Reactivity of lysine moieties toward an epoxyhydroxylinoleic acid derivative: Aminolysis versus hydrolysis

Article abstract of DOI:10.1021/jf990383o

Epoxyols are generally accepted as crucial intermediates in lipid oxidation. The reactivity of tertbutyl (9R*,10S*,11E,13S)-9,10-epoxy-13- hydroxy-11-octadecenoate (11a,b) toward lysine moieties is investigated, employing N²-acetyllysine 4-methylcoumar-7-ylamide (12) as a model for protein-bound lysine. The prefixes R* and S* denote the relative configuration at the respective stereogenic centers. Independent synthesis and unequivocal structural characterization are reported for 11a,b, its precursors, and tert-butyl (9R*,10R*,11E,13S)-10-({5-(acetylamino)-6-[(4- methyl-2-oxo-2H-chromen-7-yl)amino]-6-oxohexyl}amino)-9,13-dihydroxy-11- octadecenoate (13a-d). Reactions of 11a,b and 12 in 1-methyl-2-pyrrolidone (MP) and MP/water mixtures at pH 7.4 and 37 °C for 56 days show formation of the aminols 13a-d to be favored by an increased water content. The same trend is observed for hydrolytic cleavage of 11a,b to tert-butyl (E)-9,10,13- trihydroxy-11-octadecenoate (14) and tert-butyl (E)-9,12,13-trihydroxy-10- octadecenoate (15). Under the given conditions, aminolysis proceeds via an S(N)2 substitution, in contrast with the S(N)1 process for hydrolysis. In the MP/water (8:2) incubation, 15.8% of 12 has been transformed to 13a-d and 10.5% of 11a,b hydrolyzed to the regioisomers 14 and 15 after 8 weeks, respectively. Aminolysis of α,β-unsaturated epoxides by lysine moleties therefore is expected to be an important mode of interaction between proteins and lipid oxidation products.

Full text of DOI:10.1021/jf990383o

J. Agric. Food Chem. 1999, 47, 4611−4620
4611
Rea ctivity of Lysin e Moieties tow a r d a n Ep oxyh yd r oxylin oleic Acid
Der iva tive: Am in olysis ver su s Hyd r olysis
Markus O. Lederer,* Axel Schuler, and Marc Ohmenha¨user
Institut fu¨r Lebensmittelchemie (170), Universita¨t Hohenheim,
Garbenstrasse 28, D-70593 Stuttgart, Germany
Epoxyols are generally accepted as crucial intermediates in lipid oxidation. The reactivity of tert-
butyl (9R*,10S*,11E,13S)-9,10-epoxy-13-hydroxy-11-octadecenoate (11a ,b) toward lysine moieties
is investigated, employing N2-acetyllysine 4-methylcoumar-7-ylamide (12) as a model for protein-
bound lysine. The prefixes R* and S* denote the relative configuration at the respective stereogenic
centers. Independent synthesis and unequivocal structural characterization are reported for 11a ,b,
its precursors, and tert-butyl (9R*,10R*,11E,13S)-10-({5-(acetylamino)-6-[(4-methyl-2-oxo-2H-
chromen-7-yl)amino]-6-oxohexyl}amino)-9,13-dihydroxy-11-octadecenoate (13a -d ). Reactions of
11a ,b and 12 in 1-methyl-2-pyrrolidone (MP) and MP/water mixtures at pH 7.4 and 37 °C for 56
days show formation of the aminols 13a -d to be favored by an increased water content. The same
trend is observed for hydrolytic cleavage of 11a ,b to tert-butyl (E)-9,10,13-trihydroxy-11-octadecenoate
(14) and tert-butyl (E)-9,12,13-trihydroxy-10-octadecenoate (15). Under the given conditions,
aminolysis proceeds via an SN2 substitution, in contrast with the SN1 process for hydrolysis. In the
MP/water (8:2) incubation, 15.8% of 12 has been transformed to 13a -d and 10.5% of 11a ,b
hydrolyzed to the regioisomers 14 and 15 after 8 weeks, respectively. Aminolysis of R,â-unsaturated
epoxides by lysine moieties therefore is expected to be an important mode of interaction between
proteins and lipid oxidation products.
Keyw or d s: Lipid peroxidation; lysine model reactions; epoxyhydroxylinoleic acid; epoxide ring
cleavage; aminol formation
INTRODUCTION
9,10-epoxy-13-hydroxy-11-octadecenoic acid (1, see Fig-
ure 1) and (10E,12R*,13S*)-12,13-epoxy-9-hydroxy-10-
octadecenoic acid (2); these products have cis configura-
tion at the epoxide ring. The possibility that the epoxide
is derived from oxygen addition to the (Z)-double bond
γ,δ to the hydroperoxy group was strengthened by both
Graveland (1973) and Heimann et al. (1973). In the case
of the broad bean (Vicia faba L.), studies with 18O2-
labeled 13-LOOH showed that the epoxide ring in 1 is
formed by either inter- or intramolecular oxygen trans-
fer from the hydroperoxide through action of an epoxy-
genase (Hamberg and Hamberg, 1990).
Epoxyols have also been characterized as metabolites
of linoleic acid hydroperoxides, generated by a prostag-
landin endoperoxide synthase from adult or fetal blood
vessels (Funck and Powell, 1983). Analogous structures
have been isolated as decomposition products of 13-
LOOH from a reaction with carp intestinal acetone
powder (Hata et al., 1986). Incubation of arachidonic
acid and its 12-hydroperoxy derivative with a rat lung
preparation (Pace-Asciak et al., 1983) affords (5Z,9E,-
14Z)-11,12-epoxy-8-hydroxy-5,9,14-eicosatrienoic acid
(3) and (5Z,8E,14Z)-11,12-epoxy-10-hydroxy-5,8,14-eico-
satrienoic acid.
Among the products of lipid hydroperoxide decompo-
sition are nonvolatile fatty acid derivatives with oxy-
genated functional groups, such as hydroxyl, ketone,
and epoxide (Gardner, 1989). Formation of epoxyols
from fatty acid hydroperoxides either is metal-catalyzed
(Dix and Marnett, 1985; Hamberg, 1975; Gardner et al.,
1974) or can proceed enzymatically. The nonenzymatic
pathway predominantly leads to epoxides with trans
configuration. This is due to the fact that metal ions or
metalloproteins readily decompose (E)-R,â-(Z)-γ,δ-un-
saturated hydroperoxides by a homolytic route giving
alkoxy radicals which attack the R-carbon to close a
trans-configured epoxide ring with concomitant forma-
tion of an allylic radical. This process may, for example,
finally transform (9Z,11E)-13-hydroperoxy-9,11-octa-
decadienoic acid (13-LOOH) into (10E,12R*,13R*)-12,-
13-epoxy-9-hydroxy-10-octadecenoic acid and (9E,12R*,-
13R*)-12,13-epoxy-11-hydroxy-9-octadecenoic acid. Thus,
13S-LOOH is in part converted by hemoglobin to (10E,-
12S,13S)-12,13-epoxy-9-hydroxy-10-octadecenoic acid
(Hamberg, 1975). Such epoxides derived from alkoxy
radicals are optically active and are formed as diaste-
reoisomers of (9R,S)-hydroxy derivatives (Van Os et al.,
1982). Graveland (1970) unraveled an enzymatic oxida-
tion pathway for linoleic acid in wheat flour/water
suspensions. Successive action of lipoxygenase and a
water-insoluble “factor Y”, which is adsorbed on gluten,
transforms linoleic acid to a mixture of (9R*,10S*,11E)-
Epoxides thus represent crucial intermediates in the
metabolism of polyunsaturated fatty acids, both in
foodstuffs and under physiological conditions. Because
an epoxy function vicinal to an olefinic double bond is
smoothly hydrolyzed, aminolysis, for example, by the
ꢀ-amino function of a lysine moiety, may likewise be
envisaged. In this context, Pokorny et al. (1966) have
reported that both 9,10-epoxystearic acid and the re-
* Author to whom correspondence should be addressed [fax
+49 (0)711 459 4096; e-mail ledererm@uni-hohenheim.de].
