Analytical and Bioanalytical Chemistry

, Volume 389, Issue 5, pp 1421–1428

Fourier transform ion cyclotron resonance mass spectrometry of covalent adducts of proteins and 4-hydroxy-2-nonenal, a reactive end-product of lipid peroxidation

Authors

  • Navin Rauniyar
    • Department of Molecular Biology & ImmunologyUniversity of North Texas Health, Science Center
  • Stanley M. StevensJr
    • Department of Molecular Biology & ImmunologyUniversity of North Texas Health, Science Center
    • Department of Molecular Biology & ImmunologyUniversity of North Texas Health, Science Center
Original Paper

DOI: 10.1007/s00216-007-1534-2

Cite this article as:
Rauniyar, N., Stevens, S.M. & Prokai, L. Anal Bioanal Chem (2007) 389: 1421. doi:10.1007/s00216-007-1534-2

Abstract

Covalent adduction of the model protein apomyoglobin by 4-hydroxy-2-nonenal, a reactive end-product of lipid peroxidation, was characterized by nanoelectrospray ionization Fourier transform ion cyclotron resonance mass spectrometry (FTICR). The high mass resolving power and mass measurement accuracy of the instrument facilitated a detailed compositional analysis of the complex reaction product without the need for deconvolution and transformation to clearly show the pattern of adduction and component molecular weights. Our study has also demonstrated the value of electron capture dissociation over collision-induced dissociation for the tandem mass spectrometric determination of site modification for the 4-hydroxy-2-nonenal adduct of oxidized insulin B chain as an example.

https://static-content.springer.com/image/art%3A10.1007%2Fs00216-007-1534-2/MediaObjects/216_2007_1534_Figa_HTML.gif
Figure

FTICR allowed characterization of 4-hydroxy-2-nonenal (HNE)-modified apomyoglobin (an expanded spectrum of the +15 charge state is shown)

Keywords

Fourier transform ion cyclotron resonance mass spectrometryElectrospray ionizationProtein carbonylation4-Hydroxy-2-nonenalElectron capture dissociation

Introduction

Reactive oxygen species (ROS) such as superoxide anion \(\left( {{\text{O}}_2^{ \cdot - } } \right)\), hydrogen peroxide (H2O2), and the hydroxyl radical (HO·) that are produced at higher levels during oxidative stress can induce numerous proteomic changes in vivo [1]. Oxidative damage of proteins can result in several key events that alter cellular activity including changes in protein activity, proteasomal quality control, cellular redox-balance, and interference with the cell cycle [2]. Protein carbonylation in particular is an event caused by direct attack of ROS, metal-catalyzed oxidation, reaction with reducing sugars, and conjugation with highly reactive carbonyl compounds produced as end-products of lipid peroxidation [1]. The last of these processes, an event in which oxidation of polyunsaturated lipids present in biological membranes occurs, forms carbonyl compounds such as the α, β-unsaturated aldehyde 4-hydroxy-2-nonenal (HNE), acrolein, and others.

Aldehydes formed by lipid peroxidation react with nucleophilic groups of proteins and are found to be selective for certain amino acid residues including cysteine (Cys), lysine (Lys), and histidine (His) [3]. As shown in Fig. 1, reaction between these aldehydes and amino acids occurs via Michael-type addition or Schiff base (imine) formation [4, 5]. In HNE-based Michael-type addition, modification involves addition of the imidazole group of His (H), ɛ-amino group of Lys (L), or the sulfhydryl (–SH) group of Cys (C) to the α,β-unsaturated bond of HNE [6]. Several reported studies have demonstrated modifications of proteins at various amino acid residues by HNE leading to alteration in their structure and biological activity. For example, cytochrome c oxidase and apolipoprotein B-100 undergo modification exclusively at His by HNE [7, 8], whereas epithelial fatty acid-binding protein was modified at Cys [9]. Cathepsin B was reported to be modified both at Cys and His [10], and glyceraldehyde-3-phosphate dehydrogenase at Cys, His, and Lys residues [11]. Glucose-6-phosphate dehydrogenase undergoes selective modification at Lys by HNE [12]. Chemical modification of proteins by HNE can induce significant changes in their activity and thus knowledge of the exact modification sites provides detailed insight into oxidative stress-mediated cellular dysfunction [13].
https://static-content.springer.com/image/art%3A10.1007%2Fs00216-007-1534-2/MediaObjects/216_2007_1534_Fig1_HTML.gif
Fig. 1

