Chem. Res. Toxicol. 2002, 15, 1445-1450
1445
Covalent Modification of Amino Acid Nucleophiles by the Lipid Peroxidation Products 4-Hydroxy-2-nonenal and 4-Oxo-2-nonenal† Jonathan A. Doorn and Dennis R. Petersen* Department of Pharmaceutical Sciences, School of Pharmacy, The University of Colorado Health Sciences Center, Denver, Colorado 80262 Received August 1, 2002
Lipid peroxidation yields the aldehydes 4-hydroxynonenal (4HNE) and 4-oxononenal (4ONE). Protein adduction by 4HNE is thought to be involved in the pathogenesis of several diseases. Currently, the reactivity of 4ONE toward proteins is unknown. The purpose of this study was to identify amino acids that react with 4HNE and 4ONE, characterize the chemical structure of the adduct, and determine the preference for amino acid modification. Model peptides containing one or more nucleophilic residues (i.e., Arg, Cys, His, Met, and Lys) were reacted with 4HNE and 4ONE and analyzed using matrix-assisted laser desorption/ionization mass spectrometry. Post-source decay analysis was used to confirm peptide modification. The bimolecular rate constant for adduction of amino acids and peptides by 4HNE and 4ONE was measured. Results of this work indicate that Cys, His, and Lys are modified by 4HNE and 4ONE. In addition, Arg was adducted by 4ONE. The predominant adduct resulting from modification of peptides by 4HNE or 4ONE had a mass of 156 or 154 Da (respectively), indicating that adduction occurs via Michael addition. Reactivity of amino acids toward 4HNE and 4ONE was found to have the following order: Cys . His > Lys (> Arg for 4ONE). The presence of an Arg on a Cys-containing peptide increased the reaction rate with 4HNE and 4ONE by a factor of ∼5-6 compared to the Cys nucleophile alone. Rate constants determined for the modification of Cys by the lipid aldehydes demonstrated a >100-fold difference in reactivity between 4HNE and 4ONE toward Cys. Results of the present study indicate that both 4HNE and 4ONE modify amino acid nucleophiles; however, the reactivity between these two lipid aldehydes differs both qualitatively and quantitatively.
Introduction Cellular oxidative stress can initiate the peroxidation of lipids, yielding reactive aldehydes capable of covalently modifying biological macromolecules (i.e., proteins, DNA) (1). 4-Hydroxynonenal (4HNE;1 Figure 1A) is documented to be a cytotoxic lipid aldehyde produced from lipid peroxidation (2). In aqueous solution, 4HNE is thought to react with Cys, His, and Lys residues via Michael addition to yield an adduct that cyclizes to a hemiacetal, and the equilibrium has been shown to favor rearrangement to the cyclic hemiacetal (1, 3-5). The predicted physiologic consequence of this reaction is that minimal protein cross-linking and aggregation will occur for 4HNE-adducted proteins because only a fraction of aqueous 4HNE-protein adducts will theoretically have a free aldehyde to react with protein amines. The adduction of proteins/biomolecules by 4HNE and/ or other lipid aldehydes is thought to be an initiating or † Presented in part at the 41st Annual Meeting of the Society of Toxicology, Nashville, TN, March 17-21, 2002. * To whom correspondence should be addressed at the School of Pharmacy C-238, The University of Colorado Health Sciences Center, 4200 E. Ninth Ave., Denver, CO 80262. Telephone: (303) 315-6159. Fax: (303) 315-0274. E-mail:
[email protected]. 1 Abbreviations: 4HNE, 4-hydroxynonenal; 4ONE, 4-oxononenal; DMP, Dess-Martin periodinane; DTNB, 5,5′-dithiobis(2-nitrobenzoic acid); GSH, glutathione; MALDI-TOF-MS, matrix-assisted laser desorption/ionization time-of-flight mass spectrometry; MH+, protonated molecule; NAC, N-acetyl-cysteamine; PSD, post-source decay.
