Studies of the Chemical Selectivity of Hapten, Reactivity, and Skin

118

Chem. Res. Toxicol. 2001, 14, 118-126

Studies of the Chemical Selectivity of Hapten, Reactivity, and Skin Sensitization Potency. 2. NMR Studies of the Covalent Binding of the 13C-Labeled Skin Sensitizers 2-[13C]- and 3-[13C]Hex-1-ene- and 3-[13C]Hexane-1,3-sultones to Human Serum Albumin Emmanuel Meschkat,† Martin D. Barratt,‡,§ and Jean-Pierre Lepoittevin*,† Laboratoire de Dermatochimie associe´ au CNRS, Universite´ Louis Pasteur, Clinique Dermatologique, CHU, F-67091 Strasbourg Cedex, France, and SEAC-Toxicology Unit, Unilever Research Colworth, Sharnbrook, Bedford, U.K. Received October 25, 2000

3-[13C]- and 2-[13C]hex-1-ene-1,3-sultones (1a and 1b, respectively) and 3-[13C]hex-1-ene1,3-sultone 2a were incubated with human serum albumin in phosphate buffer at pH 8.1. In both cases, the main reaction was a hydrolysis via an SN reaction at position 3, but several adducts were also formed. Hex-1-ene-1,3-sultone, which is a strong skin sensitizer, appears also to be a strongly oxophilic molecule reacting mainly at position 3 through an SN reaction to give adducts on tyrosines. This sultone was also able to react with a single lysine residue, also via an initial SN reaction at position 3, followed by an intramolecular Michael addition at position 2 to form a mixture of aziridinium intermediates which were subsequently hydrolyzed to give an amino alcohol derivative as the final product. The same reaction carried out on acetylated human serum albumin seems to indicate that the target lysine could be Lys199, which is known to be easily acetylated. Hexane-1,3-sultone, which is a weak sensitizer, appears to be an even more oxophilic molecule, making adducts on tyrosines through an SN reaction at position 3. No reaction was observed on Lys199. The difference in skin sensitization potential seems therefore to be more related to the selective ability of modifying lysine residues than to the more general ability to modify tyrosine residues.

Introduction Hapten-protein interactions are one of the key steps in the induction and elicitation mechanisms of allergic contact dermatitis (ACD)1 (1). The processing of the hapten-protein complex by immunocompetent skin antigen-presenting cells (Langerhans cells) and the subsequent transmission of this information to T-cells in lymphatic nodes lead to the biological and clinical aspects of ACD (erythema and edema). The hapten-protein complex is formed mainly through a covalent bond between the hapten and nucleophilic groups on proteins. Most skin allergens are, therefore, electrophiles (2). In a previous paper (3), we have shown that hex-1ene-1,3-sultone (Chart 1), which is a strong skin sensitizer, is able to react in water with various model nucleophiles and that the reactive sites and reaction mechanisms involved are predicted by the hard and soft acids and bases (HSAB) theory (4). In fact, hex-1-ene1,3-sultones contain two potential reactive electrophilic centers (at positions 2 and 3), which are able to react with †

Universite´ Louis Pasteur. Unilever Research Colworth. Present address: Marlin Consultancy, 10 Beeby Way, Carlton, Bedford MK43 7LW, U.K. 1 Abbreviations: ACD, allergic contact dermatitis; DNCB, dinitrochlorobenzene; DNFB, dinitrofluorobenzene; HSA, human serum albumin; HSAB, hard and soft acids and bases; HSQC, heteronuclear single-quantum correlation; PNPA, p-nitrophenyl acetate. ‡ §

Chart 1

nucleophilic amino acids via two different mechanisms (“Michael type” addition at position 2 and SN substitution at position 3). Hard nucleophiles, such as primary and secondary amines or phenate, react at position 3 via an SN substitution reaction, while soft nucleophiles, such as thiolate or imidazole, react at position 2 via a Michael addition reaction (Chart 1). Hexane-1,3-sultone (Chart 1), which is a weak sensitizer, is characterized by a strong oxophilic character but a very poor reactivity toward amino groups such as primary or secondary amines. In this paper, we report the results of NMR studies of the covalent binding of 3-[13C]- and 2-[13C]hex-1-ene-1,3sultones 1a and 1b and of 3-[13C]hexane-1,3-sultone 2a to human serum albumin (HSA) using 13C and 13C{1H} heteronuclear single-quantum correlation (HSQC) experiments.

