HNE Michael Adducts to Histidine and Histidine-Containing Peptides

Nov 3, 2007 - (His) residues modified by reactive carbonyl species. (RCS) generated by lipid peroxidation. This approach has been applied to urines fr...
0 downloads 0 Views 281KB Size
Download PDF
Anal. Chem. 2007, 79, 9174-9184

HNE Michael Adducts to Histidine and Histidine-Containing Peptides as Biomarkers of Lipid-Derived Carbonyl Stress in Urines: LC-MS/ MS Profiling in Zucker Obese Rats Marica Orioli, Giancarlo Aldini, Maria Carmela Benfatto, Roberto Maffei Facino, and Marina Carini*

Istituto di Chimica Farmaceutica e Tossicologica “Pietro Pratesi”, Faculty of Pharmacy, University of Milan, Via Mangiagalli 25, I-20133 Milan, Italy

A new liquid chromatography-tandem mass spectrometric (LC-MS/MS) approach, based on the precursor ion scanning technique using a triple-stage quadrupole, has been developed to detect free and protein-bound histidine (His) residues modified by reactive carbonyl species (RCS) generated by lipid peroxidation. This approach has been applied to urines from Zucker obese rats, a nondiabetic animal model characterized by obesity and hyperlipidemia, where RCS formation plays a key role in the development of renal and cardiac dysfunction. The immonium ion of His at m/z 110 was used as a specific product ion of His-containing peptides to generate precursor ion spectra, followed by MS2 acquisitions of each precursor ion of interest for structural characterization. By this approach, three novel adducts, which are excreted in free form only, have been identified, two of them originating from the conjugation of 4-hydroxy-trans-2nonenal (HNE) to His, followed by reduction/oxidation of the aldehyde: His-1,4-dihydroxynonane (His-DHN), His-4-hydroxynonanoic acid (His-HNA), and carnosineHNE, this last recognized in previous in vitro studies as a new potential biomarker of carbonyl stress. No free HisHNE was found in urines, which was detected only in protein hydrolysates. The same LC-MS/MS method, working in multiple reaction monitoring (MRM) mode, has been developed, validated, and applied to quantitatively profile in Zucker urines both conventional (1,4dihydroxynonane mercapturic acid, DHN-MA) and the newly identified adducts, except His-HNA. The analytes were separated on a C12 reversed-phase column by gradient elution from 100% A (water containing 5 mM nonafluoropentanoic acid) to 80% B (acetonitrile) in 24 min at a flow rate of 0.2 mL/min and analyzed for quantification in MRM mode by applying the following precursor-to-product ion transitions m/z 322.2 f 164.1 + 130.1 (DHN-MA), m/z 314.7 f 268.2 + 110.1 (HisDHN), m/z 312.2 f 110.1 + 156.0 (His-HNE), m/z 383.1 f 266.2 + 110.1 (CAR-HNE), m/z 319.2 f 301.6 + 156.5 (H-Tyr-His-OH, internal standard). * To whom correspondence should be addressed. E-mail: [email protected]. Phone: + 39-02-50319298. Fax: +39-02-50319359.

9174 Analytical Chemistry, Vol. 79, No. 23, December 1, 2007

Precision and accuracy data, as well as the lower limits of quantification in urine, were highly satisfactory (from 0.01 nmol/mL for CAR-HNE, His-DHN, His-HNE, to 0.075 nmol/mL for DHN-MA). The method, applied to evaluate for the first time the advanced lipoxidation end products profile in urine from obese Zucker rats, an animal model for the metabolic syndrome, has proved to be suitable and sensitive enough for testing in vivo the carbonyl quenching ability of newly developed RCS sequestering agents. Reactive carbonyl species (RCS) generated by lipid peroxidation, such as 4-hydroxy-trans-2-nonenal (HNE), 2-propenal (acrolein, ACR), malondialdehyde (MDA), and glyoxal (GO), leading to covalently modified proteins, may contribute to oxidative tissue damage. They are known to be involved in the pathogenesis of several cardiovascular (atherosclerosis, cerebral or heart ischemia-reperfusion injury) and neurodegenerative (Alzheimer’s, Parkinson’s) diseases, as well as in diabetes and cancer.1-3 Most of the biological effects of intermediate lipid-derived RCS are attributed to their capacity to react with the nucleophilic sites of peptides and proteins (Cys, His, or Lys residues) forming advanced lipoxidation end products (ALEs), such as MDA-Lys (Schiff base adduct), HNE-Lys (Michael adduct),4 HNE-Lys (pyrrole derivative),5 FDP-Lys [N-(3-formyl-3,4-dehydropiperidino)lysine],6 MP-Lys [N-(3-methylpyridinium)lysine],7 levuglandin adducts (pyrrole derivatives),8 CMC [S-(carboxymethyl)(1) Aldini, G.; Dalle-Donne, I.; Colombo, R.; Maffei Facino, R.; Milzani, A.; Carini, M. ChemMedChem 2006, 1, 1045-1058. (2) Aldini, G.; Dalle-Donne, I.; Maffei Facino, R.; Milzani, A.; Carini, M. Med. Res. Rev. 2007, 27, 817-868. (3) Carini, M.; Aldini, G.; Maffei Facino, R. In Redox Proteomics: from Protein Modifications to Cellular Dysfunction and Diseases; Dalle-Donne, I., Scaloni, A., Butterfield, A., Eds.; John Wiley & Sons Inc.: Hoboken, NJ, 2006; pp 887-929. (4) Requena, J. R.; Fu, M. X.; Ahmed, M. U.; Jenkins, A. J.; Lyons, T. J.; Baynes, J. W.; Thorpe, S. R. Biochem. J. 1997, 322, 317-325. (5) Sayre, L. M.; Arora, P. K.; Iyer, R. S.; Salomon, R. G. Chem. Res. Toxicol. 1993, 6, 19-22. (6) Uchida, K. Trends Cardiovasc. Med. 1999, 9, 109-113. (7) Furuhata, A.; Ishii, T.; Kumazawa, S.; Yamada, T.; Nakayama, T.; Uchida K. J. Biol. Chem. 2003, 278, 48658-48665. (8) Davies, S. S.; Amarnath, V.; Roberts, L. J., II. Chem. Phys. Lipids 2004, 128, 85-99. 10.1021/ac7016184 CCC: $37.00

