|Home | About | Journals | Submit | Contact Us | Français|
Acrolein exposure leads to the formation of protein-acrolein adducts. Protein modification by acrolein has been associated with various chronic diseases including cardiovascular and neurodegenerative diseases. Here we report an analytical strategy that enables the quantification of Michael-type protein adducts of acrolein in mitochondrial proteome samples using liquid chromatography in combination with tandem mass spectrometry and selected ion monitoring (LC-MS/MS SRM) analysis. Our approach combines site-specific identification and relative quantification at the peptide level of protein–acrolein adducts in relation to the unmodified protein thiol pool. Treatment of 3-month old rats with CCl4, an established in vivo model of acute oxidative stress, resulted in significant increases in the ratios of distinct acrolein-adducted peptides to the corresponding unmodified thiol-peptides obtained from proteins that were isolated from cardiac mitochondria. The mitochondrial proteins that were found adducted by acrolein were malate dehydrogenase, NADH dehydrogenase [ubiquinone] flavoprotein 1, cytochrome c oxidase subunit VIb isoform 1, ATP synthase d chain, and ADP/ATP translocase 1. The findings indicate that protein modification by acrolein has potential value as an index of mitochondrial oxidative stress.
Humans are exposed to acrolein (CH2=CH-CHO) from endogenous sources, such as oxidative degradation of polyunsaturated fatty acids, polyamines and threonine, as well as exogenous sources such as heated cooking oil, heat-processed foods containing carbohydrates and amino acids, cigarette smoke, automobile exhaust and industrial emissions [1, 2]. Acrolein exposure may result in the modification of biological nucleophiles, such as nucleic acids, proteins and peptides [3–5]. Protein adducts of acrolein have been associated with various pathophysiological conditions including different types of cancer, diabetes, atherosclerosis, ischemia/reperfusion injury, spinal cord injury and neurodegenerative diseases [6–12]. Michael-type adduction by acrolein contributes to protein carbonyl content. Protein carbonylation occurs via diverse chemical pathways, such as backbone peptide bond cleavage and metal-catalyzed side-chain oxidations, and Michael-type adduction with 2-enals that include acrolein and (other) lipid peroxidation products [13, 14]. Protein carbonyls have also been extensively studied in meat products and other food proteins .
Traditionally, protein carbonyls are detected by colorimetric assays using 2,4-dinitrophenylhydrazine (DNPH)  or by immunochemical assays using anti-dinitrophenyl antibodies [17, 18]. However, these methods only report on the overall content of protein carbonyls and fail to assign modifications of distinct proteins. Protein carbonyl-specific antibodies have become available, e.g. antibodies against acrolein-protein adducts (i.e., Nε-(3-formyl-3,4-dehydropiperidino) (FDP) -lysine ), and HNE-protein adducts, which allow the detection of protein targets using gel-based approaches. However, site-specific assignments of oxidative modifications are usually not achievable with gel-based approaches. Mass spectrometry-based approaches have emerged that identify sites of oxidative protein modifications in in vitro studies [20–23]. Several groups have reported the use of tandem mass spectrometric approaches for the site-specific identification of protein modifications by 2-enals and other oxidative modifications in cell culture studies, tissues and biofluids [24–30]. Only a few studies have become available that attempt the targeted quantification of protein oxidative modifications using tandem mass spectrometry-based approaches [31, 32]. There is a need for analytical strategies that allow quantification of specific aldehydic protein adducts in relation to the unmodified protein pool in biological systems. Here we describe a targeted proteomic approach for protein carbonyls in conjunction with a selective mass spectrometry-based method for the simultaneous detection and quantification of distinct protein-acrolein adducts. The mass spectrometry platform used in this study consisted of a nanoflow liquid chromatography coupled to an electrospray triple quadrupole/linear ion trap mass spectrometer that was operated in the selected reaction monitoring mode (nanoLC-ESI-SRM analysis). This strategy was applied to quantify protein–acrolein adducts in cardiac mitochondria in an in vivo rodent model of oxidative stress (Figure 1).
