New Method for the Quantitative Determination of Major Protein

Jul 6, 2006 - Agricultural Science, Tohoku UniVersity, Sendai 981-8555, Japan, Department of Applied Molecular Bioscience,. DiVision of Biofunctional ...
1 downloads 3 Views 209KB Size
Download PDF
Chem. Res. Toxicol. 2006, 19, 1059-1065

1059

New Method for the Quantitative Determination of Major Protein Carbonyls, r-Aminoadipic and γ-Glutamic Semialdehydes: Investigation of the Formation Mechanism and Chemical Nature in Vitro and in Vivo Mitsugu Akagawa,*,† Daisuke Sasaki,‡ Yoshihisa Ishii,§ Yayoi Kurota,‡ Mari Yotsu-Yamashita,‡ Koji Uchida,§ and Kyozo Suyama‡,| Department of Biological Chemistry, DiVision of Applied Life Science, Graduate School of Life and EnVironmental Sciences, Osaka Prefecture UniVersity, Sakai 599-8531, Japan, Department of Applied Bioorganic Chemistry, DiVision of Bioscience and Biotechnology for Future Bioindustries, Graduate School of Agricultural Science, Tohoku UniVersity, Sendai 981-8555, Japan, Department of Applied Molecular Bioscience, DiVision of Biofunctional Chemistry, Graduate School of Bioagricultural Sciences, Nagoya UniVersity, Nagoya 464-8601, Japan, and Department of Sports Dietetics, Sendai UniVersity, Shibata-gun, Miyagi 989-1693, Japan ReceiVed February 9, 2006

R-Aminoadipic semialdehyde (AAS) and γ-glutamic semialdehyde (GGS) are identified as the major carbonyl products in oxidized proteins. To elucidate the formation pathway of AAS and GGS in vivo, we developed and validated a new quantification method. AAS and GGS in proteins were derivatized by reductive amination with NaCNBH3 and p-aminobenzoic acid, a fluorescent reagent, followed by acid hydrolysis. It is noteworthy that the fluorescent derivatives were completely stable during acid hydrolysis. The present method permitted the specific, accurate, and sensitive quantification of both semialdehydes by fluorometric high-performance liquid chromatography. Analysis of proteins oxidized by various oxidation systems revealed that AAS and GGS are notably generated by the reaction of proteins with • OH, which is produced by metal-catalyzed oxidation (MCO). Furthermore, exposure of transferrin and human plasma to ascorbic acid and H2O2 significantly promoted the formation of AAS and GGS in vitro, suggesting that both semialdehydes can be generated by MCO in vivo. We also demonstrated their generation through oxidative stress induced by acute iron overload in vivo. In this paper, we describe this analytical technique for simple and precise measurement of AAS and GGS and discuss their formation mechanism in vivo. Introduction Oxidative damage to proteins by reactive oxygen species can result in cleavage of the polypeptide backbone, cross-linking, and modification of the side chains of amino acids (1, 2). There is now a large body of evidence suggesting that protein oxidation plays a major role in a number of human diseases and aging (3-5). Among a wide variety of protein oxidations, introduction of carbonyl groups into amino acid residues is a hallmark for oxidative damage to proteins (6, 7). Carbonyl groups are introduced in proteins by a variety of modification pathways in vivo and in vitro, particularly metal-catalyzed oxidation (MCO)1 of specific amino acid residues (2, 3, 5) and also adduction of carbonyl-containing peroxidized lipids (4-hydroxynonenal, malondialdehyde, acrolein, etc.) (8-11) or carbohydrate (i.e., * To whom correspondence should be addressed. Fax: 81-72-254-9921. E-mail: [email protected]. † Osaka Prefecture University. ‡ Tohoku University. § Nagoya University. | Sendai University. 1 Abbreviations: AAS, R-aminoadipic semialdehyde; ABA, p-aminobenzoic acid; Asc, ascorbic acid; DNPH, 2,4-dinitrophynilhydrazine; DTPA, diethylenetriamine-N,N,N′,N′′,N′′-pentaacetic acid; Fe3+-NTA, ferric nitrilotriacetate; GGS, γ-glutamic semialdehyde; HACA, hydroxyaminocaproic acid; HAVA, hydroxyaminovaleric acid; HPLC, high-performance liquid chromatography; MCO, metal-catalyzed oxidation; MES, 2-(Nmorpholino)ethanesulfonic acid; PBS, phosphate-buffered saline; SDS, sodium dodecyl sulfate.

glycation) (12). Carbonyl derivatives can be sensitively measured by convenient methods using 2,4-dinitrophynilhydrazine (DNPH), which reacts with carbonyl groups to generate dinitrophenylhydrazones with characteristic absorbance maxima (13). Using these methods, it has been confirmed that carbonyl derivatives accumulate on tissue proteins during aging and disease development such as Alzheimer’s disease (14, 15), rheumatoid arthritis (16), amyotrophic lateral sclerosis (17), diabetes mellitus (18), and Parkinson’s disease (14). Nevertheless, the methods are limited to measuring total carbonyl derivatives formed by unspecific various pathways, and information on chemical structures and formation mechanisms is not provided. Therefore, more specific methods are required for the determination of carbonyl derivatives to understand their chemical nature, oxidation pathway, and distribution level in vivo. Recently, Requena et al. developed methods based on gas chromatography/mass spectrometry with isotopic dilution for the determination of major carbonyl derivatives, R-aminoadipic semialdehyde (AAS) and γ-glutamic semialdehyde (GGS), after their reduction with sodium borohydride (NaBH4) (19). AAS is an oxidation product of lysine residue by MCO systems, while GGS originates from arginine and proline (Figure 1). GGS and AAS in proteins are destroyed during acid hydrolysis and, therefore, need to be stabilized by derivatization. Reduction of GGS and AAS by NaBH4 produces their corresponding hydroxy amino acids, hydroxyaminovaleric acid (HAVA) and hy-

10.1021/tx060026p CCC: $33.50 © 2006 American Chemical Society Published on Web 07/06/2006

1060 Chem. Res. Toxicol., Vol. 19, No. 8, 2006

Akagawa et al.