© 2007 American Chemical Society Published on Web 11/03/2007

cysteine],9 CML [N-(carboxymethyl)lysine],10 and GOLD (GOLys dimer).11,12 In particular, CML (formed by reaction of GO with Lys residues) is now considered as a general marker of carbonylic stress and long-term damage to proteins in aging, Alzheimer’s disease (AD), atherosclerosis, and diabetes.13 Antisera for quantitative immunoassays of protein-bound ALEs in body fluids and tissues, although widely used, are questionable, because they yield only semiquantitative results. Hence, more specific and sensitive analytical approaches, mainly based on LC, GC/MS, and LCMS/MS methodologies, have been developed for their quantitative determination in proteins hydrolysates. Free CMC or CMC released from proteins by hydrolysis can be easily determined by reversed-phase HPLC with fluorescence detection after precolumn derivatization with o-phthaldialdehyde.9 CML has been estimated in modified skin collagen, plasma, or urines by reversedphase HPLC method with o-phthalaldehyde precolumn derivatization14 or by isotope dilution, selected ion monitoring gas chromatography/mass spectrometry (SIM-GC/MS).15,16 CML (in free and protein-bound form after hydrolysis) is now determined by isotope dilution LC-MS/MS.17 This method can be accurately applied to all the biological matrices (plasma, tissues, and urine). The mercapturic acid conjugate of 1,4-dihydroxynonane (DHNMA)18 is another well-established and highly stable biochemical biomarker for lipid peroxidation and carbonylation. This adduct, representing the main urinary end metabolite of HNE following conjugation with GSH, has been shown to be a physiological component of rat and human urine.19 DHN-MA is generally determined by LC-MS/MS.20 More recently, an enzyme immunoassay has been developed and validated for its determination in urines.21 Hence, the mainly used in vivo biomarkers for carbonylation are RCS-modified Cys and Lys residues. On the contrary, little attention has been devoted to covalently modified His as a possible diagnostic tool. This is quite surprising, in view of the following evidence: (a) the well-documented reactivity of His residues toward several RCS (mainly HNE, ACR, and MDA);1-3 (b) the (9) Zeng, J.; Davies, M. J. Chem. Res. Toxicol. 2005, 18, 1232-1241. (10) Wells-Knecht, K. J.; Zyzak, D. V.; Litchfield, J. E.; Thorpe, S. R.; Baynes, J. W. Biochemistry 1995, 34, 3702-3709. (11) Wells-Knecht, K. J.; Brinkmann, E.; Thorpe, S. R.; Baynes, J. W. J. Org. Chem. 1995, 60, 6246-6247. (12) Brinkmann, E.; Wells-Knecht, K. J.; Thorpe, S. R.; Baynes, J. W. J. Chem. Soc., Perkin Trans. 1 1995, 1, 2817-2818. (13) Girone`s, X.; Guimera`, A.; Cruz-Sa´nchez, C. Z.; Ortega, A.; Sasaki, N.; Makita, Z.; Lafuente, J. V.; Kalaria, R.; Cruz-Sa´nchez, F. F. Free Radical Biol. Med. 2004, 36, 1241-1247. (14) Wihler, C.; Scha¨fer, S.; Schmid, K.; Deemer, E. K.; Mu ¨ nch, G.; Bleich, M.; Busch, A. E.; Dingermann, T.; Somoza, V.; Baynes, J. W.; Huber, J. Diabetologia 2005, 48, 1645-1653. (15) Alderson, N. L.; Chachich, M. E.; Youssef, N. N.; Beatti, R. J.; Nachtigal, M.; Thorpe, S. R.; Baynes, J. W. Kidney Int. 2003, 63, 2123-2133. (16) Petrovicˇ, R.; Futas, J.; Chandoga, J.; Jakus, V. Biomed. Chromatogr. 2005, 19, 649-654. (17) Teerlink, T.; Barto, R.; ten Brink, H. J.; Schalkwijk, C. G. Clin. Chem. 2004, 50, 1222-1228. (18) Peiro, G.; Alary, J.; Cravedi, J. P.; Rathahao, E.; Steghens, J. P.; Gue´raud, F. Biofactors 2005, 24, 89-96. (19) Alary, J.; Debrauwer, L.; Fernandez, Y.; Cravedi, J. P.; Rao, D.; Bories, G. Chem. Res. Toxicol. 1998, 11, 130-135. (20) Rathahao, E.; Peiro, G.; Martins, N.; Alary, J.; Gue´raud, F.; Debrauwer, L. Anal. Bioanal. Chem. 2005, 381, 1532-1539. (21) Gue´raud, F.; Peiro, G.; Bernard, H.; Alary, J.; Cre´minon, C.; Debrauwer, L.; Rathahao, E.; Drumare, M. F.; Canlet, C.; Wal, J. M.; Bories, G. Free Radical Biol. Med. 2006, 40, 54-62.

already established order of reactivity of the nucleophilic amino acids (Cys . His > Lys) toward R,β-unsaturated aldehydes,22 as demonstrated for a number of peptides, proteins, and enzymes by mass spectrometry.23 Structural information on the nature of the HNE modification of histidine side chain was first reported by Uchida and Stadtman.24 Reduction of the aldehyde group of the primary Michael addition product with sodium borohydride converts it to the hydroxy derivative that is stable to strong acid hydrolysis. This approach formed the basis of methods for the identification and quantification of the HNE-histidine Michael adduct of proteins by conventional amino acid analytical techniques. Immunohistochemistry using monoclonal antibodies to HNE-histidine conjugates has been widely applied to detect HNE-histidine protein adducts in human astrocytomas,25 in the ischemic rat heart after transplantation,26 in the target organ of a ferric nitrilotriacetate-induced renal carcinogenesis model,27 in ischemia/reperfused rat heart,28 and in serum of type 2 diabetic patients.29 On the contrary, very few studies report free His modified by RCS as a tool to confirm involvement of carbonyl stress in different pathological conditions. Increased levels of the His-HNE adduct have been found in the middle frontal gyrus of AD patients compared with controls, but no details on the analytical method used (LC-MS/MS) are reported.30 Other endogenous histidinecontaining dipeptides have been more recently recognized as detoxifying agents against cytotoxic HNE and ACR: carnosine (β-alanyl-L-histidine, CAR), homocarnosine (γ-amino-butyryl-histidine, HCAR), and anserine (β-alanyl-L-1-methylhistidine, ANS).31-33 These compounds are widely distributed in vertebrates and particularly abundant in excitable tissues such as nervous system and skeletal muscles,34,35 only CAR and HCAR being normally present in human tissues. In particular, the Michael adduct of HNE with CAR (CAR-HNE) has been identified as a novel, early, specific, and stable marker of lipid peroxidation in those biological districts where CAR is specifically located, and an LC-MS/MS (22) Doorn, J. A.; Petersen, D. R. Chem.sBiol. Interact. 2003, 143-144, 93100. (23) Carini, M.; Aldini, G.; Maffei Facino, R. Mass Spectrom. Rev. 2004, 23, 281305. (24) Uchida, K.; Stadtman, E. R Proc. Natl. Acad. Sci. U.S.A. 1992, 89, 45444548. (25) Zarkovic, K.; Juric, G.; Waeg, G.; Kolenc, D.; Zarkovic, N. Biofactors 2005, 24, 33-40. (26) Renner, A.; Sagstetter, M. R.; Harms, H.; Lange, V.; Go¨tz, M. E.; Elert, O. J. Heart Lung Transplant. 2005, 24, 730-736. (27) Ozeki, M.; Miyagawa-Hayashino, A.; Akatsuka, S.; Shirase, T.; Lee, W. H.; Uchida, K.; Toyokuni, S. J. Chromatogr., B: Anal. Technol. Biomed. Life Sci. 2005, 827, 119-126. (28) Chen, J.; Henderson, G. I.; Freeman, G. L. J. Mol. Cell. Cardiol. 2001, 33, 1919-1927. (29) Toyokuni, S.; Yamada, S.; Kashima, M.; Ihara, Y.; Yamada, Y.; Tanaka, T.; Hiai, H.; Seino, Y.; Uchida, K. Antioxid. Redox Signaling 2000, 2, 681685. (30) Cutler, R. G.; Kelly, J.; Storie, K.; Pedersen, W. A.; Tammara, A.; Hatanpaa, K.; Troncoso, J. C.; Mattson, M. P. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 2070-2075. (31) Aldini, G.; Carini, M.; Beretta, G.; Bradamante, S.; Maffei Facino, R. Biochem. Biophys. Res. Commun. 2002, 298, 699-706. (32) Aldini, G.; Granata, P.; Carini, M. J. Mass Spectrom. 2002, 37, 12191228. (33) Carini, M.; Aldini, G.; Beretta, G.; Arlandini, E.; Maffei Facino, R. J. Mass Spectrom. 2003, 38, 996-1006. (34) Bonfanti, L.; Peretto, P.; De Marchis, S.; Fasolo, A. Prog. Neurobiol. 1999, 59, 333-353. (35) Stuerenburg, H. J. Biochemistry (Moscow, Russ. Fed.) 2000, 65, 862-865.