Sequencing grade-modified trypsin was purchased from Promega Corp. (Madison, WI). Acrolein (ACR) (≥ 99%) was obtained from Fluka (St. Louis, MO). Aldehyde-reactive probe (ARP, N-aminooxymethylcarbonylhydrazino D-biotin) was purchased from Dojindo Laboratories (Kumamoto, Japan). UltraLink-immobilized monomeric avidin, iodoacetyl-PEG2-Biotin (IPB) and Triton X-100 detergent were obtained from Pierce (Rockford, IL). Bovine serum albumin (BSA) was from Calbiochem (La, CA). Macrospin strong cation exchange (SCX) columns were from Nest Group (Southborough, MA).
The experimental protocol for the animal studies was approved by the Institutional Animal Care and Use Committee at Oregon State University (ACUP #3770). Seven Male F344 rats (Harlan, Indianapolis, IN) were housed individually in plastic cages covered with Hepa filters. The animals were allowed free access to standard animal chow and water ad libitum. After 1 week of acclimatization, one 25-month old rat was sacrificed for an analytical reproducibility study and the other six 3-month old rats were transferred to metabolism cages. These six animals were divided into two groups of three, with one group receiving an intraperitoneal dose of 1 mL/kg CCl4 (dissolved in corn oil), and the other group (control) receiving the vehicle alone. The CCl4 dose of 1 mL/kg was chosen on the basis of literature reports [33, 34]. The rats were sacrificed 24 h after the treatment.
A solution of bovine serum albumin (BSA, 2 mg) in 1.0 mL of phosphate buffer (20 mM, pH 8.2) was reacted with 25 µL of DTT (60 mM in 20 mM phosphate buffer, pH 8.3) at 95 °C for 10 min. One half of the reduced BSA solution was then mixed with 200 µL 20 mM IPB in 50 mM phosphate buffer, (pH 8.3). The mixture was kept in the dark at room temperature for 90 min to form IPB-labeled BSA. Excess reagent was removed by adding 30 µL DTT (60 mM in 20 mM phosphate buffer, pH 8.3). The other half of the reduced BSA was reacted with 60 µL of acrolein (80 mM in 20 mM phosphate buffer, pH 8.3) at room temperature for 60 min. The excess of acrolein was removed by adding 40 µL DTT (60 mM in 20 mM phosphate buffer, pH 8.3). The ACR-modified BSA was then reacted with 250 µL ARP (30 mM in H2O) at room temperature for 60 min. Both ARP-ACR-modified and IPB-labeled BSA were digested with trypsin and passed through an ultrafiltration membrane (10 kDa MWCO). The peptide encompassing the residues 286 and 297 of BSA, YIC*DNQDTISSK modified on the cysteine residue (marked with an asterisk) with the ARP-ACR and IPB moiety, respectively, was used for method development.
Mitochondria were isolated from rat hearts and separated by differential centrifugation to obtain subsarcolemmal mitochondria (SSM) . Mitochondria were stored at −80 °C. Each sample of subsarcolemmal mitochondria (SSM) containing approximately 0.5 mg total protein was washed twice with phosphate buffer (10 mM NaH2PO4 pH 7.4) at 0 °C. The mitochondria were then resuspended in 400 µL of 10 mM NaH2PO4 (pH 7.4) containing 1% Trition X-100 detergent and 3 mM DTT. DTT was added to prevent thiol modifications by reactive species during sample preparation. The mitochondria were sonicated in ice water for 5 min to solubilize the proteins. The soluble protein fraction was obtained by centrifugation at 14,000 × g for 15 min at 4 °C. The supernatant was then filtered through an Amicon Microcon centrifugal filter (10 kDa MWCO) to remove low molecular weight molecules at 4 °C. Mitochondrial proteins were re-suspended in 400 µL of 10 mM NaH2PO4 (pH 7.4) containing 1 mM DTT. Further sample preparation was as follows (Supporting Information Figure S19):
Modified peptides were purified by strong cation exchange (SCX) chromatography using Macrospin SCX columns (Nest Group, Inc., Southborough, MA). Acetonitrile was used to activate the column. An elution buffer consisting of 20 % acetonitrile in 10 mM potassium phosphate buffer/0.6 M KCl (adjusted to pH 3 with H3PO4) was applied to condition the column. A washing buffer (10 mM potassium phosphate, 10 mM KCl, pH 3.0) containing 20% acetonitrile was used to equilibrate the column. Peptide samples (~pH 3.0, adjusted with H3PO4) were applied to the conditioned column and rinsed with washing buffer three times to remove Triton X-100 detergent and unbound components. Finally, 350 µL elution buffer was applied to the columns to release the peptides.
