Aminophospholipid glycation and its inhibitor screening system: a new role of pyridoxal 5'-phosphate as the inhibitor.

Peroxidized phospholipid-mediated cytotoxity is involved in the pathophysiology of a number of diseases [i.e., the abnormal increase of phosphatidylcholine hydroperoxide (PCOOH) found in the plasma of type 2 diabetic patients]. The PCOOH accumulation may relate to Amadori-glycated phosphatidylethanolamine (deoxy-d-fructosyl PE, or Amadori-PE), because Amadori-PE causes oxidative stress. However, lipid glycation inhibitor has not been discovered yet because of the lack of a lipid glycation model useful for inhibitor screening. We optimized and developed a lipid glycation model considering various reaction conditions (glucose concentration, temperature, buffer type, and pH) between PE and glucose. Using the developed model, various protein glycation inhibitors (aminoguanidine, pyridoxamine, and carnosine), antioxidants (ascorbic acid, α-tocopherol, quercetin, and rutin), and other food compounds (l-lysine, l-cysteine, pyridoxine, pyridoxal, and pyridoxal 5′-phosphate) were evaluated for their antiglycative properties. Pyridoxal 5′-phosphate and pyridoxal (vitamin B6 derivatives) were the most effective antiglycative compounds. These pyridoxals could easily be condensed with PE before the glucose/PE reaction occurred. Because PE-pyridoxal 5′-phosphate adduct was detectable in human red blood cells and the increased plasma Amadori-PE concentration in streptozotocin-induced diabetic rats was decreased by dietary supplementation of pyridoxal 5′-phosphate, it is likely that pyridoxal 5′-phosphate acts as a lipid glycation inhibitor in vivo, which possibly contributes to diabetes prevention.

Lipid peroxidation plays a role in the pathophysiology of atherogenesis, diabetes, aging, and other conditions (1). To determine lipid hydroperoxides as a primary oxidation product, we established a chemiluminescence detectionliquid chromatography (LC) method (2). Using this method, it was confirmed that plasma phosphatidylcholine hydroperoxide (PCOOH) abnormally increases in hyperlipidemic (3) and in type 2 diabetic (4) patients. Hence, we hypothesized that PCOOH-mediated cytotoxity is closely involved in the pathophysiology of these diseases.
Recently, while investigating why PCOOH increases in diabetic plasma, we found that diabetic plasma contained an abnormal amount of glycated lipid. The glycated lipid was identified as an Amadori product of phosphatidylethanolamine (deoxy-D-fructosyl phosphatidylethanolamine, or Amadori-PE) (Fig. 1) by LC online with hybrid quadrupole/linear ion-trap mass spectrometry (QTRAP LC/MS/MS) (5,6). Amadori-PE generates reactive oxygen species and thereby triggers lipid peroxidation (7). Therefore, it is likely that PE is exposed to glycation under hyperglycemic conditions, yielding Amadori-PE in vivo. Amadori-PE causes oxidative stress (i.e., PCOOH) (5,6), leading to a disorder of cellular integrity (i.e., angiogenesis stimulation) (8). Amadori-PE and PCOOH could play a role in the development of diabetes. Several foods (i.e., infant formula and chocolate) contain high amounts of Amadori-PE (9). In these products, lipid glycation could impair food flavor, color, and nutritive value.
Despite the potential significance of Amadori-PE in pathological signaling and food deterioration, lipid glycation inhibitor has not yet been discovered. This is because of the lack of a lipid glycation model useful for inhibitor screening. PE glycation involves a Schiff base formation and its rearrangement to Amadori-PE. Therefore, to develop a lipid glycation model, it is necessary to elucidate the optimal reaction conditions (glucose concentration, temperature, buffer type, and pH) for Schiff-PE and Amadori-PE formation.
In this work, we optimized reaction conditions between PE and glucose and developed a lipid glycation model. Using this model, we found that pyridoxal 59-phosphate and pyridoxal (vitamin B 6 derivatives) are attractive antiglycative compounds. The possible therapeutic roles of vitamin B 6 derivatives in vivo were evaluated.

