Reaction of Aflatoxin B1 Oxidation Products with ... - ACS Publications

May 10, 2002 - Aflatoxin (AF) B1 exo-8,9-epoxide hydrolysis yields AFB1 dihydrodiol, which undergoes base-catalyzed rearrangement to, and is in equili...
1 downloads 8 Views 196KB Size
780

Chem. Res. Toxicol. 2002, 15, 780-792

Reaction of Aflatoxin B1 Oxidation Products with Lysine F. Peter Guengerich,*,†,‡ Kyle O. Arneson,†,‡ Kevin M. Williams,†,‡ Zhengwu Deng,‡,§ and Thomas M. Harris‡,§ Departments of Biochemistry and Chemistry and Center in Molecular Toxicology, Vanderbilt University, Nashville, Tennessee 37232-0146 Received September 28, 2001

Aflatoxin (AF) B1 exo-8,9-epoxide hydrolysis yields AFB1 dihydrodiol, which undergoes basecatalyzed rearrangement to, and is in equilibrium with, AFB1 dialdehyde. We investigated the reaction of AFB1 dialdehyde with albumin to generate a Lys adduct, previously characterized by others [Sabbioni, G., Skipper, P. L., Bu¨chi, G., and Tannenbaum, S. R. (1987) Carcinogenesis 8, 819-824; Sabbioni, G. (1990) Chem.-Biol. Interact. 75, 1-15]. Pronase digestion of bovine albumin serum treated with AFB1 dialdehyde and HPLC yielded the adduct, identified by its characteristic UV and mass spectra. The structure of the Lys-AFB1 dialdehyde adduct is concluded to be (S)-R-amino-2,3-dihydro-2-oxo-4-(1,2,3,4-tetrahydro-7-hydroxy-9-methoxy-3,4dioxocyclopenta[c][1]benzopyran-6-yl)-1H-pyrrole-1-hexanoic acid, structure B of the former paper and 8 of the latter, based on work with the methylamine adduct described in the following paper in this issue [Guengerich, F. P., Voehler, M., Williams, K. M., Deng, Z., and Harris, T. M. (2002) Chem. Res. Toxicol. 15, 793-798]. The time course of product formation at varying concentrations of AFB1 dialdehyde could be described by complexation with albumin with a Kd of 1.5 mM and a first-order reaction rate with the N6-amino group of Lys of 0.033 min-1. The reaction of AFB1 dialdehyde with N2-acetylLys was monitored by UV spectroscopy and yielded a final spectrum similar to that of the described Lys adduct. Kinetic analysis of the changes at pH 7.2 was best described with a scheme involving equilibrium of the dialdehyde with dihydrodiol and a rate-limiting reaction of AFB1 dialdehyde with the N6 atom of N2acetylLys, with an apparent second-order rate constant of 2.6 × 103 M-1 min-1, followed by putative carbinolamine formation and rearrangement, collectively described by a first-order rate constant of 7.6 min-1. Competition experiments with the hydrolysis of AFB1 exo-8,9-epoxide indicate that N2-acetylLys also reacts with the epoxide at pH 7.2 (k ) 350 M-1 min-1) and 9.5 (k ) 1.8 × 103 M-1 min-1). This reaction might contribute to the formation of protein Lys adducts, depending upon the local concentration of free or protein Lys. Mass spectral analysis of trypsin digests of bovine serum albumin modified with AFB1 dialdehyde indicated selective modification of Lys455 and Lys548. Collectively, these results provide more insight into the mechanism of formation of AFB1 dialdehyde-protein adducts and indicate that the formation of Lys adducts is a moderately efficient process. The binding of AFB1 dialdehyde to albumin or the protonation of the N6-amino group retards the reaction with Lys residues.