propagating factor in the pathophysiology of several diseases, such as athereosclerosis (6-8) and diabetesrelated complications (9, 10). Proteins important in carbohydrate metabolism, mitochondrial respiration, and DNA transcription are known to be covalently modified by 4HNE (11-13). However, additional studies are warranted to fully characterize the reactivity of 4HNE toward protein nucleophiles and elucidate the cellular targets of this lipid aldehyde. Recently, 4-oxononenal (4ONE; Figure 1B) was discovered to be a major product of lipid peroxidation (14, 15). While 4ONE has been found to modify covalently DNA bases (16, 17), the chemical reactivity of 4ONE toward protein/amino acid nucleophiles has not been previously reported. Compared to 4HNE, 4ONE is structurally analogous but contains a different functional group at the C4 position that is predicted to make the compound more reactive toward protein nucleophiles than 4HNE for the following two reasons. First, the ketone at the C4 is an additional target for nucleophiles (e.g., amines) (18, 19). Second, hydration of the aldehyde moiety (18) on 4HNE but not 4ONE results in loss of the R,β-unsaturated carbonyl, yielding the molecule unreactive as a Michael acceptor toward protein nucleophiles.2 2 The K eq for hydration of ketones is very low, ranging in values from 10-2 to 10-4 (18). Therefore, the ketone at C4 of 4ONE will be hydrated only to a minimal extent (i.e., 10-fold excess with the thiol-containing compound and stopping the reaction with the addition of excess dithiobis(nitrobenzoic acid) (28). The absorbance at 412 nm was measured over time and used to calculate [thiol] (28). k′ was determined from linear regression of ln(% 4HNE or 4ONE) versus time, and k was calculated from the slope of k′ versus [amino acid/peptide]. Each k′ was calculated from at least three independent experiments. Statistical analysis was performed using Graph Pad Prism 3.02 (Graph Pad Software, San Diego, CA).
Results Adduction of Peptides by 4HNE and 4ONE. Representative MALDI-TOF-MS spectra are shown for peptides modified by 4HNE and 4ONE in Figures 2 and 3,
Amino Acid Modification by 4HNE and 4ONE
Figure 2. Representative MALDI-TOF-MS spectra demonstrating adduction of Ac-RRWWCR-amide by 4HNE. (A) Spectrum of unmodified peptide (m/z 1004.5). (B) Spectrum of peptide treated with 4HNE (m/z 1160.6). The difference in m/z between peaks representing 4HNE-modified and unmodified peptide is 156, which corresponds to one 4HNE molecule.
Chem. Res. Toxicol., Vol. 15, No. 11, 2002 1447
Figure 3. Representative MALDI-TOF-MS spectra demonstrating adduction of Ac-RRWWCR-amide by 4ONE. (A) Spectrum of unmodified peptide (m/z 1004.1). (B) Spectrum of peptide treated with 4ONE (m/z 1158.1 and 1311.2). The differences in m/z between peaks representing 4ONE-modified and unmodified peptides are 154 and 307, which correspond to one 4HNE and one 4ONE molecule, respectively.
Table 1. Model Peptides Reacted with 4HNE and 4ONE peptide Ac-RRWWCR-amide Ac-FNleRF Ac-PFRSVQ Ac-WEHD-aldehyde Ac-EHFRWG For-NleLFNleYKe Ac-RYYRIK-amide Ac-DYMGWM-amide BOC-GWMDFe
4HNE adduct massa 4ONE adduct massa 156.5 ((0.1800) NDb NDb 156.2 ((0.1733) 156.0 ((0.3735) 156.4 ((1.293) 155.7 ((0.2970) NDb NDb
153.8 ((0.4139)c 154.2 ((0.2589) 154.9 ((0.3009) 155.0 ((0.5838) 154.1 ((0.01200) 155.6 ((1.707) 154.2 ((0.09060)d NDb NDb
a Values shown are the mean adduct mass (Da) ( SE, determined from three independent experiments. Each adduct mass represents the difference in m/z observed between the 4HNE/ 4ONE-modified peptide and the unmodified peptide. b None detected. No 4HNE- or 4ONE-adducted peptides were observed in the mass spectrum. c A second 4ONE adduct was detected on this peptide with average mass of 154.1 ((0.5521) Da. d A second 4ONE adduct was detected on this peptide with average mass of 154.1 ((0.4687) Da. e For and BOC refer to the N-blocking groups formyl and butyloxycarbonyl, respectively.
respectively. Peptides containing Cys, Lys, and His reacted with 4HNE to yield a peak in the spectra corresponding to the peptide with a 4HNE adduct (Figure 2; Table 1). Neither Arg- nor Met-containing peptides were covalently modified by 4HNE. The predominant adduct mass resulting from peptide modification with 4HNE was 156 Da, which corresponds to the theoretical mass of an adduct resulting from Michael addition. In contrast, 4ONE reacted with model peptides containing Arg, His, Cys, or Lys residues to yield a peak in the spectra corresponding to the peptide with a covalently bound 4ONE molecule (Figure 3; Table 1). Met-containing peptides were not modified by 4ONE. The main
Figure 4. Representative PSD spectra demonstrating fragmentation of Ac-RYYRIK-amide modified by 4ONE, confirming the presence of a 4ONE adduct. The peak with m/z 940.7 (MH+4ONE) corresponds to the mass of the unmodified peptide, representing loss of the 4ONE adduct. Minor peaks were observed corresponding to fragmentation of the peptide chain.
