Oxidative Fragmentation of Hydroxy Octadecadienoates Generates Biologically Active γ-Hydroxyalkenals

Article abstract of DOI:10.1021/ja038756w

Oxidative fragmentation of polyunsaturated fatty acids (PUFAs) in vivo generates cytotoxic aldehydes. Among these, 4-hydroxynon-2-enal and analogous γ-hydroxyalkenal phosphatidylcholines (PCs) have attracted attention because these oxidatively truncated lipids are biologically active and have been implicated in diseases. A previous study showed that hydroxydienes, generated by allylic oxygenation of linoleic acid, are unreactive toward oxidative fragmentation. We now show that, in the presence of hydroperoxides, hydroxydienes fragment as readily as the corresponding hydroperoxydienes, generating γ-hydroxyalkenals. In a physiomimetic model study, myeloperoxidase-promoted free radical-induced fragmentation of either hydroperoxy- or hydroxyoctecadienoate esters of 2-lyso-PC in small unilamellar vesicles produced the 9-hydroxy-12-oxododec-10-enoic acid (HODA) ester HODA-PC. Therefore, hydroxydienes, that are generally more abundant in vivo than hydroperoxydienes, are plausible intermediates in the production of oxidatively truncated lipids in vivo where a constant flux of radicals and hydroperoxides is present. Our findings also show that the formation of dioxetane intermediates through peroxyradical cyclization is not required to achieve oxidative fragmentation of PUFAs.

Full text of DOI:10.1021/ja038756w

Published on Web 04/14/2004
Oxidative Fragmentation of Hydroxy Octadecadienoates
Generates Biologically Active γ-Hydroxyalkenals
Mingjiang Sun and Robert G. Salomon*
Contribution from the Department of Chemistry, Case Western ReserVe UniVersity,
CleVeland, Ohio 44107-7078
Received September 27, 2003; E-mail: rgs@cwru.edu
Abstract: Oxidative fragmentation of polyunsaturated fatty acids (PUFAs) in vivo generates cytotoxic
aldehydes. Among these, 4-hydroxynon-2-enal and analogous γ-hydroxyalkenal phosphatidylcholines (PCs)
have attracted attention because these oxidatively truncated lipids are biologically active and have been
implicated in diseases. A previous study showed that hydroxydienes, generated by allylic oxygenation of
linoleic acid, are unreactive toward oxidative fragmentation. We now show that, in the presence of
hydroperoxides, hydroxydienes fragment as readily as the corresponding hydroperoxydienes, generating
γ-hydroxyalkenals. In a physiomimetic model study, myeloperoxidase-promoted free radical-induced
fragmentation of either hydroperoxy- or hydroxyoctecadienoate esters of 2-lyso-PC in small unilamellar
vesicles produced the 9-hydroxy-12-oxododec-10-enoic acid (HODA) ester HODA-PC. Therefore, hydroxy-
dienes, that are generally more abundant in vivo than hydroperoxydienes, are plausible intermediates in
the production of oxidatively truncated lipids in vivo where a constant flux of radicals and hydroperoxides