Figure 1. Reaction scheme for the derivatization of AAS and GGS in protein by reductive amination with ABA and NaCNBH3.

droxyaminocaproic acid (HACA), respectively. Nevertheless, HAVA is extensively converted to chloro-aminovaleric acid during acid hydrolysis with HCl, while HAVA is also converted to chloro-aminocaproic acid and partially reverted to proline (19-21). Therefore, the determination of AAS and GGS has been a long-standing major difficulty. Accordingly, the use of deuterated internal standard of HACA and HAVA is essential for the determination of HACA and HAVA to eliminate the variability introduced during hydrolysis by side reactions. Another derivatizing method based on reductive amination of AAS and GGS with fluoresceinamine and sodium cyanoborohydride (NaCNBH3) has also been reported (22). The resulting fluoresceinamine derivatives are degraded to nonfluorescent decarboxylated derivatives by acid hydrolysis. The measurements of AAS and GGS should provide important insights into the mechanism of protein oxidation in aging and disease. However, the details of the formation pathway that lead to formation of AAS and GGS in vivo remain unknown. To elucidate the mechanism underlying the formation of AAS and GGS in vivo, we developed and validated a more simple and precise method for the determination of AAS and GGS. In the present study, we investigated the production of AAS and GGS in model proteins and human plasma exposed to various oxidation systems by the present method. Moreover, to assess their formation by oxidative stress in vivo, we examined the endogenous generation in an experimental acute iron overload model. In this paper, we describe a new analytical technique for a simple and precise measurement of AAS and GGS in proteins and discuss their formation mechanism and chemical nature in vivo. A preliminary account of this work was presented at the Eighth International Symposium on the Maillard reaction (23).

Experimental Procedures Materials. Acetonitrile was of high-performance liquid chromatography (HPLC) grade from Kanto Chemicals (Tokyo, Japan). NR-Acetyl-L-lysine and KO2 were purchased from Sigma-Aldrich (St. Louis, MO). NR-Acetyl-L-ornithine was from Tokyo Kasei (Tokyo, Japan). Diethylenetriamine-N,N,N′,N′′,N′′-pentaacetic acid (DTPA) was from Wako Pure Chemical Industries (Tokyo, Japan). Human holo-transferrin was from ICN Biomedicals (Irvine, CA). H2O2 (30% aqueous solution) was purchased from Santoku Chemical Industries (Tokyo, Japan). All other chemicals were of analytical grade from Nacalai Tesque (Kyoto, Japan). Egg shell membrane isolated from fresh white leghorn hen eggs was washed thoroughly with distilled water and cut into small pieces (5 mm × 5 mm). After removal of extra moisture of the egg shell membrane with filter paper, the egg shell membrane was weighed. Synthesis of AAS-ABA and GGS-ABA. N-Acetyl-L-AAS and N-acetyl-L-GGS were prepared using lysyl oxidase activity of egg shell membrane as previously described (24). Briefly, 10 mM NRacetyl-L-lysine and 10 mM NR-acetyl-L-ornithine were individually incubated with egg shell membrane (10 g) in 100 mL of 20 mM

