rated aldehydes, especially 4-hydroxynonenal, may be directly responsible for structural modification of proteins in LDL [9], and the subsequent changes in its ...
149
Biochem. J. (1993) 289, 149-153 (Printed in Great Britain)
Aldehydes from metal ion- and lipoxygenase-induced lipid peroxidation: detection by 'H-n.m.r. spectroscopy John K. LODGE, Sunil U. PATEL and Peter J. SADLER* Depatment of Chemistry, Birkbeck College, University
of
London, Gordon House and Christopher Ingold Laboratories, 29 Gordon Square, London WC1 H OPP, U.K.
The modification of lipoproteins by reactive aldehydes formed via lipid peroxidation is thought to be a key process in the pathogenesis of atherosclerosis. We show that 'H-n.m.r. spectroscopy can readily be used to detect a variety of different aldehydes resulting from peroxidation of liposomes induced by Fenton's
reagent or lipoxygenase, and aldehydes arising from copperinduced reactions of low-density lipoprotein. There is a clear contrast between the major aldehydic products arising from metal-ion- and lipoxygenase-induced reactions.
INTRODUCTION
detection of a variety of different aldehydic products from Fenton- and lipoxygenase-induced reactions of liposomes and from copper treatment of LDL, and compare n.m.r. data with TBA assays. N.m.r. spectroscopy is shown here to be a useful method for investigating aldehyde production during lipid peroxidation.
Oxidation of low-density lipoprotein (LDL) in the artery wall appears to be a key process in the pathogenesis of atherosclerosis [1,2]. Oxidized LDL undergoes accelerated endocytosis by macrophages via their scavenger receptors, leading to cholesterol accumulation and the production of foam cells [1]. Hence there is much current interest in the mechanisms of LDL oxidation. The redox-active transition metals iron and copper may play important roles in these oxidative processes [3,4]. For example, low-molecular-mass iron complexes can be produced in blood plasma as a result of cell damage following myocardial infarction, liberation of myoglobin and attack by H202 [5]. Iron is also present at the active site of cellular lipoxygenases, enzymes which catalyse the oxidation of both esterified and non-esterified fatty acids [6]. Some cells can oxidize LDL via a lipoxygenasedependent pathway and convert LDL into a cytotoxic form in vivo [7]. Oxidation of polyunsaturated lipids is known to give rise to a variety of saturated and unsaturated aldehydes, including malonaldehyde (MDA) and 4-hydroxy-2,3-trans-nonenal [8]. Unsaturated aldehydes, especially 4-hydroxynonenal, may be directly responsible for structural modification of proteins in LDL [9], and the subsequent changes in its recognition properties [1]. Although the thiobarbituric acid (TBA) test has been widely used for detecting and quantifying aldehydes arising from lipid peroxidation, this test is known to be more sensitive to MDA than to many other aldehydes, and much of this MDA may be generated during the harsh acidic conditions of the assay [4,8]. H.p.l.c. methods have been shown to be very effective for the determination of a variety of these aldehydes [8], but prior derivatization and lengthy pre-separation steps are required. The possibility of detecting aldehyde products by n.m.r. spectroscopy is attractive, because little pretreatment is required, and other products may be detected at the same time. Also n.m.r. chemical shifts and coupling constants provide valuable information about molecular structures. Although peroxidation reactions of lipids have been studied previously by 'H-n.m.r. spectroscopy [10-14], it has been reported that aldehydic products are not detectable by this method [13]. In the present paper we describe procedures which readily allow the
MATERIALS AND METHODS Materials L-a-Phosphatidylcholine (PtdCho; type XVI-E, egg yolk, 99%), sodium linoleate (sodium cis-9,cis- 12-octadecadienoate) (LA), cholesterol, L-ascorbic acid, and lipoxygenase (linoleate: oxygen oxidoreductase; EC 1.13.11.12; type 1-B, soybean) were purchased from Sigma (Poole, Dorset, U.K.). TBA and FeCl2,4H20 were obtained from BDH (Poole, Dorset, U.K.), and 1,1,3,3-tetramethoxypropane, trans-2-nonenal, hexanal, acetaldehyde, butylated hydroxytoluene (BHT) and n.m.r. solvents from Aldrich (Gillingham, Dorset, U.K.).
