2 equivalents of lithium aluminum hydride in tetrahydrofuran to afford the diol 4 in very good yields (91%). The primary alcohol was then selectively converted, with 77% yield, into a bromide function by reaction with tetrabromomethane in the presence of 2 equivalents of triphenylphospine. The carbon 13 was easily introduced by an SN2 reaction with [13C]-sodium cyanide in dimethylformamide at 80 °C to afford with 99% yield the nitrile 6, a precursor of hydroxycitronellal 1a. The treatment of 6 with 3 equivalents of diisobutylaluminium hydride at -70 °C in dichloromethane, followed by hydrolysis of the intermediate imine with a tartrate buffer, gave 1a with 63% yield. Citral 2 (Chart 1) was prepared 13C labeled at position 3 (Michael acceptor) and position 1 (aldehyde function) following

ReactiVity of Skin Sensitizers in a Microemulsion Scheme 2. Synthetic Route for the Preparation of 1-[13C]-Citral 2a and 3-[13C]-Citral 2b

the same strategy (Scheme 2). Diethylchlorophosphate was treated with the anion derived from the treatment of 1-[13C]acetonitrile with in situ prepared lithium diisopropylamide, to form a phosphonate that was able to react with ketone 7 to form the intermediate 8a in very good yields (97%). The cyanide 8a was readily reduced by diisobutylaluminium hydride into the corresponding conjugated imine. Further hydrolysis under acidic conditions gave citral 2a labeled in position 1 with 67% yield. Introduction of a carbon 13 in position 3 was achieved according to the same strategy. Thus, bromide 9 was converted into cyanide 10b by treatment with [13C]-sodium cyanide in dimethylformamide (88% yield). Addition of methyllithium on 10b followed by an acidic hydrolysis gave the aldehyde 7b, which was converted without isolation and purification into the conjugated cyanide 8b with 37% yield. Compound 8b was then converted into citral 2b (71% yield) according to the abovedescribed procedure. Reaction Mechanisms. [13C]-Hydroxycitronellal 1a and 13 [ C]-citral 2a-b were reacted with 10 equivalents of GSH, used as the model nucleophile, in two different systems. Semiorganic conditions consisted of a 1:2 mixture of acetonitrile/ water, and microemulsion conditions consisted of a microemulsion based on water/chloroform using sodium dodecyl sulfate as surfactant and tert-butanol as cosurfactant. The selection of this microemulsion system was essentially based on the easy access to all constituents in their deuterated form. [13C]-Hydroxycitronellal 1a. First, an equilibrium was observed between the carbonyl function of 1a at 207.4 ppm in the semiorganic medium and 204.3 ppm in the microemulsion, respectively, and its hydrated form 11 at 89.3 ppm in the semiorganic medium and 90.0 ppm in the microemulsion, respectively, even if barely detectable in the microemulsion system. Thirty minutes after the addition of GSH, a first series of 4 signals (75.18-77.16 ppm in the semiorganic medium/ 75.95-77.77 ppm in the microemulsion), assigned to the 4 different diastereoisomers of hemithioacetals 12a-d, formed by the addition of the thiol group of GSH on the carbonyl function of 1a, were detected (Scheme 3). After 24 h, a second series of 4 new signals (65.09-65.90 ppm in the semiorganic medium/65.59-65.94 ppm in the microemulsion), assigned to the 4 different diastereoisomers of the intramolecular cyclization products 13a-d, were observed. Then, signals of the hemithioacetals 12a-d gradually disappeared with time to the benefit of signals corresponding to the cyclic adducts 13a-d (Figure 1). After 46 days in the semiorganic conditions and 15 days in the microemulsion system, the only signals remaining were those

