- Email: [email protected]
Determination of thiobarbituric acid-reactive substances in oxidized lipids by high-performance liquid chromatography with a postcolumn reaction system
ANALYTICALBIOCHEMISTRY
182,116-120
(1989)
Determination of Thiobarbituric Acid-Reactive Substances in Oxidized Lipids by High-Performance Liquid Chromatography with a Postcolumn Reaction System Kazuaki
Yoden’
and Toshihiro
Showa College of Pharmaceutical
Received
March
Iio
Sciences, l-8, Tsurumaki
5-chome,
Tokyo, Japan
3,1989
A new high-performance liquid chromatography procedure with a postcolumn reaction system for determination of free malondialdehyde (MDA) and other thiobarbituric acid-reactive substances (TBA-RS) in oxidized lipids in vitro has been developed. Using this procedure, both thermally oxidized methyl linoleate and the degradation products of methyl linoleate hydroperoxides revealed many kinds of lipophilic TBARS, but no free MDA was detected on the high-performance liquid chromatography. Similarly, oxidized methyl arachidonate also produced certain kinds of TBA-RS in the lipophilic phase and a small amount of free MDA in the hydrophilic phase. These results indicate that lipophilic TBA-RS produced in oxidized lipids in vitro are major TBA-RS and that the production of free MDA is small, even though the degree of lipid oxidation has previously been estimated as an MDA equivalent measured by the TBA calorimetric test. o ISSSAC~demic Press, Inc.
Lipid peroxides produced from polyunsaturated fatty acids (PUFA)’ in oxidized biological systems and food systems have been considered to show various significant biological or toxic effects both in viuo and in vitro (l-3). For example, lipid peroxides cause accumulation of fluorescent lipofuscin pigments with aging in animal tissues (4,5). Lipid peroxides are generally measured using the calorimetric test by reaction with 2-thiobarbituric acid (TBA) (6). Malondialdehyde (MDA), one of 1 To whom correspondence should be addressed. ’ Abbreviations used: heme, heme methyl ester; HPO, hydroperoxide; MA, methyl arachidonate; MDA, malondialdehyde; ML, methyl linoleate; PUFA, polyunsaturated fatty acids; TBA, 2-thiobarbituric acid; TBA-RS, TBA-reactive substances. 116
Setagaya,
the end products of PUFA peroxidation, reacts with TBA to produce a red pigment with an absorption maximum at 532 nm, and therefore lipid peroxide contents are usually estimated as an MDA equivalent by the TBA value, which has been regarded as an index of the degree of PUFA peroxidation. However, Hirayama et al. have shown the presence of certain TBA red pigments other than MDA-TBA red pigment in oxidized methyl linolenate at 80°C (7). Therefore the TBA calorimetric test is not specific for free MDA, and other compounds may give a positive reaction with TBA to produce TBA red pigments. Kosugi et al. have reported that many monomeric aldehydes are able to react with TBA to produce this characteristic red pigment (8). They reported that MDA content estimated by specific determination of MDA is always lower than that estimated from TBA values (7,9). These findings suggest that the level of lipid peroxides determined by the TBA calorimetric test may express the total amount of many kinds of TBA-reactive substances (TBA-RS), including MDA and other aldehydic breakdown compounds, in oxidized lipids. Certain kinds of TBA-RS other than MDA present in oxidized lipids may also be significantly involved in various reactions. Methods for separation of free MDA or MDA-TBA red pigment by HPLC have been reported recently for the measurement of lipid peroxides in oxidized lipids or biological samples (10-12). However, these methods are specific for determination of free and bound MDA and are not able to detect any other TBA-RS in oxidized lipids. Frankel et al. have recently reported the presence of many kinds of volatile oxidative product, oligomer, and monomer from thermal or metal-catalyzed decomposition of methyl linolenate hydroperoxide in vitro (13). These oxidative products may react with TBA to produce TBA red pigments. In this study, therefore, we de0003~2697/89
$3.00
Copyright 0 1999 by Academic Press, Inc. All rights of reproduction in any form reserved.
CHROMATOGRAPHIC
DETERMINATION
OF
THIOBARBITURIC
ACID-REACTIVE
117
SUBSTANCES
veloped an assay method for the separation and determination of free MDA and certain kinds of TBA-RS in oxidized lipids in vitro employing a new HPLC method with a postcolumn reaction system.
D
TY
R
P
M III
MATERIALS
AND
METHODS
Materials. Methyl arachidonate (MA), methyl linoleate (ML), hemin (type I), and lipoxygenase (type I) were purchased from Sigma Chemical Co. (St. Louis, MO). TBA was purchased from Merck (Darmstadt, FRG). Heme methyl ester (heme) was prepared from hemin by the method of Ortiz de Montellano et al. (14). HPLC-grade methanol (Kant0 Kagaku Co., Ltd., Tokyo) was used for HPLC separation in all experiments. Preparation of standard MDA. MDA was prepared from 1,1,3,3-tetramethoxypropane (Tokyo Kasei Kogyo Co., Ltd., Tokyo) by the method of Crawford et al. (15). 1,1,3,3-Tetramethoxypropane (20 ~1) dissolved in 0.6 ml of 0.1 N HCl was hydrolyzed at 40°C for 30 min, and to this solution, 5.6 ml of water and 6 ml of 0.1 N NaOH solution were added. This MDA solution was stocked at 4°C in a refrigerator. The MDA stock solution was diluted with 10 mM Tris-HCl buffer (pH 7.4) and applied for the measurement of uv absorption spectra and the HPLC analysis. MDA concentration was estimated using a molar absorbance value of 31,800 at 267 nm (16). Thermal oxidation of ML and MA. Ten milligrams of ML or MA was oxidized directly by heating at 100°C for 60 min in an oil bath and then dissolved in 0.1 ml of methanol. Degradation of ML-HP0 or MA-HP0 by heme. HP0 of ML and MA was prepared using lipoxygenase, according to the method previously reported (17). The concentration of each HP0 was estimated using a molar absorbance value of 24,500 at 233 nm (18). Each HP0 (1 pmol) was incubated with heme (10 nmol) in 2 ml of methanol at 37°C for 30 min, and then the incubation mixture was evaporated under an N2 stream. The thermally oxidized MA or ML and the degradation products of each HP0 were fractionated into chloroformand water-soluble compounds by extraction with a chloroform:methanol (2:l) mixture. Both the chloroform and water phases evaporated under an N, stream were dissolved in 0.1 ml of methanol or 10 mM Tris-HCl buffer (pH 7.4) and then injected into the HPLC. Oxidation of rat liver microsomes. Rat liver microsomes were prepared according to the usual method. Livers of Sprague-Dawley rats (6 weeks) were homogenized in 0.15 M KC1 solution using a homogenizer. The microsomal pellet was prepared by centrifugation at 105,OOOg for 60 min. The microsomes (10 mg protein), resuspended in 0.1 M Tris-HCl buffer (pH 7.4), were then oxidized by incubation in an NADPH-generation system in 3 ml of 0.1 M Tris-HCl buffer (pH 7.4) at 37°C for 30