Analytical Chemistry, Vol. 79, No. 23, December 1, 2007

9175

method for its quantification in biological matrices (rat skeletal muscle) has been developed.36 In view of these findings, the first aim of this work was to develop an LC-MS/MS approach, based on the precursor ion scanning mode, to detect all free and protein-bound His species modified by RCS. This technique is a valuable tool for the rapid confirmation of targeted compounds or for the nontargeted detection of compounds bearing a common moiety. It has been widely used in drug discovery and development37 and in vitro and in vivo drug metabolism studies,38-40 although it has never been applied to screen a biological matrix for the presence of covalently modified peptides. The only application in this field has been reported by Fenaille et al.,41 which used the precursor ion scanning technique to identify peptides containing hexanal-modified Lys residues within an unfractionated tryptic digest of hexanalmodified apomyoglobin (in vitro studies). We used the immonium ion of histidine (m/z 110) as a wellknown and specific fragment ion of His-containing peptides36 to generate precursor ion spectra, with the final aim to identify in urines from Zucker obese rats covalently modified His residues. The Zucker fa/fa rat has been chosen as a nondiabetic animal model characterized by obesity and hyperlipidemia, where ALEs formation plays a key role in the development of renal and cardiac dysfunction.15 The involvement of lipid peroxidation in renal injury in obese Zucker rats has been recently demonstrated by the heavy deposition of HNE adducts, using a specific rabbit polyclonal antibody to HNE adducts.42 Furthermore, although a wide number of papers have been published on determination of ALEs in biological fluids, none of them tried to simultaneously quantitate the most important biomarkers of carbonylic stress in urines from Zucker rats. In order to fill this gap, and to define for the first time the urinary profile of both conventional ALEs (DHN-MA) and the newly identified adducts to histidine, the second aim of this work was to develop an LC-MS/MS methodology in multiple reaction monitoring (MRM) for their quantitation. This approach was done with the final goal to use the obese Zucker rat as a pharmacological tool to evaluate, through the determination of a series of highly stable biochemical markers for carbonylation, the protective effects of a long-term administration of newly developed RCS sequestering agents. EXPERIMENTAL SECTION Chemicals and Reagents. All chemicals and reagents were of analytical grade and purchased from Sigma-Aldrich Chemical Co. (Milan, Italy). HPLC-grade and analytical-grade organic solvents were also purchased from Sigma-Aldrich (Milan, Italy). HPLC-grade water was prepared with a Milli-Q water purification (36) Orioli, M.; Aldini, G.; Beretta, G.; Maffei Facino, R.; Carini, M. J. Chromatogr., B: Anal. Technol. Biomed. Life Sci. 2005, 827, 109-118. (37) Papac, D. I.; Shahrokh, Z. Pharm. Res. 2001, 18, 131-145. (38) Zhang, J. Y.; Wang, Y.; Dudkowski, C.; Yang, D.; Chang, M.; Yuan, J.; Paulson, S. K.; Breau, A. P. J. Mass Spectrom. 2000, 35, 1259-1270. (39) Kostiainen, R.; Kotiaho, T.; Kuuranne, T.; Auriola, S. J. Mass Spectrom. 2003, 38, 357-372. (40) Liu, D. Q.; Hop, C. E. J. Pharm. Biomed. Anal. 2005, 37, 1-18. (41) Fenaille, F.; Tabet, J. C.; Guy, P. A. Rapid Commun. Mass Spectrom. 2004, 18, 67-76. (42) Dominguez, J. H.; Wu, P.; Hawes, J. W.; Deeg, M.; Walsh, J.; Packer, S. C.; Nagase, M.; Temm, C.; Goss, E.; Peterson, R. Kidney Int. 2006, 69, 19691976.

9176

Analytical Chemistry, Vol. 79, No. 23, December 1, 2007

system. NFPA (nonafluoropentanoic acid) was purchased from Sigma-Aldrich (Milan, Italy). Carnosine (β-alanyl-L-histidine) and the internal standard (IS) H-Tyr-His-OH were a generous gift from Flamma S.p.A. (Chignolo d’Isola, Bergamo, Italy). HNE was prepared from synthesized 4-hydroxy-non-2-enal diethylacetal as previously described36,43 and quantitated by UV spectroscopy (λmax 224 nm;  13.75 × 104 cm-1 M-1). CAR-HNE and DHN-MA adducts were synthesized according to the published methods.36,44 His-HNE was prepared as described for CAR-HNE, using histidine instead of carnosine,36 and His-DHN by NaBH4 reduction of His-HNE.45 At the end of the incubation periods, sample aliquots relative to CAR-HNE and His-HNE reaction mixtures were directly analyzed by HPLC for determination of HNE consumption, according to the method previously described,35 and by ESI-MS (infusion) to confirm adduct formation. The final concentrations of CAR-HNE, His-HNE, and His-DHN were calculated by considering that CAR and His form a 1:1 molar ratio adduct with HNE and by calculating the residual amount of HNE (0.9% for CAR; 1.0% for His). His-HNA was prepared by treatment of His-HNE with sulfamic acid and sodium chlorite as previously described.46 LC-MS/MS Instrument and Conditions. The HPLC system (Surveyor, ThermoFinnigan Italia, Milan, Italy), equipped with a quaternary pump, a Surveyor UV-vis diode array programmable detector 6000 LP, a vacuum degasser, a thermostatted column compartment, and a Surveyor autosampler (200 vials capacity), was used for solvent and sample delivery. Separations were performed on a C12 Phenomenex Sinergy polar-RP column (150 mm × 2 mm i.d.; particle size 4 µm) (Chemtek Analytica, Anzola Emilia, Italy) protected by a polar-RP guard column (4 mm × 2 mm i.d.; 4 µm) kept at 25 °C and equipped with an on-line sample preconcentration C18 cartridge (Opti-lynx, El-Chimie SrL, Cinisello Balsamo, Milan, Italy). Gradient elution was employed from 100% water containing 5 mM nonafluoropentanoic acid (A) to 80% acetonitrile (B) in 24 min at a flow rate of 0.2 mL/min (injection volume 50 µL) followed by a 6 min isocratic elution. The composition of the eluent was then restored to 100% A within 2 min, and the system was reequilibrated for 6 min. The samples rack was maintained at 4 °C. A TSQ Quantum triple-quadrupole mass spectrometer (ThermoFinnigan Italia, Milan, Italy) with electrospray ionization (ESI) source was used for mass detection and analysis. Mass spectrometric analyses were performed in positive ion mode. ESI interface parameters were set as follows: middle position; capillary temperature 270 °C; spray voltage 4.0 kV. Nitrogen was used as nebulizing gas at the following pressure: sheath gas 30 psi; auxiliary gas 5 a.u. MS conditions and tuning were performed by mixing through a T-connection the water-diluted stock solutions of analytes (flow rate 10 µL/min), with the mobile phase maintained at a flow rate of 0.2 mL/min: the intensities of the [M + H]+ ions were monitored and adjusted to the maximum by using the Quantum Tune Master software. (43) Aldini, G.; Granata, P.; Orioli, M.; Santaniello, E.; M. Carini M. J. Mass Spectrom. 2003, 38, 1160-1168. (44) Ahmed, M. U.; Thorpe, S. R.; Baynes, J. W. J. Biol. Chem. 1986, 261, 48894894. (45) Alary, J.; Bravais, F.; Cravedi, J. P.; Debrauwer, L.; Rao, D.; Bories, G. Chem. Res. Toxicol. 1995, 8, 34-39. (46) Aldini, G.; Orioli, M.; Carini, M.; Maffei Facino, R. J. Mass Spectrom. 2004, 39, 1417-1428.