Chem. Res. Toxicol., Vol. 23, No. 9, 2010 1437 Scheme 3. Reaction of 1a with GSH (10 Equiv) in a Semi-Organic or Microemulsion System and Structures of Adducts Formed 12a-d and 13a-d

of the cyclic adducts 13a-d. Structures of 11, 12, and 13 were confirmed by HMBC/HSQC experiments. [13C]-Citral 2b. Thirty minutes after the addition of GSH, a new broad signal was observed (45.99 ppm in the semiorganic medium/46.66 ppm in the microemulsion), assigned to the diastereoisomers 14a-b due to Michael addition of the thiol group of GSH on the conjugated double bond (Scheme 4). Then, a series of 4 new signals (48.30, 48.37, 48.43, and 48.50 ppm in the semiorganic medium/47.82, 47.84, 47.92, and 48.02 ppm in the macroemulsion) was detected, referring to diastereoisomers 15a-d, formed after a second addition of GSH on the carbonyl function of 14a-b. Signals of the intermediates 14a-b then decreased with time to the benefit of signals corresponding to adducts 15a-d. After 2 days in the semiorganic medium and 10 days in the microemulsion system, a complete disappearance of the initial citral peaks (168.83 and 168.44 ppm in the semiorganic medium/164.70 and 164.56 ppm in the microemulsion) was observed.

Figure 1. 13C NMR of 1a and GSH (10 equiv) in CH3CN/H2O 1:2. Adducts 12 and 13 are formed as a mixture of 4 diastereomers, 12a-d and 13a-d, respectively.

Scheme 4. Reaction of 2a-b with GSH (10 Equiv) in a Semi-Organic or Microemulsion System and Structures of Adducts Formed 14a-b and 15a-d

1438

Chem. Res. Toxicol., Vol. 23, No. 9, 2010

Merckel et al.

+ 4.818/R2 ) 0.970) than in the microemulsion (k1′; y ) -0.475x + 4.856/R2 ) 0.992). In order to approach the kinetic rate of the second addition of GSH in the microemulsion system or in the semiorganic medium, signals of 2a and 15a-d were integrated after the study of the reaction of 2a toward an excess of GSH. The percentages of 2a and 15a-d over time are reported in Figure 4c and d as well as the percentage of 14a-b, obtained by subtraction. In the microemulsion system, only a tiny amount of intermediates 14a-b was detected, indicating that the second addition of GSH on the aldehyde function of citral is very fast and that k2′ is much higher than k1′. Therefore, in the microemulsion, the limiting step is the addition of a Michael-type addition of GSH on the conjugated double bond. On the contrary, in the semiorganic medium, a larger amount of 14a-b was observed all along the reaction, indicating that the addition of GSH on the aldehyde is slower and that k1 is higher than k2. DOSY Experiments. Self-diffusion coefficients of the different constituents of the microemulsion were measured using DOSY experiments. These self-diffusion coefficients were found to be of the same order of magnitude for H2O (Dw ) 5.5 × 10-10 m2 s-1), tert-BuOH (D ) 6.0 × 10-10 m2 s-1), CHCl3 (Do ) 3.2 × 10-10 m2 s-1), and SDS (Ds ) 1.0 × 10-10 m2 s-1).

Discussion Figure 2. 13C NMR of 2a and GSH (10 equiv) in CH3CN/H2O 1:2. Adducts 14 and 15 are formed as a mixture of 2 diastereomers (14a-b) and 4 diastereomers (15a-d), respectively.