is present. Our findings also show that the formation of dioxetane intermediates through peroxyradical
cyclization is not required to achieve oxidative fragmentation of PUFAs.
Scheme 1
Lipid peroxidation generates various cytotoxic aldehydes,
among which the γ-hydroxyalkenals are the most studied.1
Oxidative fragmentation of arachidonate and linoleate phos-
pholipids has pathological significance. We recently showed that
it generates biologically active truncated γ-hydroxyalkenal
phospholipids that are abundant in atherosclerotic plaques.2,3
These oxidized phospholipids promote the formation of foam
cells, precursors of atherosclerotic plaques, because they are
ligands for the macrophage scavenger receptor CD36.4 They
also interfere with the proteolytic degradation of internalized
macromolecules by macrophages because, inter alia, they
covalently modify a lysosomal protease.3 In addition, they
activate human aortic endothelial cells to bind monocytes,
promote the expression of chemokines that are important in
monocyte entry into chronic lesions, and inhibit expression of
a major adhesion molecule that mediates neutrophil endothelial
interactions.5
The major primary products of both enzymatic and free
radical-induced oxidation of linoleic acid (LA, 1) are the
hydroperoxydienes 2 and 3 (Scheme 1). In vivo, these are
glutathione peroxidases.6,7 In the aorta of rabbits, the hydroxy-
diene 9-HODE (4) and its regioisomer 13-HODE (9) are 5-fold
more abundant than corresponding hydroperoxides 3 and 2.8
Increased levels of 9-HODE (4) are found in human low-density
lipoprotein from atherosclerotic patients as compared with
individuals who have no cardiovascular disease.9
efficiently reduced to hydroxydienes by selenium-containing
(1) Carmen Vigo-Pelfrey, E. Membrane Lipid Oxidation; CRC Press: Menlo
Park, CA, 1990; Vol. I.
(2) Podrez, E. A.; Poliakov, E.; Shen, Z.; Zhang, R.; Deng, Y.; Sun, M.; Finton,
P. J.; Shan, L.; Febbraio, M.; Hajjar, D. P.; Silverstein, R. L.; Hoff, H. F.;
Salomon, R. G.; Hazen, S. L. J. Biol. Chem. 2002, 277, 38517-38523.
(3) Hoff, H. F.; O’Neil, J.; Wu, Z.; Hoppe, G.; Salomon, R. L. Arterioscler.,
Thromb., Vasc. Biol. 2003, 23, 275-282.
(4) Podrez, E. A.; Poliakov, E.; Shen, Z.; Zhang, R.; Deng, Y.; Sun, M.; Finton,
P. J.; Shan, L.; Gugiu, B.; Fox, P. L.; Hoff, H. F.; Salomon, R. G.; Hazen,
S. L. J. Biol. Chem. 2002, 277, 38503-38516.
(6) Reinaud, O.; Delaforge, M.; Boucher, J. L.; Rocchiccioli, F.; Mansuy, D.
Biochem. Biophys. Res. Commun. 1989, 161, 883-891.
(7) Kuhn, H. Prog. Lipid Res. 1996, 35, 203-226.
(5) Subbanagounder, G.; Deng, Y.; Borromeo, C.; Dooley, A. N.; Berliner, J.
A.; Salomon, R. G. Vasc. Pharmacol. 2002, 38, 201-209.
(8) Upston, J. M.; Witting, P. K.; Brown, A. J.; Stocker, R.; Keaney, J. F., Jr.
Free Radical Biol. Med. 2001, 31, 1245-1253.
9
10.1021/ja038756w CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 5699-5708
5699