Introduction Aflatoxin (AF)1 B1 is a serious health threat to individuals in many underdeveloped countries in which this mycotoxin is present in staple foods (1-4). Most of the interest in AFB1 and related AFs has been in the area of genotoxicity and liver cancer (3, 5, 6). However, the point should be made that AFB1 is also a toxicant, lethal through noncancerous events, and it was first discovered through an incident involving the deaths of turkeys consuming contaminated meal in Britain (7). The toxicity * To whom correspondence should be addressed. Phone: (615) 322-2261. Fax: (615) 322-3141. E-mail: guengerich@ toxicology.mc.vanderbilt.edu. † Department of Biochemistry. ‡ Center in Molecular Toxicology. § Department of Chemistry. 1 Abbreviations: AF, aflatoxin; CHES, 2-(N-cyclohexylamino)ethanesulfonate; MOPS, 3-(N-morpholino)propanesulfonate; Tris, tris (hydroxymethyl)aminomethane; AFAR, AFB1 aldehyde reductase; GST, GSH transferase; ESI, electrospray ionization; MALDI, matrix-assisted laser desorption ionization; BSA, bovine serum albumin.

of AFB1 may contribute a “promotional” effect to the genotoxicity of AFB1 in liver cells. The major reactions in AFB1 metabolism are relatively well established (Scheme 1). AFB1 itself is a relatively innocuous molecule but P450 enzymes (and some other oxygenases, to a lesser extent) oxidize it to the 8,9epoxide, which has a central role in all succeeding reactions (8-11). AFB1 exo-8,9-epoxide hydrolyzes rapidly to the dihydrodiol (t1/2 1 s at 23 °C) (9). Only the exo epoxide reacts efficiently with DNA, and the kinetics of this process have been analyzed in detail (12, 13). The endo epoxide (14) is a product of some P450s [e.g., P450 1A2 (15)] but does not bind to DNA because intercalation of the coumarin ring system into DNA (16, 17) does not facilitate the requisite stereochemistry required for SN2 reaction (12). The dihydrodiol/dialdehyde equilibrium is complex (pKa 8.2) and both forms are present at physiological pH (9). AFB1 aldehyde reductase (AFAR) enzymes reduce the dialdehyde (and not the dihydrodiol) (18). Only the epoxide (and not the dialdehyde) reacts

10.1021/tx010156s CCC: $22.00 © 2002 American Chemical Society Published on Web 05/10/2002

Aflatoxin B1-Lysine Adducts

Chem. Res. Toxicol., Vol. 15, No. 6, 2002 781 Scheme 1. Major Reactions in AFB1 Metabolism

Scheme 2. A Postulated Reaction of AFB1 Dialdehyde with Lys (21)

with DNA, as documented by the nearly exclusive reaction of the exo epoxide (12). The epoxide is also a substrate for conjugation by several GSH transferase (GST) enzymes (19, 20). AFB1 dialdehyde is generally considered to be the major AFB1 derivative responsible for the formation of protein adducts (21). These adducts may be responsible for the acute AFB1 toxicity. For instance, treatment of experimental animals with chemicals (e.g., ethoxyquin) that induce AFAR has chemoprotective effects (22) (although part of the effect may be due to the induction of GSTs, which interact with the epoxide). The only protein adduct described is the Lys product, which was characterized by Sabbioni et al. (21, 23). The reaction was postulated to involve initial Schiff base formation by the N6 atom of Lys, an Amadori rearrangement, and subsequent Schiff base formation and bond rearrangement (Scheme 2) (21). An alternative structure of the Lys adduct was proposed in the work of Sabbioni (23), and the structure requires an alternate scheme because the formation of an amide is not particularly feasible follow-

ing an Amadori rearrangement (Scheme 3). Our own work with the model methylamine adduct formed by reaction with AFB1 dialdehyde supports the structure B of ref 21 and structure 8 of refs 23 and 24. Thus, Scheme 3 is preferred for formation of Lys adducts. Further details of the events associated with this reaction, e.g., kinetics, are unknown. A thorough understanding of the reaction is in order because albumin Lys adducts are used extensively as biomarkers of AFB1 exposure and risk, as well of the potential contribution of protein-AFB1 adducts to toxicity (23, 25-27). We address several questions in our own long-term effort to describe the mechanisms and kinetics of all of the major reactions involved in AFB1 metabolism (9-20, 24, 28-39). Rates of reaction of AFB1 dialdehyde with protein Lys groups, and N2-acetylLys have been measured to describe the kinetics of the process. We also addressed the question of whether Lys can directly react with AFB1 epoxide. Finally, we used MS to identify the primary site(s) of modification of bovine serum albumin (BSA) by AFB1 dialdehyde.