[13C]-Citral 2a. The same observations were made by running experiments with 1-[13C]-citral 2a. Signals of compounds 14a-b (204.64 ppm in the semiorganic medium/203.58 ppm in the microemulsion) and 15a-d (66.83, 66.98, 67.64, and 68.45 ppm in the semiorganic medium/63.54, 63.71, 64.07, and 64.27 ppm in the microemulsion) were detected. After 6 days in the acetonitrile/water medium and 11 days in the microemulsion system, signals of 2a (193.09 and 193.79 ppm in the semiorganic medium/191.00 and 191.63 ppm in the microemulsion) and 14a-b completely disappeared, and the only signals remaining were those of 15a-d (Figure 2). Reaction Rates. The NMR spectra of the reaction of [13C]hydroxycitronellal 1a with 10 equivalents of GSH in both media (semiorganic and microemulsion) were recorded approximately every other day using the same NMR acquisition conditions (NS ) 800; SWH ) 20325.203 Hz; FIDRES ) 0.620 Hz; AQ ) 0.806 s; D1 ) 0.600 s; pulse ) 30°). Signals of [13C]hydroxycitronellal 1a were integrated and normalized on the basis of the signal of tert-butanol (δ ) 29.23 ppm) or of deuterated tert-butanol (δ ) 68.46 ppm) used as the internal reference. The percentage of hydroxycitronellal over time, in both conditions, is reported in Figure 3a where it can be seen that the reaction in the microemulsion was much faster than the one in the semiorganic medium. A comparison of the observed rate constants obtained by plugging the logarithm of these values against time (Figure 3b), indicates that the observed rate constant was about 3 times higher (y ) -0.293x + 4.252/ R2 ) 0.994) in the microemulsion compared to the one in the semiorganic medium (y ) -0.101x + 4.544/R2 ) 0.988). In the case of citral, the addition of GSH on the Michael position was found to be much faster in the acetonitrile-water medium than in the microemulsion system as shown in Figure 4a and b. Indeed, the observed kinetic constant of this first step is about 4.5 higher in the semiorganic medium (k1; y ) -2.179x

The use of in chemico approaches, abiotic chemical reactivity methods as replacements for animal assays, to identify and if possible rank skin sensitizers has attracted much attention in the past few years (10). These methods, based on the chemical reactivity measurement of test molecules toward various nucleophiles, mainly GSH (13, 29, 30) or synthetic peptides (14, 15, 31), have been shown to be promising. Moreover, in terms of alternative methods, they are expected to be highly reproducible and easy to transfer from one center to another one. However, approaching reaction rates for lipophilic molecules toward peptides in aqueous solutions can be difficult for solubility issues (32). For this reason, in current peptide reactivity approaches, the reaction mixtures are made up by first dissolving the test chemical in a water-soluble organic solvent such as acetonitrile and adding this solution to the peptide dissolved in an appropriate buffer (12, 14, 32). Reactions are thus carried out in a semiorganic homogeneous system mainly based on water but with a significant amount of organic solvent. If this approach can overcome most of the problems, it cannot be excluded that, for the most lipophilic substances such as R-hexylcinnamic aldehyde, solubility issues remain critical. Indeed, depending on the peptide used, there will be a limit to how much organic solvent can be added without the peptide coming out of solution. Another issue is associated with weak sensitizers for which a low reactivity toward nucleophiles is expected. It can therefore be difficult to derive reactivity parameters for these chemicals in the time frame usually described for such in chemico assays (2-24 h). These week sensitizers are, however, very important to detect as they are usually widely used, with a high population exposure, and are therefore associated with a significant level of clinical manifestations. Microemulsions could be interesting systems to overcome some of these solubility issues. Microemulsions are excellent solvents for both hydrophobic organic compounds and hydrophilic molecules. Being macroscopically homogeneous but microscopically dispersed, they can be seen as an intermediate stage between the solvent-based one-phase and the true two-

ReactiVity of Skin Sensitizers in a Microemulsion

Chem. Res. Toxicol., Vol. 23, No. 9, 2010 1439

Figure 3. (a) Percent 1a versus time (days): 9, CH3CN/H2O 1:2; [, microemulsion. (b) ln(% 1a) versus time (days): 9, CH3CN/H2O 1:2; [, microemulsion.

