Mass Spectroscopic Characterization of Protein Modification by 4

a Asterisks denote Michael adducts, and pound signs denote dehydrated Michael adducts. The initial stage of interpretation consisted of comparing the ...
0 downloads 0 Views 186KB Size
Download PDF
Chem. Res. Toxicol. 2003, 16, 901-911

901

Mass Spectroscopic Characterization of Protein Modification by 4-Hydroxy-2-(E)-nonenal and 4-Oxo-2-(E)-nonenal Zhongfa Liu,† Paul E. Minkler,‡ and Lawrence M. Sayre*,† Department of Chemistry, Case Western Reserve University, Cleveland, Ohio 44106, and Medical Services and Geriatric Research, Education and Clinical Center, Louis Stokes Veterans Affairs Medical Center, 10701 East Boulevard, Cleveland, Ohio 44106 Received January 6, 2003

The modification of proteins by 4-hydroxy-2-nonenal (HNE) and 4-oxo-2-nonenal (ONE) was investigated using mass spectroscopic approaches. Electrospray ionization MS analysis of HNEand ONE-treated myoglobin and apomyoglobin revealed that the latter more “open” protein structure resulted in more extensive modification. Reductive methylation of Lys residues halved the extent of modification, implicating the importance of adduction of HNE and ONE to both His and Lys residues. HPLC-MS/MS analysis of tryptic and chymotryptic peptides of HNEor ONE-adducted apomyoglobin was aided by the knowledge of structures previously elucidated through model reactions. In the case of HNE, the adducts detected were the HNE-His Michael adduct (on H24, H36, H64, and H113), its dehydrated form (on H36), and the HNE-Lys pyrrole adduct (on K16, K42, K45, K145, and K147). In the case of the more reactive ONE, the adducts detected were the ONE-His Michael adduct (on H24), the ONE-Lys pyrrolinone adduct (on K16 and K145), and the ONE-His-Lys pyrrole cross-link (linking K16 to H24 in the C5 peptide). Although previous analyses of tryptic peptides yielded findings about the nature of His modification, the current chymotryptic peptide analysis produced the first structural characterization of Lys modification on intact proteins by HNE and ONE using mass spectrometry.

Introduction Substantial research has focused on the toxicologic properties of products of membrane lipid oxidation (1), in particular reactive aldehydes such as 4-hydroxy-2nonenal (HNE)1 that are capable of covalent adduction to proteins. HNE has been implicated in ischemiareperfusion damage to the heart (2-5), and is also a major product of the oxidation of the lipid component of the LDL particle (6). HNE modifies the behavior of LDL in recognition and binding to LDL receptors (7), by reaction with amino acid side chains of apolipoprotein B-100, the major protein of LDL, in large part involving neutralizing modification of Lys -amino groups (8). HNE readily forms Michael adducts 1 of Cys and His (6, 9-11), and also Lys (2), although Lys Michael adducts are reversible (11) (Scheme 1). Less frequent Schiff base 3 formation with Lys groups results in some late stage stable adducts such as the 2-pentylpyrrole 4 (12-14), and the bifunctionality of HNE allows it to induce protein cross-links (15-20) such as the 2-hydroxy-2-pentyl-1,2dihydropyrrol-3-one iminium 5 (Scheme 1) (17-20). Recently, 4-oxo-2-nonenal (ONE) has been identified as a product of lipid oxidation (21, 22), and is more reactive than HNE in modifying and cross-linking protein side * To whom correspondence should be addressed: Department of Chemistry, Case Western Reserve University, Cleveland, OH 44106. Phone: (216) 368-3704. Fax: (216) 368-3006. E-mail: [email protected]. † Case Western Reserve University. ‡ Louis Stokes Veterans Affairs Medical Center. 1 Abbreviations: LDL, low-density lipoprotein; HNE, 4-hydroxy-2nonenal; ONE, 4-oxo-2-nonenal.

chains (20, 23, 24). HNE and ONE share at least one advanced lipid peroxidation “end product” of protein modification, the Lys-Lys cross-link 5 (17, 20). Structural identification of adducts of HNE and ONE with proteins allows for generation of antibodies that can selectively recognize specific adducts in oxidized LDL (14, 25-28) and biological tissues (3-5, 14, 27, 29-31). Immunochemical methods may have high sensitivity, but are limited by possible incomplete specificity of the antibodies, and quantitation of adducts can be complicated because antibody affinity can be modulated by the protein context of the epitope. Thus, it is important to have complementary methods for identification and quantitation of protein modifications. Efforts to define the nature of HNE modification by mass spectrometric methods have provided information that has so far been limited to Michael adduction to His residues. For example, modification of particular surface histidines of apo B-100 by HNE in LDL oxidized in vitro was demonstrated using a product ion scan method for analyzing HPLC fractions of the tryptic digest following reductive (borohydride) stabilization of His Michael adducts (34). Also, tandem mass spectral analysis of the unfractionated tryptic digest of apomyoglobin modified with HNE provided evidence for Michael adduction to 10 of the 11 His residues (35). Furthermore, mass spectral analysis of both the hemoglobin R-chain and β-lactoglobulin modified with a large excess of HNE revealed that the intact proteins had incorporated up to four and up to eight molecules, respectively, of Michael-adducted HNE (observed peaks at m/z M + x156) (36). For these