782

Chem. Res. Toxicol., Vol. 15, No. 6, 2002

Guengerich et al.

Scheme 3. Reaction of AFB1 Dialdehyde with Lys (23), Supported by the Structure of the Methylamine Adduct (24)

Experimental Procedures Caution: AFB1 exo-8,9-epoxide is a potent mutagen and probable direct-acting carcinogen and should be handled only with gloves and other appropriate precautions. AFB1 and AFB1 dialdehyde should also be treated with the same care. Glassware and unused samples should be treated with commercial bleach (HOCl). Chemicals. AFB1 exo-8,9-epoxide was synthesized from AFB1 (Sigma Chemical Co., St. Louis, MO) by oxidation with dimethyldioxirane (8) and recrystallized from CHCl3/(CH3)2CO (8, 14). The material was stored desiccated at -20 °C. Samples were dissolved in dry CH3CN (distilled over P2O5 and stored over molecular sieves), and aliquots were either used directly in reactions or hydrolyzed to AFB1 dihydrodiol/dialdehyde by dilution in 20 mM sodium 2-(N-cyclohexylamino)ethanesulfonate (CHES) buffer (pH 10.0). Dilutions of AFB1 dihydrodiol were made in pH 4.6 (50 mM sodium acetate) buffer to estimate concentrations, using 360 ) 21 800 M-1 cm-1 (40). N2-Acetyl-L-Lys was a gift of H. P. Broquist, Vanderbilt University (41). AFB1 dialdehyde (3.4 nmol, in 2.0 mL of 0.10 M potassium phosphate buffer, pH 7.4) was treated with N2acetylLys (50 mg) overnight at 37 °C in the general procedure described by Sabbioni et al. (21) and treated with acylase (20 mg, from Sigma) overnight to prepare a standard Lys-AFB1 dialdehyde conjugate, which was purified by HPLC separation. The 399 value of 25 400 M-1 cm-1 (pH 7.0) reported earlier (21) was used in standardizations. Spectroscopy and HPLC of AFB1 Dialdehyde-Lys Adducts. UV spectra and kinetic traces were recorded using a Cary 14/OLIS spectrophotometer (OLIS, Bogart, GA). HPLC was done using a Spectra-Physics 8700 pumping system (Thermo-Separations, Piscataway, NJ) and either a 10 × 250 mm Beckman Ultrasphere octadecylsilane (C18) column (5 µm, San Rafael, CA) or a 6.2 × 80 mm Zorbax octysilane (C8) column (3 µm, MacMod, Chadds Ford, PA). The effluent passed through a ThermoSeparation Products 6000 detector equipped with a rapidscanning monochromator (ThermoSeparation Products, Piscataway, NJ) and a McPherson FL-750BX spectrofluorometer (McPherson, Chelmsford, MA), connected in-line. UV spectra were recorded on-line with this system, with subtraction of background absorbance as necessary. For the separation and analysis of Pronase digests of BSA-AFB1 dialdehyde conjugates, solvent A was 20 mM potassium phosphate (pH 7.2) and solvent B was a 9:1 mixture of CH3CN/H2O, v/v. The system was held at 100% A for 2 min after injection of a sample and then programmed to 70% B (and 30% A, v/v) at an elapsed time of 20 min. The gradient was held at 70% B (v/v) for another 10 min and then returned to 100% A over the next 4 min. The flow rate was 1.0 mL min-1 with the 6.2 × 80 mm column and 4.0