Liposome preparations Multilamellar liposomes were prepared [15], taking care to exclude air and light, by dissolving PtdCho and LA (2: 1, w/w) in chloroform/methanol (- 1: 1, v/v). The solution was evaporated to dryness in vacuo, and after removal of residual solvent with a stream of N2 gas, the lipids were swelled in the appropriate buffer (either 0.2 M phosphate/0.15 M NaCl, pH 7.4, or 0.1 M borate/0. 1 M KCl, pH 8.9) to give a liposomal suspension containing 5 mg of lipid/ml. These suspensions were subjected to at least five cycles of freezing and thawing [15], and were stored frozen at -20 °C until use, when they were thawed at ambient temperature. In some preparations cholesterol was incorporated at a similar level to that found in LDL (40 %O, w/w, relative to PtdCho).
Peroxidation reactions Fenton's reagent Microlitre aliquots of freshly-prepared stock solutions containing H20 FeCl2 and ascorbic acid were added to 0.5 ml of liposome
Abbreviations used: BHT, butylated hydroxytoluene; COSY, shift-correlated n.m.r. spectroscopy; HA, hexanal; HDL, high-density lipoprotein; LA, linoleic acid (cis-9,cis-12-octadecadienoic acid); LDL, low-density lipoprotein; MDA, malon(di)aldehyde; PtdCho, L-a-phosphatidylcholine; TBA, thiobarbituric acid; TMP, 1,1,3,3-tetramethoxypropane; TMS, tetramethylsilane; TNA, trans-2-nonenal. * To whom correspondence should be addressed.
150
J. K. Lodge, S. U. Patel and P. J. Sadler
suspension in buffer, giving final concentrations of 0.77 mM Fe, 0.57 mM H202 and 0.39 mM ascorbate. This was mixed thoroughly and aerated using a Pasteur pipette (for about 10 min) and then incubated at ambient temperature (- 23 °C) for various times.
chloroform (7.262 p.p.m.). The concentrations of aldehydes were determined by comparing peak areas with that of acetaldehyde as an external or an added internal standard (100 ,uM).
TBA tests
An aliquot of a fresh stock solution of lipoxygenase (30 mg/ml in 0.1 M borate buffer, pH 8.9) was added to 0.5 ml of liposome suspension, giving a final enzyme concentration of 200 units/ ,tmol of LA. This was shaken and aerated (with Pasteur pipette for about 10 min) and incubated. For control experiments, additions of buffer alone were made.
TBA-reactive substances were determined by using the same liposome suspensions as those used for the n.m.r. work. The procedure was that of Kuzuya et al. [18]. To 0.1 ml of the liposome suspension was added 1.5 ml of 20 % trichloracetic acid and 1.5 ml of 0.67 % TBA (w/v in water), and BHT (final concn. 50 ,#M). This was incubated for 20 min at 97 °C, cooled, centrifuged (5 min, 500 g) and the A532 of the supernatant was recorded. 1,1,3,3-Tetramethoxypropane (TMP) was used as a standard, with £532 1.56 x 105 M-'.cm-' [19].
Extractions
RESULTS AND DISCUSSION
We used a modification of the method of Folch et al. [7]. A solution of ['H]chloroform/['H]methanol (3: 1, v/v) was carefully layered on to the liposome suspension at various times after peroxidation reactions. Disodium EDTA solution (10 ltl of a 100 mM solution; final EDTA concentration 1.8 mM) was added to the reactions involving iron and copper salts to chelate weakly bound metals. In the case of lipoxygenase, the enzyme denatured at the solvent interface. The tube was shaken vigorously (by using a vortex mixer), and the contents were allowed to settle at 4 °C for I1 Omin. After centrifugation (5000 rev./min; 2700 g; 10 min), the top aqueous layer was carefully removed and 600 ,ul of the lower ['Hichloroform layer was transferred to a 5 mm n.m.r. tube.