10.1021/tx0300030 CCC: $25.00 © 2003 American Chemical Society Published on Web 06/27/2003

902

Chem. Res. Toxicol., Vol. 16, No. 7, 2003

Liu et al. Scheme 1

two latter proteins, amino acid analysis on the HCl hydrolysate, following reductive stabilization (NaBH4) of the HNE adducts, revealed the loss of the majority of the His residues, but only two to three Lys residues (36). Although reduced HNE-Lys Michael adducts have been quantitated by liquid chromatographic analysis of the HCl hydrolysate of NaBH4-treated, HNE-treated ribonuclease A (RNase) or copper-oxidized LDL (37), mass spectral analysis of HNE-modified proteins has so far failed to reveal details of Lys modification in the absence of reductive stabilization. Prior focus on tryptic peptides may have been adequate for detecting His Michael adducts as the most frequent HNE modification, but limits the ability of detecting Lys (or Arg) adducts such as the HNE-Lys 2-pentylpyrrole adduct (12). The goal of this study was to carry out a mass spectral analysis of both Lys and His modification by HNE and ONE. Only modest concentrations of HNE and ONE were used so as not to saturate protein nucleophilic sites, thus allowing for cross-linking when possible for these bifunctional electrophiles. Adduct identification was aided by chemical model studies that have identified end products of both oxidative and nonoxidative aging of initial adducts. Localization of modifications to specific sites has been addressed by proteolytic studies followed by HPLC-tandem mass spectral analysis, and the relative importance of different modifications has been explored by analysis of modifications of a specific peptide containing both Lys and His.

Experimental Procedures General. Myoglobin from horse heart, chymotrypsin, and sequencing grade modified trypsin (as a kit with buffer) were purchased from Sigma-Aldrich Chemical Co. (St. Louis, MO). PD-10 columns (Sephadex G-25) were purchased from Amersham Bioscience (Piscataway, NJ). Apomyoglobin was prepared from myoglobin by a reported method (35). HNE and ONE were from previous studies (11, 20). All other reagents were of the highest grade commercially available. The water that was used for all studies was purified by a Millipore system. All solutions were prepared immediately prior to being used. Definition of noncovalent interactions in apomyoglobin was obtained using the program O (38) on a Silicon Graphics workstation as well as the program DISTANG within the crystallographic computing package CCP4 (39).

Proteolytic Digestion. Lyophilized protein (100 µg) was dissolved in a solution of 8 M urea and 400 mM NH4HCO3 (25 µL). Then either 40 µL of a solution of 20 µg of modified trypsin dissolved in 200 µL of the supplied trypsin suspension buffer or 4 µL of a solution of chymotrypsin in water (1 µg/µL) was added, and the final volume was adjusted to 100 µL. The mixture was kept at 37 °C for 24 h, and the digestion was quenched by the addition of 1 µL of 90.0% aqueous formic acid. The solution was stored at -20 °C for LC-MS analysis. ESI-MS. ESI-MS studies on intact proteins were performed using a Branford electrospray source interfaced with a Thermo Finnigan (San Jose, CA) Duo mass spectrometer. Mass spectra were deconvoluted into the mass domain using the Biomass Deconvolution algorithm of the Qual Browser in Xcalibur, the Finnegan software for instrument control and data analysis. A solution of the desalted protein in 0.6% acetic acid was delivered to the electrospray unit at a rate of 5 µL/min using a syringe pump. A countercurrent of warm nitrogen sheath gas (80 bar) facilitated desolvation. The heated capillary temperature was 220 °C, the electrospray voltage 5.2 keV, and the capillary voltage set to -4 V. HPLC-ESI-MS/MS. Reverse phase HPLC of tryptic or chymotryptic digests was performed with an HP1100 system equipped with a 2 mm × 250 mm Vydac Low TFA C18 column with a pore size of 5 µm. Eluent A was 950 mL of H2O, 50 mL of MeCN, and 0.2 mL of TFA, and eluent B was 50 mL of H2O, 950 mL of MeCN, and 0.2 mL of TFA. The flow rate was 0.3 mL/min, and the gradient program was from 100% A to 100% B over the course of 60 min (100% B for 10 min and 100% A for 10 min). The injection volume was 20 µL. The LC elution was monitored at 214 nm, and the UV-vis spectrum of each component was obtained from 200 to 600 nm. ESI mass spectrometry was performed using a Finnigan LCQDeca instrument in the positive mode using nitrogen as the sheath and auxiliary gas. The heated capillary temperature was 250 °C, the electrospray voltage 5.2 keV, and the capillary voltage set to -4 V. Three scan events were used: (i) m/z 300-2000 full scan MS, (ii) data-dependent zoom scan on the most intense ion from the MS full spectrum, and (iii) data-dependent full scan MS/MS on the most intense ion from the MS full spectrum. The MS/MS collision energy was set to 35 V. This data-dependent multistage MS fragmentation was used to sequence the peptides and modified peptides. All data were processed with a Qual browser in Xcalibur. Reductive Methylation of Apomyoglobin. A solution of apomyoglobin (1 mg/mL, 3 mL) in 50 mM sodium phosphate buffer (pH 7.0) was treated with a freshly prepared solution of NaCNBH3 and formaldehyde (both 0.1 M, 0.34 mL) at room temperature for 24 h. Then the solution was dialyzed against