mL min-1 with the 10 × 250 mm column. The settings generally used for analyses were (UV) 380 nm and (fluorescence) 390 nm excitation/440 emission (filter). For the analysis of hydrolysis of AFB1 exo-8,9-epoxide and possible reaction with N2-acetylLys, the HPLC systems described previously were used with only slight modification (10 × 250 mm Beckman Ultrasphere octadecylsilane, flow 4.0 mL min-1, UV 360 nm, fluorescence 360 nm excitation/440 emission filter, increasing CH3CN gradient of 5% to 50% in H2O (v/v) with 0.1% CH3CO2H, v/v). MS of AFB1-derived Lys adducts was performed on a Finnigan TSQ-7000 triple quadrupole mass spectrometer (San Jose, CA) equipped with a standard API-1 electrospray ionization (ESI) source with a deactivated fused silica capillary heated to 200 °C. N2 was used as both the sheath and auxiliary gases (70 and 10 psi, respectively). Spectra were acquired in the positive mode using an ESI needle voltage of 4.5 kV. The tube lens was operated at 85 V, and the electron multiplier was set to 1800 V. The autosampler and HPLC system consisted of an Alliance 2690 Separations Module from Waters (Milford, MA). A PE Biosystems (Foster City, CA) 785A Programmable Absorbance UV detector was set to 360 nm. A Zorbax Rx octylsilane (C8) column (2.1 × 150 mm) was connected to the system, and 20 µL aliquots of the samples were injected. The gradient conditions were as follows: 0.2 mL min-1 flow rate; solvent A, 0.1% HCO2H in H2O; solvent B, CH3CN; 95% A/5% B (v/v) at time zero, 95% A/5% B (v/v) at 5 min, 50% A/50% B (v/v) at 50 min, 10% A/90% B (v/v) at 55 min, 10% A/90% B (v/v) at 65 min, 95% A/5% B (v/v) at 67 min, and 95% A/5% B at 75 min. Incubations of AFB1 Dialdehyde with BSA. Each incubation contained 3.0 mg of BSA dissolved in 10 mM sodium 3-(Nmorpholino)propanesulfonate (MOPS) buffer (pH 7.2), in a final reaction volume of 250 µL (after addition of the AFB1 dialdehyde). The indicated concentration of AFB1 dialdehyde, dissolved in a mixture of CH3CN/20 mM sodium CHES (pH 10.0), 4:1, v/v, was added to start the reaction at room temperature (18). At each indicated time point, 50 µL of 1.0 M tris (hydroxymethyl)aminomethane (Tris)‚HCl buffer (pH 8.0) was added to quench the reaction (i.e., ∼30-fold excess of primary amine relative to Lys in BSA). The samples were allowed to stand for 5 min, for the Tris to react with residual AFB1 dialdehyde (42), and then 2.0 mL of cold (-20 °C) C2H5OH was added to each sample to precipitate the BSA (>98%). The samples were held at -20 °C for 30 min, and then the BSA was precipitated by centrifugation at 3 × 104 × g for 15 min. The BSA pellets were washed with another 2.0 mL of cold C2H5OH and recovered by centrifugation as before. Residual C2H5OH was removed under an N2 stream (at room temperature), and the samples were dissolved in 250 µL of 10 mM sodium MOPS buffer (pH 7.2). Pronase (Fisher Biotech, Fair Lawn, NJ, 1.0 mg in 57 µL of the