10.1021/jf990383o CCC: $18.00 © 1999 American Chemical Society
Published on Web 10/12/1999

4612 J. Agric. Food Chem., Vol. 47, No. 11, 1999
Lederer et al.
Soybean lipoxygenase I (89 mg) was added, a gentle stream
of oxygen was bubbled through the solution via a Pasteur
pipet, and the mixture was stirred for 30 min. Progress of the
reaction was monitored by UV spectroscopy at 234 nm
(absorption maximum of the conjugated diene chromophore
in 6). The mixture was allowed to warm to room temperature,
sodium borohydride (1.2 g, 31.8 mmol) added, and the solution
stirred for 1 h. The pH was adjusted to 3 with HCl (7.7 N),
the emulsion was extracted with diethyl ether (4 × 150 mL),
the organic layer was dried with anhydrous Na2SO4 and
evaporated to dryness, and the residue was purified on a silica
gel column [6 (height) × 3 (i.d., L) cm, hexane/diethyl ether/
glacial acetic acid 83:30:1, v/v/v]. Fractions (6 mL each) were
tested for 7 by TLC (Rf ) 0.12, solvent mixture as above). From
the respective fractions, the solvent was stripped off and the
residue dried in high vacuum, yielding 7 (695 mg, 2.34 mmol,
71.1%). For 1H and 13C NMR (CDCl3) see Table 1. 1H NMR
and UV data are fully consistent with those in references
Gargouri and Legoy (1997) and Maguire et al. (1997); CI-GLC-
MS (silylated compound, program A): tR ) 22.83 min, m/z
(relative intensity) 441 (4, [M + H]+), 425 (16), 351 (100), 261
(33), 243 (20), 173 (20), 98 (18), 73 (53).
(9Z,11E,13S)-13-(Acet oxy)-9,11-oct a d eca d ien oic Acid
(8). Compound 7 (695 mg, 2.34 mmol) was dissolved in
pyridine (1 mL), acetic anhydride (2 mL) was added, and the
solution was kept at room temperature for 17 h. The mixture
was poured on ice (80 mL), stirred for 1 h, and, after pH
adjustment to 3 (1 N H2SO4), extracted with CH2Cl2 (3 × 20
mL). The organic layer was washed with H2SO4 (1 N, 20 mL)
and water (acidified with H2SO4, pH ∼2, 3 × 20 mL), dried
with anhydrous Na2SO4, and evaporated to dryness, and the
residue was dried in high vacuum, yielding 8 (700 mg, 2.07
mmol, 88.2%). For 1H and 13C NMR (CDCl3) see Table 1. CI-
GLC-MS (silylated compound, program A): tR ) 23.92 min,
m/z 351 (100, [M + H]+ - CH3COOH), 279 (21), 261 (63), 243
(38), 177 (12), 75 (13).
ter t-Bu tyl (9Z,11E,13S)-13-(Acetoxy)-9,11-octa d eca d i-
en oa te (9). Compound 8 (700 mg, 2.07 mmol) was reacted with
1,1′-carbonyldiimidazole (335 mg, 2.07 mmol) in dimethylfor-
mamide (DMF; 2.2 mL) at 40 °C for 4 h. tert-Butyl alcohol (313
mg, 4.22 mmol) and DBU (315 mg, 2.07 mmol) were added,
and the reaction mixture was flushed with nitrogen and kept
at 40 °C for 22 h. The mixture was diluted with CH2Cl2 (35
mL) and washed successively with HCl (3 N, 10 mL) and
aqueous K2CO3 (0.72 M, 10 mL). The K2CO3 solution was
extracted with CH2Cl2 (4 × 10 mL), the combined organic
layers were evaporated to dryness, and the residue was
purified on a silica gel column (8 × L 3 cm, hexane/ethyl
acetate 19:1, v/v). Fractions (6 mL each) were tested for 9 by
TLC (Rf ) 0.22, solvent mixture as above). From the respective
fractions the solvent was stripped off and the residue dried in
high vacuum, yielding 9 (185 mg, 0.47 mmol, 22.7%). For 1H
and 13C NMR (CDCl3) see Table 1. CI-GLC-MS (program A):
tR ) 24.20 min, m/z 335 (7, [M + H]+ - CH3COOH), 279 (100),
261 (6), 243 (3).
F igu r e 1. Structural formulas of (9R*,10S*,11E)-9,10-epoxy-
13-hydroxy-11-octadecenoic acid (1), (10E,12R*,13S*)-12,13-
epoxy-9-hydroxy-10-octadecenoic acid (2), (5Z,9E,14Z)-11,12-
epoxy-8-hydroxy-5,9,14-eicosatrienoic acid (3), and (2E,4R*,5R*)-
4,5-epoxy-2-hexen-1-ol (4).
spective methyl ester are bonded to albumin without,
however, characterizing the structure of the addition
product.
Among proteinogenic amino acids, lysine and histi-
dine moieties are the most susceptible to attack by lipid
oxidation products. Nielsen et al. (1985) reported a
significant loss of lysine (up to 71%) and histidine (up
to 57%) in a whey protein/methyl linolenate (C18:3)/water
model system. Steinbrecher (1987) has found that 32%
of the lysine residues of apolipoprotein B are modified
by lipid peroxide decomposition products from human
low-density lipoprotein (LDL). Due to the innate com-
plexity of these processes, few of the products formed
from lipid oxidation derivatives and amino acid moieties
have so far been structurally characterized (Zamora and
Hidalgo, 1995; Zamora et al., 1995; Gardner, 1979).
Recently, we have shown that (2E,4R*,5R*)-4,5-
epoxy-2-hexen-1-ol (4), a structural model for the γ-hy-
droxy-R,â-unsaturated epoxides of linoleic and ara-
chidonic acid, reacts with the ꢀ-amino group of lysine
side chains to form aminol compounds (Lederer, 1996).
We have now synthesized the tert-butyl ester of 1, tert-
butyl (9R*,10S*,11E,13S)-9,10-epoxy-13-hydroxy-11-
octadecenoate (11, see Figure 3) and reacted it with
N2-acetyllysine 4-methylcoumar-7-ylamide (12). For-
mation of the aminol tert-butyl (9R*,10R*,11E,13S)-10-
({5-(acetylamino)-6-[(4-methyl-2-oxo-2H-chromen-7-yl)-
amino]-6-oxohexyl}amino)-9,13-dihydroxy-11-octade-
cenoate (13a -d ) is shown to be a competitive process
to direct hydrolysis of 11.
ter t-Bu tyl (9R*,10S*,11E,13S)-13-(Acetoxy)-9,10-ep oxy-
11-octa d ecen oa te (10a ,b). Compound 9 (185 mg, 0.47 mmol)
was dissolved in diethyl ether (2 mL) and cooled to 0 °C in an
ice bath. m-Chloroperoxobenzoic acid (210 mg, ∼55%, ∼0.67
mmol) was added and the mixture kept at 0 °C for 24 h.
Progress of the reaction was monitored by TLC [mobile phase,
hexane/ethyl acetate 19:7 (v/v); Rf ) 0.90 (9), 0.80 (10a ,b);
detection, spray reagent 5% H2SO4 in anhydrous ethanol/10
min at 120 °C]. The mixture was diluted with diethyl ether
(30 mL) and extracted with aqueous NaOH (2.5 M, 3 × 20
mL) and water (3 × 10 mL); negative starch iodine reaction
in the washing water indicated complete removal of peroxides.
The organic layer was stripped off and the residue dried in
high vacuum, yielding 10a ,b (146 mg, 0.36 mmol, 77%). For
1H and 13C NMR (CDCl3) see Table 1. CI-GLC-MS (program
A): tR ) 26.17 min, m/z 351 (5, [M + H]+ - CH3COOH), 295
(98), 277 (100), 185 (4), 75 (14).