sodium phosphate buffer, pH 9.0, at 37 °C for 24 h with shaking. After the egg shell membrane was removed by centrifugation, the reaction mixtures were adjusted to pH 6.0 by the addition of 1 M HCl. Then, the resulting aldehydes were reductively aminated with p-aminobenzoic acid (ABA) and NaCNBH3 by a modification of the previous method (22, 25, 26) as follows. After the addition of ABA (10 g, 3 mmol), NaCNBH3 (5 g, 4.5 mmol) was slowly added to the reaction mixture with stirring, and the mixture was allowed to react at 37 °C for 2 h with stirring. Then, 12 M HCl (100 mL) was slowly added to the reaction mixture, and ABA derivatives of NR-acetyl-L-AAS and NR-acetyl-L-GGS were hydrolyzed for 10 h at 110°C. The hydrolysates were evaporated at 40 °C in vacuo to dryness. The resulting AAS-ABA and GGS-ABA were separated by silica gel column chromatography using ethyl acetate/acetic acid/ water (20:2:1, v/v/v) as the elution solvent. Furthermore, each compound was purified by preparative HPLC with a C-18 reversed phase column (LiChroprep, 30 mm × 240 mm, Lober, Merck, Darmstadt, Germany), using water/acetonitrile/acetic acid (90:10: 1) as the eluent at a flow rate of 3.0 mL/min and monitoring the eluate at 250 nm. AAS-ABA and GGS-ABA were characterized from fast atom bombardment (FAB) MS, 1H NMR, and 1H-1H correlation spectroscopy (COSY) spectra. FAB-MS was measured with a Mstation JMS-700 spectrometer (JEOL, Tokyo, Japan), and the sample was dissolved with methanol. NMR spectra were obtained on Varian Unity INOVA 500 spectrometers (Palo Alto, CA) at room temperature, and the sample was dissolved in D2O. AAS-ABA. FAB-MS 267 (M + H)+. 1H NMR (500 MHz, DDS): δ 1.60 (m, 2H, γ-CH2), 1.85 (m, 2H, δ-CH2) 1.98 (m, 2H, β-CH2), 3.48 (t, 2H, J ) 5.6 Hz, -CH), 4.05 (t, 1H, J ) 5.5 Hz, R-CH), 7.70 (d, 1H, J ) 8.5, Ar-H), 8.18 (d, 1H, J ) 8.5, ArH). GGS-ABA. FAB-MS 253 (M + H)+. 1H NMR (500 MHz, DDS): δ 1.95 (m, 2H, γ-CH2), 2.04 (m, 2H, β-CH2) 3.52 (t, 2H, J ) 5.6 Hz, δ-CH2), 4.07 (t, 1H, J ) 5.5 Hz, R-CH), 7.69 (d, 1H, J ) 8.5, Ar-H), 8.11 (d, 1H, J ) 8.5, Ar-H). Derivatization and Acid Hydrolysis of Proteins. Proteins were dialyzed at 4 °C by ultrafiltration with Ultrafree-MC Centrifugal Filter Units (10000 Da NMWL, Millipore, Billerica, MA). The dialysis was continued with three changes of phosphate-buffered saline (PBS) containing 1 mM DTPA and a final change of 0.25 M 2-(N-morpholino)ethanesulfonic acid (MES) buffer (pH 6.0) containing 1% sodium dodecyl sulfate (SDS) and 1 mM DTPA. Reductive amination of protein carbonyls was conducted by a modification of the previous method (22). After the protein concentration was measured using the commercial kit, Protein-HA Test Wako (Wako Pure Chemical Industries), 250 µL of the protein solution was transferred to a vial, and 500 µL of 50 mM ABA in 0.25 M MES buffer (pH 6.0) was added. The reaction was started by addition of 250 µL of freshly prepared 100 mM NaCNBH3 in 0.25 M MES buffer (pH 6.0), and the mixture was allowed to react at 37 °C for 90 min with shaking in the dark. After the reaction, the protein was precipitated by the addition of 500 µL of cold 50% (w/v) trichloroacetic acid (TCA). After it stood for 10 min in an ice bath, the mixture was centrifuged at 17000g for 10 min at 4 °C, and the pellet of precipitated protein was separated. The pellet was washed with 1.0 mL of cold 10% (w/v) TCA and 1.0 mL of cold ethanol twice. Then, the resulting protein was hydrolyzed for

R-Aminoadipic and γ-Glutamic Semialdehydes in Proteins 24 h at 110 °C with 1.5 mL of 6 M HCl. The hydrolysate was evaporated to dryness in vacuo followed by reconstitution in 200 µL of 50 mM sodium acetate buffer (pH 5.4). After filtration with a PVDF syringe filter (0.45 µm pore size, Whatman, Clifton, NJ), the hydrolysate was analyzed by HPLC as described below. Detection of AAS-ABA and GGS-ABA by HPLC. The sample (20 µL) was injected into an HPLC apparatus with a C-18 reversed phase column (COSMOSIL 5C18-AR-II, 250 mm × 4.6 mm, Nacalai Tesque) eluted with 50 mM sodium acetate buffer (pH 5.4). The HPLC instrument (Hitachi, Tokyo, Japan) consisted of an L-6020 pump and an L-7485 fluorescence detector in a D-2500 data station. The column oven (L-7300, Hitachi) was maintained at 40 °C, and the flow rate was 1.5 mL/min. The eluate was monitored with excitation and emission wavelengths set at 283 and 350 nm, respectively. Calibration curves were obtained for the authentic standards by plotting the peak areas. Correlation coefficients greater than 0.999 were obtained. In Vitro Oxidation of Proteins. BSA (5.0 mg/mL) was incubated with various oxidation systems: (i) 1.0 mM H2O2, (ii) 100 µM FeCl2, (iii) 100 µM FeCl3, (iv) 1.0 mM H2O2 and 100 µM FeCl2, (v) 1.0 mM ascorbic acid (Asc), (vi) 1.0 mM Asc and 100 µM FeCl3, and (vii) 10 mM KO2 in 50 mM sodium phosphate buffer (pH 7.4) at 37 °C for 24 h with shaking in the dark. Then, the oxidized proteins were dialyzed by ultrafiltration as described above. Preparation of Human Plasma Proteins and Hemoglobin. Blood samples were obtained from six healthy, normolipidemic volunteers (22-30 years old) after overnight fasting and collected into heparinized polypropylene tubes. We obtained informed consent from individuals. Plasma was immediately prepared by centrifugation at 1500g for 20 min. After the plasma and buffy coat were removed, 1 mL of erythrocyte was washed with 10 mL of PBS three times, and then 5 mL of distilled deionized water was added. The mixture was thoroughly shaken and kept at room temperature for 15 min. Then, hemoglobin was collected by centrifugation at 5000g for 15 min. Oxidation of Plasma. Freshly isolated human plasma (180 µL) was mixed with 20 µL of 10 mM H2O2 or 10 mM Asc in PBS. The mixture was incubated at 37 °C for 24 h with shaking in the dark. Then, the oxidized proteins were dialyzed by ultrafiltration as described above. Animals and Treatment. All experiments were done following the ethical procedures accepted by the Graduate School of Bioagricultural Sciences, Nagoya University. Ferric nitrilotriacetate (Fe3+-NTA) solution was prepared immediately before use by the method described in a previous study (27) with a slight modification. Briefly, ferric nitrate enneahydrate and the nitrilotriacetic acid disodium salt were each dissolved in PBS to form 300 and 600 mM solutions, respectively. They were mixed at a volume ratio of 1:2, and the pH was adjusted with sodium hydrogen carbonate to 7.4. Male ddY mice (Shizuoka Laboratory Animal Center, Shizuoka, Japan) weighing 25-35 g (8 weeks of age) were injected intraperitoneally with the Fe3+-NTA solution at a dose of 15 mg Fe/kg body weight. Control mice were injected with saline solution. Three hours after injection, blood was drawn from the abdominal aorta of mice under anesthesia with ethyl ether and placed into heparinized tubes. Plasma was immediately prepared by centrifugation at 1500g for 10 min at 4 °C. Rat liver and plasma were obtained from male Wistar rats (Shizuoka Laboratory Animal Center), weighing 180-200 g (7 weeks of age). Rat liver was homogenized in PBS containing 1 mM DTPA. Homogenates were centrifuged at 15000g for 15 min, and clear supernatants were dialyzed as described above. Determination of Protein Carbonyl Content in Plasma. Protein carbonyls were assayed by a modification of the previous method (13). Briefly, freshly isolated plasma (100 µL) was mixed with 300 µL of 6 M guanidinium chloride (pH 2.3) and 1.0 mL of 10 mM DNPH in 2 M HCl. The mixture was incubated at room temperature for 1 h, followed by the addition of 1.5 mL of cold 20% TCA. The samples were incubated on ice for 10 min. After centrifugation, the protein pellets were washed three times with 6 mL of ethanol/ethyl acetate (1:1, v/v) and dissolved in 3.0 mL of