Spectra acquisition for qualitative analysis was based on a survey precursor ion scan for m/z 110.2. The Q1 quadrupole was scanned from m/z 50 to 400 in 1 s (scan time) with resolution of 0.70 m/z, and the precursor ions were fragmented in Q2 using collision potentials from 20 to 40 V (20 V). Finally, Q3 was set to transmit only ions at m/z 110.2 with resolution 0.70 m/z. This scan mode was followed by an enhanced resolution experiment for the ions of interest and then by MS2 acquisitions. MS2 spectra were acquired using the optimized collision energy (35 V). Ionization efficiency in the presence of the mobile phase modifiers was monitored in MRM mode at 2.00 kV multiplier voltage, by selecting for each analyte the MRM transitions of [M + H]+ precursor ions f product ions and optimizing the relative collision energies by the Quantum Tune Master software. The selected MRM transitions were also used for quantitative analysis. The parameters influencing these transitions were optimized as follows: argon gas pressure in the collision Q2, 1.5 mbar; peak full width at half-maximum (fwhm), 0.70 m/z at Q1 and Q3; scan width for all MRM channels, 1 m/z; scan rate (dwell time), 0.2 s/scan. Data processing was performed by the Xcalibur 2.0 version software. The SEQUEST database (Bioworks version 3.2; Thermo Electron San Jose, CA) was used for peptide peak assignments. Animals, Sample Collection, and Preparation. Male Zucker obese (fa/fa) rats (ZO n ) 5; body mass 696.1 g ( 13.8) and lean (LN) littermates (Fa/fa) (n ) 5; body mass 454.5 ( 22.8 g) were purchased at 14 weeks of age from Charles River Italia spa (Calco, Como, Italy). The animals were housed individually in metabolic cages with a light-dark cycle of 12 h each and with free access to food and water. Additional LN rats (n ) 5; body mass 82.4 ( 4.7 g) were purchased at 4 weeks of age and used as controls. ZO rats develop with age a significant proteinuria, mainly due to albumin leakage.15 The animals were maintained in compliance with the policy on animal care expressed in National Research Council guidelines (NRC 1985). Urines from each animal were collected daily at alternate days for 3 days starting from the 24th week of age. One milliliter of a 360 mM BHT ethanolic solution was added to the urine collection tubes that were placed in a unit maintained at 0 °C during collection. The total volume of urine collected in each day was recorded, and 5 mL aliquots were frozen and stored at -80 °C until analysis. Renal disease was assessed by measurements of 24 h urinary albumin and total urinary protein, using the Sentinel 17600 and 17620 commercial kits (Sentinel, Milan, Italy). Urine samples were centrifuged at 18 000 rpm for 10 min; 200 µL of the supernatants was added with 10 µL of internal standard (1 µM dissolved in mobile phase) and immediately ultrafiltered by centrifugation at 15 000 rpm in a Microfuge through a 3 kDa molecular mass cutoff membrane (YM-3000, Millipore, Milan, Italy). The filtrate was transferred to the autosampler vial insert (50 µL injected) for quantitative analysis. For determination of protein-bound adducts (His-HNE), urine aliquots corresponding to 3.0 mg of protein were subjected to hydrolysis (6 N HCl for 20 h at 110 °C in a heating block) as described by Teerlink et al.17 The residue of the final digest was taken up in 100 µL of mobile phase containing the internal standard, filtered through 0.2 µm nylon filters, and the filtrate was injected (80 µL). The internal standard was omitted

in samples (ultrafiltered and hydrolyzed) used for qualitative analysis in precursor ion scanning mode. Preparation of Standards and Quality Control Samples. Stock solutions of 1 µmol/mL DHN-MA, His-DHN, His-HNE, CAR-HNE, and the internal standard stock solution (IS; 0.1 µmol/ mL) were prepared in PBS and stored at 4 °C for 1 week. Stock solutions were diluted further with PBS, and the solutions of each analyte except His-HNE (separate working solutions) were mixed together to obtain working solutions. The 1 nmol/mL working solutions were analyzed by LC-MS/MS to ensure that the concentrations of the original solutions were within the limits of the maximum established error (e3%). Urine samples from 4 week old LN rats were used as blank urine to prepare positive control urine samples. This urine was checked by the applied analytical procedure to ensure it did not contain the selected ALEs above the limit of detection. Calibration samples for adducts in free form were prepared by spiking blank urine samples with each working solution to provide the following final concentrations: 0.075, 0.10, 0.5, 1.0, 2.5, 5.0, 10.0 nmol/mL for DHN-MA, 0.01, 0.05, 0.10, 0.50, 1.00, 2.50, 5.00 nmol/mL for His-DHN and CAR-HNE. IS was added at the concentration of 1 nmol/mL. Quality control (QC) samples at four concentrations, 0.075, 0.5, 2.5, 10.0 nmol/mL for DHN-MA, 0.01, 0.10, 1.00, 5.00 nmol/mL for His-DHN and CAR-HNE, were prepared in the same way, using blank urine spiked with independently prepared stock standard solutions. Each calibration and QC sample was processed as described in sample preparation. Calibration and QC samples for determination of protein-bound His-HNE were prepared by spiking protein hydrolysates (3 mg of BSA) with the His-HNE working solutions (0.01, 0.05, 0.10, 0.50, 1.00, 2.50 e 5.00 nmol/mL). Method Validation. Calibration standards were prepared and analyzed in duplicate in three independent runs. The calibration curves were constructed by weighted (1/x2) least-square linear regression analysis of the peak area ratios of each analyte to the IS against nominal analyte concentration. The lower limit of quantification (LLOQ) was determined as the lowest concentration with values for precision and accuracy within (20% and a signalto-noise (S/N) ratio of the peak areas g10. Intra- and interday precisions and accuracies of the method were determined by assaying five replicates of each of the QC samples (four levels in the low, intermediate, and high concentration range) in three separate analytical runs. Precision and accuracy were determined by calculating the coefficient of variation (CV%) and the relative error (RE%). The absolute recovery of the free analytes after ultrafiltration was determined by comparing the mass spectrometric response of ultrafiltered urine standards to that of ultrafiltered urine blanks spiked with a corresponding set of concentrations (containing 100% of the theoretical concentration) over the entire calibration range. The overall absolute recovery was measured as the ratio of the slopes of the two calibration curves and expressed as percentage. The specificity of the assay was evaluated by comparison of LC-MS/MS chromatograms of the analytes at the LLOQ to those of blank urine samples in triplicate. The stability of all the analytes in stock solutions (4 weeks at 4 °C) and in processed samples, including the resident time (24 h) in the autosampler (4 °C), were determined in triplicate. The Analytical Chemistry, Vol. 79, No. 23, December 1, 2007