10.1021/tx000226f CCC: $20.00 © 2001 American Chemical Society Published on Web 12/15/2000

NMR of Labeled Sultones

Chem. Res. Toxicol., Vol. 14, No. 1, 2001 119

Materials and Methods Caution: Skin contact with sultone derivatives must be avoided; as sensitizing agents, they must be handled with care. Coupling of 3-[13C]- and 2-[13C]Hex-1-ene-1,3-sultone to HSA. To HSA (100 mg, 1.4 µmol) in phosphate buffer (10 mL, 0.1 M, pH 8.1) were added EtOH (1.2 mL) and a solution of sultone 1a or 1b (23 mg, 140 µmol) in EtOH (0.4 mL). The reaction mixture was stirred at room temperature for 17 days, dialyzed against water (3 × 5 L), and lyophilized to give the modified protein as a white solid. Coupling of 3-[13C]- and 2-[13C]Hex-1-ene-1,3-sultone to Acetylated HSA. To acetylated HSA (5) (100 mg, 1.4 µmol) in phosphate buffer (10 mL, 0.1 M, pH 8.1) were added EtOH (1.2 mL) and a solution of sultone 1a or 1b (23 mg, 140 µmol) in EtOH (0.4 mL). The reaction mixture was stirred at room temperature for 17 days, dialyzed against water (3 × 5 L), and lyophilized to give the modified protein as a white solid. Coupling of 3-[13C]Hexane-1,3-sultone to HSA. To HSA (70 mg, 1.0 µmol) in phosphate buffer (7 mL, 0.1 M, pH 8.1) were added EtOH (0.8 mL) and a solution of sultone 2a (16.7 mg, 101 µmol) in EtOH (0.3 mL). The reaction mixture was stirred at room temperature for 17 days, dialyzed against water (3 × 5 L), and lyophilized to give the modified protein as a white solid. Preparation of Model Amino Alcohol Adducts of 3-[13C]and 2-[13C]Hex-1-ene-1,3-sultone with Butylamine. The labeled sultone 1a or 1b (2 mg, 12 µmol) was added to a solution of butylamine (27 mg, 369 µmol) in water (0.4 mL), and then the reaction mixture was stirred at room temperature for 1 h and directly lyophilized. The aziridine residue was dissolved in water (1 mL), heated under reflux for 2 h, and then lyophilized. NMR Experiments. 13C NMR spectra were recorded on a Bruker AM 400 MHz spectrometer in a mixture of H2O and D2O (0.35 and 0.05 mL, respectively). 1H{13C} HSQC experiments were carried out on a Bruker ARX 500 MHz spectrometer in a mixture of H2O and t-BuOH-d9 (0.35 and 0.05 mL, respectively). Chemical shifts are reported in parts per million (δ) with respect to TMS, and a trace of CH3CN was used as an internal standard (1H, δ ) 2.06 ppm; 13C, δ ) 1.32 ppm). Structure Assignment. Structures of the different adducts were assigned using a combination of {1H}-decoupled 13C NMR and 13C DEPT 135 sequences. The measured 13C chemical shifts of the different adducts were compared with those calculated using the additivity principle and to NMR data of analogous compounds reported in the literature. These assignments were further confirmed by 1H chemical shifts obtained by heteronuclear correlations and are in accordance with the sequence of adduct formation and hydrolysis (3). Hydrolysis Reaction of an Equimolar Mixture of 2-[13C]Hex-1-ene-1,3-sultone 1b and 3-[13C]Hexane-1,3-sultone 2a. An equimolar mixture of 1b (2 mg, 12 µmol) and 2a (2 mg, 12 µmol) was added to HSA (17 mg, 0.2 µmol) in a mixture of phosphate buffer (0.35 mL, 0.1 M, pH 8.1) and D2O (0.05 mL). A trace of CH3CN was added as an internal reference and the reaction followed by 13C NMR. Competitive Reaction of an Equimolar Mixture of 2-[13C]Hex-1-ene-1,3-sultone 1b and 3-[13C]Hexane-1,3-sultone 2a with Butylamine. To an equimolar mixture of 1b (0.5 mg, 3 µmol) and 2a (0.5 mg, 3 µmol) was added a solution of n-butylamine (13.5 mg, 0.18 mmol, 60 equiv) in a mixture of H2O and D2O (0.35 and 0.05 mL, respectively). A trace of CH3CN was added as an internal reference and the reaction followed by 13C NMR. Competitive Reaction of an Equimolar Mixture of 2-[13C]Hex-1-ene-1,3-sultone 1b and 3-[13C]Hexane-1,3-sultone 2a with Sodium Phenate. To an equimolar mixture of 1b (0.5 mg, 3 µmol) and 2a (0.5 mg, 3 µmol) in 1 drop of acetone was added a solution of PhONa‚3H2O (10.5 mg, 0.06 mmol, 20 equiv) in a mixture of H2O and D2O (0.35 and 0.05 mL,