8 FIG. 1.
T
Schematic diagram of the HPLC system. P, pump; ent controller; I, injector; C, column; R, reactor; D, detector; module; S, solvent; T, TBA reagent.
G, gradiM, data
min. Water-soluble TBA-RS in the medium of oxidized microsomes were extracted with a chloroform:methanol (2:l) mixture. Assay of TBA calorimetric test. To an aliquot of each sample, 1 ml of water, 1 ml of 0.1 N HCl, and 2 ml of 0.5% TBA solution were added, and then the mixture was heated at 100°C for 20 min. After the reaction, the mixture was cooled in ice, the TBA red pigment was extracted with 4 ml of n-butanol, and its absorbance at 532 nm was measured with a Hitachi 200-10 spectrophotometer. The TBA value of the sample was estimated as an MDA equivalent on the basis of the calibration curve for standard MDA. HPLC with a postcolumn reaction system. A schematic diagram of the HPLC with a postcolumn reaction system is presented in Fig. 1. HPLC was performed on an Inertsil ODS-2 (5 km, 6 X 150 mm, Gasukuro Kogyo Co., Ltd., Tokyo) using a linear gradient of methanol in water from 50 to 100% over 30 min, and followed by methanol for 15 min, at a flow rate of 0.7 ml/min as the mobile phase. The eluate was mixed with the TBA reagent in a reaction coil (0.4 mm X 20 m, Toso Co., Ltd., Tokyo, Japan) at 55°C. Although the TBA calorimetric test is usually carried out by boiling at 100°C for 2030 min at pH 2-3 (19), the reaction temperature (55°C) enployed here is the optimum one for a postcolumn reaction system, because methanol was used as the mobile phase for HPLC. The TBA reagent was prepared by dissolving 0.5% TBA in 0.05 N HCl solution. The flow rate of the TBA reagent was 0.7 ml/min. The TBA red pigment that formed in the reaction coil was monitored at excitation and emission maxima of 515 and 553 nm (20) with a Hitachi FlOOO fluorescence spectrophotometer. RESULTS
AND
DISCUSSION
Standard MDA prepared from 1,1,3,3-tetramethoxypropane, which showed an absorption maximum at 267 nm in 10 mM Tris-HCl buffer (pH 7.4), was injected into the HPLC system, and the eluate was reacted with the TBA reagent in a reaction coil at 55°C. A typical chromatogram of the MDA-TBA red pigment that formed
YODEN
An’,-. filYlJ
--IIU TABLE
Extraction Sample
of
TBA-RS Water Thermal (nmol
I MA ML
from
1 Oxidized
phase
MA-HP0 ML-HP0
/L
5
time
(min)
FIG. 2. Elution profile of malondialdehyde. Standard MDA (1 nmol) was injected into the HPLC with a postcolumn reaction system. The MDA-TBA red pigment formed in a reaction coil (0.4 mm X 20 m) at 55°C was monitored at excitation and emission maxima of 515 and 553 nm. The HPLC condition was described under Materials and Methods.
in this postcolumn reaction system is shown in Fig. 2. Under these conditions, the MDA-TBA red pigment monitored at excitation and emission maxima of 515 and 553 nm was detected as a single peak at a retention time of 5.9 min. This chromatogram shows that free MDA was eluted as a single peak at a retention time of 5.9 min using the present HPLC system and formed MDA-TBA red pigment by reaction with TBA. The lowest amount of MDA detected by the HPLC system was 0.5 nmol/ injection, which cannot be detected by absorption at 532 nm. The calibration curve of free MDA was obtained by measuring the fluorescence peak height for the MDATBA red pigment, as shown in Fig. 3.
1.0 Malondialdehyde
FIG. 3. Calibration amounts of standard rescence peak height monitor at excitation
2.0
ML phase
1.07 (3.1%) 0.23 (3.4%)
6.30 (9.8%) 0.70 (8.9%)
32.85 (96.9%) 6.47 (96.6%) * HPO) 58.20 (90.2%) 7.20 (91.1%)
20
15
10 Retention
Chloroform
and
oxidation” MDA/mg)
HP0 degradation (nmol MDA/pm01
Ol0
MA
3.0 (nmol)
curve of malondialdehyde. The indicated MDA were iniected into the HPLC system. Fluoof the MDA-“TBA red pigment was measured by and emission maxima of 515 and 553 nm.
’ Ten milligrams of MA and ML was heated at 100°C for 60 min in an oil bath and was extracted with a chloroform:methanol mixture. *Each HP0 (1 rmol) was degraded by incubation with heme (10 nmol) at 37°C for 30 min and was extracted with a chloroform:methano1 mixture.