Aflatoxin B1-Lysine Adducts

Chem. Res. Toxicol., Vol. 15, No. 6, 2002 783

Figure 1. Spectra of purified Lys-AFB1 dialdehyde adduct. (A) UV-vis, pH 7.2, potassium phosphate buffer. (B) ESI-MS. Compare with refs 21 and 24. same buffer) was added, and the samples were incubated at 37 °C with gentle shaking (under Ar) for 14 h (21, 23). The samples were applied to Baker Bond 3 mL spe octadecyl extraction columns, which had been prewashed with CH3OH and H2O. Samples were processed as described (27) and the eluates (CH3OH/H2O, 4:1, v/v) were concentrated under N2 at 23 °C and then adjusted to a total volume of 1.0 mL. A 200 µL aliquot of each incubate was directly injected onto the HPLC column (usually 6.2 × 80 mm) for analysis, with quantitation of the Lys adduct using F390/440 measurements and an external standard (vide supra). Reaction of AFB1 Dialdehyde with N2-AcetylLys. The indicated concentration of N2-acetylLys (in each figure or text) was dissolved in 50 mM sodium MOPS buffer (pH 7.2) at 23 °C. The reaction was started by the addition of a small aliquot of AFB1 dialdehyde [in CH3CN:20 mM sodium CHES (pH 10.0), 4:1, v/v], to give a final concentration of 50 µM. The reaction was monitored either by collection of repetitive spectra (250480 nm) or single wavelength kinetic analysis (A400). Reaction of AFB1 exo-8,9-Epoxide with N2-AcetylLys. AFB1 exo-8,9-epoxide was dissolved in dry CH3CN and an aliquot (to yield a final concentration of 260 µM) was added to a 250 µL solution containing varying concentrations of N2acetylLys adjusted to either pH 7.2 or 9.5. After 10 s (at 23 °C) each reaction was quenched by the addition of 50 µL of concentrated HCO2H, and a 50 µL aliquot was immediately injected onto the HPLC column (vide supra). The amount of AFB1 dihydrodiol was quantified by F360/440 measurements (9, 18, 20). MS Analysis of Peptides Derived from BSA Modified with AFB1 Dialdehyde. BSA (2.0 mg in 100 µL of 10 mM sodium MOPS buffer, pH 7.2) was treated with 1, 2, 3, 5, and 10 M equiv of AFB1 dialdehyde (diluted into the BSA from a concentrated sodium CHES, pH 10, stock solution) and allowed to react 60 min at room temperature. The BSA was precipitated by the addition of 2.0 mL of cold C2H5OH and collected following centrifugation for 15 min at 3 × 104 × g. The pellets were washed twice with 2.0 mL of C2H5OH, with centrifugation each time as before. The modified BSA was dissolved in 100 µL of NH4HCO3 buffer (pH 9). An 8 µL aliquot of the BSA solution (i.e., 2.5 nmol of BSA) was added to 88 µL of a trypsin solution (in NH4HCO3 buffer) to give a final trypsin concentration of 0.25 mg mL-1 and incubated overnight at 37 °C. Aliquots of the resulting peptides were analyzed by matrixassisted laser desorption ionization (MALDI) MS using R-cyano4-hydroxycinnamic acid (10 mg mL-1 in CH3CN:2% aqueous CF3CO2H, 60:40, v/v) as a matrix. Samples used for MALDI analysis were diulted in 0.1% CF3CO2H (v/v) and then desalted with ZipTipC18 pipet tips (Millipore, Milford, MA). Spectra were aquired on a Perseptives Voyager Elite instrument in the positive reflector mode. Bradykinin and insulin were used as external references.