Ultralink monomeric avidin (200 µL, Pierce, Rockford, IL) was packed into Handee Mini Spin Columns (Pierce Rockford, IL) following the manufacturer’s protocol. Columns were washed with 1.5 mL of 10 mM NaH2PO4,pH 7.4. Irreversible binding sites, consisting of tetrameric avidin, were blocked by washing with 600 µL of 2 mM D-biotin. To remove excess D-biotin, columns were washed with 1 mL of 2 M glycine-HCl (pH 2.8). The columns were then re-equilibrated by washing twice with 2 mL of phosphate buffered saline (PBS, 20 mM NaH2PO4, 300 mM NaCl). The peptide samples were then slowly added to the affinity columns. To remove non-labeled and non-specifically bound peptides the column was washed twice with 1 mL PBS followed by 1 mL of 10 mM NaH2PO4 (pH 7.4) and finally 1.5 mL of 50 mM NH4HCO3 containing 20 % CH3OH. The columns were then rinsed with 1 mL of MilliQ H2O before eluting the ARP-labeled peptides with 0.4 % aqueous formic acid containing 20 % acetonitrile. Collected fractions were concentrated using a freeze dryer and stored at −20 °C prior to mass spectrometric analysis.
An Ultimate LC Packing system (Dionex, Sunnyvale, CA) was used. Peptide samples were loaded onto a 5 mm × 0.50 mm C18 trap cartridge (Dionex, Sunnyvale, CA) at a flow rate of 20 µL/min. After 4 min the trap cartridge was automatically switched in-line to a 75 µm i.d. ×15 cm C18 PepMap 100 column (Dionex, Sunnyvale, CA). Peptides were eluted using a gradient from 9 % to 18 % solvent B in A over 90 min at 0.260 µL/min. Solvent A was 1% aqueous acetonitrile containing 0.1% formic acid and solvent B was acetonitrile containing 0.1% formic acid.
All LC-MS/MS analyses were carried out on a 4000 Q-Trap hybrid tandem mass spectrometer (AB/MDS SCIEX, Concord, Ontario, Canada) equipped with a nano-ESI source. The electrospray voltage was set to 2300 V and the declustering potential was 60 V. For the identification of ARP-labeled peptides, precursor ion scanning was performed over a mass range of 400–1300 amu at 500 amu/s (Q1 and Q3 with unit resolution).
An enhanced product ion scan (MS/MS) was performed if the intensity of any of the precursors of m/z 227 exceeded the threshold value of 1000 counts/s (cps). The scan rate for MS/MS was set to 4000 amu/s. Tandem mass spectral data were analyzed using MASCOT v2.1 (Matrix Science, London, UK) as described previously [21, 36]. The Swiss Prot database v50 (270778 sequences, 99412397 residues) was searched using the following parameters: taxonomy rodentia (20991 sequences), ± 0.5 Da mass tolerances for the precursor and fragment ions, possibility of 2 missed proteolytic cleavage sites, with trypsin/P or semitrypsin selected as the digesting enzyme, and ARP-Acrolein (CHK), ARP-HNE (CHK), ARP-ONE (CHK), ARP-HHE (CHK), ARP-MDA (KR), ARP-β-hydroxyacrolein, ARP-crotonaldehyde selected as variable modifications at the residues specified in parenthesis. Fragment ion assignments were verified and probe-specific ions were annotated manually.
The SRM analyses were conducted with Q1 and Q3 set at unit resolution. Each SRM transition period was 30 ms. SRM collision energies were 50 eV and 51 eV for ARP-ACR-modified and IPB-labeled model peptides. For the mitochondrial peptides the collision energies used for the SRM analyses are listed in Table 1.