Protein Modification by HNE and ONE 50 mM sodium phosphate buffer (pH 7.0) three times in 8 h intervals. The bicinchoninic acid (BCA) assay using BSA as the standard indicated a methylated apomyoglobin concentration of 0.76 mg/mL. The trinitrobenzenesulfonate (TNBS) assay indicated that the yield of methylation was more than 90%. Assuming all 19 Lys side chains and the amino terminus were dimethylated, the expected molecular weight should increase from 16 951.5 to 17 511.5. The deconvoluted ESI mass spectrum of the methylated apomyoglobin showed a broad peak from m/z 17 500 to 17 650, with the highest point near 17 550. The lack of signal below m/z 17 500 is consistent with the high level of methylation indicated by the TNBS assay, whereas the highermolecular weight species probably arise from the expected low level of methylation of nucleophilic side chains other than Lys (40). Incubation of either ONE or HNE with Apomyoglobin, Myoglobin, and Methylated Apomyoglobin. A solution of either apomyoglobin or myoglobin (1 mg/mL, 2 mL) or methylated apomyoglobin (0.76 mg/mL, 2 mL) in 0.1 M sodium phosphate buffer (pH 7.4) was incubated with a solution of either HNE or ONE in acetonitrile (100 mM, 40 µL) or with acetonitrile (40 µL) alone. After 1 and 24 h (only for myoglobin), the incubation was quenched with 20 µL of 90.0% aqueous formic acid. The incubation mixture was immediately applied to a PD10 column to remove unbound ONE or HNE and buffer salts. The desalted protein fraction was eluted in 4 mL of 0.6% aqueous acetic acid. Aliquots of the solution were lyophilized for proteolytic digestion, and the remainder was stored at -20 °C for ESI-MS analysis to determine the stoichiometry of HNEor ONE-adducted protein. Incubation of Apomyoglobin with either HNE or ONE Followed by Quenching with Sodium Borohydride. A solution of apomyoglobin (1 mg/mL, 2 mL) in 0.1 M sodium phosphate buffer (pH 7.4) was incubated with either a solution of either ONE or HNE in acetonitrile (100 mM, 40 µL) or with acetonitrile (40 µL) alone. After 1 h, the incubation was quenched with NaBH4 (1.0 mg) for 4 h. The incubation mixture was immediately applied to a PD-10 column to remove unbound ONE and buffer salts. The desalted protein fraction was eluted in 4 mL of 0.6% acetic acid. Aliquots of the solution were lyophilized for proteolytic digestion.

Results and Discussion Preliminary Studies. On the basis of our frequent use of RNase to analyze protein modification by HNE and ONE (20, 24), we first attempted to use this protein for a comprehensive characterization of HNE or ONE modification. However, using both matrix-assisted laserinduced desorption ionization (MALDI) and electrospray ionization (ESI) mass spectroscopy to analyze RNase exposed to HNE or ONE for 24 h, a very low level of modification (predominantly 1 mol of aldehyde/mol of protein) was indicated. Since only limited information would thus be provided, we decided to use apomyoglobin, which was reported to be extensively modified by HNE (34). This protein lacks Cys residues, and should thus facilitate a focus on Lys and His modification. Stoichiometric Studies on the Modification of Apomyoglobin and Myglobin by HNE. To check the stoichiometry of modification of apomyoglobin by HNE over a short period of time, apomyoglobin (1 mg/mL) was incubated with HNE (2 mM) in 50 mM sodium phosphate buffer (pH 7.4) at room temperature for 1 h. The ESI mass spectrum shown in Figure 1A was obtained following separation of apomyoglobin from unbound HNE and desalting via gel filtration chromatography. The spectrum was recorded by signal averaging for 2 min, corresponding to the consumption of 100 pmol of HNE-adducted apomyoglobin. Multiple peaks at each charge state in the