MATERIALS AND METHODS
Ma ter ia ls. Ultrapure water, obtained from a Milli-Q 185
Plus apparatus (Millipore, Eschborn, Germany), HPLC grade
methanol, and HPLC grade acetonitrile were used for LC. For
preparative HPLC, solvents were deaerated by flushing with
helium. Linoleic acid, lipoxygenase I (EC 1.13.11.12; ∼9 units/
mg; 1 unit generates 1 µmol of 13S-LOOH/min) from soybean,
1,1′-carbonyldiimidazole, 1,8-diazabicyclo[5.4.0]-7-undecene
(DBU), m-chloroperoxobenzoic acid, p-toluenesulfonic acid, (-)-
(1R)-menthyl chloroformate, and 1-methyl-2-pyrrolidone were
purchased from Fluka (Neu-Ulm, Germany). Pb(IV) tetraac-
etate was obtained from Merck (Darmstadt, Germany).
P r ep a r a tion of (9Z,11E,13S)-13-Hyd r oxy-9,11-octa d ec-
a d ien oic Acid (7), Accor d in g to th e Meth od of Ma gu ir e
et a l. (1997). Linoleic acid (925 mg, 3.3 mmol) was dissolved
in a sodium borate buffer solution (470 mL; 0.1 M, pH 9.0)
under nitrogen atmosphere and cooled to 0 °C in an ice bath.
ter t-Bu tyl (9R*,10S*,11E,13S)-9,10-Ep oxy-13-h yd r oxy-
11-octa d ecen oa te (11a ,b). Compound 10a ,b (146 mg, 0.36
mmol) was dissolved in MeOH (4.2 mL), aqueous K2CO3 (1.2

Protein−Lipid Oxidation Product Interaction
J. Agric. Food Chem., Vol. 47, No. 11, 1999 4613
Ta ble 1. 1H a n d 13C NMR Da ta of Com p ou n d s 7-11 a n d 14 (0.1 M, in CDCl3)a
7
8/9
10a ,b
11a ,b
14
OH
R1
R2
R3
OH
H
H
CO-CH3
CO-CH3
H
1H NMR δ
C(CH3)3
2-H2
3-H2
(4-7; 15-17)-H14
CO-CH3
8-HA
8-HB
9-H
10-H
11-H
12-H
13-H
14-HA
14-HB
18-H3
-/1.44
2.44/2.20
1.66/≈1.6
1.30 br
2.05
2.17
2.17
5.46
5.94
6.50
5.56
5.28
1.58
1.66
0.88
1.44
2.20
1.44
2.20
1.44
2.20
2.34
1.62
1.31 br
≈1.6
≈1.6
≈1.6
1.30 br
2.05
1.31 br
1.30 br
2.17
2.17
5.44
5.97
6.48
5.66
4.17
1.51
1.57
0.89
≈1.5
≈1.5
3.06
1.45
1.57
3.07
3.41
≈1.45
≈1.45
3.46
3.94
5.71
5.83
4.15
≈1.55
≈1.55
0.89
3.384, 3.385b
5.56, 5.58b
5.54, 5.55b
5.947, 5.950b
4.15
5.83, 5.84b
5.272, 5.266b
≈1.5
≈1.7
0.88
1.52
1.81
0.89
J (Hz)
3J (8-HA/B, 9-H)
3J (9-H, 10-H)
3J (10-H, 11-H)
3J (11-H, 12-H)
3J (12-H, 13-H)
7.7
10.9
11.0
15.2
6.9
7.6
10.9
11.0
15.2
7.6
≈6
4.3
7.0
5.5, 6.7d
4.3
7.5
15.6
6.3
e
6.1
6.5
15.6
5.9
6.5
15.6
6.9, 6.5c
3J (13-H, 14-HA/B
4J (8-HA/B, 10-H)
4J (9-H, 11-H)
4J (10-H, 12-H)
4J (11-H, 13-H)
5J (9-H, 12-H)
)
6.4
1.9
1.3
0.6
6.2
1.7
1.0
0.8
1.1
1.0
6.7
6-7
0.8
0.7
1.2
1.1
1.2
1.1
1.1, 0.9c
13C NMR δ
C(CH3)3
C(CH3)3
C-1
C-2
C-3
CO-CH3
CO-CH3
C-(4-7)
-/28.1
-/79.9
169.6/173.3
35.3/35.6
24.2/25.1
170.5
28.1
79.9
28.1
28.1
80.0
173.4
35.6
80.0
173.3
35.6
179.5
34.0
24.6
173.2
35.6
25.1
25.0
25.1
170.2
21.4
21.3
28.84, 28.88,
28.91, 29.4
27.6
132.8
127.9
125.9
135.7
73.0
37.2
25.1
31.8
22.6
28.8/-, 29.0/29.1 × 2,
29.1/29.2, 29.5
27.7
26.3, 29.0, 29.2 × 2
(26.24,26.26),b
28.9, 29.16, 29.21
27.8
25.5, 28.9,
29.1, 29.3
32.9
75.4
74.6
129.5
136.4
72.1
37.2
25.1
C-8
C-9
27.7
133.6
127.6
128.0
131.0
74.9
34.5
24.8
31.6
58.84, 58.91b
56.14, 56.22b
126.9, 127.3b
134.5
58.90, 58.92b
56.51, 56.55b
124.6
C-10
C-11
C-12
C-13
C-14
C-15
C-16
C-17
C-18
139.7
72.2
73.7, 73.9b
34.22, 34.32b
24.7
31.5
22.5
37.10, 37.18b
25.0
31.7
22.6
14.0
31.7
22.6
14.0
22.5
14.0
14.1
14.0
a
δ, chemical shift for the indicated hydrogen/carbon; J (Hz), coupling constant between the indicated protons. Hydrogen/carbon
assignment has been validated by 1H,1H-COSY, 1H,13C-COSY, and 13C-DEPT techniques. Hydrogen/carbon chemical shifts for the
b
respective two diastereoisomers. c Coupling constants for the respective two diastereoisomers. Coupling constants between 10-H and
d
the diastereotopic protons 8-HA/B.
e Coupling constant cannot be determined, due to overlapping multiplets.
M, 1.8 mL) added, and the mixture vigorously stirred at room
temperature for 4 h. Progress of the reaction was monitored
by TLC [mobile phase: hexane/ethyl acetate 19:7 (v/v); Rf )
0.80 (10a ,b), 0.42 (11a ,b); detection as described for 10a ,b].
The mixture was diluted with water (10 mL) and extracted
with diethyl ether (3 × 30 mL). The organic layer was washed
with water (2 × 30 mL), dried with anhydrous Na2SO4, and
evaporated to dryness, yielding 11a ,b (116 mg, 0.31 mmol,
88.5%). For 1H and 13C NMR (CDCl3) see Table 1. CI-GLC-
MS (silylated compound, program A): 11a , tR ) 25.05 min,
m/z 441 (3, [M + H]+), 439 (10, [M - H]+), 367 (100), 351 (18),
295 (65), 277 (87), 185 (13), 105 (9), 73 (7); 11b, tR ) 25.13
min, m/z 439 (7, [M - H]+), 367 (85), 351 (21), 295 (71), 277
(100), 185 (11), 105 (7), 73 (8).

4614 J. Agric. Food Chem., Vol. 47, No. 11, 1999
Lederer et al.