Chem. Res. Toxicol., Vol. 19, No. 8, 2006 1061

Figure 2. Determination of AAS and GGS in native BSA by HPLC. Native BSA was derivatized by reductive amination with ABA and NaCNBH3. After acid hydrolysis, AAS-ABA and GGS-ABA were analyzed by a fluorometric HPLC with detection at Ex 283 nm/Em 350 nm. (A) AAS-ABA, (B) GGS-ABA, and (C) hydrolysate of BSA. Details are shown in the Experimental Procedures.

6 M guanidinium chloride (pH 2.3). The peak absorbance at 370 nm was measured to quantify protein carbonyls. The data are expressed as nmol/mg protein, using a molar absorption coefficient of 22000 M-1 cm-1 for the DNPH derivatives. Statistical Analysis. Statistical significance was determined by single factor ANOVA with Student’s t-test, using Statview statistical software (StatView J-4.5, Abacus Concepts, Berkeley, CA). A p value of less than 0.05 was considered significant.

Results Analytical Method for the Determination of AAS and GGS. Figure 1 shows the novel fluorescence-derivatizing strategy to determine AAS and GGS. AAS and GGS present in proteins are destroyed during acid hydrolysis and, therefore, stabilized by reductive amination with NaCNBH3 and ABA, a fluorescence reagent, to their corresponding derivatives, AASABA and GGS-ABA. After acid hydrolysis, AAS-ABA and GGS-ABA are determined by a fluorometric HPLC. The standards of AAS-ABA and GGS-ABA were synthesized from NR-acetyl-L-lysine and NR-acetyl-L-ornithine, respectively, by lysyl oxidase as described in the Experimental Procedures, and the purity and identity were confirmed using HPLC, MS, and NMR techniques. As expected, the synthesized preparations were fluorescent at Exmax 283 nm/Emmax 350 nm and were completely recovered from the acid hydrolysis in 6 M HCl at 110 °C for 24 h in both cases. The fluorometric HPLC analysis of AAS-ABA and GGS-ABA standards showed one single peak in each chromatogram (Figure 2A,B), and the quantification limits of AAS-ABA and GGS-ABA were approximately 10 and 4 fmol, respectively. Using the analytical method, the peaks of AAS-ABA and GGS-ABA could be clearly detected in the hydrolysate of a native BSA (Figure 2C). The peaks were not

1062 Chem. Res. Toxicol., Vol. 19, No. 8, 2006

Akagawa et al.

Table 1. Levels of AAS and GGS in Biological Proteinsa pmol/mg protein protein

AAS

GGS

bovine serum albumin human plasma proteins human hemoglobin rat plasma proteins rat liver proteins mouse plasma proteins

172.9 ( 4.0 21.0 ( 2.1 23.0 ( 3.8 304.5 ( 26.9 49.8 ( 7.5 20.0 ( 6.9

62.1 ( 2.1 11.0 ( 0.5 38.8 ( 5.6 26.1 ( 8.8 42.1 ( 7.6 3.4 ( 0.9

a AAS and GGS were measured by HPLC as described in the Experimental Procedures. Values are means from at least five analyses ( SD.