Chem. Res. Toxicol., Vol. 16, No. 7, 2003 903

Figure 1. Raw electrospray mass spectrum (A) and deconvoluted electrospray mass spectrum (B) of apomyoglobin modified with HNE.

mass spectrum attest to the presence of a mixture resulting from the covalent attachment of several HNE molecules to the protein. The deconvoluted real mass spectrum (Figure 1B) showed that apomyoglobin was transformed into several HNE-modified species with two to six molecules of HNE adducted to each apomyoglobin via simple addition, with additional minor peaks appearing 18 ( 2 Da lower in mass in the cases of three and four molecules of HNE bound to apomyoglobin. When the 1 h incubation was repeated with 0.5 mM HNE, the deconvoluted mass spectrum exhibited two predominant peaks corresponding to only one to two molecules of HNE adducted to each apomyoglobin. In the same manner, myoglobin (1 mg/mL) was incubated with HNE (2 mM) for 1 h. In this case, the deconvoluted real mass spectrum showed that the predominant mass peak corresponded to unmodified myoglobin, with minor peaks being seen for myoglobin containing one molecule of HNE with and without dehydration. The deconvoluted mass spectrum of myoglobin incubated with 2 mM HNE for 24 h showed that several HNE-adducted species were formed with one to two molecules of HNE adducted per myoglobin, with losses of up to three waters. Additional peaks with three or more HNE molecules attached to each myoglobin were also apparent, but the resolution was insufficient for deciphering the possible linkage. Stoichiometric Studies on the Modification of Myoglobin and Apomyoglobin by ONE. Using the same reaction condition as described above for HNE, the deconvoluted mass spectrum of myoglobin incubated with

904

Chem. Res. Toxicol., Vol. 16, No. 7, 2003

Liu et al.

Figure 3. HPLC chromatogram (UV at 214 nm) of trypsinized apomyoglobin. HPLC was performed using a Vydac Low TFA C18 column with elution by the binary system described in Experimental Procedures.

Figure 2. Deconvoluted electrospray mass spectrum of myoglobin modified with ONE for 24 h.

2 mM ONE for 24 h (Figure 2) exhibited peaks corresponding to one to two molecules of ONE adducted, with loss of zero, one, or two waters (the predominant peaks reflected the loss of one water per adducted ONE) (Figure 2). In contrast, the deconvoluted mass spectrum of apomyoglobin incubated with ONE even for 1 h exhibited a much more complex pattern of substantial modification. In the low mass range of the spectrum, the predominant peaks corresponded to two molecules of ONE adducted per apomyoglobin with loss of four waters and three molecules of ONE adducted with loss of six waters. The high mass range of the spectrum exhibited a bell shape, which was too complicated to permit analysis of the stoichiometry of adduct formation, though up to eight molecules of ONE appear to be added. Even when the ONE concentration was reduced to 0.5 mM, the deconvoluted mass spectrum for a 1 h incubation with apomyoglobin exhibited a pattern that was too complex to analyze. The greater loss of water apparent for the ONE adducts relative to HNE adducts probably reflects in part the more dehydrative nature of ONE adduction chemistry, and in part the greater tendency for the ONE adducts to undergo in-source collision-induced loss of water (35). It is clear that apomyoglobin was modified by HNE and ONE much more extensively than myoglobin, probably reflecting a less stable (more conformationally mobile) structure (41) and greater solvation of the apoprotein following removal of the heme, as supported by the observed differences in the binding affinity of halothane (42). Stoichiometric Studies on the Modification of Methylated Apomyglobin by HNE and ONE. In the observed adduction of HNE and ONE to apomyoglobin, it was not possible to ascertain to what extent modification involved His as opposed to Lys residues. To obtain information bearing on this point, the available lysines in apomyoglobin were reductively methylated using formaldehyde and cyanoborohydride. The “methylated apomyoglobin” was incubated with 2 mM HNE or ONE for 1 h under the same condition described above. The deconvoluted mass spectra of these samples indicated adduction of one to two molecules of HNE and two to four molecules of ONE per molecule of methylated apomyoglobin. Assuming this represents reaction with histidines, then the difference (up to four molecules in each case)

Figure 4. HPLC chromatogram (UV at 214 nm) of trypsinized HNE-adducted apomyoglobin. HPLC was performed using a Vydac Low TFA C18 column with elution by the binary system described in Experimental Procedures.