We first oxidized MA and ML by heating at 100°C for 60 min and fractionated TBA-RS by chloroform:methano1 extraction. Most of TBA-RS produced from thermally oxidized MA and ML was localized in the chloroform phase, and a little water-soluble TBA-RS was detected in the aqueous phase (Table 1). Similarly, about 10% of TBA-RS in the degradation products of both HP0 by heme was detected in the aqueous phase. This result indicated that the majority of TBA-RS in oxidized lipids was lipophilic compounds and that the amount of water-soluble TBA-RS such as free MDA was small. We therefore attempted to analyze the chloroformand water-soluble TBA-RS in oxidized lipids in vitro using the present HPLC with a postcolumn reaction system. First, the chloroform-soluble TBA-RS in thermally oxidized ML were injected into the HPLC, and the eluate was reacted with the TBA reagent in a reaction coil at 55°C. Elution of chloroform-soluble TBA-RS on the HPLC system revealed the presence of many kinds of TBA red pigment through reaction with TBA reagent (Fig. 4a). The major TBA red pigment was eluted at a retention time of 16.1 min. This peak height showed an increase that paralleled the rise in the TBA value with heating time. On the other hand, no MDA-TBA red pigment was detected on the HPLC at a retention time of 5.9 min. When chloroform-soluble TBA-RS in the degradation products of ML-HP0 prepared by incubation with heme were injected into the same HPLC system, three major TBA red pigments were detected at retention times of 9.5, 10.9, and 16.1 min, and no MDA-TBA red pigment was observed (Fig. 4b). Many kinds of TBA red pigment formed from thermally oxidized ML and the degradation products of ML-HP0 showed the same elution profile on the HPLC. These results indicate that thermal oxidation of ML and degradation of ML-HP0
CHROMATOGRAPHIC
DETERMINATION
OF
THIOBARBITURIC
lo(a) 8. 1 ,
1:.
.-------
-100
.' .. 52.
," ,
2 ,' boo p.i (b) %3
*
50 /-
-------100
/
0
20
10 Retention
30 time
40
(min)
FIG. 4. HPLC of chloroform-soluble TBA-RS in thermally oxidized ML and the degradation products of ML-HPO. Thermally oxidized ML (a) and the degradation products of ML-HP0 (b) were extracted with a chloroform:methanol mixture. The chloroform phase was injected into the HPLC with a postcolumn reaction system. The TBA red pigments were monitored at excitation and emission maxima of 515 and 553 nm. The HPLC condition was described under Materials and Methods.
by heme produce certain kinds of chloroform-soluble TBA-RS other than free MDA and that the TBA value estimated by the TBA calorimetric test in oxidized ML in vitro is completely dependent on some other TBA-RS apart from free MDA. We next separated the chloroform-soluble TBA-RS in the thermally oxidized MA by the same HPLC system. The chloroform-soluble TBA-RS in thermally oxidized MA were found to produce many kinds of TBA red pigment at retention times different from that of free MDA using the postcolumn reaction system (Fig. 5a). Similarly, the degradation products of MA-HP0 by heme were also found to produce many kinds of TBA red pigment at retention times of 6.2,6.9,9.8, and 16.1 min (Fig. 5b). Upon injection of the water-soluble TBA-RS in oxidized MA into the HPLC, a small peak of TBA red pigment was eluted at the same retention time (5.9 min) as that of the standard MDA-TBA red pigment (data not shown). When the compound eluted at 5.9 min was collected from the HPLC without reaction with the TBA reagent, it showed the same uv absorption maximum (267 nm) as that of the standard MDA and also reacted with TBA to produce a red pigment with an absorption maximum at 532 nm. This result indicates that the compound collected from water-soluble TBA-RS by HPLC is free MDA formed from oxidized MA. Although the presence of a small amount of free MDA was confirmed in both the thermally oxidized MA and the degradation products of MA-HPO, free MDA was not the major TBA-RS in either. Thermally oxidized methyl linolenate also produced many kinds of chloroform-soluble
ACID-REACTIVE
SUBSTANCES
119
TBA-RS and a little MDA (data not shown). Seto has recently reported determination of MDA by HPLC using a postlabeling method (21). However, he only measured MDA in meat samples and any other TBA-RS were not detected on HPLC. Rat liver microsomes (22) and liposomes prepared from rat liver phosphatidylcholine (23) are known to release TBA-RS, which may contain water-soluble compounds, into the medium from their membranes during lipid peroxidation. When rat liver microsomes were oxidized by incubation in an NADPH-generation system, TBA-RS in the medium were markedly increased, and TBA-RS in the medium were extracted with a chloroform:methanol (2:l) mixture. The amount of TBA-RS in the aqueous phase was increased about 14-fold in comparison with that of control microsomes. Each aqueous phase was injected into the HPLC system. Elution of water-soluble TBA-RS from the oxidized microsomes yielded free MDA as a single peak on HPLC with a retention time of 5.9 min (Fig. 6a). The MDA content estimated from the peak height was the major part of that estimated by the TBA calorimetric test. On the other hand, control microsomes showed only a small elution peak of MDA. This compound was collected by HPLC and its uv absorption spectrum was determined. As shown in Fig. 6b, the compound collected from watersoluble TBA-RS in oxidized microsomes showed an absorption maximum at around 267 nm. This compound also showed the formation of TBA red pigment with an absorption maximum at 532 nm by reaction with TBA. From these results, the water-soluble compound eluted at a retention time of 5.9 min was considered to be free
2 0 Retention
time
(min)
FIG. 5. HPLC of chloroform-soluble TBA-RS in thermally oxidized MA and the degradation products of MA-HPO. Thermally oxidized MA (a) and the degradation products of MA-HP0 (b) were extracted with a chloroform:methanol mixture. Chloroform-soluble TBA-RS were injected into the HPLC with a postcolumn reaction system. The TBA red pigments were monitored at excitation and emission maxima of 515 and 553 nm. The HPLC condition was described under Materials and Methods.