product resulting from the reaction of peptides with 4ONE had a mass of 154 Da and corresponds to the adduct resulting from Michael addition. Treatment of AcRRWWCR-amide and Ac-RYYRIK-amide with 4ONE resulted in both mono- and bis-4ONE adducts of each peptide (Figure 3; Table 1). PSD analysis was used to confirm peptide adduction by 4HNE and 4ONE (Figure 4). For both 4HNE- and 4ONE-modified peptides, the predominant fragment ion observed was the unmodified peptide, resulting from neutral loss of the 4HNE or 4ONE adduct (m/z 940.7 in Figure 4). Minor peaks were observed in the spectra corresponding to fragmentation of the peptide chain (see Figure 4 for example of peptide fragmentation).
1448
Chem. Res. Toxicol., Vol. 15, No. 11, 2002
Doorn and Petersen
Table 2. Rate Constants for the Reaction of 4HNE/4ONE with Amino Acids and Peptides k (M-1 s-1)a amino acid/peptide Ac-Lys-amide Ac-histamine Ac-Arg-amide Ac-cysteamine GSH Ac-RRWWCR-amide
4ONE
k4ONE/k4HNEb
4HNE 10-3
7.46 ((0.610) × 2.21 ((0.170) × 10-2 6.11 ((0.850) × 10-4 186 ((29.3) 145 ((10.1) 887 ((88.3)
10-3
1.33 ((0.0830) × 2.14 ((0.312) × 10-3 NDc 1.21 ((0.0268) 1.33 ((0.0500)d 6.79 ((0.646)
5.61 ((0.577) 10.3 ((1.70) NDc 153 ((24.3) 109 ((8.63) 130 ((18.0)
a Values shown are mean ( SE. All rate constants were determined at 23 °C and pH 7.4. b Values shown are mean ( SE. SE calculated using standard rules of error propagation for quotients (29). c Not determined. Mass spectral analysis of Arg-containing peptides demonstrated that Arg does not react with 4HNE to yield a covalent adduct. d Value is in agreement with that calculated by Esterbauer et al. under similar experimental conditions (i.e., 1.09 M-1 s-1; 1).
Figure 6. Time-course for the reaction of GSH with 4ONE demonstrating that the reaction is bimolecular, proceeding via a 1:1 titration.
Figure 5. Representative kinetic plots used for determination of the bimolecular rate constant (k). Each data point represents the pseudo-first-order rate constant (k′) for the reaction of 4HNE and 4ONE with NAC, calculated from 3 experiments. k is determined from the slope of each kinetic plot. (A) 4HNE and NAC. (B) 4ONE and NAC.
Kinetics of Amino Acid and Peptide Adduction by 4HNE and 4ONE. The rate of amino acid/peptide modification by 4HNE and 4ONE was measured, and the second-order rate constant for the reaction was calculated (Figure 5A,B; Table 2). As shown in Table 2, the reactivity of amino acids toward 4HNE was found to have the following order: Cys . His > Lys. The rate constants for the reaction of N-acetyl-cysteamine (NAC) and glutathione (GSH) with 4HNE were not significantly different (p > 0.05). However, the Arg-containing peptide (AcRRWWCR-amide) had a 5-6-fold higher rate constant than NAC and GSH for modification by 4HNE. The reactivity of amino acids toward 4ONE was found to have the following order: Cys . His > Lys > Arg (Table 2). The rate constant for the reaction of 4ONE with NAC was higher than that for the reaction of GSH and 4ONE. However, the means were not significantly different (p > 0.05) based on one-way ANOVA analysis of the rate constants for the three thiol-containing compounds. The rate constant for modification of Ac-RRWWCR-amide by 4ONE was 5-6-fold higher than that for the reaction of 4ONE with NAC and GSH. Because the reaction order of 4ONE with nucleophiles has not been previously characterized, an experiment was
performed to verify that the reaction of 4ONE with nucleophiles was bimolecular. Furthermore, Esterbauer et al. demonstrated that 1 mol of 4HNE can react with 2 mol of cysteine under certain conditions (4). The reaction of 4ONE with GSH was found to proceed via a 1:1 titration under the conditions used, and, therefore, be bimolecular (Figure 6). A comparison of rate constants revealed that 4ONE is more reactive than 4HNE toward amino acid nucleophiles (Table 2). For the thiol nucleophiles, the k4ONE/ k4HNE ratio varied from 110 ((8.7) for GSH to 153 ((24) for NAC. 4ONE was also found to be more reactive toward the amine nucleophiles, with the k4ONE/k4HNE ratio ranging from 6 to 10.