between these values and the stoichiometries seen for the nonmethylated apoprotein represents modification of lysines. This conclusion assumes that Lys methylation does not alter the average reactivity of the protein histidines [there is little direct Lys-His interaction in native myoglobin (43)]. In summary, independent of the amino acid that is modified, the observed mass differences suggest that the predominant modifications involve simple HNE or ONE Michael adducts on protein nucleophiles, though occasional adducts involving dehydration undoubtedly explain some of the minor -n18 Da peaks. Characterization of Tryptic Peptides from HNETreated Apomyoglobin. Figure 3 shows the reverse phase HPLC separation of trypsinized unmodified apomyoglobin, with identification of eight known peptides. Table 1 provides a listing of the observed peptides with molecular masses of more than 500 Da (excluding the smaller tryptic peptides), representing 65% coverage of the sequence of the protein. Incubation of apomyoglobin with 2 mM HNE for 1 h followed by tryptic digestion and HPLC-MS analysis revealed four new peaks (UV detection) designated N0-N3 (Figure 4) that were not seen in the tryptic map of the parent peptide. No electrospray ion current was seen in the case of N0, and thus, identification of this peak was not pursued. The initial stage of interpretation consisted of comparing the m/z values of the ions in the precursor spectrum with those values calculated from the predicted tryptic peptides of apomyoglobin where every His residue was replaced with an HNE-His Michael adduct (M + 156). For tryptic peptides that contained more than one His residue, values were calculated with all possible combinations of native and adducted residues. If m/z values are more than 2000, their doubly charged ions were checked. According to the specified possibilities, Table 2 provides a listing of the singly charged peptide ions observed that were consistent with HNE modification as calculated. Under the conditions that were used, corre-

Protein Modification by HNE and ONE

Chem. Res. Toxicol., Vol. 16, No. 7, 2003 905

Table 1. Observed Peptides in the Tryptic Digestion of Apomyoglobin positions

T

peptide sequence

retention time (min)

calcd (M + H)+

obsd (M + H)+

1-16 17-31 32-42 64-77 103-118 119-133 134-139 148-153

T1 T2 T3 T10 T16 T17 T18 T21

GLSDGEWQQVLNVWGK VEADIAGHGQEVLIR LFTGHPETLEK HGTVVLTALGGILK YLEFISDAIIHVLHSK HPGDFGADAQGAMTK ALELFR ELGFQG (C-terminus)

20.81 12.63 10.68 19.11 23.39 9.57 15.03 11.27

1815.9 1606.9 1271.6 1378.8 1885.0 1502.7 748.4 650.3

1815.9 1606.8 1271.6 1378.8 1884.8 1502.6 748.6 650.3

Table 2. Observed Modified Peptides in the Tryptic Digest of HNE-Treated Apomyoglobin modified peptide (peak)

peptide sequencea

retention time (min)

obsd (M + H)+

modified residue

T2HNE (N2) T3HNE (N1) T3HNE - 18 (N3) T10HNE (N3)

VEADIAGH*GQEVLIR LFTGH*PETLEK LFTGH#PETLEK H*GTVVLTALGGILK

16.97 15.53 22.92 22.72

1762.7 1427.7 1409.6 1534.8

H24 H36 H36 H64

a

Asterisks denote Michael adducts, and pound signs denote dehydrated Michael adducts.

Figure 5. ESI-tandem mass spectrum of the (M + H)+ ion of the T10HNE histidine Michael adduct at m/z 1534.7 and the amino acid sequence and ions in the tandem mass spectrum.

sponding doubly charged ions were weaker or not observed. Three mono-HNE adducts were seen, displaying molecular masses 156 Da higher than those of the native sequences of T2, T3, and T10, and corresponded to the newly formed peaks N1, N2, and one of the two components of N3. A fourth mono-HNE adduct, corresponding to the second N3 component, displayed a molecular mass 138 Da (+ HNE - 18) higher than that of the native sequence of T3. No mass signals corresponding to a tryptic peptide + 130 (+ HNE - 2 × 18) were observed. Figure 5 shows the tandem mass spectrum of the singly charged ion at m/z 1534.8, corresponding to the predicted value of (M + H)+ for an HNE Michael adduct of the tryptic fragment T10. The base peak at m/z 1378.5, corresponding to the (M + H)+ ion of the parent peptide, was generated from a neutral loss of the HNE molecule from the HNE-adducted peptide. Two sets of b ions (44), b9-b14 and b5*-b14*, were observed, and the m/z values of the b5*-b14* ions were 156 Da higher than those of the corresponding b ions. The observed b5*-b14* ions were readily interpreted in terms of the amino acid sequence of the first 10 C-terminal residues of the T10HNE adduct with one molecule of HNE adducted to one

of the first five N-terminal amino acids. The five Nterminal amino acids are HGTVV, so the HNE adduct was presumed to represent Michael addition to the imidazole group of H64 in T10. The observation of the normal b series ions reflects neutral loss of HNE from the b* ions or fragmentation of the parent peptide base peak in the mass spectrometer. The tandem mass spectra of the other two mono-HNE adducts (T2HNE and T3HNE, Table 2) also gave their parent peptides as the base peak and two sets of b ions. The b8*-b15* ions in the tandem mass spectrum of the singly charged ion at 1762.7 Da, corresponding to the predicted value for (M + H)+ for an HNE adduct of tryptic fragment T2, are 156 Da higher than the corresponding b ions in the tandem mass spectrum of the parent peptide T2. The observed b8*-b15* ions were readily interpreted in terms of the amino acid sequence of the first eight C-terminal residues of the T2HNE adduct with HNE adducted to the first eight N-terminal amino acids. The eight N-terminal amino acids are VEADIAGH, and no b* ions shorter than b8* were observed. The HNE was thus presumed to be Michael adducted to the imidazole group of H24 in T2. In the same manner, the b6*-b11*