Primary mechanisms observed for the reaction of amine groups of proteins and peptides with 4-hydroxy-2-nonenal (HNE) are Michael-type addition (a) and formation of Schiff base adducts (b). HNE can also form Michael-type adducts (+156) with the imidazole =NH group of histidine (His) and sulfhydryl (–SH) groups of Cys side-chains (products not shown)

For the identification of post-translational modifications such as carbonylation, Fourier transform ion cyclotron resonance (FTICR) mass spectrometry [14] has unprecedented advantages over other instrumentation including ultra-high resolution, exceptionally high mass accuracy, and unique ion manipulation capabilities such as electron capture dissociation (ECD) [15]. Characterization of protein carbonylation using FTICR/MS can be particularly effective, since the increase in molecular weight of intact proteins due to addition of aldehyde moieties such as HNE to Lys, His, or Cys residues is measured with ultra-high accuracy. Tandem mass spectrometry (MS/MS) can also be employed using both top-down (MS/MS of intact proteins typically by ECD) and bottom-up approaches (MS/MS of peptides derived from protease digestion) to provide further information on the exact site of protein modification. Identification of specific targets of covalent adduction of reactive aldehydes such as HNE is important to understand the chemistry behind the modification and subsequent variation in protein activity mediated by the modification.

In CID and IRMPD methods of dissociation, the amide bond of the peptide backbone is cleaved forming primarily b- or y-type fragment ions [16]. During CID-based fragmentation, however, there is a probability of neutral loss of the modifying group, which is dependent on the chemistry of the modification and its residue interaction, rendering it frequently difficult to correctly identify the peptide modification site. In comparison to CID where the elimination of a labile covalent modification is sometimes favored over peptide backbone fragmentation, ECD provides complementary (N–Cα amine-bond cleavage which results in the formation of c- and z-type fragment ions) and perhaps more valuable sequence information, since unstable modifications are typically retained on the amino acid residue during the fragmentation process [17].

The purpose of this study was to detect the extent of HNE modification of apomyoglobin by FTICR electrospray ionization mass spectrometry (ESI-MS). Bolgar and Gaskell have reported the formation of three to ten HNE adducts per protein molecule in apomyoglobin with modification occurring only at His residues [17]. Fenaille et al. have also reported nine His residues in apomyoglobin modified by HNE, after tryptic digestion and immunoaffinity purification of HNE-labeled peptides [18]. In this study we provide evidence of Schiff base addition of HNE to apomyoglobin in addition to Michael addition using FTICR/MS. We also show the advantages of electron capture dissociation (ECD) for gas-phase peptide fragmentation compared to conventional collision-induced dissociation (CID) using HNE-modified oxidized insulin B chain as a model peptide.

Experimental

Chemicals

Apomyoglobin from horse skeletal muscle and oxidized insulin B chain were purchased from Sigma–Aldrich (St. Louis, MO). 4-Hydroxy-2-nonenal was obtained from Alpha Diagnostic (San Antonio, TX). Water and acetonitrile were of high-performance liquid chromatography grade and purchased from Honeywell Burdick and Jackson (Morristown, NJ). All other chemicals were obtained from Sigma–Aldrich.

Preparation of apomyoglobin–HNE adduct

HNE adducts of apomyoglobin were generated via the method described by Bolgar and Gaskell [17] with slight modification. Apomyoglobin (137 μM) was incubated in an aqueous solution (2 mM) of HNE, buffered with 25 mM ammonium bicarbonate (final volume of 0.5 mL). The temperature was maintained at 37 °C for 2 h. The reaction was then terminated by the addition of 0.25 mL of 1% formic acid. The volume of the reaction mixture was adjusted to 2.5 mL with 25 mM ammonium bicarbonate and subjected to gel filtration by G-25 column (PD-10, GE Healthcare, Piscataway, NJ) in order to remove the unreacted substances. The desalted protein fraction was eluted in 3.5 mL of 0.1% formic acid and dried in a vacufuge concentrator (Eppendorf Scientific, Westbury, NY). The sample was resuspended in 49.5% acetonitrile/49.5% H2O/1% acetic acid and used directly for nanoelectrospray ionization mass spectrometric analysis.