Initially we studied 'H n.m.r. spectra of intact PtdCho/LA and PtdCho/LA/cholesterol multilamellar liposomes after treatment with Fenton's reagent and soybean lipoxygenase. However, the spectra were relatively broad, owing to the restricted mobility of the components, and it was not possible to observe resonances for oxidation products. Therefore reactions were followed by taking extracts into ['H]chloroform/[2H]methanol (3: 1, v/v) at various times. Reactions at pH 7.4 and pH 8.9 were studied, the latter being close to the optimum pH for enzymic activity of soybean lipoxygenase [6]. Typical spectra of the aliphatic and olefinic regions of extracts from PtdCho/LA liposomes are shown in Figure 1. For both lipoxygenase and Fenton's reagent, there is a selective decrease in intensity of peaks for allylic and bisallylic methylenes of LA at 2.03 and 2.75 p.p.m. respectively, and olefinic protons at 5.33 p.p.m. New peaks for hydroperoxides containing conjugated dienes are seen at 6.54-5.42 p.p.m. [10,11]. These are most clearly resolved with lipoxygenase, partly due to the specificity of the enzymic reaction, and partly due to the broadening of peaks with the Fenton reaction, presumably due to extraction of paramagnetic iron, even after treatment with EDTA (i.e. some iron remains tightly bound). The multiplets labelled b (5.56 p.p.m.), c (6.54 p.p.m.) and d (6.00 p.p.m.) with associated coupling constants 3J(Hb -H.) 8.0 Hz, 3J(Hb- H) 15.1 Hz, 3J(H, - Hd) 11.2 Hz, and 3J(Hd- He) 11.2 Hz are prominent, and assignable to the conjugated diene 13(S)-hydroperoxy-cis-9,trans-11octadecadienoic acid. The peaks for protons a and e are overlapped by other peaks, but are seen in two-dimensional COSY spectra (Figure 2b) or spectra obtained at higher pH (8.9). Minor peaks from the isomer 9(RIS)-hydroperoxy-trans-IO,cis12-octadecadienoic acid are also seen at 6.25 (c') and 5.90 (d') p.p.m. These octadecadienoic acid derivatives are known products from reactions of lipoxygenase with LA [6]. Reactions of Fenton's reagent and lipoxygenase with lipids both lead initially to the production of hydroperoxides [3], in the former case via non-specific attack of hydroxyl radicals on double bonds, and in the latter case via specific attack on cis,cis- 1 ,4-pentadiene groups. These hydroperoxides are known to fragment, via /J-scission of alkoxyl radicals, to give a variety of aldehydic products [3,20]. Figure 2(a) shows the aldehydic region of 'H n.m.r. spectra of extracts from similar reactions using .PtdCho/LA/cholesterol liposomes, pH 8.9. With lipoxygenase, a prominent doublet at 9.57 p.p.m. (J 8.1 Hz, peak D) and a smaller one at 9.49 p.p.m. (J 8.1 Hz, peak E) are observed. In some reactions with lipoxygenase, under slightly different conditions (e.g. Figure 2b), apparent singlets are also seen at 9.72 (peak B) and 9.65 (peak C) p.p.m. When Fenton's reagent is used, the most prominent
Lipoxygenase
Lipoproteins LDL was isolated from heparinized blood of non-fasted healthy volunteers by using f.p.l.c. [16,16a] (Pharmacia, two Superose 6 HR 10/30 columns in series eluted with 0.1 M phosphate/0. 15 M NaCl/0.03 % NaN, pH 7.4). Fractions were analysed for their cholesterol, triacylglycerol and protein contents, and those corresponding to LDL were further concentrated with Centricon filters (Amicon, Stonehouse, Glos., U.K.; cut-off 5 kDa), giving a final protein concentration of 1.9 mg/ml. These samples were stored at 4 °C and used within 10 days. Solutions of LDL were incubated at 37 °C for 24 h after addition of 5 ,1 of water (control) or 5 u1 of 0.1 M CuSO4 solution (final Cu concentration 0.8 mM; Cu/apoB molar ratio 218:1) to initiate copper-catalysed decomposition. Extractions using f2H]chloroform/[2H]methanol were then carried out as described above.