seen when NaCNBH3 was omitted from the reaction mixture, indicating that these compounds were not artifactual products formed during derivatization and acid hydrolysis of BSA and ABA. We examined the effects of reaction time on the derivatization using different amounts of derivatizing chemicals and confirmed that the derivatization of both aldehydic residues in oxidized BSA would be sufficient for the reaction with 25 mM ABA and 25 mM NaCNBH3 in 0.25 M MES buffer (pH 6.0) containing 0.5% SDS at 37 °C for 90 min as described in the Experimental Procedures. Using this method, the contents of each carbonyl in native BSA were estimated at 172.9 ( 4.0 pmol/mg protein (10.06 ( 0.23 mmol/mol) of AAS and 62.1 ( 2.1 pmol/mg protein (5.21 ( 0.15 mmol/mol) of GGS, in agreement with a previous study (19). Protein carbonyl contents in native BSA were estimated at 1.03 nmol/mg protein by the DNPH method, and the sum of AAS and GGS accounted for about 23% of the total carbonyl groups. Determination of AAS and GGS in Biological Samples. We evaluated the distribution of AAS and GGS in various biological proteins by the present method and observed the accumulation of both oxidation products in all native proteins analyzed (Table 1). The amount of AAS was generally higher than that of GGS with the exception of human hemoglobin, in agreement with a previous study (19). Rat plasma proteins and BSA contained a much higher amount of AAS than other proteins. Interestingly, there was a significant difference in each level of AAS and GGS in plasma proteins among mammalian species. Generation of AAS and GGS in Proteins Exposed to Various Oxidation Systems. To investigate the generating mechanism of AAS and GGS, we incubated BSA in phosphate buffer (pH 7.4) at 37 °C for 24 h under various oxidation systems and then determined the formation of both oxidation products by the present method (Figure 3). The incubation of BSA with 100 µM Fe2+ significantly increased both AAS and GGS contents but not that with 100 µM Fe3+, probably due to the generation of O2•- and subsequent dismutation to H2O2. A marked increase in both oxidation products was observed during incubation with the Fenton reaction systems (100 µM Fe2+ and 1.0 mM H2O2), but incubation with H2O2 alone resulted in no increase, indicating that •OH can oxidize amino acid residues to generate AAS and GGS. Figure 4 shows H2O2-dependent formation of AAS and GGS in BSA exposed to the Fenton reaction system. The formation of both semialdehydes was increased in a dose-dependent manner. In the presence and absence of 100 µM Fe3+, the incubation of BSA with 1.0 mM Asc also significantly yielded both oxidation products, and there was a tendency for GGS to be more preferentially formed by Asc, in agreement with a previous study (Figure 3) (19). Accumulation of both oxidation products during incubation with 10 mM KO2 was lower than that in the MCO systems. Hence, these results indicate that •OH is more effectively involved in

Figure 3. Formation of AAS and GGS in BSA exposed to various oxidation systems in vitro. BSA (5.0 mg/mL) was incubated with various oxidation systems: (i) 100 µM FeCl2, (ii) 1.0 mM H2O2, (iii) 1.0 mM H2O2 and 100 µM FeCl2, (iv) 100 µM FeCl3, (v) 1.0 mM Asc, (vi) 1.0 mM Asc and 100 µM FeCl3, and (vii) 10 mM KO2 in 50 mM sodium phosphate buffer (pH 7.4) at 37 °C for 24 h. AAS and GGS were measured by HPLC as described in the Experimental Procedures. Each value is the mean ( SD from at least three experiments. *Denotes values that are significantly different (p < 0.05) from the control.

Figure 4. H2O2-dependent formation of AAS and GGS in BSA exposed to the Fenton reaction system. BSA (5.0 mg/mL) was incubated with 100 µM FeCl2 and 0-2.0 mM H2O2 in 50 mM sodium phosphate buffer (pH 7.4) at 37 °C for 24 h. AAS and GGS were measured by HPLC as described in the Experimental Procedures.

the oxidation of lysine, arginine, and proline residues in protein than O2•- and H2O2. Furthermore, we examined whether protein-bound metals catalyze the oxidation of amino acids, because transition metal ions in vivo are almost entirely bound to proteins. Therefore, human holo-transferrin, which is an iron-binding plasma protein, was incubated with 1.0 mM H2O2 or 1.0 mM Asc in 50 mM phosphate buffer, pH 7.4, at 37 °C for 24 h. As shown in Figure 5, both semialdehydes in transferrin increased significantly during the incubation with H2O2 and Asc. To gain further insight into their generation by the MCO systems under physiological conditions, 1.0 mM H2O2 or 1.0 mM Asc was incubated in human plasma at 37 °C for 24 h, and then, the content of both semialdehydes in plasma proteins was measured (Figure 6). Thus, the level of AAS and GGS in plasma proteins was found to significantly increase by the addition of H2O2 and Asc to plasma, suggesting that AAS and GGS can be generated by the MCO systems in vivo. Formation of AAS and GGS in Mice Plasma Subjected to Oxidative Stress by Exposure to Fe3+-NTA. The endogenous formation of AAS and GGS by oxidative stress in vivo was assessed in a mouse model of acute iron overload with Fe3+-NTA. Iron overload induces oxidation of lipid, DNA, and protein in plasma (28, 29). As shown in Figure 7, the levels of AAS and GGS were found to increase about 2- and 9-fold,

R-Aminoadipic and γ-Glutamic Semialdehydes in Proteins

Figure 5. Formation of AAS and GGS in human transferrin exposed to H2O2 and Asc in vitro. Human transferrin (2.0 mg/mL) was incubated with 1.0 mM H2O2 or 1.0 mM Asc in 50 mM sodium phosphate buffer (pH 7.4) at 37 °C for 24 h. AAS and GGS were measured by HPLC as described in the Experimental Procedures. Each value is the mean ( SD from at least three experiments. *Denotes values that are significantly different (p < 0.05) from the control.