Figure 4. (a) Percent 2b versus time (days): 9, CH3CN/H2O 1:2; [, microemulsion. (b) ln(% 2b) versus time (days): 9, CH3CN/H2O 1:2; [, microemulsion. (c) Percent 2a, 14a-b, 15a-d versus time (days) in a microemulsion system: 9, 2a; [, 14a-b; 2, 15 a-d. (d) Percent 2a, 14a-b, 15a-d versus time (days) in a semiorganic medium: 9, 2a; [, 14a-b; 2, 15 a-d.

phase systems (18, 19). Menger was one of the pioneers in the exploration of microemulsions as media for organic reactions (21). Since, microemulsions have been shown to be valuable systems for macrocyclic lactone formation (33), Diels-Alder reactions (34), oxidations with H2O2, nitration of aromatic compounds, and catalytic reductions and oxidations (35). However, we have not found uses of microemulsions for the alkylation of peptides by haptens. Hydroxycitronellal and isomers of citral are aldehydes present in many essential oils and therefore widely used as fragrances in cosmetics. These molecules are classified as weak sensitizers (36) according to the local lymph node assay (LLNA) but of major clinical relevance due to the high exposure of consumers. Hydroxycitronellal and citral are therefore classified, by the EU cosmetics directive, among the 26 fragrance ingredients to be labeled on cosmetics due to their sensitizing properties (5). Hydroxycitronellal is also part of Fragrance Mix I, a mixture of 7 major fragrance chemical sensitizers plus one natural extract, and citral is part of Fragrance Mix II, a complementary mixture of 6 fragrance chemical sensitizers, used as diagnostic tools to detect patients allergic to fragrances (37).

Hydroxycitronellal has been shown to induce a significant depletion of GSH (used as a model nucleophile in reactivity assay) when present in excess (14) even if it has been suggested that this depletion could be more related to oxido-reduction processes than to alkylation (31). Following the reaction by 13C NMR it is clear that, in addition to the equilibrium between the carbonyl function and its hydrated form, the first adducts result from the nucleophilic addition of the thiol group of GSH on the aldehyde to form a mixture of 4 diastereomeric hemithioacetals 12a-d. These adducts were already present after 30 min. After 24 h, a second adduct was observed arising from the intramolecular reaction of the N-terminal amino group to form 4 diastereomeric cyclic products 13a-d. Such a secondary reaction of the R-NH2 position of GSH on a first adduct formed by a primary reaction with the thiol group is not unusual and has already been reported for other haptens of the isothiazolone type (38). Hydroxycitronellal being present as a racemate, adducts formed with GSH were observed as mixtures of 4 diastereomers 12a-d and 13a-d. It is interesting to note that this aldehyde reacts with GSH to form an cyclic hemithioacetal and not a Schiff’s base as very often reported.