Ta ble 2. 1H a n d 13C NMR Da ta of Com p ou n d s 13a -d (0.1 M, in DMSO-d 6)a
13a ,b
13c,d
13a ,b
1.35
13c,d
1H NMR δ
C(CH3)3
2-H2
1.36
2.13
1.39
2.15
1.25
3.36
4′-H2
5′-H2
6′-HA
6′-HB
-NHCOCH3
-NHCOCH 3
3′′-H
4′′-CH3
5′′-H
1.35
1.50
2.69
2.52
8.37, 8.38b
1.88
6.26
2.40
1.50
2.60
2.45
8.31
1.85
6.27
2.38
7.72
7.51
8-H2
1.27
9-H
3.33
10-H
11-H
12-H
13-H
14-H2
18-H3
2′-H
2.97, 2.99b
5.32
3.02, 3.03b
5.29
5.68
3.94
1.35
0.80
5.68
3.91
1.40
0.81
4.39
1.70
1.61
7.71
7.53
6′′-H
4.39
7′′-NH
8′′-H
HCOO-
10.62, 10.64b 10.68, 10.70b
7.80
8.34
3′-HA
3′-HB
1.68
1.57
7.81
8.39
J (Hz)
3J (9-H, 10-H)
3J (10-H, 11-H)
7.5
9.2
7.2
9.5
15.5
6.3
7.5
5.6
3J (2′-H, 3′-HB)
3J (5′′-H, 6′′-H)
4J (11-H, 13-H)
4J (3′′-H, 4′′-CH3)
4J (6′′-H, 8′′-H)
8.5
8.7
1.0
1.3
2.1
9.0
8.7
1.0
1.3
2.1
3J (11-H, 12-H) 15.5
3J (12-H, 13-H)
3J (2′-H, NH)
3J (2′-H, 3′-HA)
5.5
7.5
5.7
13C NMR δ
C(CH3)3
C(CH3)3
C-1
27.7
79.2
172.2
34.7
27.7
79.2
172.2
34.7
C-4′
C-5′
C-6′
-NHCOCH3
22.9
25.8
45.1
169.4
22.3
22.9
25.8
44.9
169.5
22.3
C-2
C-(3-7; 15, 16) 24.3, 24.5, 24.8, 28.3 28.6, 28.9, 31.2 24.46, 24.52, 24.7, 28.3 28.6, 28.9, 31.2 -NHCOCH3
C-8
C-9
33.3
70.9
65.05, 65.12b
124.9
140.2
69.81, 69.86 b
37.0
22.0
13.7
33.4
70.5
C-2′′
C-3′′
C-4′′
-CH3 (4′′)
C-4a′′
C-5′′
C-6′′
C-7′′
159.9
112.2
153.6c
17.9
114.9
125.8
115.2
142.2
105.6
153.0c
164.8
159.9
112.2
153.0d
17.9
114.9
125.8
115.2
142.3
105.7
153.6d
165.9
C-10
C-11
C-12
C-13
C-14
C-17
C-18
C-1′
C-2′
C-3′
65.11, 65.16b
124.8
140.7, 141.0b
70.29, 70.33 b
37.0
22.1
13.8
171.8
53.6
C-8′′
C-8a′′
HCOO-
171.7
53.5
31.5
31.5
a
δ, chemical shift for the indicated hydrogen/carbon; J (Hz), coupling constant between the indicated protons. Hydrogen/carbon
assignment has been validated by 1H,1H-COSY, 1H,13C-COSY, and 13C-DEPT techniques. Hydrogen/carbon chemical shifts for the
b
respective two diastereoisomers. c ,dAssignment may have to be reversed.
ter t-Bu tyl (9R*,10R*,11E,13S)-10-({5-(Acetyla m in o)-6-
[(4-m et h yl-2-oxo-2H -ch r om en -7-yl)a m in o]-6-oxoh exyl}-
a m in o)-9,13-d ih yd r oxy-11-octa d ecen oa te (13a-d ). A mix-
ture of 11a ,b (120 mg, 0.33 mmol) and 12 [130 mg, 0.38 mmol;
prepared according to the method of Lederer (1996)] in
1-methyl-2-pyrrolidone (0.5 mL) was kept at 100 °C for 48 h.
The solution was diluted with eluent (5.5 mL) and purified by
preparative HPLC. Fractions at tR ) 13.5 and 14.2 min were
collected and the methanol was removed in vacuo. The aqueous
layer was lyophilized, yielding 13a ,b (42.2 mg, 0.056 mmol,
16.8%) and 13c,d (84.3 mg, 0.111 mmol, 33.7%) as the
respective formate salts. For 1H and 13C NMR (CDCl3) see
Table 2. UV and fluorescence characteristics are identical with
those of compound 2 in Lederer (1996). HPLC-DAD: tR ) 30.87
min (13a ,b), 31.12 min (13c,d ). FAB-HRMS (m-nitrobenzyl
alcohol): m/z 714.4694 [M + H]+ (714.4693, calcd for C40H64
-
N3O8).
Hyd r olysis of 11a ,b a n d P b(IV) Tetr a a ceta te Clea va ge
of th e Resu ltin g Tr iols ter t-Bu tyl (E)-9,10,13-Tr ih yd r oxy-
11-octa d ecen oa te (14) a n d ter t-Bu tyl (E)-9,12,13-Tr ih y-
d r oxy-10-octa d ecen oa te (15). Compound 11a ,b (40 mg, 0.11
mmol) was dissolved in THF (1.6 mL), p-toluenesulfonic acid
(0.1 N, 0.4 mL) added, and the mixture stirred at room
temperature for 1 h. Progress of the reaction was monitored
by TLC [mobile phase, hexane/ethyl acetate 1:1 (v/v); Rf ) 0.78
(11a ,b), 0.15, 0.12 (14, 15); spray reagent, (1) as described for
10a ,b and (2) for specific detection of 14 and 15, (a) 1% Pb-
(Ac)4 in toluene (w/v) and (b) 0.05% fuchsine in MeOH (w/v)].

Protein−Lipid Oxidation Product Interaction
J. Agric. Food Chem., Vol. 47, No. 11, 1999 4615
The mixture was diluted with water (10 mL) and the pH
adjusted to 8 with solid NaHCO3. The solution was extracted
with CH2Cl2 (4 × 10 mL), and the organic phase was dried
with anhydrous Na2SO4 and evaporated to dryness. The
residue was purified on a silica gel column [6 × L 0.8 cm,
hexane/ethyl acetate 1:1 (v/v)]. Fractions (1.5 mL each) were
collected and tested by TLC (see above). Two groups of isomers
could be isolated for 14 and 15; however, no complete separa-
tion of the regioisomers could be achieved. The respective
fractions yielded (I) (Rf ) 0.15, 6.8 mg, 0.018 mmol) and (II)
(Rf ) 0.12, 6.4 mg, 0.017 mmol); both products contained 14
and 15 in 3:1 (I) and 2:1 (II) ratio, respectively. For 1H and
13C NMR (CDCl3) for the major diastereoisomer of 14, data
obtained from product I, see Table 1. CI-GLC-MS (silylated
compounds, program A): 14, tR ) 23.02 min, m/z 603 (5; [M +
H]+), 513 (4), 439 (68), 367 (100), 349 (19), 317 (78), 301 (4),
277 (24), 245 (98), 155 (37), 109 (28), 89 (17), 73 (69); tR ) 23.11
min, m/z 603 (4; [M + H]+), 513 (3), 439 (81), 367 (100), 349
(17), 317 (94), 301 (5), 277 (26), 245 (80), 155 (33), 109 (26), 89
(16), 73 (83); tR ) 23.40 min, m/z 603 (4; [M + H]+), 513 (3),
439 (81), 367 (94), 351 (20), 317 (87), 301 (7), 277 (26), 245
(100), 155 (39), 109 (31), 89 (22), 73 (93). 15, tR ) 23.35 min,
m/z 603 (2; [M + H]+), 513 (4), 439 (82), 429 (11), 413 (10),
367 (71), 357 (25), 349 (15), 317 (68), 277 (18), 245 (63), 173
(38), 155 (17), 105 (13), 83 (25), 73 (100); tR ) 23.50 min, m/z
603 (1; [M + H]+), 513 (4), 439 (89), 429 (30), 413 (15), 367
(70), 357 (39), 349 (15), 317 (33), 277 (24), 245 (22), 173 (50),
155 (15), 105 (15), 83 (36), 73 (100).
Products I and II (6 mg each) were dissolved in diethyl ether
(1 mL), and Pb(Ac)4 (13 mg, 0.025 mmol) was added. The
mixtures were kept at 0 °C for 1 h and at room temperature
for 2 h, diluted with diethyl ether (25 mL), and extracted
successively with water (3 × 5 mL), aqueous NaOH (0.25 M,
5 mL), and water (3 × 5 mL). The organic layer was dried
with anhydrous Na2SO4 and analyzed by GLC-FID (program
B) and GLC-MS (program B). CI-GLC-MS: 16, tR ) 19.80 min,
m/z 157 (100; [M + H]+), 139 (48), 121 (39), 97 (39), 93 (28),
81 (31); 17, tR ) 23.73 min, m/z 155 (100; [M + H]+ - tert-
butyl alcohol), 137 (6), 109 (5); 18, tR ) 6.29 min, m/z 101 (19;
[M + H]+), 83 (100); 19, tR ) 31.51 min, m/z 211 (100; [M +
H]+ - tert-butyl alcohol), 193 (7).
at 250/500 and 63/126 MHz nominal frequency, respectively.