120
YODEN
(b)
250
200
Wavelength
300 (nm)
I
LJ!!!!
20
10
Retention
110
lipophilic TBA-RS in oxidized lipids determined by the present HPLC method may play a similar significant role. After collection of TBA-RS from oxidized ML or MA, we may be able to further clarify the significance of each of the TBA-RS. This approach should also shed light on the mechanism of production of TBA-RS during PUFA peroxidation in uitro.
r
0
AND
30
time
40
(min)
FIG. 6.
HPLC and an absorption spectrum of water-soluble TBARS in oxidized microsomes. Rat liver microsomes were oxidized by incubation in an NADPH-generation system and then TBA-RS in the medium were extracted with a chloroform:methanol (2:l) mixture. Water-soluble TBA-RS from the oxidized (-) and control (---) microsomes were injected into the HPLC (a). The HPLC condition was described under Materials and Methods. The TBA-RS eluted at a retention time of 5.9 min were collected by HPLC and its uv absorption spectrum was measured (b).
REFERENCES 1. Frankel, 2. Kappus, Academic
Oil Chem. Sot. 61,1908-1917. Stress (Sies, H., Ed.), pp. 273-310,
3. O’Brien, P. J. (1987) in Autoxidation of Unsaturated Lipids (Chan, H. W.-S., Ed.), pp. 233-281, Academic Press, London. 4. Kikugawa, K., and Beppu, M. (1987) Chem. Phys. Lipids 44,277296. 5. Tsuchida, M., ids 44,297-325.
Miura,
T., and Aibara,
6. Slater, T. F. (1984) in Methods Vol. 105, pp. 283-293, Academic 7. Hirayama, Sci. Food 8. Kosugi,
MDA produced by an NADPH-dependent reaction in oxidized microsomes. It is well known that oxidized PUFA react with TBA to produce TBA red pigments with an absorption maximum at 532 nm. MA and methyl linolenate, which contain three or four double bonds, are considered to produce free MDA during their peroxidation. However, in this study, while oxidized MA produced a small amount of free MDA, no free MDA was produced in oxidized ML. The major TBA-RS present in oxidized lipids in uitro were not free MDA but many kinds of other TBARS solubilized in the lipophilic phase. Some reports have suggested that the MDA contents of oxidized lipids are much lower than those obtained by the TBA colorimetric test (7,9). Hirayama et al. reported that MDA contents measured by reaction with dansyl hydrazine were only 30% of the TBA value for oxidized lipids (7). Also, Lee and Csallany have shown that free and bound MDA levels measured by HPLC are lower than those estimated by the TBA calorimetric test and that some other TBA-RS seem to exist in liver tissue (24). We also detected the presence of the lipophilic TBA-RS in oxidized liposomal membranes and the water-soluble TBA-RS in the medium. We have thus separated free MDA and some major TBA-RS from oxidized ML and MA in vitro using the present HPLC system. Bound-type MDA in biological samples (24) may be able to elute on the HPLC after acid hydrolysis. MDA and monomeric aldehydes have been considered to react with proteins or nucleic acids and to cause various biological effects (25,26). Similarly, certain kinds of
E. N. (1984) J. Amer. H. (1985) in Oxidative Press, London.
T., Yamada, N., Agric. 35,338-344. H., Kato,
K. (1987)
Chem.
in Enzymology Press, London. Nohara,
M.,
T., and Kikugawa,
Phys.
(Packer,
and
Fukui,
K. (1988)
Lip-
L., Ed.), S. (1984)
Lipids
23,
J.
1024-
1031. 9. Kikugawa, 512-521.
K., Kato,
T., and Iwata,
10. Csallany, A. S., Guan, (1984) Anal. Biochem. 11. Esterbauer, Methods Academic 12. Largilliere, 123-126. 13. Frankel, Lipids.
M. D., Manwaring,
E. N., Neff, 22,322-327.
14. Ortiz de Montellano, col. l&128-135.
Melancon,
21. 22.
J. D., and Addis,
S. B. (1988)
R. P., and Mica,
Anal.
E., and Brooks, B. A. (1980)
Sinnhuber,
P. B.
170,
Biochem.
D. D. (1987) Mol.
R. 0. (1966)
Kwon, T. W., and Watts, B. M. (1963) J. Food Iio, T., and Yoden, K. (1988) Lipids 23,65-67.
18. Johnston, J. Amer. 19. Bird, R. (Packer, 20. Yagi, K. 105, pp.
174,
Biochem.
S., and Slater, T. F. (1984) in L., Ed.), Vol. 105, pp. 319-328,
W. E., Selke,
15. Crawford, D. L., Yu, T. C., and Food Chem. 14,182-184. 16. 17.
Anal.
142,277-283.
H., Lang, J., Zadravec, in Enzymology (Packer, Press, London. C., and
A. (1988)
PharmaJ. Agric.
Sci. 28,627-630.
A. E., Zilch, K. T., Selke, E., and Dutton, H. J. (1961) Oil Chem. Sot. 38,367-371. P., and Draper, H. H. (1984) in Methods in Enzymology L., Ed.), Vol. 105, pp. 299-305, Academic Press, London. (1984) in Methods in Enzymology (Packer, L., Ed.), Vol. 328-331, Academic Press, London.
Seto, H. (1987) Eisei Kagaku 33,436-441. Itoh, F., Horie, T., and Awazu, S. (1988) Arch. Biochem. 264,184-191. 23. Shimasaki, H., Ueta, N., Mowri, H., and Inoue, K. (1984) Biophys. Acta 792,123-129. 24. Lee, H-S., and Csallany, A. S. (1987) Lipids 22,104-107.
Biophys. Biochim.
25. Ames, B. N., Hollstein, M. C., and Cathcart, R. (1982) in Lipid Peroxidation in Biology and Medicine (Yagi, K., Ed.), pp. 339351, Academic Press, London. 26. Kikugawa, K. (1986) Adv. Free Radical Biol. Med. 2,389-417.