Figure 1. (A) 13C NMR spectra of HSA in a mixture of H2O and D2O. (B) 13C NMR spectra of HSA incubated with sultone 1a in a mixture of H2O and D2O after standard dialysis. (C) 13C NMR spectra of HSA in a mixture of H2O and D2O incubated with sultone 1a after extensive dialysis. respectively). A trace of CH3CN was added as an internal reference and the reaction followed by 13C NMR.

Results Human serum albumin was incubated in phosphate buffer (pH 8.1) with one of the hex-1-ene-1,3-sultone 1a or 1b, labeled at position 3 or 2, respectively, or hexane1,3-sultone 2a, labeled at position 3. After dialysis and lyophilization, 13C NMR spectra of the modified proteins were recorded and compared with the spectrum of the native protein (Figures 1A, 2A, and 3A) for new signals. Hex-1-ene-1,3-sultone Labeled at position 3 (Figure 1). After incubation of HSA with sultone 1a, two sets of signals were observed (Figure 1B). The first, containing three peaks at 71.4, 68.4, and 66.6 ppm, is characteristic of adducts and/or products resulting from an SN substitution reaction at position 3, with the formation of new carbon-oxygen bonds, while the second, containing two peaks at 49.7 and 47.7 ppm, is characteristic of the same type of reaction, but involving the formation of carbon-

120

Chem. Res. Toxicol., Vol. 14, No. 1, 2001

Figure 2. (A) 13C NMR spectra of HSA in a mixture of H2O and D2O. (B) 13C NMR spectra of HSA incubated with sultone 1b in a mixture of H2O and D2O after standard dialysis. (C) 13C NMR spectra of HSA incubated with sultone 1b in a mixture of H2O and D2O after extensive dialysis.

nitrogen bonds. When the material was extensively dialyzed (15 × 5 L), to exclude the presence of noncovalent sultone derivatives, the peak at 66.6 ppm and the two small peaks at 49.9 and 47.7 ppm were lost (Figure 1C), suggesting that these were due either to the presence of noncovalently bound labeled molecules or to hydrolysis of the protein-sultone adducts. Hex-1-ene-1,3-sultones Labeled at Position 2 (Figure 2). After incubation of HSA with 1b, two sets of signals were observed (Figure 2B). The first, containing three peaks at 141.1, 140.0, and 136.9 ppm, is characteristic of the formation of adducts at position 3 (SN reaction) with a conserved sp2 carbon at position 2, while the second, with lower-intensity signals at 59.6, 45.1, and 44.4 ppm, is characteristic of the formation of carbonnitrogen bonds through Michael addition at position 2. Extensive dialysis (15 × 5 L) of the protein led to the loss of the signals at 141.1 and 140.0 ppm, a markedly decreased intensity of peaks at 45.1 and 44.4 ppm, and an increased intensity of the signal at 59.6 ppm (Figure 2C).

Meschkat et al.

Figure 3. (A) 13C NMR spectra of HSA in a mixture of H2O and D2O. (B) 13C NMR spectra of HSA incubated with sultone 2a in a mixture of H2O and D2O after standard dialysis. (C) 13C NMR spectra of HSA incubated with sultone 2a in a mixture of H2O and D2O after extensive dialysis.

Hexane-1,3-sultone Labeled at Position 3 (Figure 3). After incubation of HSA with sultone 2a, a rather large peak centered at 74.6 ppm and a sharper one at 70.4 ppm were observed (Figure 3B). They are both characteristic of adducts and/or products resulting from an SN substitution reaction at position 3, with the formation of new carbon-oxygen bonds. When the material was extensively dialyzed (15 × 5 L), to exclude the presence of noncovalent sultone derivatives, the peak at 70.4 ppm was lost (Figure 3C), suggesting that this was due to the presence of a noncovalently bound labeled molecule. Determination of 1H Chemical Shifts of Model Adducts. The adducts were further characterized by combining 1H and 13C information with the data on chemical shifts of protons at position 2 and/or 3 for the previously prepared model adducts (3) determined using 1H{13C} HSQC experiments. The adducts and chemical shifts are listed in Table 1. 1 H{13C} HSQC Experiments with Sultone 1a- or Sultone 1b-Modified HSA. 1H{13C} HSQC experiments



Table 1. 1H and 13C Chemical Shifts of Model Adducts on Hex-1-ene-3-sultones and Hexane-3-sultones Determined by 1H{13C} HSQC Experiments

Figure 5. 1H{13C} HSQC of HSA incubated with 2a after extensive dialysis. Correlations corresponding to the 13C signal at 74.6 ppm.

a 13C/1H

chemical shifts. b Stereochemistry of the aziridine.