9177

mean values of the triplicate samples were compared with those of the initial condition or those of freshly prepared QC samples, and the percentage deviation was calculated. In addition, for HisDHN and His-HNE, three replicates of urine were used for stability studies at two concentration levels (0.05 and 2.50 nmol/ mL) in different conditions: at laboratory room temperature (25 °C), storage temperature (-20 °C), and after freeze-thaw cycles. The stability of His-HNE under hydrolytic conditions was established only indirectly, by adding known amounts of HisHNE (10 nmol/mL) to a BSA solution (3 mg in 100 µL of ammonium acetate buffer pH 5.5). The samples, after protein precipitation, were submitted to acidic hydrolysis and treated as described above. Percentage recoveries were calculated by comparing the mass spectrometric response of BSA spiked with His-HNE hydrolyzed samples to that of hydrolyzed samples spiked with the corresponding concentration (containing 100% of the theoretical concentration). RESULTS AND DISCUSSION Optimization of LC and MS/MS Conditions. In previous studies36 we have developed a sensitive and selective method for the simultaneous determination in rat skeletal muscle of the Michael adducts between HNE and glutathione (GS-HNE) and the endogenous histidine-containing dipeptides carnosine and anserine (CAR-HNE, ANS-HNE), using a gradient elution with HFBA as ion-pairing agent and H-Tyr-His-OH as internal standard for quantitative analysis. H-Tyr-His-OH was maintained as IS as an alternative approach to isotope-labeled analytes because they are not readily and commercially available. In addition, it was found to fulfill some basic criteria: it is not an endogenous dipeptide, it is stable and matches the chromatographic retention, recovery, matrix effects and ionization properties with all the analytes.36,45,47 The same chromatographic conditions were not suited for separation of IS, CAR-HNE, and DHN-MA in the presence of a C18 preconcentration cartridge (used to increase sensitivity), since IS coelutes near the void volume. In addition, the UV-DAD profile of blank urines (data not shown) indicates that the bulk of endogenous interfering material eluted within 9 and 11 min and coeluted with all the other analytes. Because these conditions caused loss of sensitivity by reduced ionization as a result of ion suppression, the gradient elution was suitably modified, by increasing the polarity of solvent A, the chromatographic run, and by replacing HFBA (0.1%) with NFPA (5 mM; 0.08%),17 to increase retention on the reversed phase of very hydrophilic compounds (IS) and to shift beyond the matrix components the more lipophilic analytes (DHN-MA, CARHNE). NFPA was already successfully employed as ion-pairing agent for analysis of CML in human plasma protein by tandem mass spectrometry.17 Hence, we studied the effect of NFPA on the ionization efficiency of all the other adducts and compared it to that of HFBA, previously used for CAR-HNE,36 and to that of HCOOH, used for DHN-MA.48,49 The results (Supporting Infor(47) Liang, Y.; Xie, L.; Liu, X. D.; Lu, T.; Wang, G. J.; Hu, Y. Z. J. Pharm. Biomed. Anal. 2005, 39, 1031-1035. (48) Rathahao, E.; Peiro, G.; Martins, N.; Alary, J.; Gue´raud, F.; Debrauwer, L. Anal. Bioanal. Chem. 2005, 381, 1532-1539. (49) Peiro, G.; Alary, J.; Cravedi, J. P.; Rathahao, E.; Steghens, J. P.; Gue´raud, F. Biofactors 2005, 24, 89-96.

9178

Analytical Chemistry, Vol. 79, No. 23, December 1, 2007

mation Figure S-1) indicate that NFPA significantly suppressed, in respect to HCOOH, the extracted ion current for all the analytes (monitored in MRM mode), except for DHN-MA, showing a poor ionization efficiency in all experimental conditions. Anyway, this suppressing effect was less than that induced by HFBA: a 40% and a 15% increase was in fact observed for DHN-MA and CARHNE, respectively. Overall, these results led us to replace HFBA with NFPA as the ion-pairing agent. Infusion experiments were performed in order to obtain the full mass spectra of DHN-MA, CAR-HNE, and IS, showing the [M + H]+ at m/z 322, 383, and 319, respectively, and to establish the optimal fragmentation reactions for MRM. Under the optimized LC and MS conditions, no interferences from endogenous compounds were found in 4 week old LN urines, and the retention times for the analytes were 24.80 min (DHNMA), 27.84 min (CAR-HNE), and 23.16 min (IS). Preliminary LC-MS/MS profiles, obtained by spiking LN urines with known amounts of the analytes (data not shown), indicate that they can be selectively detected and quantitated, even if they are simultaneously present, by the following precursor-product ion combinations:

m/z 322.2 f 130.1 + 164.1

(collision energy 15 eV) DHN-MA

m/z 383.1 f 110.1 + 266.2

(collision energy 40 eV) CAR-HNE

m/z 319.2 f 156.5 + 301.6

(collision energy 25 eV) H-Tyr-His-OH (IS)

Precursor Ion Monitoring for Qualitative Analysis of Zucker Rat Urines. The first step was to apply to urine from ZO rats the precursor ion scanning technique, a valuable tool for the rapid confirmation of targeted compounds or for the nontargeted detection of compounds bearing a common moiety. We used the immonium ion of histidine (m/z 110) as a well-known and specific fragment ion of His-containing peptides36 to generate precursor ion spectra in order to screen ZO urines for the presence of covalently modified His residues, both in free (analysis of ultrafiltrates) and protein-bound (analysis of protein hydrolysates) form. The precursor ion mass spectra were evaluated in order to identify the m/z of the precursor ions. Figure 1 reports the bidimensional plot (m/z values against retention times) of urine ultrafiltrates from 4 week old LN (panel a), 6 month old LN (panel b), and ZO (panel c) rats. This latter only contains three distinct [M + H]+ signals at m/z 328, 314, and 383, with retention times of 16.71 min, 25.23, and 27.84 min, respectively, two of them being detectable also in urines from 6 month old LN rats. Retention times and molecular weights were confirmed by the extracted ion currents corresponding to the identified precursor ions (Figure 2). Each [M + H]+ ion of interest was then selected for MS2 acquisitions obtained by applying the optimized collision energy for each species, in order to obtain structural characterization (Figure 3). In detail, the MS2 of the [M + H]+ at m/z 383 (Figure 3a) contains two main product ions at m/z 266.2 and 110.2. The first is diagnostic for HNE-modified His residues (immonium ion),36,50,51 and the second (His immonium ion) is due to a retro-

Figure 3. LC-MS/MS analysis of urine ultrafiltrates from ZO rats: MS2 spectra of the precursor ions at (a) m/z 383.1; (b) m/z 314.7; (c) m/z 328.2. Figure 1. LC-ESI-MS/MS analysis of urine ultrafiltrates (free adducts). Data were acquired in precursor ion scanning mode (product ion m/z 110) and reported as two-dimensional plots (m/z values against retention times): (a) 4 week old LN; (b) 6 month old LN; (c) ZO rats.

Figure 2. Extracted ion chromatograms (SICs) of the identified precursor ions in urines ultrafiltrates from ZO rats: (a) m/z 328.2; (b) m/z 314.7; (c) m/z 383.1.

Michael reaction from the ion at m/z 266. The same retro-Michael reaction occurs also on the [M + H]+ at m/z 383 (neutral loss of the HNE moiety, 156 Da) and formation of the product ion at m/z

227, corresponding to the molecular mass of the peptide moiety (carnosine). All these data (MW, retention time, and MS2) allowed assigning the peak with retention time of 27.8 min to the structure of CAR-HNE, previously identified in rat skeletal muscle as a marker of carbonylation36 but never before identified in urines. The same approach was applied to the other two peaks recognized by the precursor ion scan experiments. The MS2 spectrum of the [M + H]+ species at m/z 314 (Figure 3b) is dominated, besides the common ion at m/z 110.2, by a main product ion at m/z 268.2 (+2 Da with respect to CAR-HNE), corresponding to the immonium ion of 1,4-dihydroxynonanemodified His residue. Hence, this adduct was preliminarily identified as His-1,4-dihydroxynonane (His-DHN), arising by HNE Michael adduction to His followed by metabolic conversion of the aldehyde into the alcohol function. Finally, the MS2 spectrum relative to the [M + H]+ species at m/z 328.2 (Figure 3c) is characterized by the product ion at m/z 282.1 [M + H - HCOOH]+, the typical immonium ion of HNEmodified His residues increased by 16 Da (m/z 266.2 in CARHNE adduct). This, indicating aldehyde metabolic oxidation to the corresponding carboxylic derivative, allowed assigning to the [M + H]+ adduct at m/z 328.2 the structure of His-4-hydroxynonanoic acid (His-HNA). Protein hydrolysates from LN and ZO urine were analyzed by the same methodology to recognize in the biological matrix (50) Bolgar, M. S.; Gaskell, S. J. Anal. Chem. 1996, 68, 2325-2330. (51) Fenaille, F.; Guy, P. A.; Tabet, J. C. J. Am. Soc. Mass Spectrom. 2003, 14, 215-226.