182,116-120
(1989)
Determination of Thiobarbituric Acid-Reactive Substances in Oxidized Lipids by High-Performance Liquid Chromatography with a Postcolumn Reaction System Kazuaki
Yoden’
and Toshihiro
Showa College of Pharmaceutical
Received
March
Iio
Sciences, l-8, Tsurumaki
5-chome,
Tokyo, Japan
3,1989
A new high-performance liquid chromatography procedure with a postcolumn reaction system for determination of free malondialdehyde (MDA) and other thiobarbituric acid-reactive substances (TBA-RS) in oxidized lipids in vitro has been developed. Using this procedure, both thermally oxidized methyl linoleate and the degradation products of methyl linoleate hydroperoxides revealed many kinds of lipophilic TBARS, but no free MDA was detected on the high-performance liquid chromatography. Similarly, oxidized methyl arachidonate also produced certain kinds of TBA-RS in the lipophilic phase and a small amount of free MDA in the hydrophilic phase. These results indicate that lipophilic TBA-RS produced in oxidized lipids in vitro are major TBA-RS and that the production of free MDA is small, even though the degree of lipid oxidation has previously been estimated as an MDA equivalent measured by the TBA calorimetric test. o ISSSAC~demic Press, Inc.
Lipid peroxides produced from polyunsaturated fatty acids (PUFA)’ in oxidized biological systems and food systems have been considered to show various significant biological or toxic effects both in viuo and in vitro (l-3). For example, lipid peroxides cause accumulation of fluorescent lipofuscin pigments with aging in animal tissues (4,5). Lipid peroxides are generally measured using the calorimetric test by reaction with 2-thiobarbituric acid (TBA) (6). Malondialdehyde (MDA), one of 1 To whom correspondence should be addressed. ’ Abbreviations used: heme, heme methyl ester; HPO, hydroperoxide; MA, methyl arachidonate; MDA, malondialdehyde; ML, methyl linoleate; PUFA, polyunsaturated fatty acids; TBA, 2-thiobarbituric acid; TBA-RS, TBA-reactive substances. 116
Setagaya,
the end products of PUFA peroxidation, reacts with TBA to produce a red pigment with an absorption maximum at 532 nm, and therefore lipid peroxide contents are usually estimated as an MDA equivalent by the TBA value, which has been regarded as an index of the degree of PUFA peroxidation. However, Hirayama et al. have shown the presence of certain TBA red pigments other than MDA-TBA red pigment in oxidized methyl linolenate at 80°C (7). Therefore the TBA calorimetric test is not specific for free MDA, and other compounds may give a positive reaction with TBA to produce TBA red pigments. Kosugi et al. have reported that many monomeric aldehydes are able to react with TBA to produce this characteristic red pigment (8). They reported that MDA content estimated by specific determination of MDA is always lower than that estimated from TBA values (7,9). These findings suggest that the level of lipid peroxides determined by the TBA calorimetric test may express the total amount of many kinds of TBA-reactive substances (TBA-RS), including MDA and other aldehydic breakdown compounds, in oxidized lipids. Certain kinds of TBA-RS other than MDA present in oxidized lipids may also be significantly involved in various reactions. Methods for separation of free MDA or MDA-TBA red pigment by HPLC have been reported recently for the measurement of lipid peroxides in oxidized lipids or biological samples (10-12). However, these methods are specific for determination of free and bound MDA and are not able to detect any other TBA-RS in oxidized lipids. Frankel et al. have recently reported the presence of many kinds of volatile oxidative product, oligomer, and monomer from thermal or metal-catalyzed decomposition of methyl linolenate hydroperoxide in vitro (13). These oxidative products may react with TBA to produce TBA red pigments. In this study, therefore, we de0003~2697/89
$3.00
Copyright 0 1999 by Academic Press, Inc. All rights of reproduction in any form reserved.
CHROMATOGRAPHIC
DETERMINATION
OF
THIOBARBITURIC
ACID-REACTIVE
117
SUBSTANCES
veloped an assay method for the separation and determination of free MDA and certain kinds of TBA-RS in oxidized lipids in vitro employing a new HPLC method with a postcolumn reaction system.
D
TY
R
P
M III
MATERIALS
AND
METHODS
Materials. Methyl arachidonate (MA), methyl linoleate (ML), hemin (type I), and lipoxygenase (type I) were purchased from Sigma Chemical Co. (St. Louis, MO). TBA was purchased from Merck (Darmstadt, FRG). Heme methyl ester (heme) was prepared from hemin by the method of Ortiz de Montellano et al. (14). HPLC-grade methanol (Kant0 Kagaku Co., Ltd., Tokyo) was used for HPLC separation in all experiments. Preparation of standard MDA. MDA was prepared from 1,1,3,3-tetramethoxypropane (Tokyo Kasei Kogyo Co., Ltd., Tokyo) by the method of Crawford et al. (15). 1,1,3,3-Tetramethoxypropane (20 ~1) dissolved in 0.6 ml of 0.1 N HCl was hydrolyzed at 40°C for 30 min, and to this solution, 5.6 ml of water and 6 ml of 0.1 N NaOH solution were added. This MDA solution was stocked at 4°C in a refrigerator. The MDA stock solution was diluted with 10 mM Tris-HCl buffer (pH 7.4) and applied for the measurement of uv absorption spectra and the HPLC analysis. MDA concentration was estimated using a molar absorbance value of 31,800 at 267 nm (16). Thermal oxidation of ML and MA. Ten milligrams of ML or MA was oxidized directly by heating at 100°C for 60 min in an oil bath and then dissolved in 0.1 ml of methanol. Degradation of ML-HP0 or MA-HP0 by heme. HP0 of ML and MA was prepared using lipoxygenase, according to the method previously reported (17). The concentration of each HP0 was estimated using a molar absorbance value of 24,500 at 233 nm (18). Each HP0 (1 pmol) was incubated with heme (10 nmol) in 2 ml of methanol at 37°C for 30 min, and then the incubation mixture was evaporated under an N2 stream. The thermally oxidized MA or ML and the degradation products of each HP0 were fractionated into chloroformand water-soluble compounds by extraction with a chloroform:methanol (2:l) mixture. Both the chloroform and water phases evaporated under an N, stream were dissolved in 0.1 ml of methanol or 10 mM Tris-HCl buffer (pH 7.4) and then injected into the HPLC. Oxidation of rat liver microsomes. Rat liver microsomes were prepared according to the usual method. Livers of Sprague-Dawley rats (6 weeks) were homogenized in 0.15 M KC1 solution using a homogenizer. The microsomal pellet was prepared by centrifugation at 105,OOOg for 60 min. The microsomes (10 mg protein), resuspended in 0.1 M Tris-HCl buffer (pH 7.4), were then oxidized by incubation in an NADPH-generation system in 3 ml of 0.1 M Tris-HCl buffer (pH 7.4) at 37°C for 30