MALDI mass spectrometry was performed with an ABI 4700 Proteomics Analyzer with TOF/TOF optics and equipped with a 200-Hz frequency-tripled Nd:YAG laser operating at a wavelength of 355 nm (Applied Biosystems, Inc., Framingham, MA). Mass Spectra were obtained over a range of m/z 700–4000 in the reflectron mode. External mass calibration was applied using the ABI 4700 calibration mixture consisting of the following peptides des-Arg1-bradykinin ([M+H]+, m/zcalc 904.4675), angiotensin I ([M+H]+, m/zcalc 1296.6847), Glu1-fibrinopeptide B ([M+H]+, m/zcalc 1570.6768), and ACTH 18–39 ([M+H]+, m/zcalc 2465.1983). Peptide samples were clean up prior to MS analyses with C18 ZipTips (Millipore, Billerica, MA) following the manufacture’s protocol. Peptides were mixed with α-cyano-4-hydroxycinnamic acid (2 mg/mL in 50 % acetonitrile containing 0.1 % TFA) and 0.5 µL of the mixture was spotted onto a 144-spot stainless steel target plate.
We have recently completed a qualitative profiling study of protein adducts of lipoxidation products in the subsarcolemmal mitochondria from hearts of 24-month old rats . This study identified a diverse set of protein lipoxidation adducts. The majority of the assigned adducts were Michael-type acrolein adducts with cysteine residues . Here we report a strategy that enables the quantification of distinct cysteine-acrolein adducts at the peptide level which is based on differential chemical tagging of the acrolein adducts and the corresponding thiol-containing peptides (Figure 1). In order to determine the ratios of acrolein-modified peptides to their corresponding, i.e. unmodified, thiol-peptides in biological matrices we used N'-aminooxymethylcarbonyl- hydrazino-D-biotin as an aldehyde/keto reactive probe (ARP)as described by us previously  and iodoacetyl-PEO2-biotin (IPB) as a chemoselective tagging reagent for thiol-peptides (Figure 2). The combined use of ARP and IPB allowed biotin-based enrichment and analysis by nano-LC MS/MS using selected reaction monitoring (SRM) of both analyte groups concomitantly. The use of SRM mass spectrometry enabled the targeted relative quantification of both analyte groups, i.e. ARP-labeled cysteine acrolein adducts and the corresponding IPB-labeled thiols, at the peptide level with high specificity and sensitivity (Figure 1).
We first tested the reaction yields for the two labeling reactions using a small, commercially available cysteine-containing model peptide, CLLLSAPRR (MH+, m/z 1028.6). The model peptide was reacted with acrolein to yield an ACR-modified peptide, which was further reacted with ARP to form an ARP-ACR modified peptide (MH+, m/z 1397.8). The model peptide was also reacted with IPB to form an IPB-labeled peptide (MH+, m/z 1442.8). The MALDI MS spectra of the two crude products indicated that the thiol and ACR-modified peptide were converted into the IPB and ARP derivatives, respectively (Figure S1 in the supporting information).
We next tested our analytical strategy using a tryptic digest of BSA containing ARP-ACR modified and IPB-labeled peptides. For this purpose, BSA was reduced with DTT and the reduced BSA solution was divided into two aliquots. One aliquot was treated with IPB, while the other aliquot was treated with acrolein and reacted with ARP. Both modified BSA samples were then digested with trypsin. The peptide YICDNQDTISSK, residues 286–297 of BSA with ARP-ACR and IPB modification, respectively, was used for the proof-of-concept experiment. The other tryptic peptides were considered as background matrix. The nanoLC-ESI-MS/MS spectra of the IPB-labeled peptide and the ARP-ACR-modified peptides are depicted in Figures S2 and S3. Three fragment ions (y3, y5, and y6) were selected for nanoLC-ESI-MS/MS SRM analysis. Figure S4 shows the SRM chromatograms for 100 fmol of ARP-labeled peptide acrolein adduct and 100 fmol IPB-labeled peptide. The SRM peak areas of the ARP-ACR-modified peptide were equivalent to those obtained for the IPB-labeled peptide, implying that the two differently modified peptides had comparable ionization and fragmentation characteristics.