Preparation of insulin–HNE adduct

Oxidized insulin B chain was incubated in an aqueous solution of HNE (2 mM), buffered with 50 mM K2HPO4 (pH 7.4). The temperature was maintained at 37 °C for 2 h (final volume 0.5 mL). The resulting modified peptides were desalted by octadecylsilica (C18) ZipTip (Millipore, Billerica, MA) microcolumns. The bound peptides were washed with 0.1% acetic acid and then recovered by elution with 20 μL of 49.5% acetonitrile/49.5% H2O/1% acetic acid and used directly for nanoelectrospray mass spectrometric analysis.

Mass spectrometry

Electrospray ionization mass spectrometry was performed on a hybrid linear ion trap-7-Tesla FTICR mass spectrometer (LTQ-FT, Thermo, San Jose, CA) equipped with a nanoelectrospray ionization source and operated with the Xcalibur (version 2.0) data acquisition software. HNE-adducted proteins or peptides were directly infused at 500 nL min−1 through a 50-μm-I.D. (pulled to 30-μm I.D.) New Objective (Woburn, MA) PicoTip. ESI spray voltage and capillary temperature were maintained at 2.0 kV and 250 °C, respectively. FTICR full-scan mass spectra were generally acquired at 100,000 nominal mass resolving power (MM at m/z 400 and taking the full width at half maximum intensity, FWHM, as ΔM) from m/z 500 to 2,000 using the automatic gain control mode of ion trapping. CID in the linear ion trap was performed using a 3.0-u isolation width and 35% normalized collision energy with helium as the target gas. ECD characterization of HNE-labeled peptides in the FTICR cell was carried out using an electrode current of 1.2 A and an ion irradiation time of 30 ms.

Results and discussion

HNE–apomyoglobin adduct characterization

The high mass resolving power provided by FTICR/MS was employed to accurately measure HNE adduct formation on apomyoglobin, since the various protein-derived charge state isotope clusters were easily resolved. To determine the extent of modification, apomyoglobin was resuspended in ESI spray solvent and analyzed by ESI-FTICR/MS as shown in Fig. 2. The charge state distribution of this protein was centered around the [M+15H]15+ ion upon ESI. The inset in Fig. 2 (upper panel) shows the isotopic resolution of the +15 charge state of native apomyoglobin obtained at 100,000 nominal mass resolving power. The mass measurement accuracy of the instrument was calculated by comparing the measured isotope mass values of the +15 charge state of apomyoglobin with the corresponding isotope mass values of the theoretical simulation of the same charge state (bottom panel). The average mass measurement accuracy obtained for the analysis of native apomyoglobin was approximately 1.5 ppm after external calibration of the FTICR instrument (mass measurement accuracies observed for full scan FTICR mass spectra were routinely less than 4 ppm).
https://static-content.springer.com/image/art%3A10.1007%2Fs00216-007-1534-2/MediaObjects/216_2007_1534_Fig2_HTML.gif
Fig. 2

High resolution mass spectrum of unmodified apomyoglobin obtained by nanoESI-hybrid linear ion trap-FTICR/MS (LTQ-FT, ThermoFisher). The top panel in the inset is an expanded m/z region showing isotopic resolution of the +15 charge state of unmodified apomyogobin. The mass accuracy and intensity distribution of this isotope envelope are in good agreement with a theoretical simulation (bottom panel)

Figure 3a represents a full-scan FTICR mass spectrum obtained for HNE-modified apomyoglobin which shows a complex pattern of peaks at each charge state of the modified protein. This complexity is due in part to the varying number and distribution of Schiff base or Michael addition HNE adducts. Figure 3b is an expanded m/z range of the +15 charge state of HNE-adducted apomyoglobin. The extent of modification was determined by comparing the experimentally derived peaks with isotopic simulations corresponding to various distributions of Michael-type (+156 Da) and/or Schiff base (+138 Da) adducts. FTICR/MS analysis allowed for the identification of three to nine sites of HNE modification in apomyoglobin. The increase in mass sequentially by 156 corroborates the proposed Michael-type addition mechanism for modification of apomyoglobin by HNE (see Fig. 1). Interestingly, ions corresponding to Schiff base addition of HNE to apomyoglobin could also be detected (Fig. 3c, upper), although the intensity of these ions is low compared to those of the apomyoglobin–HNE Michael-type adducts. Theoretical simulation of each HNE adduct within this particular m/z range was performed to support this finding (Fig. 3c, lower panel). Future studies will incorporate a bottom-up approach to characterize Schiff base adducts putatively identified from this analysis.
https://static-content.springer.com/image/art%3A10.1007%2Fs00216-007-1534-2/MediaObjects/216_2007_1534_Fig3_HTML.gif
Fig. 3