HPLC/MS of tryptic peptides was performed using a Zorbax Rx C8 column (2.1 × 150 mm) with the same gradient described above; 20 µL aliquots of the samples were injected. A needle voltage of 4 kV and a tube lens voltage of 100 V were used, and the UV detector was set at 310 nm. Kinetic Analysis and Reaction Modeling. Changes in A400 were recorded at 23 °C using a Cary 14/OLIS instrument, saved as ASCII files, and converted to Cricket Graph and then Graphpad Prism files (Graphpad, San Diego, CA) for analysis. Biexponential reactions (decrease followed by increase) were divided into two parts by inspection and each phase was fit to a single exponential. Plots of formation of BSA-AFB1 dialdehyde adduct formation vs time were obtained from HPLC data with the Lys-AFB1 dialdehyde adducts, converted to product molarity [assuming complete recovery in the methods (21, 23)]. Several plots were converted to Word text files and fit with the program DynaFit using a Macintosh G4 computer (Apple Computer, Cupertino, CA). DynaFit is a fitting and simulation system used to analyze kinetic data (43) (available from P. Kuzmic, http://www.biokin.com) by nonlinear regression methods, in the same general manner as the program FITSIM (43). In this work, data files were fit to mechanisms (as indicated), and the indicated constants were derived from the process. Some relevant examples are presented in the Supporting Information, complete with text files used to run the software. In the system, one or more parameters may be indicated as of unknown value (with a trial value and a “?”) and the system will provide estimates (to the mechanism) by reiterative fitting. Error estimates are somewhat artificial in that a mechanism is assumed/tested in each case. However the “% error” and other variance parameters are provided in the Supporting Information (“% error” for Figure 3, K2 33% and k3 23%; Figure 5B, k2 20% and k3 9%; Figure 6C, k2 19%).

Results Preparation and Characterization of AFB1 Dialdehyde-Lys Adduct. AFB1 dialdehyde was reacted with N2-acetylLys and the resulting mixture was treated with acylase. HPLC of the mixture yielded two major fluorescent peaks (F390/440 at pH 7.2), the latter of which was found to have same tR as the major Pronase digestion product recovered from AFB1 dialdehyde:BSA adducts. On the basis of the UV and mass spectra of the recovered adduct (Figure 1), this compound is assigned the same structure as the major Lys conjugate reported by Sabbioni et al.. (21) as B and by Sabbioni (23) as 8, derived in that work from either (i) Pronase treatment of serum

784

Chem. Res. Toxicol., Vol. 15, No. 6, 2002

Guengerich et al.

Figure 3. Formation of Lys:AFB1 dialdehyde adducts in BSA as a function of time and AFB1 dialdehyde concentration. Time course plots at AFB1 concentrations of 0.2 (O), 0.4 (0), 0.7 (b), and 1.6 (4) mM AFB1 dialdehyde and BSA at 3 mg mL-1, with fits made using simulation analysis yielding K2 ) 1.5 mM and k3 ) 0.033 min-1 (see Supporting Information).

Figure 4. Spectral intermediates in the reaction of AFB1 dialdehyde with N2-acetylLys. The reaction was initiated (23 °C) with the addition of 15 µL of a 3.4 mM solution of AFB1 dialdehyde in 20 mM sodium CHES (pH 10.0) to 985 µL of a solution of 100 mM N2-acetylLys in 50 mM sodium MOPS buffer (pH 7.2) and spectrum 1 was recorded immediately. Subsequent spectra (2, 3, 4, ...) were recorded every 10 min, and successive spectra are shown with alternate solid and dashed lines.

Figure 2. HPLC analysis of Pronase digest of BSA:AFB1 dialdehyde conjugate. HPLC was done on a 10 × 250 mm Beckman semipreparative (5 µm) octadecylsilane column as described under Experimental Procedures. (A) A380 trace. (B) F390/440 trace. (C) UV spectrum of peak eluting at tR 19.7 min (rapid-scanning monochromator), after subtraction of absorbance at tR 19.4 min. MS of the tR 19.7 min peak of part C (tR 26.4 in HPLC/MS system) yielded an apparent MH+ ion at m/z 457 (results not presented).

albumin derived from rats treated with AFB1 or (ii) acylase digestion of the product of the reaction of N2acetylLys with 8,9-dibromo-8,9-dihydro-AFB1, an electrophilic model of AFB1 8,9-epoxide. The UV spectrum at pH 7.2 showed peaks at 400, 333, and 256 nm (Figure 1A). The band at 400 nm disappeared when the pH was lowered to 5 and the 333 nm based was shifted (21, 24). The mass spectrum showed the expected MH+ ion at m/z 457 (Figure 1B). The structure of the methylamine adduct is presented in the following (24) and the identity of the Lys adduct is based on this work, particularly the 1H NMR spectra. Characterization of Lys Adduct in BSA Treated with AFB1 Dialdehyde. BSA was incubated with AFB1 dialdehyde (prepared at pH 10 and then added to a pH 7.2 solution of BSA). The reaction was quenched at various times with a large excess (30× in free amine) of