906

Chem. Res. Toxicol., Vol. 16, No. 7, 2003

ions in the tandem mass spectrum of the singly charged ion at 1427.7 Da, corresponding to the predicted value for (M + H)+ for an HNE adduct of tryptic fragment T3, are 156 Da higher than the corresponding b ions in the tandem mass spectrum of the parent peptide T3. The observed b6*-b11* ions were readily interpreted in terms of the amino acid sequence of the first six C-terminal residues of the T3HNE adduct with one HNE adducted to one of the first six N-terminal amino acids. The six N-terminal amino acids are LFTGHP, so the adduct was presumed to form through Michael addition of HNE to the imidazole group H36 in T3. The singly charged ion at m/z 1409.7 was found to correspond to the predicted (M + H)+ ion for a dehydrated mono-HNE adduct on the T3 peptide (T3HNE - 18). The tandem mass spectrum of this ion exhibited the parent unmodified peptide as the base peak, a series of normal b ions (b7-b11), and a series of b* ions (b7*-b11*) with masses 138 Da higher than those of the corresponding b ions. The observed b7*-b11* ions were readily interpreted in terms of the amino acid sequence of the first five C-terminal residues of the T3HNE - 18 adduct with one HNE adducted to one of the first seven N-terminal amino acids. Since the seven N-terminal amino acids are LFTGHPE, the adduct was assigned as a dehydrated form of the HNE Michael adduct on the His of T3 (vide infra). The observation of two sets of b ions indicates that the dehydrated Michael adduct is also subject to a facile neutral loss (of dehydrated HNE) in the mass spectrometer. Overall, although the HPLC-MS analysis chosen here excluded the smaller tryptic fragments, some of which contain histidine, the finding of HNE Michael adducts on H24, H36, and H64 confirms the analysis by Bolgar and Gaskell on the unfractionated tryptic digest of HNEtreated apomyoglobin (35), although we did not find the HNE adduct on H113 or H116 in T16, or on H119 in T17. Characterization of Tryptic Peptides from HNETreated Apomyoglobin following Reductive Stabilization. It was argued above that the normal b series ions observed in modified tryptic fragments arose from retro Michael addition of HNE in the mass spectrometer. Retro Michael addition can also occur under acidic conditions, though it is known that HNE Michael adducts can be stabilized against elimination by hydride reduction (11, 20) (an example is shown in eq 1). Thus, following reductive stabilization, there should be no normal b series ions observed. After being treated with HNE and then NaBH4, apomyoglobin was subjected to tryptic digestion. Three adduct ions at m/z 1536.8, 1429.7, and 1764.7 were observed, corresponding to values 2 Da higher than those of T10HNE, T3HNE, and T2HNE, respectively.

The tandem mass spectrum of the singly charged ion at 1536.8 Da exhibited one set of b* ions (b4*-b14*) with masses 2 Da higher than those of the corresponding b*

Liu et al.

ions in the tandem mass spectrum of the T10HNE adduct, and no normal b series ions were observed. This localized the modification to the first four N-terminal amino acids (HGTV), and it is thus presumed that the modification represents the reduced HNE Michael adduct (1,4nonanediol) attached through C3 to the imidazole of H64. In the same manner, the reduced HNE Michael adducts were localized to H36 in T3 and H24 in T2. These results support the conclusion that the normal b series ions in the tandem mass spectra of the three HNE-adducted tryptic peptides represent neutral loss of HNE (retro Michael addition) in the mass spectrometer. The dehydrated Michael adduct T3HNE - 18 ion at m/z 1409.7 discussed above was still observed following treatment with NaBH4, indicating that this adduct is inert to reduction. On this basis, its structure was tentatively assigned as the dihydrofuran resulting from dehydration of the cyclic hemiacetal form of the HNE Michael adduct (eq 2). The dihydrofuran is inert to NaBH4 reduction and should easily fragment in the mass spectrometer to the parent peptide by loss of 2-pentylfuran, thereby explaining the facile loss of 138 Da seen in the tandem mass spectrum.