a FTICR mass spectrum of HNE-modified apomyoglobin. The charge state distribution ranged from [M+12H]11+ to [M+21H]21+. b Expanded m/z range of the +15 charge state of HNE-adducted apomyoglobin. Three to nine HNE adducts on apomyoglobin were observed. c Expanded m/z range of the +15 charge state of HNE-adducted apomyoglobin that contains between three to four Michael-type adducts (MA) of HNE. In addition to Michael-type addition of HNE to apomyoglobin, the formation of Schiff base adducts are also indicated (SB) in the upper panel. Theoretical isotopic simulation attests the presence of these adducts (bottom)

In this study, we investigated the extent of modification and mechanism by which adduct formation occurred by directly profiling the HNE reaction with apomyoglobin using the high mass resolving power and mass measurement accuracy of FTICR/MS. Extensive modification of apomyoglobin by HNE has previously been reported to occur [17] where the complexity of the ESI mass spectra obtained required processing and transformation (deconvolution) to clearly show the pattern of adduction and component molecular weights using a triple quadrupole mass analyzer [17]. However, deconvolution techniques may not be ideal when characterizing modifications of intact proteins, since artifact peaks and peak distortions may be introduced into the analysis [19].

The ESI-FTICR mass spectrum also showed a complex pattern of apomyoglobin–HNE adducts with multiple peaks observed at each charge state; however, evaluation of the pattern of adduction and the determination of component molecular weights were possible without deconvolution and additional data manipulation. We were able to identify the covalent attachment of up to nine HNE molecules to the protein. Most peaks could be assigned to products of Michael-type addition that represented the predominant mechanism of HNE reaction [20] over the competing Schiff base condensation. In horse apomyoglobin, 11 His residues are present and can be assumed to be targets for Michael-type addition, as indicated in previous studies [17, 18]. In addition to confirming earlier results by accurate m/z measurement at high mass resolving power, we have also detected the possible formation of Schiff base adducts on the protein apomyoglobin.

CID and ECD tandem mass spectrometry of HNE–insulin adducts

In contrast to CID, the ECD technique has encouraging potential for the characterization of protein carbonylation, since labile post-translational modifications such as HNE attachment are retained during MS/MS fragmentation. The CID and ECD product-ion spectra of oxidized insulin B chain modified by one and two HNEs are shown in Figs. 4 and 5, respectively, along with the corresponding MS/MS spectra of the unmodified peptides. The CID-derived MS/MS spectrum of the singly HNE-modified insulin peptide was complicated by numerous fragments produced due to the overall size of the peptide (>3 kDa) as well as the presence of HNE neutral loss fragment ions (Fig. 4b). From our experience with CID fragmentation of HNE-modified peptides, it is often difficult to assign the exact site(s) of modification. Frequently, no signature mass tag remains in the spectra due to HNE neutral loss from the peptide, which can be easily recognized upon comparing the CID product-ion spectra of the unmodified and modified peptides (Fig. 4a and b). As indicated in Fig. 4b, the predominant peaks observed in the MS/MS spectrum of the singly HNE-modified oxidized insulin B chain correspond to b-type fragment ions with and, mostly, without the HNE moiety (occurrence of neutral loss denoted by asterisk). Nevertheless, the high mass resolving power of the FTICR instrument was particularly effective in distinguishing isobaric fragment ions generated from the front-end linear ion trap. For example, an expanded m/z range of the CID product-ion spectrum shows clear separation of the \({\text{b}}_{16}^{2 + }\) and y8 monoisotopic peaks which differ by m/z 0.031. A theoretical simulation of these two fragment ions at 20,000 mass resolving power is superimposed on the expanded m/z region in Fig. 4c, demonstrating that some ambiguity with fragment ion assignment would exist with conventional high-performance instrumentation such as time-of-flight (TOF) mass spectrometers.
https://static-content.springer.com/image/art%3A10.1007%2Fs00216-007-1534-2/MediaObjects/216_2007_1534_Fig4_HTML.gif
Fig. 4