To test the suitability of our strategy for quantification, three SRM standard curves were constructed using 100 fmol IPB-labeled BSA peptide (aa 286–297) as a reference. The amount of ARP-ACR-modified peptide (analyte) was varied between 10.0 fmol to 1000 fmol (Figure S5). LC-SRM experiments were conducted by monitoring three transitions for the analyte peptide and the corresponding transitions for the reference peptide. The following transitions were monitored for the ARP-ACR-modified peptide (analyte): m/z 878.4 (MH22+) → m/z 321.2 (y3), m/z 878.4 (MH22+) → m/z 535.3 (y5) and m/z 878.4 (MH22+) → m/z 650.3 (y6). For the corresponding IPB-labeled peptide the following SRM transitions were used: m/z 900.9 (MH22+) → m/z 321.3 (y3), m/z 900.9 (MH22+) → m/z 535.3 (y5), and m/z 900.9 (MH22+) → m/z 650.3 (y6). From the SRM chromatograms the peak area for each transition pair were extracted and the peak area ratios determined. To construct the standard curves for each of the three transition pairs monitored the peak area ratio (analyte-to-reference) were plotted as a function of amount ratio (analyte-to-reference). Linear regression analysis yielded slopes of 0.979, 1.102 and 1.100 for each transition pair monitored (Figure S5). The R2 was higher than 0.997. For the subsequent mitochondrial proteome analyses, we therefore assumed for the transition pairs monitored a response factor of 1. Thus, the amount ratio between ARP-ACR and IPB-labeled peptides was directly derivable from the ratio of the peak areas extracted from the SRM chromatograms.
In our previous work, profiling of oxidative modification sites was based on tandem mass spectrometry of ARP-tagged aldehydic protein adducts [29, 37]. Tandem mass spectra of ARP-labeled protein lipoxidation adducts showed a fragment ion at m/z 227.1 which originates from the ARP moiety (F1, Figure S18A) . Thus, we used this probe-specific feature to identify protein adducts of lipoxidation products in the current mitochondrial preparations. We conducted precursor ion scanning experiments to trigger the acquisition of full scan tandem mass spectra for identifying the targets of lipoxidation. This assay consistently yielded seven acrolein-modified peptides with acrolein modifications on distinct cysteine residues of five mitochondrial proteins. The acrolein-modified Cys-containing peptides are listed in Table 2 and are labeled as a through g. The modified sites were also found as targets of acrolein adduction in our previously described study . The modified proteins are involved in mitochondrial energy production (TCA cycle and oxidative phosphorylation) and ADP/ATP transport (Table 2) [38, 39]. The peptide adducts a through g were subsequently selected for quantification using LC-MS/MS SRM analysis.
In order to use these distinct modification sites for quantification we first conducted a series of tandem mass spectral experiments to obtain optimized CID parameters for the ARP-labeled peptide–acrolein adducts and the IPB-alkylated thiol peptides. The tandem spectrum for the EFNGLGDCLTK peptide derived from ADP/ATP translocase 1 is shown in Figure 3. The fragmentation spectrum of the doubly charged ion of the ARP-labeled peptide-acrolein adduct (MH22+, m/z 783.5) showed mainly yn and bn-type ions, which are typically observed in ESI-MS/MS studies with low-energy CID conditions. Similarly, the tandem mass spectrum of the IPB-labeled EFNGLGDCLTK peptide (MH22+, m/z 805.8) showed the corresponding yn and bn-type ions. Examination of the mass spectral data indicated that the fragmentation pattern of the peptide modified by ARP-ACR (+ 369.2 Da) moiety was analogous to the fragmentation pattern of the corresponding peptide alkylated with IPB (+ 414.2 Da). The relative fragment ion intensity of the yn or bn-type ions was also comparable for the ARP-ACR and IPB-labeled peptide. Both fragment ion spectra showed probe-specific ions. The ARP-ACR-labeled peptide yielded two characteristic ions, F1 (m/z 227.1) and F2 (m/z 332.1) (Figure 3A, and Figure S18A). Whereas in the spectra of the IPB-labeled peptide the IPB-specific fragment ions, Fa (m/z 270.1) and Fb (m/z 332.2) were observed (Figure 3B, Figure S18B). The tandem mass spectra for the other ARP-labeled peptide-acrolein adducts and the respective IPB-labeled peptides are provided in the supplemental material (Figures S6-S17).