Figure 4. 1H{13C} HSQC of HSA incubated with 1a after extensive dialysis: (A) correlations corresponding to the 13C signal at 68.4 ppm and (B) correlations corresponding to the 13C signals at 71.2-71.6 ppm. 1H{13C} HSQC of HSA incubated with 1b after extensive dialysis: (C) correlations corresponding to the 13C signal at 137.0-137.1 ppm and (D) correlations corresponding to the 13C signals at 59.7 ppm.

were performed on samples previously analyzed by 13C NMR and subjected to extensive dialysis to remove noncovalently associated labeled products (Figures 1C and 2C). Characteristic correlations were selected from their 13C chemical shifts. HSA incubated with sultone 1a labeled at position 3 (Figure 1C) exhibited two characteristic peaks, the first of which, corresponding to a single correlation (68.6/4.12 ppm) (Figure 4A), confirmed the substitution and the presence of a carbon-oxygen bond at position 3, while the second, corresponding to at least three correlations (71.6/6.26, 71.6/6.23, and 71.2/6.19 ppm) (Figure 4B), confirmed the substitution at position 3, with very high 1H chemical shifts characteristic of carbon-oxygen bonds of the phenol type. When HSA was incubated with sultone 1b labeled at position 2, two characteristic peaks were observed (Figure 2C), the first of which, corresponding to four different adducts (137.1/

5.95, 136.9/5.93, 137.0/5.90, and 137.0/5.86 ppm) (Figure 4C), confirmed the presence of an sp2 carbon at position 2 and therefore an SN susbtitution at position 3, while the second, corresponding to a single 1H chemical shift (59.7/3.63 ppm) (Figure 4D), is characteristic of a Michael adduct at position 2 and the formation of a carbonnitrogen bond. 1H{13C} HSQC Experiments with Sultone 2aModified HSA. 1H{13C} HSQC experiments were performed on samples previously analyzed by 13C NMR and subjected to extensive dialysis to remove noncovalently associated labeled products (Figure 3C). Characteristic correlations were selected from their 13C chemical shifts. HSA incubated with sultone 2a labeled at position 3 (Figure 3C) exhibited one characteristic peak corresponding to at least four correlations (74.65/5.06, 74.5/5.03, 74.4/4.99, and 74.5/4.97 ppm) (Figure 5). This confirmed the substitution at position 3, with very high 1H chemical shifts characteristic of carbon-oxygen bonds of the phenol type. Incubation of Acetylated HSA with Sultone Labeled at Position 3 (Figure 6). After incubation of HSA acetylated with p-nitrophenyl acetate (PNPA) (5) with sultone 1a, only one set of signals containing three peaks at 71.9, 71.4, and 66.6 ppm was observed (Figure 6A). These are characteristic of adducts and/or products resulting from an SN substitution reaction at position 3, with the formation of new carbon-oxygen bonds of the phenolic type. When the material was extensively dialyzed (15 × 5 L) to exclude the presence of noncovalent sultone derivatives, the peaks at 71.9 and 66.6 ppm were lost (Figure 6B), suggesting that these were due either to the presence of noncovalently bound labeled molecules or to hydrolysis of the protein-sultone adducts. Incubation of Acetylated HSA with Sultone Labeled at Position 2 (Figure 7). After incubation of acetylated HSA with 1b, only one sets of signals containing three peaks at 141.1, 140.0, and 136.9 ppm was observed (Figure 7A). These are characteristic of the formation of adducts at position 3 (SN reaction) with a conserved sp2 carbon at position 2. Extensive dialysis (15 × 5 L) of the protein led to the loss of the signals at 141.1 and 140.0 ppm (Figure 7B). Competitive Reaction of an Equimolar Mixture of 2-[13C]Hex-1-ene-1,3-sultone 1b and 3-[13C]Hexane-1,3-sultone 2a with Butylamine. To compare better the reactivity of both sultones toward primary amino nucleophiles of the lysine type, we have performed a competitive reaction of an equimolar mixture of unsaturated and saturated sultones in water in the presence

122


Meschkat et al.