Lipid glycation model
The stock buffer solutions (0.1 M acetate buffer, pH 4.0-5.5, 0.1 M phosphate buffer, pH 6.0-8.0, 0.1 M Tris-HCl buffer, pH 7.5-9.0, and 0.1 M sodium carbonate buffer, pH 9.5-10.0) were prepared. Solutions 1-3 were prepared before use. For solution 1 (dioleoyl-PE solution), 3 mM dioleoyl-PE was dissolved in methanol; for solution 2 (glucose solution), 0.5-2.5 M D-glucose was dissolved in stock buffer; and for solution 3 (working buffer solution), the buffer was prepared by diluting stock buffer with methanol at a ratio of 3:4, 2:5, 1:6, or 0:7 (v/v). To initiate lipid glycation, dioleoyl-PE solution (100 ml of solution 1), D-glucose solution (200 ml of solution 2), and working buffer solution (700 ml of solution 3) were mixed. The mixture was incubated at 37-608C for 0-48 h under an air atmosphere. After incubation, to extract lipid-soluble products, chloroform, methanol, and water were added to the incubated mixture so that the final ratio of chloroform-methanol-water was 2:1:1 (10). After centrifugation at 1,600 g for 10 min, the chloroform layer (lipid fraction) was collected and evaporated. The lipid residue was dissolved in 1 ml of methanol, and a portion (50 ml) was subjected to LC coupled to an evaporative light-scattering detection (ELSD) device to monitor the yield of Schiff-PE and Amadori-PE, as described below.

LC-ELSD analysis
An ODS column (TSK gel ODS-80-Ts, 4.6 3 150 mm; Tosoh, Tokyo, Japan) was used with a mobile phase of methanol-water (99:1) containing 5 mM ammonium acetate at a flow rate of 1 ml/min. The column temperature was maintained at 358C. At the postcolumn, the eluent was split. One of the split eluents (flow rate, 0.95 ml/min) was sent to a Sedex 55 ELSD device (Sedere, Alfortville, France). The temperature of the drift tube was set at 608C. Nitrogen gas was used as nebulizer gas at a pressure of 2.0 bar, and the gain was set at 8. Peak areas of Schiff-PE (5.6 min), Amadori-PE (7.5 min), and dioleoyl-PE (9.2 min) were recorded using a Chromatocorder 12 (System Instruments, Tokyo, Japan). Schiff-PE, Amadori-PE, and dioleoyl-PE concentrations were calculated based on their calibration curves.
To identify Schiff-PE and Amadori-PE, the other eluent (flow rate, 0.05 ml/min) was sent to a Mariner electrospray ionization time-of-flight mass spectrometer (Applied Biosystems, Foster City, CA). Mass spectrometry (MS) was carried out in the positive ion measurement mode with spray voltage of 2,900 V, nozzle potential of 220 V, and nozzle temperature of 1408C. The flow rate of nebulizer gas was 0.3 ml/min. Full-scan spectra were obtained by scanning masses between m/z 400 and 1,000 at 3 s/scan.
Extraction of PE-pyridoxal 59-phosphate adduct from human red blood cells Healthy male volunteers (mean age, 24 years; n 5 3) participated in this study. Written informed consent was obtained from all volunteers. Blood (20 ml) was collected from all subjects into tubes containing 0.1% EDTA-2Na as anticoagulant and then subjected to centrifugation at 1,000 g for 15 min at 48C. After plasma and buffy coat were removed, red blood cells (RBCs) were washed three times with phosphate-buffered saline (pH 7.4) to prepare packed cells. RBC total lipids were extracted from packed cells (1 ml) with 2-propanol/chloroform by the method of Rose and Oklander (11). The lipid extract was redissolved in 1 ml of methanol.
To reduce PE-pyridoxal 59-phosphate adduct present in RBCs, solid NaBH 4 (1 mmol) was added to the RBC lipid extract (1 ml). After incubation at 48C for 1 h, the mixture was evaporated and dissolved in 4 ml of chloroform. The chloroform solution was washed with water to remove NaBH 4 , evaporated, and finally dissolved in 1 ml of methanol-water (99:1; containing 5 mM ammonium acetate). A portion (10 ml) was subjected to QTRAP LC/ MS/MS, as described below.