Chemical shifts are given in δ relative to Me4Si (tetrameth-
ylsilane) as internal standard; coupling constants J are given
in hertz. Liquid secondary-ion high-resolution mass spectra
(SIMS-HRMS, analogous to FAB-HRMS) were obtained on a
Finnigan MAT 95 (Bremen, Germany).
P ola r im etr y. Optical rotation was determined on a Perkin-
Elmer 241 polarimeter.
An a lytica l High -P er for m a n ce Liqu id Ch r om a togr a -
p h y (HP LC). The analytical HPLC system comprised an
HP1100 autosampler, HP1100 gradient pump, HP1100 diode
array detector (DAD), and HP1046A fluorescence detector
(FLD) module (Hewlett-Packard, Waldbronn, Germany). For
data acquisition and processing, the HP Chem Station (rev A
04.02) software was used. The column (Bischoff, Leonberg,
Germany) used was a Nucleosil C18, 5 µm, 100 Å (column, 250
× 4 mm; guard column, 10 × 4 mm); flow rate ) 0.8 mL/min;
injection volume ) 10 µL; MeOH/phosphate buffer (0.01 M,
pH 7.4) gradient, % MeOH [t (min)] 5(0), 100(35-40), 5(50-
60); DAD detection wavelengths ) 230 and 326 nm; spectral
bandwidth (SBW) ) 4 nm, ref 500 nm (SBW 100 nm); FLD
λex ) 326 nm, λem ) 390 nm.
P r ep a r a tive HP LC P u r ifica tion . The preparative HPLC
system consisted of a Knauer (Berlin, Germany) 64 liquid
chromatograph combined with an A0293 variable-wavelength
detector and a Kronlab HPLC column (Nucleosil C18, 7 µm,
100 Å, column 250 × 20 mm, guard column 50 × 20 mm); flow
rate ) 10 mL/min; eluent, MeOH/NH4HCOO buffer (0.01 M,
pH 4.0) 80:20 (v/v); injection volume ) 0.75 mL; detection
wavelength ) 230 nm. Solutions were filtered (membrane
filter, 0.45 µm) before preparative HPLC.
Ga s-Liqu id Ch r om a togr a p h y (GLC). GLC-FID was run
on a Carlo Erba (Hofheim, Germany) HRGC 5160 instrument
equipped with an on-column (OC) injector and a flame ioniza-
tion detector (FID). The operating conditions were as follows:
carrier gas, hydrogen; carrier linear velocity, 62 cm/s at 100
°C; oven temperature programs, (A) 100 °C, raised at 10 °C/
min to 200 °C, raised at 3 °C/min to 280 °C with 5 min
isothermal at 280 °C, and (B) 5 min at 35 °C, raised at 8 °C/
min to 270 °C with 5 min isothermal at 270 °C; FID temper-
ature, 300 °C; capillary column, Supelco PTE 5 (Steinheim,
Germany), 30 m × 0.32 mm i.d.; film thickness, 0.3 µm.
GLC-Ma ss Sp ectr om etr y An a lysis. GLC-MS was per-
formed using a Finnigan MAT (Bremen, Germany) Ion Trap
800 equipped with a Perkin-Elmer (U¨ berlingen, Germany)
8420 gas-liquid chromatograph. GC conditions were as fol-
lows: carrier gas, helium; carrier linear velocity, 28 cm/s at
100 °C; oven temperature programs, (A) 150 °C, raised at 6
°C/min to 290 °C with 15 min isothermal at 290 °C, and (B) 5
min at 40 °C, raised at 8 °C/min to 270 °C with 10 min
isothermal at 270 °C; split injector and transfer line temper-
ature, 290 °C, split ratio, 1:30; capillary column, PVMS 54
(Perkin-Elmer), 30 m × 0.25 mm i.d.; film thickness, 0.25 µm.
MS conditions were as follows: positive MeOH-CI mode; ion
source temperature, 220 °C.
In cu b a t ion s of ter t-Bu t yl (9R*,10S*,11E,13S)-9,10-
E p oxy-13-h yd r oxy-11-oct a d ecen oa t e (11a ,b ) w it h N2-
Acetyllysin e 4-Meth ylcou m a r -7-yla m id e (12) a t p H 7.4
a n d 37 °C. Compound 12 (70 mg, 0.20 mmol) was dissolved
in water (10 mL), the pH adjusted to 7.4 with H3PO4 (0.1 N),
and the solution lyophilized, yielding 12‚H3PO4 (87 mg).
Compounds 12‚H3PO4 (11.1 mg; 0.025 mmol) and 11a ,b (18.4
mg; 0.05 mmol) were dissolved in (A) 1-methyl-2-pyrrolidone
(MP), (B) MP/water (9:1, v/v), (C) MP/water (8:2, v/v), and (D)
MP/water (1:1, v/v) (0.5 mL each) and the mixtures kept at 37
°C under nitrogen atmosphere for 56 days; incubation D was
stirred vigorously. In regular intervals aliquots (5 µL) of each
incubation were diluted with MeOH (1 mL) and 10 µL was
injected into the HPLC system.
Qu a n tifica tion of ter t-Bu tyl (E)-9,10,13-Tr ih yd r oxy-11-
octadecen oate (14) an d ter t-Bu tyl (E)-9,12,13-Tr ih ydr oxy-
10-octa d ecen oa te (15) in In cu ba tion s A-D. Incubations
A-D were diluted with water (10 mL) and extracted with CH2-
Cl2 (3 × 10 mL). The solutions were filled to a final volume of
50 mL with CH2Cl2 and divided into two portions, and the
solvent was stripped off. N,O-Bis(trimethylsilyl)acetamide
(BSA, 150 µL) was added to one aliquot, the mixture was kept
at room temperature for 2 h, and the volume was filled up to
100 mL with toluene. One microliter was injected into the
GLC-FID system (program A). The other aliquot was dissolved
in diethyl ether (1 mL), Pb(Ac)4 (20 mg, 0.038 mmol) was
added, and the solution was kept at 0 °C for 1 h and at room
temperature for 2 h and diluted with diethyl ether (25 mL).
Further workup and analysis of compounds 16-19 followed
the procedure given above.
Liqu id Ch r om a togr a p h y. Silica gel 60 F254 (Merck, Darm-
stadt, Germany) was used for thin-layer chromatography
(TLC) and silica gel (63-200 µm) (Baker, Gross-Gerau,
Germany) for column chromatography.
Lyop h iliza tion . A Leybold-Heraeus (Ko¨ln, Germany) Lyo-
vac GT 2 was applied.
RESULTS AND DISCUSSION
Because no independent synthesis is reported for
either (9R*,10S*,11E)-9,10-epoxy-13-hydroxy-11-octa-
decenoic acid (1) or its tert-butyl ester 11 in the
literature, we elaborated such a protocol, starting from
linoleic acid (5, see Figure 2). Compound 5 was oxidized
by soybean lipoxygenase I to (9Z,11E,13S)-13-hydro-
peroxy-9,11-octadecadienoic acid (6, 13S-LOOH) and
reduced in situ with sodium borohydride to (9Z,11E,-
13S)-13-hydroxy-9,11-octadecadienoic acid (7), which is
also designated (+)-coriolic acid (Maguire et al., 1991,
Sp ectr a . Ultraviolet (UV) spectra were measured with a
1
Perkin-Elmer Lambda 2 (U¨ berlingen, Germany). H and 13C
nuclear magnetic resonance (NMR) spectra were recorded on
Bruker (Karlsruhe, Germany) AC-250/ARX-500 spectrometers

4616 J. Agric. Food Chem., Vol. 47, No. 11, 1999
Lederer et al.