Figure 6. (A) 13C NMR spectra of acetylated HSA incubated with sultone 1a in a mixture of H2O and D2O after standard dialysis. (B) 13C NMR spectra of acetylated HSA incubated with sultone 1a in a mixture of H2O and D2O after extensive dialysis.

Figure 8. 13C NMR spectra of an equimolar solution (H2O and D2O) of sultone 1b and 2a in the presence of 30 equiv of butylamine after 11 h (A), DEPT 135 after 11 h (A′), and after 35 h (B). Peaks labeled with asterisks correspond to carbons of butylamine.

Figure 7. (A) 13C NMR spectra of acetylated HSA incubated with sultone 1b in a mixture of H2O and D2O after standard dialysis. (B) 13C NMR spectra of acetylated HSA incubated with sultone 1b in a mixture of H2O and D2O after extensive dialysis.

of butylamine. The reaction was followed by 13C NMR (Figure 8). After 11 h (Figure 8A,A′), the unsaturated sultone 1b had already completely reacted and been

converted into a mixture of trans and cis aziridines 9a and 9b (40.5 and 37.9 ppm, respectively) together with some hydrolysis product 3 (141.4 ppm). The saturated sultone 2a (86.8 ppm) was still present even if it had mainly been hydrolyzed to form 7 (70.4 ppm) or had reacted with butylamine to give adduct 10 (55.7 ppm). Traces of the elimination product 11 were also present at 126.2 ppm. After 35 h (Figure 8B), the reaction was complete (no starting material left) and it can be seen that the unsaturated sultone 1b has mainly reacted with butylamine to form aziridine derivatives 9a and 9b while the saturated sultone 2a has been mainly hydrolyzed into 7. Competitive Reaction of an Equimolar Mixture of 2-[13C]Hex-1-ene-1,3-sultone 1b and 3-[13C]Hexane-1,3-sultone 2a with Sodium Phenate. As the main targets of sultones were tyrosines, we compared the reactivity of both sultones toward sodium phenate (Figure 9). After 9 h (Figure 9A), it can be seen that the saturated sultone 2a (86.8 ppm) has already reacted to



Figure 9. 13C NMR spectra of an equimolar solution (H2O and D2O) of sultone 1b and 2a in the presence of 10 equiv of sodium phenate after 9 h (A), 35 h (B), 10 days (C), and 15 days (D). Peaks labeled with asterisks correspond to carbons of sodium phenate and the residual signal of acetone.

give the hydrolysis product 7 (70.4 ppm) and the phenate adduct 13 (77.9 ppm). The unsaturated sultone 1b (144.3 ppm) has just started to react to form tiny amounts of hydrolysis product 3 (141.4 ppm) and phenate adduct 12 (139.5 ppm). After 35 h (Figure 9B), the saturated sultone 2a has completely reacted, while the unsaturated sultone 1b is still slowly reacting. After 7 days (Figure 9C), some sultone 1b was still present, even if minor compared to hydrolysis product 3 and phenate adduct 12. The reaction was complete after 15 days (Figure 9D). Hydrolysis Reaction of an Equimolar Mixture of 2-[13C]Hex-1-ene-1,3-sultone 1b and 3-[13C]Hexane1,3-sultone 2a in the Presence of HSA. The reaction was followed by 13C NMR, and after 4 h (Figure 10A), the signal of 2a at 86.6 ppm has almost disappeared while the signal of 1b at 144.2 ppm was still present. We can observe a new signal at 70.3 ppm assigned to 7, the hydrolysis product of sultone 2a. After 24 h (Figure 10B), the saturated sultone 2a has been completely hydrolyzed with a total disappearance of the signal at 86.6 ppm while the unsaturated sultone 1b remains almost intact. After 34 days (Figure 10C), some unsaturated sultone was still present (144.2 ppm) together with its hydrolysis product 3 (141.3 ppm).