Discussion Previous studies found that 4HNE is reactive primarily toward Cys, His, and Lys amino acids (1, 3-5). The same result was found in the present study in that peptides containing Cys, His, and Lys residues were modified by 4HNE to yield a peptide with a 156 Da adduct. The mass of the adduct corresponds to 1 molecule of 4HNE and indicates that the reaction of 4HNE with amino acid nucleophiles occurs via a 1:1 titration via Michael addition. Such a result is in agreement with a previous report that the Michael adduct is the most stable modification in an aqueous environment (3). Under the conditions used in the present study (i.e., pH 7.4, 37 °C), neither Arg- nor Met-containing peptides reacted with 4HNE to yield a stable covalent adduct. It is conceivable that 4HNE could modify nucleophilic residues such as Arg, Met, Ser, and Tyr; however, conditions needed for such reactions would most likely be nonphysiologic (e.g., pH . 7.4, temperature . 37 °C). While it has been shown that 4ONE modifies DNA bases (16, 17), the present study is the first to demonstrate and characterize the reactivity of 4ONE toward amino acids. In contrast to 4HNE reactivity, 4ONE was found to modify peptides containing Arg, Cys, His, and
Amino Acid Modification by 4HNE and 4ONE
Lys residues to yield a stable, covalent adduct. The observed adduct mass was 154 Da, which corresponds to one molecule of 4ONE and indicates that the reaction of 4ONE with peptides occurs via Michael addition. Modification of Arg by 4ONE (see Figure 1) but not 4HNE may be due to the reactivity of Arg toward dicarbonyl compounds (e.g., phenylglyoxal, 2,3-butanedione) (3032). Arg is therefore a unique target for 4ONE, and the Arg-4ONE adduct could potentially represent a biomarker enabling discrimination between 4HNE- versus 4ONE-protein adduction. For the peptides Ac-RRWWCR-amide and Ac-RYYRIKamide treated with 4ONE, peaks were observed corresponding to each peptide with one or two 4ONE molecules covalently adducted. This result was not observed for 4HNE and implies that an Arg was modified on each peptide in addition to the Cys or Lys for Ac-RRWWCRamide or Ac-RYYRIK-amide (respectively). Such a finding for these short-chain polypeptides implies that covalent binding of a 4ONE molecule to a peptide may not hinder (via steric or electronic effects) the adduction of the peptide by a second 4ONE molecule. The reactivity of 4HNE toward amino acids was found to have the following order: Cys . His > Lys. Results of kinetic experiments for 4ONE revealed a similar order of reactivity: Cys . His > Lys > Arg. Such findings indicate that Cys is the most likely residue adducted by 4HNE and 4ONE in a system containing various nucleophilic amino acids (e.g., biological sample). However, the experiments performed in the present study utilized peptides and single amino acid derivatives in an aqueous, buffered environment, and it remains to be demonstrated that the same order of reactivity for 4HNE and 4ONE occurs in protein microenvironments that have low dielectric constants (i.e., ) 3-5) (30). While the order of reactivity for amino acid adduction is similar for 4HNE and 4ONE, the difference in absolute reactivity (i.e., rate constants) for each amino acid between these two lipid aldehydes is remarkable. The k4ONE/k4HNE ratio for the thiol nucleophiles ranged from110 (GSH) to 153 (NAC). In terms of chemical structure, 4HNE and 4ONE vary only in functional moieties at the C4 position (i.e., hydroxyl versus carbonyl, respectively), amounting to a 2 Da mass difference, and such disparity in rate constants has not been previously reported for structurally analogous R,β-unsaturated aldehydes. A possible explanation for the difference in reactivity between 4HNE and 4ONE stems from the observation that aldehydes are hydrated in an aqueous environment to yield a geminal diol (18).2 Hydration of 4HNE but not 4ONE would yield a molecule inactive as a Michael acceptor due to the absence of an R,β-unsaturated carbonyl (see Figure 1). The reactivity of GSH modification by 4ONE suggests that spontaneous GSH conjugation may be a primary route of 4ONE metabolism/detoxification. In the presence of 5 mM GSH, 4ONE is predicted to have a half-life