8 FIG. 1.
T
Schematic diagram of the HPLC system. P, pump; ent controller; I, injector; C, column; R, reactor; D, detector; module; S, solvent; T, TBA reagent.
G, gradiM, data
min. Water-soluble TBA-RS in the medium of oxidized microsomes were extracted with a chloroform:methanol (2:l) mixture. Assay of TBA calorimetric test. To an aliquot of each sample, 1 ml of water, 1 ml of 0.1 N HCl, and 2 ml of 0.5% TBA solution were added, and then the mixture was heated at 100°C for 20 min. After the reaction, the mixture was cooled in ice, the TBA red pigment was extracted with 4 ml of n-butanol, and its absorbance at 532 nm was measured with a Hitachi 200-10 spectrophotometer. The TBA value of the sample was estimated as an MDA equivalent on the basis of the calibration curve for standard MDA. HPLC with a postcolumn reaction system. A schematic diagram of the HPLC with a postcolumn reaction system is presented in Fig. 1. HPLC was performed on an Inertsil ODS-2 (5 km, 6 X 150 mm, Gasukuro Kogyo Co., Ltd., Tokyo) using a linear gradient of methanol in water from 50 to 100% over 30 min, and followed by methanol for 15 min, at a flow rate of 0.7 ml/min as the mobile phase. The eluate was mixed with the TBA reagent in a reaction coil (0.4 mm X 20 m, Toso Co., Ltd., Tokyo, Japan) at 55°C. Although the TBA calorimetric test is usually carried out by boiling at 100°C for 2030 min at pH 2-3 (19), the reaction temperature (55°C) enployed here is the optimum one for a postcolumn reaction system, because methanol was used as the mobile phase for HPLC. The TBA reagent was prepared by dissolving 0.5% TBA in 0.05 N HCl solution. The flow rate of the TBA reagent was 0.7 ml/min. The TBA red pigment that formed in the reaction coil was monitored at excitation and emission maxima of 515 and 553 nm (20) with a Hitachi FlOOO fluorescence spectrophotometer. RESULTS
AND
DISCUSSION
Standard MDA prepared from 1,1,3,3-tetramethoxypropane, which showed an absorption maximum at 267 nm in 10 mM Tris-HCl buffer (pH 7.4), was injected into the HPLC system, and the eluate was reacted with the TBA reagent in a reaction coil at 55°C. A typical chromatogram of the MDA-TBA red pigment that formed
YODEN
An’,-. filYlJ
--IIU TABLE
Extraction Sample
of
TBA-RS Water Thermal (nmol
I MA ML
from
1 Oxidized
phase
MA-HP0 ML-HP0
/L
5
time
(min)
FIG. 2. Elution profile of malondialdehyde. Standard MDA (1 nmol) was injected into the HPLC with a postcolumn reaction system. The MDA-TBA red pigment formed in a reaction coil (0.4 mm X 20 m) at 55°C was monitored at excitation and emission maxima of 515 and 553 nm. The HPLC condition was described under Materials and Methods.
in this postcolumn reaction system is shown in Fig. 2. Under these conditions, the MDA-TBA red pigment monitored at excitation and emission maxima of 515 and 553 nm was detected as a single peak at a retention time of 5.9 min. This chromatogram shows that free MDA was eluted as a single peak at a retention time of 5.9 min using the present HPLC system and formed MDA-TBA red pigment by reaction with TBA. The lowest amount of MDA detected by the HPLC system was 0.5 nmol/ injection, which cannot be detected by absorption at 532 nm. The calibration curve of free MDA was obtained by measuring the fluorescence peak height for the MDATBA red pigment, as shown in Fig. 3.
1.0 Malondialdehyde
FIG. 3. Calibration amounts of standard rescence peak height monitor at excitation
2.0
ML phase
1.07 (3.1%) 0.23 (3.4%)
6.30 (9.8%) 0.70 (8.9%)
32.85 (96.9%) 6.47 (96.6%) * HPO) 58.20 (90.2%) 7.20 (91.1%)
20
15
10 Retention
Chloroform
and
oxidation” MDA/mg)
HP0 degradation (nmol MDA/pm01
Ol0
MA
3.0 (nmol)
curve of malondialdehyde. The indicated MDA were iniected into the HPLC system. Fluoof the MDA-“TBA red pigment was measured by and emission maxima of 515 and 553 nm.
’ Ten milligrams of MA and ML was heated at 100°C for 60 min in an oil bath and was extracted with a chloroform:methanol mixture. *Each HP0 (1 rmol) was degraded by incubation with heme (10 nmol) at 37°C for 30 min and was extracted with a chloroform:methano1 mixture.