F igu r e 3. Reaction scheme for the formation of tert-butyl
(9R*,10R*,11E,13S)-10-({5-(acetylamino)-6-[(4-methyl-2-oxo-
2H-chromen-7-yl)amino]-6-oxohexyl}amino)-9,13-dihydroxy-
11-octadecenoate (13a -d ) from tert-butyl (9R*,10S*,11E,13S)-
9,10-epoxy-13-hydroxy-11-octadecenoate (11a ,b) and N2-
acetyllysine 4-methylcoumar-7-ylamide (12).
(9R*,10S*,11E,13S)-13-(acetoxy)-9,10-epoxy-11-octade-
cenoate (10) as a pair of diastereoisomers 10a and 10b.
Because the regioisomeric tert-butyl (9Z,11R*,12R*,-
13S)-13-(acetoxy)-11,12-epoxy-9-octadecenoate was for-
med only in negligible amounts, no chromatographic
purification of 10a ,b on silica gel was necessary; this
slightly acidic stationary phase causes partial hydrolysis
of the epoxide ring and substantially decreases the yield
of 10. The acetate group was cleaved with potassium
carbonate in methanol/water, and the structure of tert-
butyl (9R*,10S*,11E,13S)-9,10-epoxy-13-hydroxy-11-oc-
tadecenoate (11a ,b), as for all precursors, was unequivo-
cally characterized by 1H, 13C NMR, and MS (see below).
For an in-depth study of the reaction of 11 with lysine
moieties, we prepared N2-acetyllysine 4-methylcoumar-
7-ylamide (12). This lysine derivative incorporates a
fluorophore and thus allows direct and sensitive HPLC
monitoring of the reaction (Lederer, 1996). The synthe-
sis of tert-butyl (9R*,10R*,11E,13S)-10-({5-(acetylamino)-
6-[(4-methyl-2-oxo-2H-chromen-7-yl)amino]-6-oxohexyl}-
amino)-9,13-dihydroxy-11-octadecenoate (13a -d , see
Figure 3) from 11 and 12 was carried out at 100 °C in
1-methyl-2-pyrrolidone. The progress of the reaction was
monitored by HPLC, showing almost quantitative turn-
over within 48 h. Two products could be isolated by
preparative HPLC in a 1:2 ratio. FAB-HRMS analysis
gave a quasimolecular ion ([M + H]+) at m/z 714.4694
for both compounds, corresponding to an elemental
composition of C40H64N3O8. Thus, the products obtained
represent 1:1 adducts of 11 and 12. The structure of the
isomers was definitely characterized as 13a -d ; that is,
four diastereoisomers have been formed from which two
F igu r e 2. Reaction pathway for the synthesis of tert-butyl
(9R*,10S*,11E,13S)-9,10-epoxy-13-hydroxy-11-octadecenoate
(11a ,b).
1997; Gargouri and Legoy, 1997). The 13-OH group of
7 was acetylated to avoid interference in the subsequent
esterification step and also to direct the epoxidation to
the γ,δ-double bond relative to C-13. Such a derivati-
zation procedure has already been successfully em-
ployed for the synthesis of (2E,4R*,5R*)-4,5-epoxy-2-
hexen-1-ol (4) from (2E,4E)-2,4-hexadien-1-ol (Lederer,
1996). (9Z,11E,13S)-13-(Acetoxy)-9,11-octadecadienoic
acid (8) was transformed to its tert-butyl ester 9 accord-
ing to a method by Ohta et al. (1982). The tert-butyl
group was chosen for protection of the COOH function
because this ester is stable under mildly basic conditions
as well as toward aminolysis. It is thus unaffected
during saponification of the 13-acetoxy protective group,
and amide formation with lysine side chains can be
definitely ruled out. Epoxidation of 9 with 3-chloroper-
oxobenzoic acid (Prilezhaev reaction) yielded tert-butyl

Protein−Lipid Oxidation Product Interaction
J. Agric. Food Chem., Vol. 47, No. 11, 1999 4617
pairs can be separated chromatographically (see Figure
3).
As shown for the formation of N2-acetyl-N6-[(E)-1,5-
dihydroxyethyl-2-hexen-4-yl]lysine 4-methylcoumar-7-
ylamide from 4 and 12 (Lederer, 1996), nucleophilic
attack of the amine to the epoxide occurs exclusively in
allylic position and strictly follows an SN2 mechanism.
This also seems to hold for the reaction of 11a ,b and
12. An SN1 process, as a mechanistic alternative, would
proceed via the allylic carbocation, which then could be
attacked by the amine at either C-10 or C-12. This
pathway would by necessity yield two regioisomers,
probably with a ratio far from 1:1. Because not even
traces of positional isomers could be detected for 13, an
SN1 mechanism can definitely be excluded. The absence
of a compound resulting from an allylic shift likewise
rules out an SN2′ mechanism as the third possible route.
The preference for nucleophilic attack at C-10 rather
than at C-9 is probably due to resonance stabilization
of the SN2 transition state vicinal to a double bond. With
the SN2 mechanism established, formation of the four
diastereoisomeric structures 13a -d can be rationalized
only as follows: The stereochemistry of the oxirane ring
in 11 has unequivocally been proven as cis; the relative
configuration consequently must be 9R*/10S*. In the
course of an SN2 process, with Walden inversion, the
ring cleavage products 13a -d then are formed as 9R*/
10R*; that is, the two possible absolute configurations
are enantiotopic. If the other two stereogenic centers
in 13a -d at C-13 and C-2′ (for the numbering see Table
2) were both (S)-configured, one pair of diastereoisomers
only would result. Therefore, racemization must have
occurred at C-13S and/or C-2′S during the respective
synthesis of 11 and/or 12. Chiral analysis of both educts
clearly proves compound 12 to have suffered racemiza-
tion, whereas the C-13 stereogenic center in 11 has
retained the (S)-configuration of its precursor 13S-
LOOH (see Structural Assignment).
F igu r e 4. Time course for the formation of 13 at pH 7.4 and
37 °C from 100 mM 11 and 50 mM 12, depending on the water
content of the incubation mixture.
To probe whether and to which extent 13 is formed
under mild reaction conditions and in the presence of
water, compounds 11a ,b (100 mM) and 12 (50 mM) were
incubated in MP and MP/water mixtures at pH 7.4 and
37 °C for 8 weeks. Due to the poor solubility in these
mixtures, buffer salts such as phosphate or borate
cannot be used directly to hold the pH stable; therefore,
an aqueous solution of 12, as the free base, was adjusted
to pH 7.4 with orthophosphoric acid and lyophilized, and
the resulting phosphate salt 12‚H3PO4 was employed
for the incubations. Progress of the reaction was moni-
tored by HPLC equipped with both a DAD and an
FLD. Fluorophore-labeled 12 allows for a straight-
forward correlation between the HPLC peak areas of
the lysine educt and eventual reaction products. The
time-dependent formation of tert-butyl (9R*,10R*,11E,-
13S)-10-({5-(acetylamino)-6-[(4-methyl-2-oxo-2H-chromen-
7-yl)amino]-6-oxohexyl}amino)-9,13-dihydroxy-11-octa-
decenoate (13a -d ) was observed, and the results are
shown in Figure 4. In pure MP, conversion to 13 is <1%
after 56 days. Product formation is favored significantly,
however, when water is added to the reaction mixture.
At the end of the reaction period, 5.8 and 15.8% of 12
have been converted to 13 in incubations with 10 and
20% (v/v) water, respectively. The higher conversion
rate, correlating with an increased solvent polarity,
corresponds to Ingold’s rules for solvent effects in
nucleophilic substitution (Ingold, 1969). The initial state
for an SN2 reaction of an amine with an epoxide, both
F igu r e 5. SN2 mechanism for the cleavage of an epoxide by
an amine [charge-type II according to the classification of
Ingold (1969)].
of which are neutral, is less polar than the transition
state (see Figure 5); Ingold has classified this case as
charge-type II. Consequently, a higher solvent polarity
stabilizes the transition state more than the initial state
and thus enhances the exothermicity of the reaction.