1440

Chem. Res. Toxicol., Vol. 23, No. 9, 2010

In the GSH depletion assay, citral induces a high depletion when present in excess and even the complete depletion of a peptide containing a cysteine residue (14). 13C NMR studies of the reaction of 2a and 2b toward an excess of GSH confirmed a high reactivity of citral toward thiol groups. In fact, citral initially undergoes a Michael-type addition of the thiol group on the conjugated double bond to form 2 enantiomers 14a-b. These intermediates could then be subject to an addition of a thiol group of another GSH molecule on the carbonyl function to form 4 diastereomers 15a-d that can be detected after less than 24 h. Thus, the formation of such double adducts can explain that citral generates more GSH depletion than hydroxycitronellal does even if both molecules belong to the same sensitization class. In order to approach the reaction rate of 1 with an excess of GSH in the semiorganic medium or the microemulsion, 13C signals of 1a were integrated over time and normalized on the basis of the signal of the tert-butanol or the deuterated tertbutanol used as the internal reference. The percentage of hydroxycitronellal over time, in both conditions, is reported in Figure 3a, where it can be seen that the reaction in the microemulsion is much faster than the one in the semiorganic medium. In the microemulsion, all hydroxycitronellal 1a has been converted in 15 days, while in the semiorganic medium, 1a is still detectable after 40 days. The logarithm of theses values against time (Figure 3b) indicates that the reaction can be approximated to a pseudofirst order and that the observed rate constant is about 3 times higher (y ) -0.293x + 4.252/R2 ) 0.994) in the microemulsion compared to the one in the semiorganic medium (y ) -0.101x + 4.544/R2 ) 0.988). The same procedures were used with citral. Percentages of 2b over time in both conditions are plotted in Figure 4a. It can be easily noticed that the first reaction, the Michael addition of the thiol group, is much faster in the semiorganic medium than in the microemulsion. In fact, no signal corresponding to 2b can be detected after 2 days in the acetonitrile-water medium, whereas complete disappearance of 2b is only observed after 11 days using the microemulsion as the solvent. The logarithm of these percentages, reported in Figure 4b, showed that the reaction could also be considered as a pseudofirst order reaction and that GSH addition is about 4.5 higher in the semiorganic medium (k1; y ) -2.179x + 4.818/R2 ) 0.970) than in the microemulsion (k1′; y ) -0.475x + 4.856/R2 ) 0.992). To consider the addition of GSH on the carbonyl function of citral, percentages of 13C signals of 2a, 15a-d, and 14a-b, obtained after subtraction, are reported in Figure 4c and d in the microemulsion and in the semiorganic medium, respectively. It can be seen that in the microemulsion the intermediate adducts 14a-b were only present in very small amounts, hardly detectable in the NMR spectra, even during the 2 first days of the reaction. This reflects the fact that the addition of GSH on 14a-b is very fast and that the rate constant of the reaction k2′ is much higher than k1′. On the contrary, in the semiorganic medium 14a-b was found in a much larger amount and all along the reaction that ended after 6 days. The accumulation of intermediate 14a-b indicates that the addition of GSH on the carbonyl is slower than the 1,4-Michael addition and that k1 is higher than k2. As for hydroxycitronellal, the reaction of the thiol group of GSH on the carbonyl chemical function is faster in the microemulsion system than in the semiorganic medium, even though the Michael addition is the first step of the reaction of GSH with citral. According to the results reported in here, the use of microemulsions can be a way to overcome compatibility