Figure 6. Formation of AAS and GGS in human plasma proteins exposed to H2O2 and Asc in vitro. Human plasma was incubated with 1.0 mM H2O2 or 1.0 mM Asc at 37 °C for 24 h. AAS and GGS were measured by HPLC as described in the Experimental Procedures. Each value is the mean ( SD from at least three experiments, using plasma obtained from the same subject. *Denotes values that are significantly different (p < 0.05) from the control.

Figure 7. Formation of AAS and GGS in mice plasma proteins subjected to oxidative stress by exposure to Fe3+-NTA. Mice were intraperitoneally treated with Fe3+-NTA (15 mg Fe/kg body weight). Control mice were injected with saline solution. Three hours after injection, the level of AAS and GGS in plasma proteins was measured by HPLC as described in the Experimental Procedures. Data are expressed as means ( SD (n ) 6 or 7). *Denotes values that are significantly different (p < 0.05) from the control.

respectively, 3 h after a single i.p. dose of Fe3+-NTA (15 mg Fe/kg of body weight), suggesting a correlation between the formation of semialdehydes and oxidative stress.

Discussion There is a large body of evidence implicating the role of oxidative damage to proteins in both normal aging and

Chem. Res. Toxicol., Vol. 19, No. 8, 2006 1063

pathogenesis of diverse diseases (3-5), and measurement of protein carbonyl groups has become one of the most widely used markers to determine oxidative damage (6, 7). The quantitative and analytical importance of protein carbonyls has prompted the development of methods to analyze specific carbonyl residues because of understanding of their chemical nature, formation pathway, and distribution level in vivo. However, chemical difficulties, that is, both semialdehydes and their reduced products, are destroyed and modified during acid hydrolysis and have prevented a general quantification of major carbonyl products, AAS and GGS, in proteins. In the present study, we developed a novel specific method to quantitate AAS and GGS after their derivatization by reductive amination with ABA and NaCNBH3 under a mild condition. It is noteworthy that the resulting ABA derivatives, AAS-ABA and GGS-ABA, were completely stable during acid hydrolysis with HCl at 110 °C for 24 h. Thus, the mild derivatizing condition and derivatives resistant to acid hydrolysis permitted the reliable and accurate quantification of both semialdehydes. In addition, the derivatives were fluorescent after acid hydrolysis and, therefore, could be determined sensitively and rapidly by a fluorometric HPLC. Another advantage of the method described here is that the standards can be easily prepared from commercially available chemicals using lysyl oxidase activity of egg shell membrane (23). Accordingly, the present method is thought to be more reliable, simple, and available than other published methods. The analysis of oxidized proteins by the present method provided important insight into the chemical pathways that lead to formation of both semialdehydes under physiological conditions. As shown in Figure 3, we measured AAS and GGS in BSA subjected to various oxidation systems in vitro. Indeed, the MCO systems markedly enhanced their formation in BSA in agreement with previous reports (19). The reaction of H2O2 with Fe2+, i.e., the Fenton reaction, gives rise to a powerful oxidant, •OH. Asc can also produce •OH and alkoxyl radical in the presence of transition metal ions. Therefore, it seems that •OH is a primary oxidant of lysine, arginine, and proline residues to form both semialdehydes. Furthermore, it has been previously reported that primary amines are sensitively oxidized to the corresponding aldehydes by •OH (30). Considering these results, their generation is most likely to be free radical-mediated reactions. Possible reaction schemes for the oxidation of lysine, arginine, and proline by •OH are depicted in Figure 8. An essential step for initiating reaction of the formation of both compounds could be a hydrogen abstraction of each amino acid residues by •OH. The hydrogen atom at the 6-position carbon of the lysine residue may be abstracted by •OH to form a carboncentered radical. Immediately, transition metal ions may accept the lone electron of the carbon radical to generate the lysine imine. Finally, spontaneous hydrolysis of the imino group can lead to the release of ammonium ion and the formation of AAS. Correspondingly, •OH may undergo abstraction of a hydrogen atom from the carbon adjacent to the guanidium group of the arginine residue to form an arginine radical. Then, the arginine radical would form an imino group followed by the release of guanidine and the formation of GGS. GGS may also arise via initial hydrogen abstraction at the 5-position carbon of the Pro residue and formation of a carbon-centered radical. Because of the presence of various metal ion-binding proteins such as transferrin, ferritin, and ceruloplasmin, the availability of free metal ions in plasma and other biological fluids is extremely low. Therefore, we examined whether protein-bound metal ions catalyze the oxidation of proteins. As shown in Figure 5, ironcontaining transferrin exposed to H2O2 and Asc contained

1064 Chem. Res. Toxicol., Vol. 19, No. 8, 2006

Akagawa et al.

Figure 8. Proposed reaction scheme for the formation of AAS and GGS in protein.