Analytical Chemistry, Vol. 79, No. 23, December 1, 2007

9179

Figure 5. LC-MS/MS analysis of protein hydrolysates from ZO rats: MS2 spectra of the precursor ions at m/z 269.4 (a) and m/z 312.2 (b).

Figure 4. Protein-bound covalently modified His residues: LCESI-MS/MS analysis of protein hydrolysates from lean and Zucker obese rats. Panels a and b report the total ion current (TIC) profiles of the precursor ion scanning of m/z 110 in LN and ZO urine, respectively; panels c and d show the SICs of the identified precursor ions in protein hydrolysates from ZO rats at m/z 269.4 (#) and 312.2 (/).

protein-bound covalently modified His residues. Figure 4a (LN) and Figure 4b (ZO) report the total ion current (TIC) profiles of the precursor ions which give a product ion of m/z 110. The abundant peak in lean (retention time 21.24 min; [M + H]+ at m/z 255), which almost disappeared in Zucker urine, was not considered as a potential biomarker of carbonylation, because its MS2 spectrum (data not shown) indicates Val-His (Bioworks analysis). Two chromatographic peaks were significantly elevated in ZO rats (retention times 23.41 and 26.12 min). The corresponding mass spectra indicate the m/z of the precursor ions at 269 and 312, respectively, further confirmed by the extracted ion traces reported in Figure 4, parts c and d. The MS2 spectrum of the [M + H]+ at m/z 269.4 contains several significant product ions only at 25 eV collision energy (Figure 5a): the ion at m/z 227.3 [M + H - CH2dCO]+ typical of N-acetylated derivatives, and the ions at m/z 209.2/210.2, 156.0, and 110.2, which are characteristic of the fragmentation pattern of carnosine.36 Hence, MW and fragmentation pattern indicate for the peak eluting at 23.41 min the structure of N-acetylcarnosine, 9180 Analytical Chemistry, Vol. 79, No. 23, December 1, 2007

which was confirmed by comparison with the retention time of the authentic standard treated under the same hydrolytic conditions. These results seem to indicate protein carnosinylation, a hypothesis already postulated by Hipkiss et al. who, using model systems in vitro, demonstrated that carnosine, by reacting with protein carbonyls with generation of “carnosinylated“ polypeptides, may intracellularly suppress the deleterious effects of protein carbonyls.52 It has been speculated that carnosinylation may facilitate inactivation/disposal of carbonylated proteins through the formation of inert lipofuscin, proteolysis via the proteasome system, and exocytosis following interaction with receptors.53,54 Because this proposal pointing to a hitherto unrecognized mechanism by which cells normally defend themselves against protein carbonyls has never been supported by experimental data in vivo, our preliminary results, if confirmed on a larger sample of animals and/or other experimental models, may represent the first evidence for protein carnosinylation. The MS2 spectrum of the [M + H]+ species at m/z 312.2 (Figure 5b), 2 Da over that of the His-DHN adduct (dehydrogenation), contains the diagnostic product ion at m/z 266, corresponding, as previously described for CAR-HNE, to the immonium ion of HNE-modified His. Another significant product ion, although of lower relative abundance, is that at m/z 156 [His + H]+, arising from the [M + H]+ by neutral loss of the HNE moiety (typical retro-Michael reaction). All this information indicates for the precursor ion at m/z 312.2 the structure of the His-HNE Michael adduct. Identity of all the adducts (free and protein-bound) was finally supported by chemical synthesis and by the match of the (52) Hipkiss, A. R.; Brownson, C.; Bertani, M. F.; Ruiz, E.; Ferro, A. Ann. N.Y. Acad. Sci. 2002, 959, 285-294. (53) Hipkiss, A. R.; Brownson, C.; Carrier, M. J. Biogerontology 2000, 1, 217223. (54) Quinn, P. J.; Boldyrev, A. A. Biochemistry (Moscow, Russ. Fed.) 2000, 65, 771-778.

Standard curves for the analytes constructed on different working days showed good linearity over the entire calibration ranges (0.075-10 nmol/mL for DHN-MA; 0.01-5.00 nmol/mL for all the other analytes), with coefficients of correlation (r2) greater than 0.998. The equations for the calibration lines were

y ) 0.3808((0.005480)x - 0.05222((0.02384) DHN-MA y ) 2.963((0.01589)x - 0.02089((0.03204) His-DHN y ) 1.433((0.02425)x + 0.00367((0.00452) His-HNE y ) 1.435((0.02115)x - 0.01559((0.04265) CAR-HNE

Figure 6. LC-MS/MS profiles (MRM traces) of blank (bold lines) and spiked (thin lines) urine with (a) IS (1 nmol/mL) and the analytes at LLOQs: (b) DHN-MA (0.075 nmol/mL), (c) His-DHN, (d) HisHNE, and (e) CAR-HNE (0.01 nmol/mL).

chromatographic and mass spectrometric properties. Presently, the His-HNA adduct is available in minimal amount (due to the low yields of reaction), just enough for structure confirmation, but not sufficient to develop and validate a quantitative method. Hence, further infusion experiments were performed to evaluate the ionization efficiency of His-DHN and His-HNE (results in Supporting Information Figure S-1) and to optimize their fragmentation reactions for MRM analysis. The following precursor-product ion combinations were used for quantitative analysis:

m/z 314.7 f 268.2 + 110.1

(collision energy 26 eV) His-DHN

m/z 312.2 f 110.1 + 156.0

(collision energy 30 eV) His-HNE

Method Validation. The LC-MS/MS approach was first applied to urine from 1 month old LN rats to ensure their use as a blank matrix for calibration: none of the analytes were detected in the MRM mode (bold lines in Figure 6). A full validation procedure (specificity, linearity, LOD, LOQ, intra- and interday precision and accuracy, recovery, and stability) was then performed on urine from young LN rats (free adducts) or on a reconstituted protein matrix (protein-bound adducts).

The LLOQs were 0.075 nmol/mL for DHN-MA, 0.01 nmol/mL for His-DHN, His-HNE, and CAR-HNE; the lower LOQ found for DHN-MA with respect to that described by Rathahao et al.48 may be explained on the basis of the decreased ionization efficiency induced by NFPA with respect to HCOOH. Representative MRMs traces relative to all the adducts and IS (1 nmol/mL) obtained from blank urine spiked with LLOQ concentrations (Figure 6) indicate that they can be selectively detected and quantitated, even if they are simultaneously present, by the above-reported precursor-product ion combinations. The specificity of the method was demonstrated by the absence of interfering peaks in the mass and time ranges for the ALEs and IS in individual urine samples from five animals (triplicate injection; free analytes) or in reconstituted protein matrix (protein-bound His-HNE). The intra- and interassay precision and accuracy of the method were determined on QC samples prepared separately from calibration standards, by analyzing five replicates at four concentration levels, and the data are reported in Supporting Information Tables S-1 and S-2. The intraday precision (CV%) was less than 3.98% (e8.72% at the LLOQs), and accuracy ranged from -4.40% to +11.00% of nominal concentrations (within (16% at the LLOQs); the interday CV values were less than 3.84% (e9.22% at the LLOQs), and accuracy was in the range of -3.79% to +6.51% (within (10.25% at the LLOQs). The extraction recoveries for free adducts were determined by comparing the peak area ratio of ultrafiltered urine standards at three concentration levels (low, intermediate, and high) to those of post-ultrafiltered blank urines spiked with the corresponding concentrations. The mean extraction recovery was satisfactory, being 96.5%, 92.7%, and 101.5% for DHN-MA, His-DHN, and CAR-HNE, respectively. Recovery of His-HNE after the hydrolytic procedure was only 55%, and studies are in progress to optimize this step. The high stability of DHN-MA and CAR-HNE in aqueous solutions, urines, or tissue extracts at room and lower temperatures has already been widely documented.36,49 We therefore checked their stability in QC samples and in the spiked urine ultrafiltrates in the experimental conditions used. The stability was guaranteed for at least 24 h at 4 °C, and no significant differences (t test) were found between freshly prepared urine ultrafiltrates and ultrafiltrates prepared from urine stored in liquid N2 (data Analytical Chemistry, Vol. 79, No. 23, December 1, 2007