Characterization of Chymotryptic Peptides of HNE-Treated Apomyoglobin. On the basis of the ability of HPLC-MS analysis of tryptic digests to localize His Michael adducts, we next switched to chymotryptic digestion to permit analysis of HNE modification of lysines residing within chymotryptic cleavage sites (at Trp, Leu, Phe, Tyr, and Met residues). Since cleavage ordinarily does not occur at every possible site, many observed chymotryptic peptides span two or more theoretical fragments and are designated as Cx-y. A complex chromatogram was observed for the digest of the native protein, and Table 3 provides a listing of the various peptides with molecular masses of more than 400 Da, containing up to two missed cleavages, which were identified by reverse phase HPLC-MS. The chymotryptic map contains two gaps and accounts for 76% coverage of the amino acid sequence. The UV chromatogram of the chymotryptic digest of apomyoglobin incubated with HNE for 1 h was even more complex, though three new peaks at retention times of 23.2-23.8 (P1), 26.20 (P2), and 27.40 min (P3) could be clearly discerned (data not shown). The initial stage of interpretation consisted of comparing the m/z values of the ions in the precursor spectrum with those values calculated from the predicted chymotryptic peptides of apomyoglobin where every His residue was replaced with an HNE-His Michael adduct (M + 156) and every Lys residue was replaced with either a Lys-HNE Michael adduct (M + 156), a Lys-HNE Schiff base (M + 156 18), or a Lys-HNE derived pyrrole (M + 156 - 36). Table 4 provides a listing of the HNE-modified peptides found by selected ion extraction. Two of the modified peptides accounted for the new UV peaks P2 and P3,

Protein Modification by HNE and ONE

Chem. Res. Toxicol., Vol. 16, No. 7, 2003 907

Table 3. Observed Peptides in the Chymotryptic Digest of Apomyoglobin positions

C

peptide sequence

retention time (min)

calcd (M + H)+

obsd (M + H)+

1-11 12-14 15-29 12-29 30-33 34-43 34-46 56-69 70-76 104-115 105-115 116-131 124-131 132-138 139-146 147-153

C1-3 C4 C5 C4-5 C6-7 C8-9 C8-10 C13-14 C15-16 C20-22 C21-22 C23-24 C24 C25-27 C28 C29-31

GLSDGEWQQVL NVW GKVEADIAGHGQEVL NVWGKVEADIAGHGQEVL IRLF TGHPETLEKF TGHPETLEKFDKF KASEDLKKHGTVVL TALGGIL LEFISDAIIHVL EFISDAIIHVL HSKHPGDFGADAQGAM GADAQGAM TKALEFL RNDIAAKY KELGFQH (C-terminal)

17.92 10.36 12.51 18.85 15.18 10.57 13.2 9.66 16.8 23.73 21.51 8.44 16.14 16.55 6.28 9.21

1231.6 418.2 1522.8 1922.0 548.4 1158.6 1548.8 1524.9 644.4 1369.8 1256.7 1625.7 720.3 821.5 950.5 778.4

1231.4 418.2 1522.7 1921.7 548.3 1158.5 1548.7 1524.9 644.4 1369.7 1256.5 1625.7 720.3 821.4 950.5 778.4

Table 4. Observed Modified Peptides in the Chymotryptic Digest of HNE-Treated Apomyoglobin

a

modified peptide

peptide sequencea

retention time (min)

obsd (M + H)+

modified residue

C5HNE - 36 C5HNE C8-10HNE C8-10HNE - 18 C8-10HNE - 36 C13-14HNE C20-22HNE C28HNE - 36 C29-31HNE - 36

GK*VEADIAGHGQEVL GKVEADIAGH#GQEVL TGH#PETLEKFDKF TGH$PETLEKFDKF TGHPETLEK*FDK*F KASEDLKKH#GTVVL LEFISDAIIH#VL RNDIAAK*Y K*ELGFQG

23.46 17.21 16.90 17.75 23.80 17.19 27.38 23.57 26.20

1642.8 1678.7 1704.6 1086.8 1668.6 1680.8 1525.6 1070.7 898.2

K16 H24 H36 H36 K42, K45 H64 H113 K145 K147

Pound signs denote Michael adducts. The dollar sign denotes a dehydrated Michael adduct. Asterisks denote pyrrole adducts.