Merckel et al.

problems. Also, the capability of microemulsions to compartmentalize and concentrate reactants can lead to rate enhancement compared to one-phase systems even if not a general principle. Microemulsions are macroscopically homogeneous but microscopically heterogeneous as they contain oil and water domains of nanometer-sized dimensions. The structure is highly dynamic, and each interface disintegrates and reforms in the time scale of milliseconds. In a first approximation, it is often considered that an oil rich microemulsion will mainly consist of water-inoil (w/o) structures, while the water rich microemulsion will mainly consist of oil-in-water (o/w) structures. When the oil and water content is very similar, a sponge-like structure is obtained, the coherent domains of water and oil being about equal in size. One can then see such microemulsions as bicontinuous systems. For microemulsions, dynamic parameters such as the molecular self-diffusion coefficient, obtained by NMR, have proven to be most informative (35). Long-range translational mobilities of molecules can be mapped, and these properties are very different if molecules are confined in closed structural domains or if molecules occur in regions that can be extended over macroscopic distances. As an example, in w/o structures, water molecules are confined to closed structures and will therefore have a restricted motion, and their diffusion will be slow compared to the that of the oil matrix (typically 2 orders of magnitude difference) and approximately equal to the diffusion of the droplet (d). Moreover, if all surfactant (s) molecules are located at the w/o interface, they will diffuse with similar rates as the droplets too. It is therefore expected to have Do . Dw ≈ Ds ≈ Dd. The opposite situation will be observed for o/w structures. For bicontinuous systems, oil and water form domains that are continuous over macroscopic distances. Characteristic for these structures is that the surfactant molecules will be located at the interface and that all species have similar (same order of magnitude) self-diffusion coefficients Do ≈ Dw ≈ Ds. As a consequence, both water and hydrocarbon constituents are, in principle, free to diffuse over macroscopic distances, and hence, their self-diffusion coefficient will be high and of the order of 10-9 m2 s-1. DOSY experiments carried out on the microemulsion used in this study indicated self-diffusion coefficients of 5.5, 6.0, 3.2, and 1.0 × 10-10 m2 s-1 for H2O, tert-BuOH, CHCl3, and SDS, respectively. A value of Ds ≈ 10-10 m2 s-1 is in very good agreement with a bicontinuous structure where the state of the surfactant will be reminiscent of that in a lamellar liquid crystalline phase. In this model, water containing the peptide and hydrocarbon containing the hapten are supposed to have unrestricted diffusion in their respective channels inside the network, while the surfactant will constitute a monolayer separating the continuous domains. However the impact of such a structure on the chemical reactivity is still unclear.