Unfortunately, at water contents >20% (v/v) 11a ,b is
no longer completely dissolved but forms emulsions.
Therefore, the incubation in 1:1 MP/water cannot be
directly compared to the others. The conversion graph
for this reaction mixture coincides with that for 9:1 MP/
water (see Figure 4).
Because formation of 13 is favored in the presence of
water, comparison of the aminolysis and the competing
hydrolysis of 11 is of interest. We have therefore
extracted the incubation mixtures with CH2Cl2 at the
end of the reaction period and determined the triol
content by GLC analysis of the respective trimethylsilyl
derivatives. tert-Butyl (E)-9,10,13-trihydroxy-11-octa-
decenoate (14) and tert-butyl (E)-9,12,13-trihydroxy-10-
octadecenoate (15) were synthesized from 11a ,b, as
authentic standards, in a 0.1 N p-toluenesulfonic acid/
THF mixture (see Figure 6). The acid-catalyzed hy-
drolysis proceeds via an SN1 mechanism, yielding two
regioisomers with the stereochemical correlation at C-9

4618 J. Agric. Food Chem., Vol. 47, No. 11, 1999
Lederer et al.
F igu r e 6. Acid-catalyzed hydrolysis of 11a ,b, yielding tert-butyl (E)-9,10,13-trihydroxy-11-octadecenoate (14) and tert-butyl (E)-
9,12,13-trihydroxy-10-octadecenoate (15), with (E)-4-hydroxy-2-nonenal (16) and tert-butyl 9-oxononanoate (17) or hexanal (18)
and tert-butyl (E)-9-hydroxy-12-oxo-10-dodecenoate (19) being formed by subsequent Pb(IV) tetraacetate cleavage of the triol 14
or 15.
and C-10 lost in the intermediate allylic carbocation.
We could not effectively separate this complex isomer
mixture. The NMR data for the major diastereoisomer
of 14, listed in Table 1, were obtained from a fraction
separated by LC on silica gel, which contained ∼75%
14 besides ∼25% 15. Differentiation between the regio-
isomers 14 and 15 is not possible on the basis of the
NMR data and rests on GLC analysis of the carbalde-
hydes 16-19, obtained by Pb(IV) tetraacetate cleavage
of the isolated fraction (see Figure 6). Unequivocal
identification of the fragment structures 16-19 by GLC-
MS confirms the location of the OH groups in 14 and
15.
GLC analyses showed, <1, 3.2, 10.4, and 20.5% of the
epoxide 11 to be hydrolyzed in incubations A-D after 8
weeks, respectively. As expected, conversion to the triols
correlates with the water content of the reaction mix-
tures. However, except for incubation D, formation of
the aminol 13 is enhanced to more or less the same
extent. Aminolysis therefore must be considered as an
important mode of interaction between amines and R,â-
unsaturated epoxides. Interestingly, both regioisomers
14 and 15 were identified from all incubations by GLC-
MS analysis of the carbaldehydes 16-19, after Pb(IV)
tetraacetate cleavage of an aliquot of the respective
incubation. This implies that hydrolysis, other than
aminol formation, at least partially takes an SN1 course
at pH 7.4.
tively. The acetoxy group at C-13 exerts a very effective
control over the regioselectivity of the epoxidation, and
it seems curious that this function does not influence
the diastereoselectivity for the oxirane ring formation.
Therefore, we have tested whether C-13 has suffered
racemization during the synthetic procedure, following
a method described by Hamberg (1991). The diastere-
oisomers 11a ,b were reacted with (-)-(1R)-mentyl chlo-
roformate and the (-)-(1R)-menthoxycarbonyl (MC)
derivatives isolated by TLC. These were oxidized with
KMnO4 in glacial acetic acid at 37 °C (Hamberg et al.,
1986), and the trimethylsilyl derivatives of the resulting
2-({[(-)-(1R)-menthoxy]carbonyl}oxy)heptanoic acids were
analyzed by CI-GLC-MS. The total ion current (TIC)
chromatogram shows two peaks (1:50 relative intensity)
with almost identical fragmentation patterns for the
diastereoisomeric MC derivatives of trimethylsilyl 2-hy-
droxyheptanoate. Together with GLC analyses of 9 and
10a ,b at various chiral stationary phases (not detailed
under Materials and Methods) that show no loss of
stereochemical purity, this result definitely proves that
no racemization has occurred at the C-13 stereogenic
center in the synthesis of 11a ,b. From the NMR
evidence, regio- and stereoselectivity of the Prilezhaev
reaction is proven unequivocally: The configurations of
the double bonds C9dC10 and C11dC12 in compounds
7-9 are definitely established as (Z) and (E), respec-
3
tively, by the vicinal coupling constants J (9-H, 10-H)
3
) 10.9 Hz and J (11-H, 12-H) ) 15.2 Hz. Because the
STRUCTURAL ASSIGNMENT
two residual olefinic protons in 10 are coupled by 15.6
Hz, the (Z) double bond must have been transformed
to the oxirane. This assignment was validated by
correlation spectroscopy (1H,1H-COSY). The 4.3 Hz
value for the coupling constant between 9-H and 10-H
in 10 and 11 constitutes an independent proof for
epoxidation of the (Z)-configured double bond, via a true
syn-addition mechanism, yielding a cis-configured ox-
Special effort was invested to definitely establish the
structures of compounds 11a ,b and 13a -d because the
mechanistic arguments rest primarily on the respective
product structures. The 1H and 13C NMR chemical shifts
(δ) and coupling constants (J ) of 7-11a ,b and 14 are
collected in Table 1 and those of 13a -d in Table 2.
In contrast to their precursors, compounds 10 and 11
show two sets of NMR signals with almost 1:1 relative
intensity for several proton and carbon sites; in Table
1, the different chemical shifts or coupling constants are
marked with index b or c for 10a ,b and 11a ,b, respec-
3
irane. For an alternative trans orientation, as in 4, a J
value of ∼2 Hz is expected (Lederer, 1996; Gardner et
al., 1974) The relative configuration at C-9 and C-10
therefore can only be (R*/ S*), and due to the stereo-

Protein−Lipid Oxidation Product Interaction
J. Agric. Food Chem., Vol. 47, No. 11, 1999 4619
genic center at C-13, 10 and 11 thus exist each as a pair
of diastereoisomers. From the structure of compound 13,
the mechanism of the nucleophilic cleavage of the
epoxide ring in 11a ,b can be derived in a straightfor-
ward manner. The 9-H and 10-H resonances in 13 were
assigned on the basis of the respective connectivities for
mittelchemie, Universita¨t Hohenheim, for the excellent
working conditions at his Institute. We are grateful to
J ochen Rebell, Institut fu¨r Organische Chemie, Uni-
versita¨t Stuttgart, for the recording of the NMR spectra.
LITERATURE CITED
1
the proton spin system (positions 8-14) and by H,1H-
Dix, T. A.; Marnett, L. J . Conversion of linoleic acid hydrop-
eroxide to hydroxy, keto, epoxyhydroxy, and trihydroxy fatty
acids by hematin. J . Biol. Chem. 1985, 260, 5351-5357.
Funck, C. D.; Powell, W. S. Metabolism of linoleic acid by
prostaglandin endoperoxide synthase from adult and fetal
blood vessels. Biochim. Biophys. Acta 1983, 754, 57-71.
Gardner, H. W. Lipid hydroperoxide reactivity with proteins
and amino acids: A review. J . Agric. Food Chem. 1979, 27,
220-229.
Gardner, H. W. Oxygen radical chemistry of polyunsaturated
fatty acids. Free Radical Biol. Med. 1989, 7, 65-86.
Gardner, H. W.; Kleiman, R.; Weisleder, D. Homolytic decom-
position of linoleic acid hydroperoxide: Identification of fatty
acid products. Lipids 1974, 9, 696-706.
Gargouri, M.; Legoy, M. D. Chemoenzymatic production of (+)-
coriolic acid from trilinolein-coupled synthesis and extrac-
tion. J . Am. Oil Chem. Soc. 1997, 74, 641-645.