Discussion Alkyl-1-ene-1,3-sultones are very potent skin sensitizers, and this has been attributed to the presence of an electron-deficient double bond that allows the formation of protein adducts through Michael type nucleophilic additions (6). This explanation is mainly supported by the significantly lower sensitizing potential of saturated analogues of sultones. In fact, alkyl-1-ene-1,3-sultones can potentially react with nucleophilic amino acids by two types of reactions, i.e., SN substitution at position 3 and Michael addition at position 2. We have shown in a previous paper (3) that alkyl-1-ene-1,3-sultones were

Figure 10. 13C NMR spectra of an equimolar solution (H2O and D2O) of sultone 1b and 2a in the presence of HSA after 4 h (A), 24 h (B), and 34 days (C).

mainly reacting at position 3 through an SN substitution with model nucleophiles in water (Chart 1), the only exception being propanethiolate, a model for cysteine, and imidazole, a model for histidine. Alkane-1,3-sultones which are weak sensitizers were found to have a low reactivity toward amino groups and to react mainly with propanethiolate and sodium phenate through an SN substitution at position 3. These findings therefore did not support the hypothesis of a Michael addition being

124


Scheme 1. Reaction Pathways of Sultone 1 and 2 with HSA in Phosphate Buffer at pH 8.1

the key step in the induction of a sensitization to alkenesultones. To gain more insights into the reactivity of these molecules toward biological nucleophiles and to see if the mechanism and/or selectivity for specific amino acids could explain the difference in reactivity between unsaturated and saturated sultones, we have reacted hex-1ene-1,3-sultones 1a and 1b and hexane-1,3-sultone 2a with HSA in phosphate buffer. After dialysis and lyophilization, the combined use of 13C and 1H chemical shift data for labeled sites, together with the chemical shifts of model adducts (Table 1), made it possible to accurately assign a structure to each adduct. Reactivity of Sultone 1. Thus, in the case of hex-1ene-1,3-sultone, the single correlation at 68.6/4.12 ppm (Figure 3A), characteristic of a carbon-oxygen bond at position 3, is probably associated with the single correlation seen at 59.7/3.63 ppm (Figure 3D), characteristic of a carbon-nitrogen bond at position 2. The presence of these two carbon-heteroatom bonds is compatible with the formation of an amino alcohol. The assignment was confirmed by comparison of the 13C and 1H chemical shifts of the model adduct (Table 1, entry 3, X ) BuNH2), prepared by hydrolysis of aziridine at 100 °C (7). The correlations at 71.6/6.26, 71.6/6.23, and 71.2/6.19 ppm (Figure 3B), characteristic of a carbon-oxygen bond of the phenolic type (high 1H chemical shifts), are probably associated with the correlations at 137.1/5.95, 136.9/5.93, 137.0/5.90, and 137.0/5.86 ppm (Figure 3C), characteristic of protons on double bonds. These data are in good agreement with the formation of phenolic adducts at position 3 through an SN reaction, with no reaction at position 2, thus preserving the original R,β-unsaturation. This assignment is also supported by the values of 73.9/ 5.81 and 139.4/5.99 ppm observed with the model adduct (Table 1, entry 1, X ) OPh) and the calculated values of 72.4/141.0 ppm. Thus, two types of modification were seen when HSA was incubated with an excess of labeled hex-1-ene-1,3sultone (Scheme 1); the first appears to involve modifica-

Meschkat et al. Scheme 2

tion of a lysine residue and the second modification of at least four tyrosines. In a previous study, we have shown that, in water, primary amines readily react with sultone 1 via an SN reaction at position 3 and that the intermediate is immediately cyclized, via an intramolecular Michael addition, to form a 1/1 mixture of cis and trans aziridine. Aziridines are quite stable, but can be hydrolyzed in water (100 °C) to amino alcohols (7). The presence of two small signals at 49.7 and 47.7 ppm (Figure 1B) and 45.1 and 44.4 ppm (Figure 2B) suggests the formation of aziridine, or more likely aziridinium intermediates 4a and 4b. The subsequent reaction of these intermediates with water would then lead to the formation of the amino alcohol 5 and thus explain the loss of these small peaks after extensive dialysis and the increase in intensity of the peaks corresponding to the related amino alcohol. Although the very low intensity of these peaks has not allowed us to characterize their 1H cross chemical shifts, we postulate that 5 is formed through a mixture of aziridinium betaines 4a and 4b. In HSA, Lys199 has been shown to have a special reactivity (5) and to be selectively acetylated by aspirin and other acetylating agents such as PNPA. To check if the lysine reacting with sultone 1 could be this special reactive lysine, we have incubated HSA, previously acetylated with PNPA, with sultones 1a and 1b. 13C NMR spectra obtained after standard dialysis (Figures 5A and 6A) and after extensive dialysis (Figures 5B and 6B) support the idea of Lys199 being the modification site. Thus, while signals assigned to tyrosine modifications remained unchanged, peaks corresponding to a lysine adduct (aziridinium 4a and 4b) were lost as well as peaks assigned to the subsequent hydrolysis product (amino alcohol 5). Reactivity of Sultone 2. The reaction of hexane-1,3sultone 2 with only one reactive site at position 3 was more simple to analyze, and only one new peak was observed after incubation with HSA. 1H{13C} HSQC experiments (Figure 5) exhibited at least four correlations (74.65/5.06, 74.5/5.03, 74.4/4.99, and 74.5/4.97 ppm) characteristic of a carbon-oxygen bond of the phenolic type (high 1H chemical shifts). The assignment was confirmed by comparison of the 13C and 1H chemical shifts