greater amounts of both semialdehydes than that of native transferrin, indicating that transferrin-bound iron can catalyze the formation of •OH. This result can also lead to speculation that the oxidation of lysine, arginine, and/or proline residues may occur at the metal ion-binding site of transferrin in a sitespecific manner. To gain further insight into the generation of AAS and GGS under physiological conditions, we also supplemented freshly isolated human plasma, used as a representative biological fluid, with H2O2 and Asc. Plasma protein levels of AAS and GGS were significantly increased by the incubation with 1.0 mM H2O2 and 1.0 mM Asc at 37 °C for 24 h (Figure 6). Taken together, these results suggest that both semialdehydes can be produced by MCO in plasma, and oxidative stress may cause protein oxidation. Conversely, Suh et al. (31) have shown that supplementation of Asc to human plasma does not increase the protein carbonyl content even in the presence of iron or copper and H2O2. Although the in vivo relevance of the observation is unclear, our result suggests that an extremely high dose of Asc may exert a pro-oxidant effect. Protein carbonylation, which is an irreversible oxidative process, alters its disposition, leading to the inactivation. In addition, it is possible that the AAS and GGS residues in proteins can spontaneously condense with the -amino group of neighboring lysine residues to form inter- and intramolecular covalent cross-links by means of the Schiff base formation (32, 33). Cross-linking is associated with protein aggregation and deposition and, therefore, is considered to cause neurodegenerative disorders (1). To investigate this possibility, we analyzed the Schiff bases of AAS and GGS with the lysine residue, after reduction with NaBH4 to stable secondary amines, lysinonorleucine and lysinonorvaline. However, no cross-linking amino acids were detected in hydrolysates of native or oxidized protein (data are not shown). This result suggests that AAS and GGS residues exist as the free aldehyde or stable cyclic aminal, which is implied to form in proteins (Figure 8). An alternative explanation for the observation is that a favorable stereochemical

distance between protein semialdehydes and lysine residues is required for the formation of the Schiff base because of its instability. In the present study, using the method, we could identify accumulations of both semialdehydes in plasma proteins, hemoglobin, and liver proteins (Table 1). Plasma proteins contained a much higher amount of AAS than GGS, suggesting that the susceptibility for oxidation of lysine, proline, and arginine residues varies with the different proteins and oxidation systems. Further studies may elucidate their individual sensitivity to oxidations. Interestingly, there was a significant difference in each level of AAS and GGS in plasma proteins among mammalian species. This observation may be explained by differences in age, protein metabolic rate, and maximum life span of mammalian species. Further work will be needed to elucidate the correlation between the level of semialdehydes and these factors. In this study, the generation of AAS and GGS in vivo was also demonstrated in plasma of a mouse model of acute iron overload with Fe3+-NTA (Figure 7), in agreement with a previous report that protein carbonyl content was significantly increased by iron overload (28). Iron is postulated as a major catalytic metal for the formation of reactive oxygen species in vivo, and iron overload can cause cirrhosis, diabetes, cancer, heart failure, and arthritis (29). It has been previously shown that the production of free radical-associated modified molecules such as lipid peroxidation products, aldehydemodified proteins, and 8-hydroxy-2′-deoxyguanosine (8-OHdG) becomes maximal as early as 3 h after a single Fe3+-NTA treatment (28). The present study also revealed that major protein carbonyls, AAS and GGS, could be generated by oxidative stress in vivo, suggesting that their measurement by the present method can provide useful information on the direct contribution of oxidative stress to various pathogenesis and aging. In summary, we developed and validated a novel specific method to quantitate AAS and GGS after their derivatization

R-Aminoadipic and γ-Glutamic Semialdehydes in Proteins

by reductive amination with ABA and NaCNBH3 under a mild condition. Analysis of oxidized proteins revealed that both semialdehydes can be generated by MCO in vivo. Moreover, the production of these two products was associated with oxidative stress induced by acute iron overload with Fe3+-NTA. The results of this study and the method developed are therefore useful to future investigations aimed at elucidating the contribution of oxidative stress to aging and disease. Acknowledgment. This work was supported by the Grantin-Aid for JSPS Fellows, the Ministry of Education, Culture, Sports, Science and Technology, Japan.

References (1) Davies, M. J. (2005) The oxidative environment and protein damage. Biochim. Biophys. Acta 1703, 93-109. (2) Levine, R. L. (2002) Carbonyl modified proteins in cellular regulation, aging and disease. Free Radical Biol. Med. 32, 790-796. (3) Stadtman, E. R. (1992) Protein oxidation and aging. Science 257, 1220-1224. (4) Beal, M. F. (2002) Oxidatively modified proteins in aging and disease. Free Radical Biol. Med. 32, 797-803. (5) Stadtman, E. R. (2001) Protein oxidation in aging and age-related diseases. Ann. N. Y. Acad. Sci. 928, 22-38. (6) Dalle-Donne, I., Rossi, R., Giustarini, D., Milzani, A., and Colombo, R. (2003) Protein carbonyl groups as biomarkers of oxidative stress. Clin. Chim. Acta 329, 23-38. (7) Nystro¨m, T. (2005) Role of oxidative carbonylation in protein quality control and senescence. EMBO J. 24, 1311-1317. (8) Uchida, K., Szweda, L. I., Chae, H. Z., and Stadtman, E. R. (1993) Immunochemical detection of 4-hydroxy-2-nonenal modified proteins in oxidized hepatocytes. Proc. Natl. Acad. Sci. U.S.A. 90, 8742-8746. (9) Furuhata, A., Nakamura, M., Osawa, T., and Uchida, K. (2002) Thiolation of protein-bound carcinogenic aldehyde: An electrophilic acrolein-lysine adduct that covalently binds to thiols. J. Biol. Chem. 277, 27919-27926. (10) Uchida, K. (2003) Histidine and lysine as targets of oxidative modification. Amino Acids 25, 249-257. (11) Yamada, S., Kumazawa, S., Ishii, T., Nakayama, T., Itakura, K., Shibata, N., Kobayashi, M., Sakai, K., Osawa, T., and Uchida, K. (2001) Immunochemical detection of a lipofuscin-like fluorophore derived from malondialdehyde and lysine. J. Lipid Res. 42, 11871196. (12) Liggins, J., and Furth, A. J. (1997) Role of protein-bound carbonyl groups in the formation of advanced glycation endproducts. Biochim. Biophys. Acta 1361, 123-130. (13) Levine, R. L., Williams, J. A., Stadtman, E. R., and Shacter, E. (1994) Carbonyl assays for determination of oxidatively modified proteins. Methods Enzymol. 233, 346-357. (14) Choi, J., Rees, H. D., Weintraub, S. T., Levey, A. I., Chin, L. S., and Li, L. (2005) Oxidative modifications and aggregation of Cu, Znsuperoxide dismutase associated with Alzheimer and Parkinson diseases. J. Biol. Chem. 280, 11648-11655. (15) Choi, J., Malakowsky, C. A., Talent, J. M., Conrad, C. C., and Gracy, R. W. (2002) Identification of oxidized plasma proteins in Alzheimer’s disease. Biochem. Biophys. Res. Commun. 293, 1566-1570. (16) Mantle, D., Falkous, G., and Walker, D. (1999) Quantification of protease activities in synovial fluid from rheumatoid and osteoarthritis cases: Comparison with antioxidant and free radical damage markers. Clin. Chim. Acta 284, 45-58.