whereas the broad peak P1 contained three components. The retention times of the remaining peptides found by selected ion extraction are listed, but no clear peaks in these cases were seen in the UV chromatogram. Four cases in Table 4 involved Michael adduct modification of His residues already identified by tryptic digestion: H24 in the C5 fragment (and T2 fragment), H36 in the C8-10 fragment (and T3 fragment), the dehydrated H36 adduct, and H64 in the C13-14 fragment (and T10 fragment). A new HNE His Michael adduct was now found at H113 in the C20-22 fragment. Importantly, the use of chymotrypsin instead of trypsin now permitted identification of the doubly dehydrated HNE-Lys adduct corresponding to the previously described 2-pentylpyrrole (12, 13). No simple HNE-Lys Michael adducts were seen, presumably because these latter adducts are formed reversibly and are unstable (11). No evidence was found for simple HNE-Lys Schiff base adducts, though any initially formed HNE-Lys Schiff bases would probably evolve to the pyrrole adduct during the 24 h proteolytic incubation. Five pyrrole adducts were observed: K16 in C5, K42 and K45 in C8-10, K145 in C28, and K147 in C29-31, as described in further detail below. Consistent with the apolar nature of the pyrrole group, the Lys-modified peptides uniformly appeared at much longer retention times in the reverse phase HPLC chromatogram than the unmodified peptides. The C28 peptide gives rise to a singly charged ion at m/z 950.5, and the observed b series ions in the tandem mass spectrum (b3-b8) are easily interpreted, confirming the first six C-terminal amino acid residues. The mass spectrum of C28HNE - 36 gives a peak at m/z 1070.5, corresponding to the predicted value of (M + H)+ for a doubly dehydrated HNE adduct of the chymotryptic fragment C28. Although the tandem mass spectrum of the singly charged ion at 1070.5 Da exhibited normal b3-b6

ions, the b7 and b8 ions were replaced by b7* and b8* ions, respectively, appearing at m/z values that were 120 Da higher. The observed b* ions indicate a doubly dehydrated HNE adduct on one of the first two C-terminal amino acids. The two C-terminal amino acids are KY, so the HNE adduct was presumed to represent pyrrolation of the -amino group of K145 in C28. The tandem mass spectrum of the singly charged molecular ion at 1548.6 Da for the parent C8-10 peptide showed a series of b ions (b7-b13), confirming the amino acid sequence of the first seven C-terminal residues, and a series of y ions (y6-y11), confirming the amino acid sequence of the six residues starting at the third from the N-terminus. Figure 6 shows the tandem mass spectrum of the singly charged ion at m/z 1668.6 arising from chymotryptic cleavage of the HNE-modified protein, corresponding to the predicted value for (M + H)+ for the peptide C8-10HNE - 36. The spectrum shows both normal b ions (b7-b11) and modified b ions (b9*-b11*) at 120 Da higher masses. No normal b12 or b13 ions were observed, but instead, b12* and b13* ions were seen at 120 Da higher masses. These results indicate that HNE was adducted to the first five C-terminal amino acid residues, where there are two lysines, K42 and K45, either of which can be converted to pyrroles to give the HNE - 36 modification. The finding of both normal and modified b9-b11 ions suggests there are two isomeric peptides bearing the pyrrole at either K42 or K45 that coeluted and were trapped together in the collision cell. When the intensities of the b10*/b11* ions and the b10/b11 ions in the tandem mass spectrum are compared, the two lysines appear to have equal tendencies to be modified by HNE. In the same manner, the other two pyrrolated peptides were confirmed by their tandem mass spectra, localizing HNE pyrrole modifications also at K16 and K147.

908

Chem. Res. Toxicol., Vol. 16, No. 7, 2003

Liu et al.

Figure 6. Tandem mass spectrum of the HNE-modified chymotrytic peptide C8-10 (C8-10HNE - 36) at m/z 1668.6.

The finding of three different modifications in the C8-10 peptide provided an opportunity to estimate their relative formation tendencies. The ion current integrations relative to the unmodified peptide were 0.64 for the His Michael adduct, 0.26 for the dehydrated His Michael adduct, and 0.34 for the Lys pyrrole adduct. Assuming that the relative ion current absorptions for the three modified C8-10 peptides do not vary greatly, one can roughly estimate that the His Michael adduct is ∼2 times as prominent as either the dehydrated His Michael adduct or the Lys pyrrole adduct. This is consistent with the data provided from mass spectral analysis of intact HNE-modified apomyoglobin, where the major peaks represented addition of multiple Michael-adducted HNE molecules, with -18 and -36 peaks being minor. Further confirmation of the HNE-Lys pyrrole adduct is provided by LC-MS amino acid analysis following complete proteolysis of apomyoglobin incubated with 2 mM HNE for 24 h (2 ( 1% of Lys; G. Zhou, Z. Liu, and L. M. Sayre, unpublished studies). Of the 19 Lys residues in apomyoglobin, the 76% chymotryptic map coverage we observed missed two Lys residues in the peptide fragment of residues 47-55 (C11-12) and seven Lys residues in the peptide fragment of residues 77-103 (C17-19). Thus, of the 10 Lys residues that were observed, only half of these (K16, K42, K45, K145, and K147) appeared to become HNE-modified. To explore a possible explanation for this observation, first the solvent accessibility of all Lys residues in myoglobin was calculated on the basis of the atomic coordinates of myoglobin from horse heart (PDB entry 1AZI) (45). However, no apparent correlation was found for modified versus unmodified residues. Then all possible hydrogen bonding and ionic interactions of each atom in the protein with other atoms were calculated. It was found that Lys residues present in tight salt bridge interactions (