Tris buffer, which should compete effectively for reaction with any residual AFB1 dialdehyde (42), and then the BSA was precipitated by the addition of cold C2H5OH. HPLC of the Pronase digest (following initial cleanup with a reversed-phase cartridge) yielded a complex pattern when absorbance was monitored (Figure 2A). Fluorescence monitoring yielded a simpler pattern (Figure 2B), with about five peaks detected (a-e). Of these only peaks d and e increased with the reaction time prior to the addition of Tris to the reactions, and the peaks seen prior to d and e were present in samples that were not treated with Pronase. The UV spectrum of peak d of Figure 2B is the same as that of the conjugate described by Sabbioni et al. (21) and the methylamine adduct we describe in the following paper in this issue (24). MS analysis yielded the MH+ ion (m/z 457). Peak a is a Tris conjugate, dependent upon the addition of Tris buffer. Peaks b and c, associated with the broad UV band of Figure 2A, are probably noncovalently bound AFB1 dialdehyde derivatives or Tris rearrangement products, because they were present in samples that had not been treated with Pronase. Peak e may be an unknown amino acid adduct formed from AFB1 dialdehyde, but we were unsuccessful in obtaining a mass spectrum. Rates of Reaction of AFB1 Dialdehyde with BSA. The approach developed in the work presented in Figure

Aflatoxin B1-Lysine Adducts

Chem. Res. Toxicol., Vol. 15, No. 6, 2002 785

Figure 5. Kinetic analysis of spectral changes in the reaction of N2-acetylLys with AFB1 dialdehyde. (A) A400 traces (continuous) are shown for experiments done with a final concentration of 52 µM AFB1 dialdehyde, using the experimental protocol described in Figure 4, with either no added N2-acetylLys (- - -) or 5 mM N2-acetylLys (s). (B) Plot of rate constants for the second phase of the reactions shown in part A (see also Figure 4) as a function of [N2-acetylLys]. Simulation analysis of the points yielded k2 ) 2.6 × 103 M-1 min-1 and k3 ) 7.6 min-1. See Supporting Information for Dynafit file.

2 was utilized in kinetic analysis. The amount of adduct (fluorescent peak d of Figure 2B) increased with reaction time and with the concentration of AFB1 dialdehyde (Figure 3). The concentration of adduct formed could be quantified by comparison with the synthetic standard, the concentration of which was calculated from 399 ) 25 400 M-1 cm-1 at pH 7 (21). Initial inspection of the results might suggest a secondorder process, but the intercepts are not zero and the reaction would not be expected to be strictly second-order. The situation is complicated by (i) the known conversion of AFB1 dialdehyde to dihydrodiol at pH 7 [k ) 0.30 min-1 (9, 18)], which is presumed to hinder reaction with nucleophiles, and (ii) the known affinity of AFB1 for albumin (44-46). The mechanism used for the analysis of the time course of the formation of AFB1 dialdehyde:(BSA)Lys adducts at varying AFB1 dialdehyde concentrations was

k1 AFB1 dialdehyde {\} AFB1 dihydrodiol k-1

adduct recovered as AFB1 dialdehyde-Lys adduct, k1 ) 0.30 min-1 [(9, 18) and vide infra], and k-1 ) 0.012 min-1 (9). Reversible binding to BSA (K2) was included in the mechanism. The parameters K2 and k3 were varied to find the most reasonable values when a reiterative fitting algorithm was applied (Figure 3) (a sample text used in the fitting and the associated file are included in the Supporting Information). The analysis yielded estimates of K2 (Kd) ) 1.5 mM and k3 ) 0.033 min-1. With lower values of K2 (