Because the fraction of a distinct acrolein adduct in the endogenous protein thiol pool is very low, an analytical workflow was needed that enabled the quantification of acrolein-modified peptides and the respective peptide thiols in one LC-MS/MS SRM-run. Figure S19 outlines the adopted workflow to circumvent the substantial concentration difference between acrolein-modified peptides and their corresponding peptide thiols, and to avoid problems with the dynamic range of the mass spectrometer and overloading of the nano-LC column. Accordingly, 800 µL of ARP-labeled ACR-peptide digest was used, but only 35 µl of IPB-modified digest was used. Thus, only a fraction of the IPB-modified digest (0.8 µl; derived from 40 µL × 35 µL/1750 µL = 0.8 µL) was ultimately used for nanoLC SRM analysis. Consequently, to calculate the ratio of the ARP-ACR-modified peptides to the IPB-labeled peptides in the original mitochondrial samples the peak area ratios of ARP-ACR-modified peptide to IPB-labeled peptides obtained from nanoLC-SRM analysis in the combined sample were divided by 1000 (derived from 800 µL/0.8 µL).
For the quantitative analyses, for each peptide three SRM transitions were constructed based on fragment ions that were observed for the ARP-labeled peptide-acrolein adduct as well as for the respective IPB-labeled thiol peptide. In Figure 3 and Figures S6 to S17 the fragment ions that were used for the SRM transitions are encircled. For each LC-MS/MS SRM analysis 42 transitions were monitored. Representative SRM chromatograms for the seven ARP-ACR-modified peptides and their corresponding IPB-labeled peptides are shown in Figure S20. The ARP-ACR-modified peptides usually eluted at shorter retention times than the corresponding IPB-labeled peptides. The chromatographic peak widths of the ARP-ACR-modified peptides were usually larger than the peak widths observed for the corresponding IPB-labeled peptides. To compensate for the mass difference of 45 Da between ARP-ACR and IPB-labeled peptides, the collision energy for the IPB-labeled peptide ions was set 1eV higher than the collision energy used for the corresponding ARP-ACR-modified peptide ions. Detailed information on the SRM parameters used for the quantitative analysis of the APR-ACR and IPB-labeled peptides are given in Table 1.
We also evaluated the precision of our method. Three mitochondrial preparations from one 25-month old rat were carried through sample derivatization, enrichment and nano-LC SRM analysis. The results are shown in Table S1. The coefficient of variation (CV) for these three samples was around 5%. The ratio of acrolein-modified peptide to the corresponding thiol peptide was found to vary from 4×10−4 to 1×10−3. Noteworthy, for ADP/ATP translocase 1 and malate dehydrogenase two distinct acrolein adduction sites were observed and we were able to determine acrolein adduction levels at discrete sites within a protein. ADP/ATP-translocase 1 (ANT1) is a well described target of oxidative modification reactions [36, 40, 41]. ANT1 has four cysteine residues, and in this study we were able to assess acrolein adduction levels at Cys-159 and Cys-256. ANT1 undergoes complex conformational transitions during transport of adenine nucleotides through the inner mitochondrial membrane . Earlier studies in which thiol reagents were used as conformational probes indicated that that Cys-159 was highly reactive and accessible to a diverse set of thiol probes in contrast to Cys-256 . Indeed, acrolein adduction at Cys-159 was found to be higher than on Cys-256. We and others have previously identified malate dehydrogenase as a protein which is susceptible to modifications by 2-enals [25, 36, 43, 44]. Malate dehydrogenase is a TCA cycle protein and has six cysteine residues. In this study, acrolein adduction was consistently observed at Cys-93 and Cys-212. As part of our earlier study, we have extracted B-Factors, an index of on the protein dynamics, and surface accessible area (SAA) values for the cysteine residues from the published X-ray structure of malate dehydrogenase (pdb 1MLD) . The higher adduction levels observed for Cys-212 are consistent with the higher B-Factor and SAA value observed for Cys-212 (B-Factor 33.71, SAA 45.95) compared to Cys-93 (B-Factor 26.27, SAA 20.55).