N.m.r. spectroscopy 'H-n.m.r. spectra (400 MHz) were recorded on a Bruker AM400 spectrometer (MRC Biomedical N.m.r. Centre, Mill Hill, London, U. K.), typically using 0.6 ml of solution in a 5 mm tube at 298 K. Typical pulsing conditions were: 256 transients, 550 pulses, relaxation delay of 1.4 s, 16384 data points, acquisition time 1.36 s, spectral width 60(24 Hz. Two-dimensional shiftcorrelated n.m.r. (COSY) spectra were recorded using the standard sequence [17], with 2048 data points in the t2 dimension, 512 increments of tp, relaxation delay 1.2 s, and 64 transients. Exponential functions equivalent to line-broadenings of 0.5-2 Hz were used for processing. The chemical shift reference was tetramethylsilane (TMS) (O p.p.m., internal) usually via residual
'H-n.m.r. detection of aldehydes from lipid peroxidation
151
HOO cis
C
a
b
d
CH2C= C e
trans
NMe3+ CH2CO
Fenton
x
x4
=CCH2C
~
~
~
CH2|P
Lipoxygenase a
I -fLA
CH20COR
CH=CH
Control
CH2N
POCH2
x4
CHCO
I
6.5
Figure
1
'H-n.m.r. spectra
6.0
of
5.5
1
I
5.0 Chemical shift (b) (p.p.m.)
extracts from liposomes treated with Fenton's reagent
or
4.0
3.0
2.0
lipoxygenase
LA/PC liposomes were treated with lipoxygenase (30 min) or Fenton's reagent (24 h) at pH 7.4, and extracted with [2H]chloroform/[2H]methanol (3:1, v/v). For both reactions, decreases in the intensity of peaks for olefinic and adjacent protons are seen. Assignments (peaks a-e) for the major cis-trans oxidation product, 13-hydroperoxy-cis-9,trans-11 -octadecadierioic acid, are shown. Peaks for the trans-cis isomer are also present (c' and d'). The differing specificities of the enzymic and the Fenton reactions are evident, although, for the latter, many peaks are broadened. The aldehydic regions for similar reactions are- shown -n Figure 2. Peak X is unassigned.
aldehydic peak is B, but peaks A (9.76 p.p.m.), D and E are also present.
None of the n.m.r. spectral changes associated with lipid peroxidation were detected in control preparations of fresh liposomes or those treated with either H202, Fe3", Fe2l or Cu2+ ions alone. The presence of cholesterol in liposomes appeared to have little or no effect on the production of aldehydes. In a two-dimensional COSY 'H-n.m.r. spectrum of products from lipoxygenase-treated PtdCho/LA liposomes,the aldehydic peak D had clear connectivities to resonances at 7A7 (doublet of doublets, J 11.4, 15.1 Hz) and 6.10 p.p.m. (Figure 2b). These features are similar to those seen for a standard sample of trans2-nonenal [doublet 9.50 p.p.m. (Figure 2d), J 8.2 Hz; doublet of triplets 6.85 p.p.m., J 6.9, 15.3 Hz; doublet of doublets 6.1 p.p.m., J 15.3, 7.7 Hz], although the shift of the aldehydic proton of the latter is closer to that of peak E, and therefore peak D probably arises from a related aldehyde, such as a hydroxynonenal. On the basis of its shift, peak A can be assigned to an alkanal such as hexanal (9.77 p.p.m., Figure 2d), which has only a small 3J value (1.7 Hz), and therefore may appear as a singlet. Both alkanals and alkenals have previously been reported as