Graveland, A. Enzymatic oxidations of linoleic acid and
glycerol-1-monolinoleate in doughs and flour-water suspen-
sions. J . Am. Oil Chem. Soc. 1970, 47, 352-361.
Graveland, A. Enzymatic oxidation of linolenic acid in aqueous
wheat flour suspensions. Lipids 1973, 8, 606-611.
Hamberg, M. Decomposition of unsaturated fatty acid hydro-
peroxides by hemoglobin: Structures of major products of
13L-hydroperoxy-9,10-octadecadienoic acid. Lipids 1975, 10,
87-92.
COSY. The 9-H resonance (3.33 ppm for 13a ,b and 3.36
ppm for 13c,d ; see Table 2) appears to be less shielded
by 0.4 ppm relative to that of the proton at C-10 (2.97,
2.99 ppm for 13a ,b, and 3.02, 3.03 ppm for 13c,d ,
respectively), clearly establishing the OH function to be
located at C-9 and the amine moiety to have been
introduced at C-10. An allylic shift, via an SN2′ mech-
anism, is likewise ruled out definitely by the spectro-
scopic evidence. As mentioned above, preparative HPLC
of 13 allows separation of this isomer mixture into two
fractions. However, a number of resonances in the
spectra of each fraction, marked with the index b in
Table 2, appear doubled, indicating that no homoge-
neous compounds have been isolated and that 13 exists
as two pairs of diastereoisomers, 13a -d . Because the
(S)-configuration at C-13 has been firmly established,
the only remaining rationale is loss of stereochemical
purity of the lysine moiety. Compound 12 in fact shows
no optical rotation; that is, the chiral carbon has
obviously suffered complete racemization in the course
of the synthetic procedure. This agrees with the 1:1
relative intensity for the doubled resonances for the
diastereoisomers 13a ,b and 13c,d , respectively. We
presume racemization to have occurred during the
amide formation of N2-acetyl-N6-(tert-butoxycarbonyl)-
L-lysine with 7-amino-4-methylcoumarin via an imida-
zolide, requiring incubation for 9 days at 50 °C (Lederer,
1996). According to Staab (1962), the loss of stereo-
chemical purity is negligible for such processes only at
low temperatures (-10 °C). Hence, the configuration of
12 had erroneously been given as pure (S) in the
literature (Lederer, 1996).
Hamberg, M. Regio- and stereochemical analysis of trihy-
droxydecenoic acids derived from lineloic acid 9- and 13-
hydroperoxides. Lipids 1991, 26, 407-415.
Hamberg, M.; Hamberg, G. Hydroperoxide-dependent epoxi-
dation of unsaturated fatty acids in the broad bean (Vicia
faba L.). Arch. Biochem. Biophys. 1990, 283, 409-416.
Hamberg, M.; Herman, R. P.; J acobsson, U. Stereochemistry
of two epoxy alcohols from Saprolegnia parasitica. Biochim.
Biophys. Acta 1986, 879, 410-418.
Hata, K.; Fujimoto, K.; Kaneda, T. Absorption of lipid hydro-
peroxides in carp. Nippon Suisan Gakkaishi 1986, 52, 677-
684.
CONCLUSIONS
The findings presented above clearly demonstrate
that lysine moieties, incorporated in a protein, may react
with γ-hydroxy-R,â-unsaturated epoxides generated in
the course of lipid oxidation. Formation of aminol
structures such as 13 (Figure 3) competes with hydroly-
sis of the epoxide; however, the reaction rate for both
processes increases with the water content [up to 20%
(v/v)] in our model incubations, and the respective
products are formed in comparable extent. In foodstuffs
with medium or low water activity, aminolysis of R,â-
unsaturated epoxides by lysine moieties therefore must
be considered as an important mode of interaction
between proteins and lipid oxidation products. Because
formation of aminols takes place under mild conditions
(pH 7.4; 37 °C), lipoproteins may be expected to be
especially susceptible to such transformations in vivo,
due to the proximity of the respective reaction partners
and the low water activity in the hydrophobic regions
of such globular proteins. Further investigations are in
progress to transfer the experiences of the model study
to biological systems and to either verify or disprove the
hypotheses outlined here.
Heimann, W.; Dresen, P.; Klaiber, V. On the formation and
degradation of linoleic acid hydroperoxides in cereals. Z.
Lebensm. Unters. Forsch. 1973, 153, 1-5.
Ingold, C. K. Structure and Mechanism in Organic Chemistry,
2nd ed.; Cornell University Press: Ithaca, NY, 1969; pp
457-463.
Lederer, M. O. Reactivity of lysine moieties toward γ-hydroxy-
R,â-unsaturated epoxidessa model study on protein-lipid
oxidation product interaction. J . Agric. Food Chem. 1996,
44, 2531-2537.
Maguire, N. M.; Mahon, M. F.; Molloy, K. C.; Read, G.; Roberts,
S. M.; Sik, V. Chemoenzymatic synthesis of some macrocy-
clic C13-lactones. J . Chem. Soc., Perkin Trans. 1 1991,
2054-2056.
Maguire, N. M.; Read, G.; Richardson, P. F.; Roberts, S. M.
Hydroperoxidation of double bonds: Preparation of (13S)-
HODE via lipoxygenase catalysed peroxidation of linoleic
acid. In Preparative Biotransformations: Whole Cell and
Isolated Enzymes in Organic Synthesis; Roberts, S. M.,
Wiggins, K., Casy, G., Phythian, S., Todd, C., Eds.; Wiley:
New York, 1997; pp 3:2.1-3:2.10.
Nielsen, H. K.; Loeliger, J .; Hurrell, R. F. Reactions of proteins
with oxidizing lipids; 1. Analytical measurement of lipid
oxidation and of amino acid losses in a whey protein-methyl
linolenate system. Br. J . Nutr. 1985, 53, 61-73.
Ohta, S.; Shimabayashi, A.; Aono, M.; Okamoto, M. A general
convenient one-pot procedure for the conversion of carboxylic
acids into their tert-butyl esters which is also applicable to
aliphatic carboxylic acids. Synthesis 1982, 833-834.
ACKNOWLEDGMENT
We thank Peter Fischer, Institut fu¨r Organische
Chemie, Universita¨t Stuttgart, for many helpful dis-
cussions and Wolfgang Schwack, Institut fu¨r Lebens-

4620 J. Agric. Food Chem., Vol. 47, No. 11, 1999
Lederer et al.
Pace-Asciak, C. R.; Granstrom, E.; Samuelsson, B. Arachidonic
acid epoxides. J . Biol. Chem. 1983, 258, 6835-6840.
Pokorny, J .; Klein, S.; Koren, J . Reaction of oxidized lipids with
protein. Nahrung 1966, 10, 321-325.
Zamora, R.; Hidalgo, F. J . Linoleic acid oxidation in the
presence of amino compounds produces pyrroles by carbonyl
amine reactions. Biochim. Biophys. Acta 1995, 1258, 319-
327.
Zamora, R.; Navarro, J . L.; Hidalgo, F. J . Determination of
lysine modification product ꢀ-N-pyrrolylnorleucine in hy-
drolyzed proteins and trout muscle microsomes by micellar
electrokinetic capillary chromatography. Lipids 1995, 30,
477-483.
Staab, H. A. Synthesis with heterocyclic amides (azolides).
Angew. Chem. 1962, 74, 407-423.
Steinbrecher, U. P. Oxidation of human low-density lipoprotein
results in derivatization of lysine residues of apolipoprotein
B by lipid peroxide decomposition products. J . Biol. Chem.
1987, 262, 3603-3608.
Van Os, C. P. A.; Vliegenhart, J . F. G.; Crawford, C. G.;
Gardner, H. W. Structural analysis of diastereoisomeric
m et h yl-9-h ydr oxy-tr a n s-12,13-epoxy-10-tr a n s-oct a de-
cenoates. Biochim. Biophys. Acta 1982, 713, 173-176.
Received for review April 19, 1999. Revised manuscript
received August 24, 1999. Accepted August 31, 1999.
J F990383O