9181

Table 1. Renal Function and Urinary Levels of ALEs in Lean and Zucker Obese Ratsa lean

Zucker

14.5 ( 1.1

20.9 ( 2.4c

total protein mg/mL mg/24 h

6.45 ( 1.20 93.52 ( 17.4

30.7 ( 6.0c 641.6 ( 125.4c

albumin mg/mL mg/24 h

2.75 ( 1.4 39.9 ( 20.3

27.2 ( 2.8c 568.5 ( 58.5c

DHN-MA nmol/mL nmol/24 h

1.116 ( 0.058 16.182 ( 0.841

2.868 ( 0.174c 59.941 ( 3.637c

His-DHN nmol/mL nmol/24 h

0.061 ( 0.005 0.884 ( 0.073

0.184 ( 0.015c 3.846 ( 0.314c

CAR-HNE nmol/mL nmol/24 h

0.053 ( 0.008 0.768 ( 0.116

0.097 ( 0.008b 2.027 ( 0.167c

0.112 ( 0.009 10.474 ( 0. 842

3.160 ( 0.628b 2027.4 ( 402.9c

urinary volume (mL/24 h)

His-HNE (protein-bound) nmol/mg protein nmol/24 h

a Urinary levels were determined in every 24 h urine fraction (n ) 3) of five animals. Values are the mean ( SD relative to five animals. b p < 0.005 vs lean (t test). c p < 0.001 vs lean (t test).

not shown). Although His-DHN has been shown to be highly stable under conditions required to release HNE-histidine Michael adduct from proteins (strong acid hydrolysis),24,55 no stability data in urines are available from the literature, or for HisHNE under the same conditions. Stability data for His-HNE and His-DHN, summarized in Supporting Information Table S-3, indicated that the analytes was stable for at least the length of time under different conditions listed, all the values ranging from 92.54% to 102.01% of the initial value, except for the second freezethaw cycle, where a 21-28% decrease was observed for both the analytes. Stock solutions, when stored in PBS at a nominal temperature of 4 °C, were stable for at least 4 weeks. Renal Function and Profiling of ALEs in Zucker Rats. Total urinary protein and albumin were measured to evaluate renal function. Urinary volumes were 14.5 ( 1.1 mL/24 h and 20.9 ( 2.4 mL/24 h, respectively, in LN and ZO (p < 0.001 between groups). A 10-fold increase in urinary albumin was observed in ZO versus LN rats, which was paralleled by a similar increase in total urinary protein (Table 1). The HPLC-MS/MS method was then applied to profile the ALEs content in rat urines. The results, summarized in Table 1, indicate that all the adducts found in free form can be determined in lean animals. As reasonably expected, DHN-MA, the main indicator of HNE formation in vivo, represents the most abundant excretion product. When the data are expressed as nmol/mL urine, a significant increase (2-3-fold; p < 0.005 and p < 0.001) in the urinary excretion of all the adducts was observed in Zucker rats. If we consider the urinary volumes and the total daily excretion, these increments become much more consistent, ranging from 2- to 4-fold increase for all the adducts. The pattern of DHN-MA excretion, the main indicator of HNE formation in (55) Hashimoto, M.; Sibata, T.; Wasada, H.; Toyokuni, S.; Uchida, K. J. Biol. Chem. 2003, 278, 5044-5051.

9182 Analytical Chemistry, Vol. 79, No. 23, December 1, 2007

vivo, has never been described in ZO rats but substantially confirms previous findings in different experimental models and conditions.48,49 Also the urinary levels of protein-bound adducts (His-HNE) were significantly elevated in ZO rats, to reach values (when expressed as total daily excretion tacking into account the protein urinary content) greater than 2 µmol. CONCLUSIONS The LC-MS/MS approach developed in the first part of this study, based on the precursor ion scanning technique, was applied for the first time for the rapid detection and identification in biological matrices of new biomarkers of carbonylation in obese Zucker rats, an experimental model where carbonyl stress plays a key role in the development of renal and cardiac dysfunction.15 Our interest was focused to recognize covalent modifications of His and His-containing peptides by RCS (especially in free form), in view of the lack of such information. Although His-HNE protein adducts have been immunohistochemically detected in several biological systems, experimental animal models and in humans,25-29 no studies report free His modified by RCS as a tool to detect involvement of carbonyl stress in different pathological conditions. Our approach allowed identifying for the first time in urine from ZO rats two novel adducts, both originating from the conjugation of HNE to His, which are excreted in free form only: His-DHN, arising from the subsequent reduction of the aldehyde by a member of aldo-keto reductase superfamily, and His-HNA, arising from the conjugation of HNE followed by the oxidation of the aldehyde by aldehyde dehydrogenase.56 No free His-HNE was found in urines, but it was detected and quantified only in protein hydrolysates. Very likely albumin, which accounts for more than 90% of total excreted proteins, is the target protein for HNE. This assumption is supported by the already demonstrated reactivity of some His residues on human serum albumin (HSA), namely, His242 (located in a fatty acid and drug binding cavity in HSA subdomain IIa), according to Szapacs et al.,57 and His146, according to Aldini et al.58 Studies are in progress to identify His residues modified by HNE in albumin isolated from ZO rat urine, using the LC-MS/MS approach recently developed to characterize the covalent modifications of HSA by HNE.58 Presently, we do not know (for the lack of a suitable standard) the stability of His-HNE bound to proteins during the hydrolytic step, but the high levels found seem to indicate it to be stable enough. This is substantiated by the fact that HNE has been shown to react with His to form a stable Michael addition-type adduct in the cyclic hemiacetal structure, which is stabilized toward retro-Michael reaction in acidic conditions. Stability of the adduct, readily isolable, is due to the poor leaving group ability of imidazole.55 Even considering the 50% loss during the experimental procedure (determined only indirectly, working on HisHNE added to a reconstituted protein mixture), the results of the quantitative analysis indicate a massive formation of HNE-protein adducts also in obese animals. (56) Alary, J.; Fernandez, Y.; Debrauwer, L.; Perdu, E.; Gueraud, F. Chem. Res. Toxicol. 2003, 16, 320-327. (57) Szapacs, M. E.; Riggins, J. N.; Zimmerman, L. J.; Liebler, D. C. Biochemistry 2006, 45, 10521-10528. (58) Aldini, G.; Gamberoni, L.; Orioli, M.; Beretta, G.; Regazzoni, L.; Maffei, Facino, R.; Carini, M. J. Mass Spectrom. 2006, 41, 1149-1161.