of the model adduct (Table 1, entry 5, X ) OPh). Here again, hexane-1,3-sultone seemed to react mainly with tyrosine to form adducts of structure 8 (Scheme 1) which are not usually considered in the induction of allergic contact dermatitis. Skin Sensitization Potential versus Reactivity toward Proteins. HSA is a large protein (8) that allows a wide range of modifications and is therefore used as a model to mimic the general reactivity of proteins. It is generally accepted, mainly for historical reasons, that the main modifications involved in the mechanism of ACD occur on lysine and/or cysteine. The former modifications are implicated because of studies on the reactivity of dinitrohalobenzene derivatives, such as dinitrochlorobenzene (DNCB) or dinitrofluorobenzene (DNFB), models of strong sensitizers used in immunology for decades (9), which react with primary amines, and the latter because of the known nucleophilicity of thiols and of studies on detoxication in which glutathione plays a major role. In fact, although lysine residues have been shown to play a role in some hapten-protein modifications (10), the level of involvement of cysteine residues is probably much lower, due to their being present in much lower amounts in proteins compared with other potential nucleophiles. In recent years, histidine has also been shown to play an important role in the modification of proteins by allergising metals, such as nickel (11), or by methyl alkanesulfonates (12). In the case of sultones 1 and 2, it seems that the major amino acid that is modified is tyrosine, not often considered a nucleophile in the modification of proteins by haptens. The only significant difference between 1 and 2 was the ability of hex-1-ene1,3-sultone to react with one extra lysine residue. In our previous study on model nucleophiles (3), we have found that the unsaturated sultone 1 was highly reactive toward n-butylamine, while the saturated sultone 2 only slowly reacted with primary amine. We have thus compared the reactivity of a 1/1 mixture of sultones 1 and 2 toward butylamine and sodium phenate under competitive conditions. Thus, after 11 h (Figure 8A,A′), the unsaturated sultone 1 had completely reacted with butylamine while the saturated sultone 2 was still present. On the other hand, in the presence of sodium phenate, the saturated sultone 2 has already reacted after 9 h (Figure 9A) while the unsaturated sultone 1 has just started to react. The reaction of 2 was complete after 35 h, while it took 15 days to realize a complete reaction of sultone 1. Protein should be regarded as a mixture of nucleophiles with different electronic characteristics, which are modified by their environment, accessibility, pH, etc. Modifications occur on those amino acids with the most suitable “affinity” in terms of electronic interactions with the hapten. This principle, developed a few years ago and widely used in organic chemsitry to explain reactivity, is known as the HSAB theory (4) which states that a hard electrophile will have an increased reactivity toward hard nucleophiles and a decreased reactivity toward soft nucleophiles and, conversely, a soft electrophile will have a decreased reactivity toward hard nucleophiles and an increased reactivity toward soft nucleophiles. Hard electrophiles and nucleophiles react through low-energy vacant and occupied orbitals, respectively, and have low polarizabilities, while soft electrophiles and nucleophiles react through higher-energy orbitals and have high polarizabilities.