Collision-induced dissociation (CID) mass spectrum of oxidized insulin B chain unmodified (a) and modified with one HNE molecule (b). Fragment ions representing neutral loss of the HNE modification were apparent (indicated by asterisks). An expanded m/z range of the CID product-ion spectrum (c) shows clear separation of the \({\text{b}}_{16}^{2 + }\) and y8 monoisotopic peaks (Δmeasured = 0.03123 Th, Δtheoretical = 0.03109 Th) by FTICR (solid line). Theoretical simulations of these two fragment ions are shown at mass resolving powers of 75,000 (dashed line) and 20,000 (dotted line), respectively

https://static-content.springer.com/image/art%3A10.1007%2Fs00216-007-1534-2/MediaObjects/216_2007_1534_Fig5_HTML.gif
Fig. 5

Electron capture dissociation (ECD) mass spectrum of oxidized insulin B chain, [M+4H]4+ precursor ions, unmodified (a) and modified with two HNE molecules (b). Retention of the HNE modification on the fragment ions allowed for definitive sequence and modification assignment as demonstrated by the increase in mass of 156 Da on the c5 and c10 ions (indicated by #)

Although the improved mass resolving power and mass measurement accuracy of the FTICR instrument facilitated tandem mass spectrometric characterization of the HNE-modified insulin peptide, the exact HNE modification site was difficult to determine due to limited sequence coverage and neutral loss fragmentation obtained by the CID process. Consequently, ECD was employed as an alternative method for the MS/MS analysis of the HNE-modified insulin peptide. Figure 5 shows the ECD-derived fragmentation spectra of unmodified (Fig. 5a) and doubly HNE-modified (Fig. 5b) oxidized insulin B chain. Upon comparison of these two spectra, a shift in mass of +156 Da for the c5 ion with an additional increase of 156 Da for the c10 ion was observed. Since ECD preserves the HNE modification during MS/MS fragmentation, the amino acid residue containing the intact modification was identified. Specifically, the mass spectrum in Fig. 5b shows that oxidized insulin B chain is modified by HNE through the formation of Michael-type adducts at His-5 and His-10 (Fig. 6). To our knowledge, this is the first reported example demonstrating the potential of ECD for an improved characterization of a protein– or peptide–HNE adducts. Future work will include the implementation of data-dependent neutral loss-driven MS3 and ECD-based tandem mass spectrometric methods for the characterization of protein carbonylation sites.
https://static-content.springer.com/image/art%3A10.1007%2Fs00216-007-1534-2/MediaObjects/216_2007_1534_Fig6_HTML.gif
Fig. 6

Amino acid sequence of the oxidized insulin B chain with potential HNE-modification sites by Michael-type addition indicated by the number sign (#). (SO3H denotes Cys oxidation)

Sequence information obtained from MS/MS spectra using CID and ECD were complementary and hence both can be beneficial for the identification of PTMs. For example, the neutral loss of HNE observed in CID fragmentation spectra could serve as a signature tag for the presence of that group in the peptide. ECD can then be utilized to provide enhanced information regarding peptide sequence and site(s) of modification. The use of ECD for tandem mass spectrometric analysis is not, however, widely implemented for this purpose, since only FTICR mass spectrometers would allow for the reliable and efficient application of this technique to induce dissociation of protonated peptides and proteins.

Conclusion

The reported study has demonstrated the value of FTICR instrumentation in addressing the chemical aspects of the impact of oxidative stress on proteins. Oxidative stress has been implicated in aging and in almost every major chronic disease [21]. Oxidative stress-associated damage to proteins can significantly affect their activity, localization, turnover, and interaction with other molecules. Protein inactivation by formation of covalent adducts with HNE has been particularly well documented [1013]. All ROS examined thus far, including reactive nitrogen species, also give rise to protein carbonyls which, therefore, have been regarded as broad biomarkers of oxidative stress [22]. Understanding the mechanisms and selectivity of protein carbonylation should provide new diagnostic biomarkers for oxidative damage, and yield basic information to aid the development of an efficacious antioxidant therapy.

Acknowledgment

This research has been supported by the grant AG025384 from the National Institutes of Health. Laszlo Prokai is the Robert A. Welch Professor at the University of North Texas Health Science Center.

Copyright information

© Springer-Verlag 2007