products from lipid peroxidation, as well as hydroxyalkenals and. MDA [3,8]. Peak B has the same shift (9.72 p.p.m.) as a product from the hydrolysis of TMP (Figure 2d), although this is probably not MDA, which is strongly hydrophilic and may not partition into chloroform. Our n.m.r. data show that the hydrolysis of TMP follows a complicated course, and definite assignments of the products cannot be made at present. There are only a few previous reports of n.m.r. data for MDA itself [21,22], which do not provide conclusive assignments. This area is worthy of further n.nir. investigation. Oxidation of LDL induced by Cu2+ is thought to be promoted by the presence of lipoxygenase-derived. hydroperoxides [23] and mediated by Cu2+ complexation to the proteins of LDL 118]. Aldehydic products corresponding to peaks A, B and D were observed by n.m.r. spectroscopy after treatment of LDL with Cu2+ (Figure 2c), resembling the reaction of Fenton's reagent with liposomes (see above). In the high-field region of the spectrum (not shown), there were accompanying decreases in the intensities of resonances for olefinic protons, and allylic and bisallylic methylenes. No peaks for hydroperoxides were detected, suggesting that these species had decomposed after 24 h of
152
_
J. K. Lodge, S. U. Patel and P. J. Sadler 140
I~
6.0-
5.0-
'
9.0
(a) Liposomes
1
Ie : 5.0
(d)
20
-
0
-
600
-
500
-
7 400
-
300
-
100 _ 80 cn < 60-
D
A
-
Liposomes
fb)
9.4 9.8 Chemical shift (3) (p.p.m.)
3 (h)
4
5
Figure 3 Time courses of aldehyde production as determined by (a) TBA and (b) n.m.r. tests Plots of the concentrations of extracted aldehydes as determined by n.m.r. (comparison of areas of peaks A-E in Figure 2 with that of an acetaldehyde standard) versus time, and comparison with assays for TBA-reactive substances (TBARS). For these experiments PtdCho/LA liposomes in borate bufter, pH 8.9, were used and were treated with either Fenton's reagent or lipoxygenase at 23 °C. In this case the control sample contained a small amount of oxidized lipid (aged LA).
LDL
9.6
9.4
Figure 2 Aldehydic region of 1H-n.m.r. spectra of extracts from
liposomes and LDL
(a) LA/PtdCho/cholesterol liposomes (pH 8.9) treated with Fenton's reagent (24 h) or lipoxygenase (30 min), followed by extraction with [2H]chloroform/[2H]methanol. Major aldehydic peaks are labelled A-E (peak C is prominent in some other preparations, e.g. that used for spectrum b). (b) shows cross-peaks for peak D in a two-dimensional COSY spectrum from a similar lipoxygenase reaction (LA/PtdCho liposomes pH 8.9), together with those for some hydroperoxy products (broken lines; see Figure 1). No aldehydic peaks were observed from extracts of lipoxygenase reactions at pH 7.4. (c) shows the effect of Cu2+ treatment (0.8 mM, 24 h, 37 °C) on spectra of extracts of LDL. (d) shows the spectrum of a [2H]chloroform extract from TMP after reaction in 2H20/1H20 (1:3:7, 45 min, ambient temperature), together with those of TNA and HA in [2H]chloroform.
possibility that any MDA which is produced does not partition into [2H]chloroform/[2H]methanol must still be borne in mind. The procedure described here for the detection of aldehydic products from lipid peroxidation reactions by n.m.r. spectroscopy is relatively simple to carry out, and in a short time also provides information about other intermediates, whereas most currently used methods are more limited in the range of products detected [4]. With high-frequency spectrometers (400-600 MHz), detection limits of 50 ,tM are readily attainable. It should now be possible to use n.m.r. spectroscopy to study the attack of a variety of oxidants on the different lipoprotein classes and to gain a deeper understanding of the pathways of hydroperoxide decomposition and aldehyde production. -
We thank CORDA (the heart charity), the Medical Research Council and the Wolfson Foundation for their support for this work. We are grateful to the MRC Biomedical N.m.r. Centre for the provision of n.m.r. facilities.