Conclusions Hydroxycitronellal 1 was found to react with the thiol group of GSH in a pseudofirst order rate to form hemithioacetal intermediates, followed by an intramolecular cyclization, leading to more stable cyclic adducts. The observed reaction rate was found to be about 3 times higher in the microemulsion compared to that in the classical semiorganic mixture. As for hydroxycitronellal 1, the reaction of the thiol group of GSH on the carbonyl chemical function of citral 2 was faster in the microemulsion system than in the semiorganic medium. However the 1,4-Michael addition, the first step of the reaction, was faster in the semiorganic medium compared to the microemulsion system. Thus, this chloroform/water/tertio-butanol/sodium

ReactiVity of Skin Sensitizers in a Microemulsion

dodecylsulphate microemulsion, apparently of the bicontinuous type according to DOSY data, could be very valuable for the in chemico evaluation of lipophilic chemicals toward peptides even if the effect of this kind of medium on the kinetic rate needs to be further investigated. Acknowledgment. We thank COLIPA (The European Cosmetic Toiletry and Perfumery Association) for funding F.M. and J.M., and the European Commission 5th PCRD project “Fragrance chemical allergy: a major environmental and consumer health problem in Europe” (contract QLK4-CT-199901558) for funding G.B.

References (1) Russel, W. H. S., and Burch, R. L. (1959) The Principle of the Humane Experimental Technique, p 238, Methuen, London. (2) EEC (1967) EEC Council Directive 67/548/EEC on the approximation of the laws, regulations and administrative provisions relating to the classification, packaging and labeling of dangerous substances. Off. J. Eur. Commun. 196, 1-98. (3) EU (1999) Directive 1999/45/EC of the European Parliament and the Council of 31 May 1999 concerning the approximation of the laws, regulation and administrative provisions of the Member States relating to the classification, packaging and labeling of dangerous preparations. Off. J. Eur. Commun. L200, 1-68. (4) EC (2006) Regulation (EC) No 1907/2006 of the European Parliament and of the Council of the 18 December 2006 concerning the Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH), establishing a European Chemicals Agency. Off. J. Eur. Union L396, 1-849. (5) EU (2003) Directive 2003/15/EC of the European Parliament and of the Council of 27 February 2003 amending Council Directive 76/ 768/EEC on the approximation of the laws of the Member States relating to cosmetic products, 7th Amendment to the European Cosmetics Directive. Off. J. Eur. Union L66, 26-35. (6) Magnusson, B., and Kligman, A. M. (1970) Allergic Contact Dermatitis in the Guinea Pig pp 141, Charles C. Thomas, Springfield, IL. (7) Buehler, E. V. (1965) Delayed contact hypersensitivity in the guinea pig. Arch. Dermatol. 91, 171–177. (8) Kimber, I., and Basketter, D. A. (1992) The murine local lymph node assay; collaborative studies and new directions: A commentary. Food Chem. Toxicol. 30, 165–169. (9) Casati, S., Aeby, P., Basketter, D. A., Cavani, A., Gennari, A., Gerberick, G. F., Griem, P., Hartung, T., Kimber, I., Lepoittevin, J.P., Meade, B. J., Pallardy, M., Rougier, N., Rousset, F., Rubinstenn, G., Sallusto, F., Verheyen, G. R., and Zuang, V. (2005) Dendritic cells as a tool for the predictive identification of skin sensitization hazard. Altern. Lab. Anim. 33, 47–62. (10) Gerberick, G. F., Aleksic, M., Basketter, D. A., Casati, S., Karlberg, A.-T., Kern, P., Kimber, I., Lepoittevin, J.-P., Natsch, A., Ovigne, J.-M., Rovida, C., Sakagushi, H., and Schultz, T. (2008) Chemical reactivity measurement and the predictive identification of skin sensitizers. Altern. Lab. Anim. 36, 215–242. (11) Kato, H., Okamoto, M., Yamashita, K., Nakamura, Y., Fukumori, Y., Nakai, K., and Kaneko, H. (2003) Peptide-binding assessment using mass spectrometry as a new screening method for skin sensitization. J. Toxicol. Sci. 28, 19–24. (12) Schultz, T. W., Yarbrough, J. W., and Johnson, E. L. (2005) Structureactivity relationships for reactivity of carbonyl-containing compounds with glutathione. SAR QSAR EnViron. Res. 16, 313–322. (13) Aptula, A. O., Patlewicz, G., Roberts, D. W., and Schultz, T. W. (2006) Non-enzymatic glutathione reactivity and in vitro toxicity: a non-animal approach to skin sensitization. Toxicol. in Vitro 20, 239–247. (14) Gerberick, G. F., Vassallo, J. D., Bailey, R. E., Chaney, J. G., Morrall, S. W., and Lepoittevin, J.-P. (2004) Development of a peptide reactivity assay for screening contact allergens. Toxicol. Sci. 81, 332– 343. (15) Gerberick, G. F., Vassallo, J. D., Foertsch, L. M., Price, B. B., Chaney, J. G., and Lepoittevin, J.-P. (2007) Quantification of chemical peptide reactivity for screening contact allergens: A classification tree model approach. Toxicol. Sci. 97, 417–427. (16) Natsch, A., and Gfeller, H. (2008) LC-MS-based characterization of peptide reactivity of chemicals to improve the in vitro prediction of skin sensitization potential. Toxicol. Sci. 106, 464–478.