We first oxidized MA and ML by heating at 100°C for 60 min and fractionated TBA-RS by chloroform:methano1 extraction. Most of TBA-RS produced from thermally oxidized MA and ML was localized in the chloroform phase, and a little water-soluble TBA-RS was detected in the aqueous phase (Table 1). Similarly, about 10% of TBA-RS in the degradation products of both HP0 by heme was detected in the aqueous phase. This result indicated that the majority of TBA-RS in oxidized lipids was lipophilic compounds and that the amount of water-soluble TBA-RS such as free MDA was small. We therefore attempted to analyze the chloroformand water-soluble TBA-RS in oxidized lipids in vitro using the present HPLC with a postcolumn reaction system. First, the chloroform-soluble TBA-RS in thermally oxidized ML were injected into the HPLC, and the eluate was reacted with the TBA reagent in a reaction coil at 55°C. Elution of chloroform-soluble TBA-RS on the HPLC system revealed the presence of many kinds of TBA red pigment through reaction with TBA reagent (Fig. 4a). The major TBA red pigment was eluted at a retention time of 16.1 min. This peak height showed an increase that paralleled the rise in the TBA value with heating time. On the other hand, no MDA-TBA red pigment was detected on the HPLC at a retention time of 5.9 min. When chloroform-soluble TBA-RS in the degradation products of ML-HP0 prepared by incubation with heme were injected into the same HPLC system, three major TBA red pigments were detected at retention times of 9.5, 10.9, and 16.1 min, and no MDA-TBA red pigment was observed (Fig. 4b). Many kinds of TBA red pigment formed from thermally oxidized ML and the degradation products of ML-HP0 showed the same elution profile on the HPLC. These results indicate that thermal oxidation of ML and degradation of ML-HP0
CHROMATOGRAPHIC
DETERMINATION
OF
THIOBARBITURIC
lo(a) 8. 1 ,
1:.
.-------
-100
.' .. 52.
," ,
2 ,' boo p.i (b) %3
*
50 /-
-------100
/
0
20
10 Retention
30 time
40
(min)
FIG. 4. HPLC of chloroform-soluble TBA-RS in thermally oxidized ML and the degradation products of ML-HPO. Thermally oxidized ML (a) and the degradation products of ML-HP0 (b) were extracted with a chloroform:methanol mixture. The chloroform phase was injected into the HPLC with a postcolumn reaction system. The TBA red pigments were monitored at excitation and emission maxima of 515 and 553 nm. The HPLC condition was described under Materials and Methods.
by heme produce certain kinds of chloroform-soluble TBA-RS other than free MDA and that the TBA value estimated by the TBA calorimetric test in oxidized ML in vitro is completely dependent on some other TBA-RS apart from free MDA. We next separated the chloroform-soluble TBA-RS in the thermally oxidized MA by the same HPLC system. The chloroform-soluble TBA-RS in thermally oxidized MA were found to produce many kinds of TBA red pigment at retention times different from that of free MDA using the postcolumn reaction system (Fig. 5a). Similarly, the degradation products of MA-HP0 by heme were also found to produce many kinds of TBA red pigment at retention times of 6.2,6.9,9.8, and 16.1 min (Fig. 5b). Upon injection of the water-soluble TBA-RS in oxidized MA into the HPLC, a small peak of TBA red pigment was eluted at the same retention time (5.9 min) as that of the standard MDA-TBA red pigment (data not shown). When the compound eluted at 5.9 min was collected from the HPLC without reaction with the TBA reagent, it showed the same uv absorption maximum (267 nm) as that of the standard MDA and also reacted with TBA to produce a red pigment with an absorption maximum at 532 nm. This result indicates that the compound collected from water-soluble TBA-RS by HPLC is free MDA formed from oxidized MA. Although the presence of a small amount of free MDA was confirmed in both the thermally oxidized MA and the degradation products of MA-HPO, free MDA was not the major TBA-RS in either. Thermally oxidized methyl linolenate also produced many kinds of chloroform-soluble
ACID-REACTIVE
SUBSTANCES
119
TBA-RS and a little MDA (data not shown). Seto has recently reported determination of MDA by HPLC using a postlabeling method (21). However, he only measured MDA in meat samples and any other TBA-RS were not detected on HPLC. Rat liver microsomes (22) and liposomes prepared from rat liver phosphatidylcholine (23) are known to release TBA-RS, which may contain water-soluble compounds, into the medium from their membranes during lipid peroxidation. When rat liver microsomes were oxidized by incubation in an NADPH-generation system, TBA-RS in the medium were markedly increased, and TBA-RS in the medium were extracted with a chloroform:methanol (2:l) mixture. The amount of TBA-RS in the aqueous phase was increased about 14-fold in comparison with that of control microsomes. Each aqueous phase was injected into the HPLC system. Elution of water-soluble TBA-RS from the oxidized microsomes yielded free MDA as a single peak on HPLC with a retention time of 5.9 min (Fig. 6a). The MDA content estimated from the peak height was the major part of that estimated by the TBA calorimetric test. On the other hand, control microsomes showed only a small elution peak of MDA. This compound was collected by HPLC and its uv absorption spectrum was determined. As shown in Fig. 6b, the compound collected from watersoluble TBA-RS in oxidized microsomes showed an absorption maximum at around 267 nm. This compound also showed the formation of TBA red pigment with an absorption maximum at 532 nm by reaction with TBA. From these results, the water-soluble compound eluted at a retention time of 5.9 min was considered to be free
2 0 Retention
time
(min)
FIG. 5. HPLC of chloroform-soluble TBA-RS in thermally oxidized MA and the degradation products of MA-HPO. Thermally oxidized MA (a) and the degradation products of MA-HP0 (b) were extracted with a chloroform:methanol mixture. Chloroform-soluble TBA-RS were injected into the HPLC with a postcolumn reaction system. The TBA red pigments were monitored at excitation and emission maxima of 515 and 553 nm. The HPLC condition was described under Materials and Methods.
120
YODEN
(b)
250
200
Wavelength
300 (nm)
I
LJ!!!!