REFERENCES reaction with Cu2+. Resonances for cholesterol epoxides (not shown) were seen at 0.62 and 0.58 p.p.m. (C18 methyl peaks) as reported previously [13]. In the case of LDL, the total aldehyde concentration in the extract was determined to be 0.31 mM. For liposomes, n.m.r. determinations of aldehydic products were compared with TBA tests. It can be seen from Figure 3 that TBA tests detect only about one-sixth of the aldehydes detected by n.m.r. spectroscopy. This implies that, under our conditions, the major products of the oxidation reactions are mono-alkanals and -alkenals and not dialdehydes such as MDA, in line with the initial n.m.r. assignments. Indeed polyunsaturated fatty acids with three or more double bonds are generally believed to be the major source of MDA [24], and therefore very little would be expected to be produced from linoleic acid. However, the
1 2 3 4 5 6 7 8
Steinbrecher, U. P., Zhang, H. and Lougheed, M. (1990) Free Radicals Biol. Med. 9, 155-168 Steinberg, D., Parthasarathy, S., Carew, T. E., Khoo, J. C. and Witztum, J. D. (1989) New Engl. J. Med. 320, 915-924 Halliwell, B. and Gutteridge, J. M. C. (1989) Free Radicals in Biology and Medicine, Clarendon Press, Oxford Gutteridge, J. M. C. and Halliwell, B. (1990) Trends Biochem. Sci. 15, 129-135 Turner, J. J. O., Rice-Evans, C. A., Davies, M. J. and Newman, E. S. (1990) Biochem. Soc. Trans. 18, 1056-1059 Vick, B. A. and Zimmerman, D. C. (1987) Biochem. Plants 9, 53-90 Folch, J., Lees, M. and Sloane Stanley, G. H. (1957) J. Biol. Chem. 226, 497-509 Esterbauer, H., JOrgens, G., Quenberger, 0. and Koller, E. (1987) J. Lipids Res. 28,
495-509 9 Jurgens, G., Lang, J. and Esterbauer, H. (1986) Biochim. Biophys. Acta 875, 103-114 10 Gardner, H. W. and Weisleder, D. (1972) Lipids 7, 191-193
'H-n.m.r. detection of aldehydes from lipid peroxidation 11 Neff, W. W., Frankel, E. N. and Miyashita, K. (1990) Lipids 25, 33-39 12 Barenghi, L., Bradamante, S., Giudici, G. A. and Vergani, C. (1990) Free Radical Res. Commun. 8, 175-183 13 Bradamante, S., Barenghi, L., Giudici, G. A. and Vergani, C. (1992) Free Radicals Biol. Med. 12, 193-203 14 Pollesello, P., Eriksson, 0., Kvam, B. J., Paoletti, S. and Saris, N.-E. L. (1991) Biochem. Biophys. Res. Commun. 179, 904-911 15 Hope, M. J., Bally, M. B., Mayer, L. D., Janoff, A. S. and Cullis, P. R. (1986) Chem. Phys. Lipids 40, 89-107 16 Cole, T., Kitchens, R., Daugherty, A. and Schonfeld, G. (1988) FPLC Biocommun. 4, 4-6 Received 8 June 1992/29 June 1992; accepted 9 July 1992
153
16a Ha, Y. C. and Barter, P. J. (1985) J. Chromatogr. Biomed. Appl. 341, 154-159 17 Aue, W. P., Bartholdi, E. and Ernst, R. R. (1976) J. Chem. Phys. 64, 2229-2246 18 Kuzuya, M., Yamada, K., Hayashi, T., Funaki, C., Naito, M., Asai, K. and Kuzuya, F. (1992) Biochim. Biophys. Acta 1123, 334-341 19 Wills, E. D. (1969) Biochem. J. 113, 315 20 Kubow, S. (1992) Free Radicals Biol. Med. 13, 63-81 21 George, W. 0. and Mansell, V. G. (1968) J. Chem. Soc. B 132-134 22 Nair, V., Vietti, D. E. and Cooper, C. S. (1981) J. Am. Chem. Soc. 103, 3030-3036 23 O'Leary, V. J., Darley-Usmar, V. M., Russell, L. J. and Stone, D. (1992) Biochem. J. 282, 631-634 24 Pryor, W. A., Stanley, J. P. and Blair, E. (1976) Lipids 11, 370-379