Analysis of PE-pyridoxal 59-phosphate adduct
The QTRAP LC/MS/MS system consisted of an Agilent 1100 series liquid chromatograph (Palo Alto, CA), including a vacuum degasser, a quaternary pump, and an autosampler, coupled with an Applied Biosystems 4000 QTRAP tandem mass spectrometer (Foster City, CA) equipped with a turbo ion spray source. This instrument is based on a triple-quadrupole ion path in which the final quadrupole can be used as a hybrid quadrupole/linear ion trap mass spectrometer (6).
Neutral loss scan was used for the qualitative analysis of the reduced form of PE-pyridoxal 59-phosphate adduct. Standard PEpyridoxal 59-phosphate adduct (reduced form) was synthesized from dioleoyl-PE and pyridoxal 59-phosphate, according to a previous report (12). Either RBC extract (100 ml) or standard (10 pmol/100 ml) was infused directly to the QTRAP by a flow syringe pump at 0.01 ml/min. Positive ionization mode was adopted. QTRAP parameters, including collision energy and spraying conditions, were optimized with standard (dioleoyl-PEpyridoxal 59-phosphate). The QTRAP was programmed to scan the protonated molecules that had a neutral loss of 372 Da after collision-activated dissociation at the collision cell (Q2; collision gas, N 2 , 4 p.s.i.; collision energy, 30 eV) over the m/z range from 100 to 1,200. Turbo gas temperature was kept at 4008C, and the spray voltage was 5,500 V. Nitrogen values as the turbo, nebulizer, and curtain gas were set at 70, 50, and 20 p.s.i., respectively.
Multiple reaction monitoring (MRM) was used for the detection of predominant molecular species of PE-pyridoxal 59-phosphate adduct in RBCs. Briefly, RBC extract (10 ml) was separated on a TSK gel ODS 100s (2.0 mm 3 150 mm; Tosoh) fitted with a guard column (Inertsil ODS-3, 3.0 mm 3 10 mm; GL Science, Tokyo, Japan). The column was eluted with a mixture of methanol-water (99:1; containing 5 mM ammonium acetate) at a flow rate of 0.3 ml/min, and the column temperature was maintained at 308C. At the postcolumn, predominant molecular species of PE-pyridoxal 59-phosphate adduct in RBCs were individually detected by the QTRAP using MRM for the transition of parent ion to product ion. The collision energy for the transition was 30 eV. The dwell time for the transition was 100 ms. Other QTRAP conditions were as described above.

Animal study
Four week old male Sprague-Dawley rats weighing 70-80 g (CLEA, Tokyo, Japan) were used. Rats were given free access to standard laboratory chow (MF pellet; Oriental Yeast, Tokyo, Japan) and water for 1 week before the experiment began. Rats were housed in a temperature-and humidity-controlled room with a 12 h light cycle. The study was conducted in accordance with the Animal Experimentation Guidelines of Tohoku University.
Diabetes was induced in rats by a single intraperitoneal injection of 80 mg/kg STZ dissolved in 0.1 M sodium citrate buffer (pH 4.5), whereas control rats received an injection of citrate buffer. Blood glucose was measured 3 days after STZ injection. Rats with blood glucose levels of $15 mM were considered diabetic and were used as diabetic rats in the study.
One week after the administration of STZ and/or citrate buffer, rats were divided into four groups: normal rats (control; n 5 6), normal rats receiving 2 mM aqueous pyridoxal 59phosphate (equivalent to 300 mg/kg body weight/day; n 5 6), diabetic STZ rats (n 5 6), and diabetic rats receiving 2 mM aqueous pyridoxal 59-phosphate (n 5 6). After feeding for 10 weeks, rats were fasted for 12 h and anesthetized with diethyl ether, and blood plasma samples were collected.