Several studies on HNE metabolism in vitro (using rat hepatic subcellular fractions or hepatocytes) and in vivo demonstrated that HNE undergo phase I metabolism (reduction/oxidation of the aldehydes), with formation of 1,4-dihydroxy-trans-nonene (DHN) and 4-hydroxy-trans-2-nonenoic acid (HNA), and phase II conjugation (GSH) and subsequent mercapturic acid formation. This latter already has been recognized for a long time as the mostefficientdetoxificationsystemtowardelectrophiliccompounds,45,56,59-63 DHN-MA being the most important excretion product. The high urinary levels of DHN-MA found in LN and ZO rats confirm that the GSH conjugative pathway plays a key role also in this animal model. The results of this study unequivocally indicate that HNE, in addition to GSH-dependent metabolism, undergoes histidinedependent metabolism, following the same detoxification pathways involving reductive/oxidative metabolic conversion. This in turn implies that His-HNE, as HNE itself or its conjugation product with GSH, is a good substrate for both dehydrogenases and reductases. The same MS/MS approach in precursor ion mode allowed also identifying for the first time in urine the Michael adduct between HNE and the endogenous dipeptide carnosine. This confirms our previous in vitro studies performed on rat skeletal muscle,36 i.e., that CAR-HNE is a stable and reliable biomarker of lipid peroxidation and carbonylation in vivo. Unlike His-HNE, CAR-HNE does not undergo further metabolism: no CAR-DHN or CAR-HNA was in fact detected in LN or ZO rat urines, thus to indicate that CAR-HNE is a poor substrate for aldehyde dehydrogenases/reductases. In this context, it must be stressed that this new histidinedependent detoxification pathway of HNE could be particularly relevant in those pathological conditions (included those induced by xenobiotics) involving a massive depletion of the endogenous pool of GSH, which represents the first line of defense toward HNE. This assumption has been already demonstrated in HNEspiked rat skeletal muscle.36 By LC-MS/MS monitoring of GSHHNE, CAR-HNE, and ANS-HNE adducts, conjugation with GSH was confirmed to play a key role in HNE detoxification. Coincubation with the alkylating agent N-ethylmaleimide (to mimic an abrupt decay in GSH pool) resulted in a drastic fall of GSH-HNE levels, with a concomitant significant elevation in the levels of both CAR-HNE and ANS-HNE. Hence, inhibition of the thioldependent pathway shifts HNE detoxification toward the histidinedependent pathway. The decline of cellular thiol reserve, as a consequence of the increased oxidative stress, has been shown to occur also in obese Zucker rats, where a 40% decrease in hepatic GSH levels in respect to the control lean animals was found.64 This could explain the higher levels of HNE adducts to His and His-peptides in obese animals than in the controls. Evidence of obesity-induced oxidative (59) Esterbauer, H.; Zollner, H.; Lang, J. Biochem. J. 1985, 228, 363-373. (60) Mitchell, D. Y.; Petersen, D. R. Toxicol. Appl. Pharmacol. 1987, 87, 403410. (61) Siems, W. G.; Grune, T.; Beierl, B.; Zollner, H.; Esterbauer, H. EXS 1992, 62, 124-135. (62) Siems, W. G.; Zollner, H.; Grune, T.; Esterbauer, H. J. Lipid Res. 1997, 38, 312-322. (63) de Zwart, L. L.; Hermanns, R. C.; Meerman, J. H.; Commandeur, J. N.; Vermeulen, N. P. Xenobiotica 1996, 26, 1087-1100. (64) Chang, S. P.; Chen, Y. H.; Chang, W. C.; Liu, I. M.; Cheng, J. T. Clin. Exp. Pharmacol. Physiol. 2004, 31, 506-511.

stress and inadequate antioxidant defenses in humans has been accumulating over the past few years, and oxidative stress biomarkers have been found to be elevated in several tissues, including skeletal muscle, plasma, and erythrocytes.65 The occurrence of the alternative mechanism of detoxification of unsaturated aldehyde via His residues in obese human subjects remains, however, to be established. Also the LC-MS/MS method developed in the second phase of this study for the quantitative determination of HNE adducts to His residues is novel since it permits, in a single chromatographic run, the simultaneous determination of adducts that, although structurally unrelated, have lipid peroxidation as a common origin. Our method was not optimized to discriminate between diastereomers (it is well-known, for example, that HisHNE adduct, possessing three chiral centers in the cyclic hemiacetal structure, gives rise to four diastereomers),55 because the characterization of the possible configurational isomers of each adduct and the knowledge of their relative amounts in vivo was beyond the aim of this work. The method has been validated not only for the newly identified His adducts but also for DHN-MA, the most widely used biomarker of carbonyl stress in vivo. DHNMA, a physiological component of rat and human urine, is to be considered as the main metabolic end product of HNE, and this is confirmed by the high urinary levels found in both LN and ZO rats, never reported before. From an analytical point of view, the LC-MS/MS method requires minimal sample manipulation, including a very simple ultrafiltration step without further purification. This guarantees high recovery rates for analytes bearing different chemical structures and polarity, as well as high reproducibility, and compensates the relatively long chromatographic run. Highly satisfactory detection limits have been obtained for all the HNE adducts to His (0.01 nmol/mL). By contrast, the LOQ for DHNMA was found to be slightly greater than that recently reported by Rathahao et al.48 and Peiro et al.49 using an ion trap instead of a triple-quadrupole MS system (performing quantification by MS3 acquisitions), and formic acid instead of NFPA as modifier, the main factor responsible for the decreased ionization efficiency. Also this apparent drawback is compensated by the high excretion levels of DHN-MA and by the suitability of the method for the simultaneous determination of more than one analyte. From a biological point of view, it is important to stress that this analytical method, specifically developed to profile in urines His residues modified by RCS, allowed reporting for the first time the excretion profile of some new ALEs (and DHN-MA) in ZO rats. In previous studies carried out on this animal model, only few markers of oxidative stress in plasma (lipid hydroperoxides and 8-epi-PGF2R)66 and carbonyl stress in renal tissue (HNE protein adducts, using a specific rabbit polyclonal antibody to HNE adducts) have been reported.67 Our results confirm HNE adduction to proteins in Zucker rats and should represent a useful tool to better understand the biochemical role of the newly identified biomarkers in different physiopathological conditions. (65) Vincent, H. K.; Taylor, A. G. Int. J. Obes. 2006, 30, 400-418. (66) Rosen, P.; Osmers, A. Horm. Metab. Res. 2006, 38, 575-586. (67) Dominguez, J. H.; Wu, P.; Hawes, J. W.; Deeg, M.; Walsh, J.; Packer, S. C.; Nagase, M.; Temm, C.; Goss, E.; Peterson, R. Kidney Int. 2006, 69, 19691976.

Analytical Chemistry, Vol. 79, No. 23, December 1, 2007

9183

Moreover, the results of this study show that ALEs excretion, being negligible in young LN rats, becomes evident in adult LN animals and significantly increased in the pathological model (ZO rats). This from one side confirms, working on a simple biological matrix (noninvasive withdrawal and easy to handle), that ALEs formation takes place in physiological conditions and is strictly associated with aging.2,68 On the other hand, these results clearly indicate that the ZO rat represents an animal model where carbonyl stress plays a key role in the development of long-term complications of the metabolic syndrome. This makes the ZO rat a highly suitable model for testing in vivo the carbonyl quenching ability of newly developed RCS sequestering agents, by measuring (68) Hamelin, M.; Borot-Laloi, C.; Friguet, B.; Bakala, H. Arch. Biochem. Biophys. 2003, 411, 215-222.

9184

Analytical Chemistry, Vol. 79, No. 23, December 1, 2007

the urinary levels of the highly stable biomarkers considered in this study. ACKNOWLEDGMENT This study was supported by COFIN2004 (Cofinanziamento Programma Nazionale 2004, Ministero dell’Istruzione, dell’Universita` e della Ricerca) and FIRST 2006-2007 (Fondo Interno Ricerca Scientifica e Tecnologica, University of Milan). SUPPORTING INFORMATION AVAILABLE Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org. Received for review July 31, 2007. Accepted September 27, 2007. AC7016184