20
10
Retention
110
lipophilic TBA-RS in oxidized lipids determined by the present HPLC method may play a similar significant role. After collection of TBA-RS from oxidized ML or MA, we may be able to further clarify the significance of each of the TBA-RS. This approach should also shed light on the mechanism of production of TBA-RS during PUFA peroxidation in uitro.
r
0
AND
30
time
40
(min)
FIG. 6.
HPLC and an absorption spectrum of water-soluble TBARS in oxidized microsomes. Rat liver microsomes were oxidized by incubation in an NADPH-generation system and then TBA-RS in the medium were extracted with a chloroform:methanol (2:l) mixture. Water-soluble TBA-RS from the oxidized (-) and control (---) microsomes were injected into the HPLC (a). The HPLC condition was described under Materials and Methods. The TBA-RS eluted at a retention time of 5.9 min were collected by HPLC and its uv absorption spectrum was measured (b).
REFERENCES 1. Frankel, 2. Kappus, Academic
Oil Chem. Sot. 61,1908-1917. Stress (Sies, H., Ed.), pp. 273-310,
3. O’Brien, P. J. (1987) in Autoxidation of Unsaturated Lipids (Chan, H. W.-S., Ed.), pp. 233-281, Academic Press, London. 4. Kikugawa, K., and Beppu, M. (1987) Chem. Phys. Lipids 44,277296. 5. Tsuchida, M., ids 44,297-325.
Miura,
T., and Aibara,
6. Slater, T. F. (1984) in Methods Vol. 105, pp. 283-293, Academic 7. Hirayama, Sci. Food 8. Kosugi,
MDA produced by an NADPH-dependent reaction in oxidized microsomes. It is well known that oxidized PUFA react with TBA to produce TBA red pigments with an absorption maximum at 532 nm. MA and methyl linolenate, which contain three or four double bonds, are considered to produce free MDA during their peroxidation. However, in this study, while oxidized MA produced a small amount of free MDA, no free MDA was produced in oxidized ML. The major TBA-RS present in oxidized lipids in uitro were not free MDA but many kinds of other TBARS solubilized in the lipophilic phase. Some reports have suggested that the MDA contents of oxidized lipids are much lower than those obtained by the TBA colorimetric test (7,9). Hirayama et al. reported that MDA contents measured by reaction with dansyl hydrazine were only 30% of the TBA value for oxidized lipids (7). Also, Lee and Csallany have shown that free and bound MDA levels measured by HPLC are lower than those estimated by the TBA calorimetric test and that some other TBA-RS seem to exist in liver tissue (24). We also detected the presence of the lipophilic TBA-RS in oxidized liposomal membranes and the water-soluble TBA-RS in the medium. We have thus separated free MDA and some major TBA-RS from oxidized ML and MA in vitro using the present HPLC system. Bound-type MDA in biological samples (24) may be able to elute on the HPLC after acid hydrolysis. MDA and monomeric aldehydes have been considered to react with proteins or nucleic acids and to cause various biological effects (25,26). Similarly, certain kinds of
E. N. (1984) J. Amer. H. (1985) in Oxidative Press, London.
T., Yamada, N., Agric. 35,338-344. H., Kato,
K. (1987)
Chem.
in Enzymology Press, London. Nohara,
M.,
T., and Kikugawa,
Phys.
(Packer,
and
Fukui,
K. (1988)
Lip-
L., Ed.), S. (1984)
Lipids
23,
J.
1024-
1031. 9. Kikugawa, 512-521.
K., Kato,
T., and Iwata,
10. Csallany, A. S., Guan, (1984) Anal. Biochem. 11. Esterbauer, Methods Academic 12. Largilliere, 123-126. 13. Frankel, Lipids.
M. D., Manwaring,
E. N., Neff, 22,322-327.
14. Ortiz de Montellano, col. l&128-135.
Melancon,
21. 22.
J. D., and Addis,
S. B. (1988)
R. P., and Mica,
Anal.
E., and Brooks, B. A. (1980)
Sinnhuber,
P. B.
170,
Biochem.
D. D. (1987) Mol.
R. 0. (1966)
Kwon, T. W., and Watts, B. M. (1963) J. Food Iio, T., and Yoden, K. (1988) Lipids 23,65-67.
18. Johnston, J. Amer. 19. Bird, R. (Packer, 20. Yagi, K. 105, pp.
174,
Biochem.
S., and Slater, T. F. (1984) in L., Ed.), Vol. 105, pp. 319-328,
W. E., Selke,
15. Crawford, D. L., Yu, T. C., and Food Chem. 14,182-184. 16. 17.
Anal.
142,277-283.
H., Lang, J., Zadravec, in Enzymology (Packer, Press, London. C., and
A. (1988)
PharmaJ. Agric.
Sci. 28,627-630.
A. E., Zilch, K. T., Selke, E., and Dutton, H. J. (1961) Oil Chem. Sot. 38,367-371. P., and Draper, H. H. (1984) in Methods in Enzymology L., Ed.), Vol. 105, pp. 299-305, Academic Press, London. (1984) in Methods in Enzymology (Packer, L., Ed.), Vol. 328-331, Academic Press, London.
Seto, H. (1987) Eisei Kagaku 33,436-441. Itoh, F., Horie, T., and Awazu, S. (1988) Arch. Biochem. 264,184-191. 23. Shimasaki, H., Ueta, N., Mowri, H., and Inoue, K. (1984) Biophys. Acta 792,123-129. 24. Lee, H-S., and Csallany, A. S. (1987) Lipids 22,104-107.
Biophys. Biochim.
25. Ames, B. N., Hollstein, M. C., and Cathcart, R. (1982) in Lipid Peroxidation in Biology and Medicine (Yagi, K., Ed.), pp. 339351, Academic Press, London. 26. Kikugawa, K. (1986) Adv. Free Radical Biol. Med. 2,389-417.