Next, we used our differential labeling strategy in combination with LC-MS/MS SRM for determining relative levels of protein–acrolein adducts in cardiac mitochondria from 3-month old rats that were administered CCl4 intraperitoneally, an accepted in vivo model of acute oxidative stress . Urine samples collected from the same rats in this experiment had elevated levels of metabolites of 4-hydroxy-2-nonenal and 4-oxo-2-nonenal compared to control animals, indicating that these urinary metabolites are indeed products of CCl4-induced lipid peroxidation . Metabolites of acrolein were not analyzed for in this study. The relative levels of peptide ACR-adducts and peptide thiols were determined in the control and the CCl4-exposed animal group. Figure 4 shows that the mean ratios of ACR-modified peptide to the unmodified peptide thiol were higher in the CCl4-treated animals than those in control rats for all seven peptides a-g listed in Table 2. CCl4-treatment resulted in significant ratio differences for peptides b, c, d, f, and g with p-values < 0.05. There was insufficient statistical difference for the other peptides (a and e), but, the trend was conserved that CCl4-exposure resulted in increased acrolein adduction levels in relation to the respective peptide thiol pool. These results indicate that ACR-modified proteins may have value as mitochondrial markers of oxidative stress.
Acrolein is formed via multiple routes, e.g. from metabolism of allyl compounds, lipid peroxidation of polyunsaturated fatty acids, and from oxidative metabolism of polyamines by amine oxidases. Acrolein is known to be the most reactive of the α,β-unsaturated aldehydes. Although the ARP labeling strategy is generally applicable to aldehydic protein modification, only ARP-labeled ACR adducts were reproducibly detected and quantified in this study. A possible reason for the selective detection of protein–acrolein adducts is that they, unlike lipid-derived 4-hydroxy-2-enals, cannot form cyclic hemiacetals that would interfere with ARP labeling. Michael-type protein adducts of 4-hydroxy-2-enals, e.g. 4-hydroxy-2-nonenal (HNE), may partially escape ARP labeling due to their formation of cyclic hemiacetals with reduced reactivity toward ARP . Alternatively, the greater reactivity of acrolein towards protein thiols compared with other alkenals may also contribute to the bias towards cysteine-acrolein adducts in the current dataset [5, 13]. The reported data only represent a subset of possible acrolein modifications, namely Michael-type adducts of acrolein to cysteine residues. For this reason, a limitation of the reported method is that it fails to quantify other possible acrolein modifications, such as Schiff’s base-type modification, FDP-lysine modification and possible cross-links . For comparison, Judge et al. reported 0.26 nmol protein carbonyl/mg protein for subsarcolemmal mitochondrial sample using an immunoassay .
Taken together, we have developed an analytical strategy that enables the quantification of Michael-type protein adducts of acrolein in complex biological mixtures. Our approach combines the identification and relative quantification of ACR-modified proteins by nano-LC MS/MS SRM analysis. In addition, the method provides information on specific amino acids that are modified by acrolein, their location and their extent of modification in a protein. We have shown that the ratios of acrolein adduction level to thiol level at a distinct site have value as a marker of oxidative stress. This strategy is not limited to acrolein adducts of proteins but could also be applied to determine levels of adduction by other 2-enals. The method has made inroads into assessment of the in vivo acrolein status as a function of oxidant stress in health and disease. The analytical strategy should be equally applicable to heat-processed foods in which thermally produced acrolein forms adducts with proteins.
We thank Dr. Cristobal Miranda for animal husbandry. This work was supported by NIH grants R01 AG025372 (Maier) and R01 HL081721 (Stevens). The OSU EHSC mass spectrometry facility and core is supported in part by the Environmental Health Sciences Center (P30 ES00210, S10 RR022589).