Chem. Res. Toxicol., Vol. 19, No. 8, 2006 1065 (17) Bowling, A. C., Schulz, J. B., Brown, R. H., Jr., and Beal, M. F. (1993) Superoxide dismutase activity, oxidative damage, and mitochondrial energy metabolism in familial and sporadic amyotrophic lateral sclerosis. J. Neurochem. 61, 2322-2325. (18) Grattagliano, I., Vendemiale, G., Boscia, F., Micelli-Ferrari, T., Cardia, L., and Altomare, E. (1998) Oxidative retinal products and ocular damages in diabetic patients. Free Radical Biol. Med. 25, 369-372. (19) Requena, J. R., Chao, C.-C., Levine, R. L., and Stadtman, E. R. (2001) Glutamic and aminoadipic semialdehydes are the main carbonyl products of metal-catalyzed oxidation of proteins. Proc. Natl. Acad. Sci. U.S.A. 98, 69-74. (20) Ayala, A., and Cutler, R. G. (1996) The utilization of 5-hydroxyl-2amino valeic acid as a specific marker of oxidized arginine and proline residues in proteins. Free Radical Biol. Med. 21, 65-80. (21) Lent, R. W., Smith, B., Salcedo, L. L., Faris, B., and Franzblau, C. (1969) Studies on the reduction of elastin. II. Evidence for the presence of R-aminoadipic acid δ-semialdehyde and its aldol condensation product. Biochemistry 8, 2837-2845. (22) Daneshvar, B., Frandsen, H., Autrup, H., and Dragsted, L. O. (1997) γ-Glutamyl semialdehyde and 2-amino-adipic semialdehyde: Biomarkers of oxidative damage to proteins. Biomarkers 2, 117-123. (23) Akagawa, M., Sasaki, D., Kurota, Y., and Suyama, K. (2005) Formation of R-aminoadipic and γ-glutamic semialdehydes in proteins by the Maillard reaction. Ann. N. Y. Acad. Sci. 1043, 129-134. (24) Akagawa, M., Wako, Y., and Suyama, K. (1999) Lysyl oxidase coupled with catalase in egg shell membrane. Biochim. Biophys. Acta 1434, 151-160. (25) Yuen, C.-T., Gee, C. K., and Jones, C. (2002) High-performance liquid chromatographic profiling of fluorescent labeled N-glycans on glycoproteins. Biomed. Chromatogr. 16, 247-254. (26) Charlwood, J., Tolson, D., Dwek, M., and Camilleri, P. (1999) A detailed analysis of neutral and acidic carbohydrates in human milk. Anal. Biochem. 273, 261-277. (27) Toyokuni, S., Uchida, K., Okamoto, K., Hattori-Nakakuki, Y., Hiai, H., and Stadtman, E. R. (1994) Formation of 4-hydroxy-2-nonenalmodified proteins in the renal proximal tubules of rats treated with a renal carcinogen, ferric nitrilotriacetate. Proc. Natl. Acad. Sci. U.S.A. 91, 2616-2620. (28) Okada, K., Wangpoengtrakul, C., Tanaka, T., Toyokuni, S., Uchida, K., and Osawa, T. (2001) Curcumin and especially tetrahydrocurcumin ameliorate oxidative stress-induced renal injury in mice. J. Nutr. 131, 2090-2095. (29) Lucesoli, F., Caligiuri, M., Roberti, M. F., Perazzo, J. C., and Fraga, C. G. (1999) Dose-dependent increase of oxidative damage in the testes of rats subjected to acute iron overload. Arch. Biochem. Biophys. 372, 37-43. (30) Akagawa, M., and Suyama, K. (2002) Oxidative deamination by hydrogen peroxide in the presence of metals. Free Radical Res. 36, 13-21. (31) Suh, J., Zhu, B. Z., and Frei, B. (2003) Ascorbate does not act as a pro-oxidant towards lipids and proteins in human plasma exposed to redox-active transition metal ions and hydrogen peroxide. Free Radical Biol. Med. 34, 1306-1314. (32) Saito, M., Marumo, K., Fujii, K., and Ishioka, N. (1997) Single-column high-performance liquid chromatographic-fluorescence detection of immature, mature, and senescent cross-links of collagen. Anal. Biochem. 253, 26-32. (33) Sims, T. J., and Bailey, A. J. (1992) Quantitative analysis of collagen and elastin cross-links using a single-column system. J. Chromatogr. Biomed. Appl. 582, 49-55.

TX060026P