There has been some debate about whether skin sensitization potential can be related to the extent of protein modification or to the type of amino acid modified. Thus, hex-1-ene-1,3-sultone 1 and hexane-1,3-sultone 2 seem to have a different behavior toward nucleophiles. It seems clear they are highly oxophilic compounds as shown by their very high susceptibility to hydrolysis and reactivity toward tyrosine. Although HSA contains 59 lysine residues versus only 18 tyrosine residues and 16 histidine residues, most of the lysines are protonated and thus nonreactive at pH 8.1. It appears that only one lysine residue is modified by hex-1-ene-1,3-sultone 1, whereas at least four tyrosine residues appear to be modified. In this study, we have an example of a very strong skin sensitizer that seems to react mainly with tyrosine and, to a lesser degree, with lysine. The saturated sultone, which is a weak sensitizer, seems to react also with tyrosine and probably at a higher rate as shown by the competitive experiments. Modification of tyrosine is not usually considered in hapten modification of proteins and does not seem to play a major role in the induction of sensitization to sultones. Therefore, it seems that the selective reactivity of sultone 1 toward amino groups could play a larger role, in terms of skin sensitization, than the wide reactivity toward tyrosine. Another parameter could also explain the difference in sensitizing potential between sultones 1 and 2. When the reaction between HSA and a 1/1 mixture of sultones 1b and 2a in water at pH 8.1 was monitored in an NMR tube for >1 month, the major reaction appeared to be rapid hydrolysis (in this experiment, adduct signals are two small to be detected). Sultone 2 seems to be much more susceptible to hydrolysis (Figure 10) with an almost complete reaction after only 4 h, while after 34 days, the unsaturated sultone 1 was still present. In conclusion, the skin sensitization potential of sultone 1 seems to be related to its hability to react with amino groups such as lysine rather than to modify tyrosine residues. This finding does not support the widly accepted concept that links the sensitizing potential to the extent of protein modification by haptens. The selectivity for amino acids could play a major role in the potency of skin sensitizers.

Acknowledgment. We thank the Centre National de la Recherche Scientifique (CNRS-France) and Unilever Research for a fellowship to E.M.

References (1) Rycroft, R. J. G., Menne´, T., and Frosch, P. J., Eds. (1995) Textbook of contact dermatitis, Springer-Verlag,. (2) Lepoittevin, J. P., Basketter, D. A., Goossens, A., and Karlberg, A. T., Eds. (1997) Allergic contact dermatitis: the molecular basis, Springer, Berlin. (3) Meschkat, E., Barratt, M. D., and Lepoittevin, J. P. (2001) Studies of the chemical selectivity of hapten, reactivity and skin sensitization potency. 1. Synthesis and studies on the reactivity toward model nucleophiles of the 13C-labeled skin sensitizers hex-1-eneand hexane-1,3-sultones. Chem. Res. Toxicol. 14, 110-117. (4) Pearson, R. G. (1963) Hard and soft acids and bases. J. Am. Chem. Soc. 85, 3533-3539. (5) Kurono, Y., Maki, T., Yotsuyanagi, T., and Ikeda, K. (1979) Esterase-like activity of human serum albumin: Structureactivity relationships for the reactions with phenyl acetates and p-nitrophenyl esters. Chem. Pharm. Bull. 27, 2781-2786. (6) Ritz, H. L., Connor, D. S., and Sauter, E. D. (1975) Contact sensitization of guinea-pigs with unsaturated and halogenated sultones. Contact Dermatitis 1, 349-358.

126


(7) Wagner-Jauregg, T., and Zirndibl, L. (1963) A ¨ thylenimin-carbanilide-(2) und damit isomere pyrozolidone-(3). Justus Liebigs Ann. Chem. 668, 30-50. (8) Minghetti, P. P., Ruffner, D. E., Kuang, W. J., Dennison, O. E., Hawkins, J. W., Beattie, W. G., and Dugaiczyk, A. (1986) Molecular structure of human albumin gene is revealed by nucleotide sequence within q11-22 of chromosome 4. J. Biol. Chem. 261, 6747-6757. (9) Parker, D., Long, P. V., Phil, D., and Turk, J. L. (1983) A comparison of the conjugation of DNTB and other dinitrobenzenes with free protein radicals and their ability to sensitize or tolerize. J. Invest. Dermatol. 81, 198-201.

Meschkat et al. (10) Franot, C., Benezra, C., and Lepoittevin, J. P. (1993) Synthesis and interaction studies of 13C labeled lactone derivatives with a model perotein using 13C NMR. Bioorg. Med. Chem. 1, 389-397. (11) Romagnoli, P., Labhardt, A. M., and Sinigaglia, F. (1991) Selective interaction of Ni with an MHC bound peptide. EMBO J. 10, 1303-1306. (12) Lepoittevin, J. P., and Benezra, C. (1992) 13C-enriched methyl alkanesulfonates: new lipophilic methylating agents for the identification of nucleophilic amino acids of protein by NMR. Tetrahedron Lett. 33, 3875-3878.

TX000226F

Studies of the Chemical Selectivity of Hapten, Reactivity, and Skin

Recommend Documents