Statistical analysis
The data are expressed as means 6 SD. Statistical analysis was performed using one-way ANOVA followed by the Bonferroni posthoc test. P , 0.05 was accepted as significant. Figure 2 shows a typical ELSD chromatogram when glucose-treated dioleoyl-PE in methanolic phosphate buffer was subjected to LC-ELSD. After glycation, dioleoyl-PE itself (9.2 min) was decreased, and two peak components (5.6 and 7.5 min) appeared as the major lipid glycation products. Despite the same molecular ion [M 1 H] 1 at m/z 906.6, these peaks could be identified as Schiff-PE (5.6 min) and Amadori-PE (7.5 min) based on a comparison of retention times with standards and the reactivity of reducing agents (i.e., NaBH 3 CN or NaBH 4 ) (data not shown).
Based on these results, we chose the best lipid glycation system: dioleoyl-PE (0.3 mM) was incubated with glucose (500 mM) in 1 ml of 70% methanolic phosphate buffer (0.1 M, pH 7.4) at 378C. In this system, Schiff-PE formation reached its maximum level (0.18 mM) at 6 h after incubation and then decreased gradually (Fig. 4). During 12-48 h of incubation, Amadori-PE formation occurred consistently. These findings are well rationalized by the Maillard theory: PE reacts with glucose to form Schiff-PE, followed by rearrangement to Amadori-PE. Therefore, if Schiff-PE formation can be minimized, the responsible compound may be used as a lipid glycation inhibitor.

Occurrence of PE-pyridoxal 59-phosphate adduct in human RBCs
Despite the wide distribution of pyridoxal 59-phosphate in human tissues and body fluids (13), the occurrence of PE-pyridoxal 59-phosphate adduct in vivo has never been investigated. As shown in Fig. 9, in positive ion mode, collision-induced dissociation of a reduced form of PEpyridoxal 59-phosphate adduct produced a diglyceride ion ([M 1 H 2 372] + ) permitting neutral loss scanning and MRM. When NaBH 4 -treated RBC extract was infused into the QTRAP, molecular species of PE-pyridoxal 59-phosphate adduct (reduced forms) could be screened out by neutral loss scanning (Fig. 10). It was found that the predominant molecular species of PE-pyridoxal 59-phosphate adduct in RBCs were 16   Interfacing LC with the QTRAP enabled the separation and detection of predominant molecular species of PEpyridoxal 59-phosphate adduct (reduced forms) in RBCs (Fig. 11). The concentration of PE-pyridoxal 59-phosphate adduct [16:0-18:2 (m/z 947.6), one of the most predominant molecular species] was z78 6 12 pmol/ml packed cells. In contrast, the concentration of PE-pyridoxal adduct was below the detection limit (,1 pmol/ml packed cells). These findings suggest a new role of pyridoxal 59phosphate in preventing lipid glycation in vivo.