Chem. Res. Toxicol., Vol. 23, No. 9, 2010 1441 (17) Kimber, I., Basketter, D. A., Butler, M., Gamer, A., Garrigue, J.-L., Gerberick, G. F., Newsome, C., Steiling, W., and Vohr, H. W. (2003) Classification of contact allergens according to potency: proposals. Food Chem. Toxicol. 41, 1799–1809. (18) Schwuger, M. J., and Stickdorn, K. (1995) Microemulsions in technical processes. Chem. ReV. 95, 849–864. (19) Holmberg, K. (2007) Organic reactions in microemulsions. Eur. J. Org. Chem. 731–742. (20) Paul, B. K., and Moulik, S. P. (2001) Uses and applications of microemulsions. Curr. Sci. 80, 990–1001. (21) Menger, F. M., and Elrington, A. R. (1991) Organic reactivity in microemulsion systems. J. Am. Chem. Soc. 113, 9621–9624. (22) Frosch, P. J., Pirker, C., Rastogi, S., Andersen, K. E., Bruze, M., Svedman, C., Goossens, A., White, I., Uter, W., Gime´nez-Arnau, E., Lepoittevin, J.-P., Johansen, J. D., and Menne´, T. (2005) Patch testing with a new fragrance mix detects additional patients sensitive to perfumes and missed by the current fragrance mix. Contact Dermatitis 52, 207–215. (23) Tanner, J. E. (1970) Use of stimulated echo in NMR diffusion studies. J. Chem. Phys. 52, 2523–26. (24) Cotts, R. M., Hoch, M. J. R., Sun, T., and Markert, J. T. (1989) Pulsed field gradient stimulated echo methods for improved NMR diffusion measurements in heterogeneous systems. J. Magn. Reson. 83, 252– 266. (25) Delsuc, M. A., and Malliavin, T. E. (1998) Maximum entropy processing of DOSY NMR spectra. Anal. Chem. 70, 2146–48. (26) Meschkat, E., Barratt, M., and Lepoittevin, J.-P. (2001) Studies of chemical selectivity of haptens, reactivity and skin sensitization potency. NMR studies of the covalent binding of the 13C-labeled skin sensitizers, 2-[13C] and 3-[13C]-hex-1-ene- and 3-[13C]-hexane-1,3sultones to human serum albumin. Chem. Res. Toxicol. 14, 118–126. (27) Alvarez-Sanchez, R., Divkovic, M., Basketter, D., Pease, C., Panico, M., Dell, A., Morris, H., and Lepoittevin, J.-P. (2004) Effect of glutathione on the covalent binding of the 13C-labeled skin sensitizer 5-chloro-2-methylisothiazol-3-one (MCI) to human serum albumin: identification of adducts by NMR, MALDI-MS and nano-ES MS/ MS. Chem. Res. Toxicol. 17, 1280–1288. (28) Eilstein, J., Gime´nez-Arnau, E., Duche´, D., Rousset, F., and Lepoittevin, J.-P. (2006) Synthesis and reactivity towards nucleophilic aminoacids of 2,5-[13C]-dimethyl-para-benzoquinonediimine. Chem. Res. Toxicol. 19, 1248–1256. (29) Ahlfors, S. R., Sterner, O., and Hansson, C. (2003) Reactivity of contact allergenic haptens to amino acid residues in a model carrier peptide, and characterization of formed peptide-hapten adducts. Skin Pharmacol. Appl. Skin Physiol. 16, 59–68. (30) Bergstrom, M. A., Ott, H., Carlson, A., Neis, M., Zwadlo-Klarwasser, G., Jonsson, C. A., Merk, H. F., Karlberg, A.-T., and Baron, J. M. (2007) A skin-like cytochrome P450 cocktail activates prohaptens to contact allergenic metabolites. J. InVest. Dermatol. 127, 1145–1153. (31) Natsch, A., Gfeller, H., Rothaupt, M., and Ellis, G. (2007) Utility and limitations of a peptide reactivity assay to predict fragrance allergens in vitro. Toxicol. in Vitro 21, 1220–1226. (32) Roberts, D. W., and Natsch, A. (2009) High throughput kinetic profiling approach for covalent binding to peptides: application to skin sensitization potency of Michael acceptor electrophiles. Chem. Res. Toxicol. 22, 592–603. (33) Gonzalez, A., and Holt, S. L. (1981) Formation of macrocyclic lactones in microemulsions. J. Org. Chem. 46, 2594–2596. (34) Gonzalez, A., and Holt, S. L. (1982) Effect of microemulsions on the Diels-Alder reaction: Endo/exo ratios in the reaction of cyclopentadiene and methyl methacrylate. J. Org. Chem. 47, 3186–3188. (35) Sjo¨blom, J., Lindberg, R., and Friberg, S. E. (1996) Microemulsions: phase equilibria characterization, structures, applications and chemical reactions. AdV. Colloid Interface Sci. 95, 125–287. (36) Gerberick, G. F., Ryan, C. A., Kern, P. S., Schlatter, H., Dearman, R. J., Kimber, I., Patlewicz, G. Y., and Basketter, D. A. (2005) Compilation of historical lymph node data for evaluation of skin sensitization alternative methods. Dermatitis 16, 157–202. (37) Andersen, K. E., White, I. R., and Goossens, A. (2006) In Contact Dermatitis (Frosch, P. J., Menne´, T., and Lepoittevin, J.-P., Eds.) 5th ed., pp 453-492, Springer, Berlin, Germany. (38) Alvarez-Sanchez, R., Basketter, D. A., Pease, C., and Lepoittevin, J.P. (2004) Covalent binding of the 13C-labeled skin sensitizers 5-chloro-2-methylisothiazol-3-one (MCI) and 2-methylisothiazol-3one (MI) to a model peptide and glutathione. Bioorg. Med. Chem. Lett. 14, 365–368.

TX100043B