DISCUSSION
Protein nonenzymatically reacts with glucose, leading to the production of Schiff base and ultimately Amadori product. These early glycation products are further transformed into advanced glycation end products (AGEs) (14). Numerous studies have described the formation and accumulation of Amadori products, such as glycated hemoglobin and glycated albumin, and AGEs, such as N e -(carboxymethyl)lysine and N e -(carboxyethyl)lysine, in blood and a range of tissues (15)(16)(17)(18). Protein glycation has been implicated in the development of diabetes and thereby contributes to the physical disability and high mortality rate found in diabetic patients (19,20).
Although protein glycation has been investigated thoroughly, little attention has been paid to lipid glycation. Because of the presence of a free amino group, it seems logical to assume that aminophospholipid can be modified by glycation (21). In vivo lipid glycation is likely to induce changes in the biosynthesis and turnover of membrane phospholipids, the physical properties of membrane, the activities of membrane-bound enzymes, and the susceptibility of oxidative stress. These changes may contribute to the pathology of aging and chronic disease, such as diabetes and atherosclerosis. Although the importance of lipid glycation to human physiology is increasing, a lipid glycation inhibitor has not been discovered. To discover an effective inhibitor, a screening system is essential. Therefore, the objectives of this study were to optimize the reaction conditions between PE and glucose and to develop a lipid glycation model useful for inhibitor screening.
Initially, we investigated the optimal reaction conditions (glucose concentration, temperature, buffer type, and pH) for dioleoyl-PE glycation (Fig. 3). High glucose concentration (300-500 mM) stimulated Schiff-PE formation, whereas temperature did not. Similar results were reported by Ravandi et al. (22). Therefore, lipid glycation kinetics would depend on glucose concentration, especially the open-chain form of glucose (Fig. 1). Methanolic buffer (70-80% methanol) enhanced Schiff-PE formation. Because excess methanol (.80%) in buffer causes instability of pH, we decided to use 70% methanolic buffer. According to buffer type and pH, phosphate buffer (pH 7.0-7.4) was the most effective for Schiff-PE formation. The accelerating effect of phosphate anion on the Maillard reaction is well known (23)(24)(25). It is likely that phosphate anion could enhance lipid glycation by a neighboring catalyst effect.
Once the conditions described above were fully optimized, the nonenzymatic conversion of dioleoyl-PE to Amadori-PE via Schiff base was seen (Fig. 4). Interestingly, rearrangement of Schiff-PE to Amadori-PE began at 12 h after incubation. In contrast, protein glycation (Amadori formation) generally occurs after incubation for several weeks (16). Therefore, the kinetics of lipid glycation is faster than that of protein glycation.
Vitamin B 6 refers to a collective of six biologically interconvertible 3-hydroxy-2-methylpyridine compounds: pyridoxal, pyridoxine, pyridoxamine, and their respective 59-phosphates. Of these, pyridoxal 59-phosphate is the active form of vitamin B 6 , serving as a coenzyme in numerous processes, including tryptophan-niacin conversion, heme synthesis, gluconeogenesis, neurotransmitter synthesis, and amino acid metabolism (31,32). In addition, pyridoxal 59-phosphate acts as a coenzyme in the catabolism of homocysteine to cystathionine and ultimately cysteine (32). As a result, vitamin B 6 played a role in decreasing homocysteine, which contributes to atherosclerosis and coronary heart disease prevention (33,34). Apart from its role as a coenzyme, recent studies suggested that vitamin B 6 can act as a chemopreventive agent. It was reported that pyridoxal 59-phosphate suppresses tumor growth (35) and that pyridoxamine prevents protein glycation (AGE formation) (28). Thus, the biological role and chemopreventive effect of vitamin B 6 have attracted considerable attention. In this study, PEpyridoxal 59-phosphate adduct was detectable in human RBCs (Figs. 9-11), suggesting a new role of vitamin B 6 as a lipid glycation inhibitor. Accordingly, we conducted an animal study and found that increased plasma Amadori-PE and PCOOH concentrations in STZ-induced diabetic rats could be decreased by dietary supplementation of pyridoxal 59-phosphate (Fig. 12). Furthermore, dietary pyridoxal 59-phosphate tended to improve the lipid metabolism (i.e., the abnormal increase of neutral lipids in plasma) of STZ rats (Fig. 13). Therefore, it is likely that dietary pyridoxal 59-phosphate can inhibit both lipid glycation and lipid peroxidation in vivo, which might contribute to the prevention of diabetes. Nevertheless, more studies are required to further elucidate the effects of lipid glycation inhibitors on lipid peroxidation (i.e., PCOOH formation), oxidative stress, the pathophysiology of human disease, and food deterioration.
In the field of protein glycation, it has been recognized that the discovery or design of a protein glycation inhibitor could provide a therapeutic approach for diabetes prevention (26). For example, amino compounds (e.g., alanine) can react with glucose and minimize the potentially deleterious consequences of hyperglycemia (36).
Hydrazine compounds trap reactive carbonyls formed during the Maillard reaction, especially Amadori intermediates, thus preventing their conversion to AGEs (36). Similarly, aminoguanidine is well known as an AGE inhibitor (27). We believe that our screening system could lead to the discovery of a new type of glycation inhibitor (i.e., pyridoxal 59-phosphate) for the purpose of lipid glycation prevention.