A R T I C L E S
Sun and Salomon
Scheme 2
5 (Scheme 1), but that 8 is not generated through the analogous
formation and fragmentation of a hydroxy hydroperoxide 6. This
might be because two hydroperoxy groups are somehow
required for the fragmentation process, or because hydroxy
hydroperoxide 6 is not readily generated from the hydroxydiene
4. The latter scenario, which requires that allylic hydrogen atom
abstraction from 3 occurs much more readily than from 4,
seemed unlikely. We reasoned that the behavior of the pure
hydroxydiene 4 may have little bearing on the biologically
relevant chemistry of these lipids. Rather, the chemistry of 4 in
vivo would be influenced by the existence of a constant flux of
lipid peroxy radicals or hydroperoxides, generated by autoxi-
dation, that seemed likely to promote autoxidation of 4. We
now provide evidence that hydroxy octadecadienoates are not
stable end-products, but rather readily undergo oxidative
fragmentation that is promoted by lipid hydroperoxides such
as 13-hydroperoxyoctadeca-9,11-dienoate (13-HPODE, 2)
(Scheme 1). The physiological relevance of this chemistry is
supported by the observation that free radical-induced fragmen-
tation of either hydroxy- (HODE) or hydroperoxyoctecadienoate
(HPODE) esters of 2-lyso-PC in a model membrane (small
unilamellar vesicles) produces the 9-hydroxy-12-oxododec-10-
enoic acid (HODA) ester HODA-PC. Our findings also show
that the formation of dioxetane intermediates through peroxy-
radical cyclization is not required to achieve oxidative frag-
mentation of polyunsaturated fatty acids.
Because the γ-hydroxyalkenals produced through lipid per-
oxidation in vivo have important biological activities, a funda-
mental understanding of the mechanism of their formation is
of considerable interest. It was suggested that a γ-hydroxy-
alkenal, 9-hydroxy-12-oxododec-10-enoic acid (HODA, 8),
could be generated through â-scission of a hydroxyalkoxyl
radical derived from a hydroxy hydroperoxide intermediate 6
(Scheme 1).10 It was later found that 9-HODE (4) is not readily
oxidized under a variety of conditions that cause rapid oxidation
of LA (1), and it was concluded that hydroxy octadecadienoates
are stable end-products that can serve as “markers for lipid
peroxidation” in vivo.11 A possible explanation for the apparent
stability of 4 is that oxidative cleavage involves the formation
and fragmentation of a hydroperoxydioxetane intermediate that
is generated by cyclization of a peroxydiene radical 10 (Scheme
2). No such peroxydiene radical intermediate is accessible from
the hydroxydiene 4. The possible involvement of dioxetane
intermediates in the oxidative fragmentation of polyunsaturated
fatty acids was suggested previously,12-14 and the aldehydes
11-13 are known products from the autoxidation of linoleic
acid (1).15-17
Results and Discussion
Autoxidation of Both Hydroperoxydienes 2 and 3 Gener-
ates Hydroxyalkenal 8. Autoxidations of 13-HPODE (2) and
9-HPODE (3) were monitored by LC/ESI/MS/MS. This tech-
nique allowed simultaneous quantification of specific isomers
based upon monitoring unique transitions between the mass-
to-charge ratio of the parent ion [M - H]- and characteristic
daughter ions for each species as well as their appropriate
retention times. Hydroperoxydiene 2 or 3 was heated as a dry
film in an open glass tube at 37 °C, and the amounts of 2 and
3 as well as HODA (8) were monitored simultaneously. 13-
HPODE (2) and 9-HPODE (3) behave similarly. After 1 h, only
40-60% of starting compound remains (Figure 1A and B).
Furthermore, interconversion of 3-HPODE (2) and 9-HPODE
(3) was observed. This phenomena had been observed previously
when the hydroperoxides were stored in organic solvents.18,19
HODA (8) was produced from both hydroperoxides (Figure 1C),
in similar yield (5%), and at similar rates.
Another possibility is that HODA (8) is only generated by
reduction of the known16 precursor HPODA (7) that is produced
through the formation and fragmentation of a bishydroperoxide
The isomerization presumably involves reversible generation
and rearrangement of peroxy radical intermediates (Scheme 3).
It results in loss of regiospecificity and complicates analysis of
the fragmentation product distribution. It is difficult to discern
which hydroperoxydiene is the immediate precursor of HODA
(8).
Hydroperoxides Promote Oxidation of Hydroxydienes 4
and 9. Pure 9-HODE (4) is not readily oxidized under a variety
of conditions that cause rapid oxidation of LA (1).11 These
include exposure to air in the presence of Fe2+, or Fe3+, or Fe2+
and ascorbate, or Fe2+ and H2O2. We postulated that alkylper-
oxyl or alkoxyl radicals, that are generated from alkyl hydro-
(9) Jira, W.; Spiteller, G.; Carson, W.; Schramm, A. Chem. Phys. Lipids 1998,
91, 1-11.
(10) Loidl-Stahlhofen, A.; Hannemann, K.; Spiteller, G. Biochim. Biophys. Acta
1994, 1213, 140-148.
(11) Spiteller, P.; Spiteller, G. Chem. Phys. Lipids 1997, 89, 131-139.
(12) Esterbauer, H.; Zollner, H.; Schaur, R. J. Membrane Lipid Oxidation; CRC
Press: Boca Raton, FL, 1990; Vol. 1, p 250.
(13) Kaur, K.; Salomon, R. G.; O’Neil, J.; Hoff, H. F. Chem. Res. Toxicol.
1997, 10, 1387-1396.
(14) Timmins, G. S.; dos Santos, R. E.; Whitwood, A. C.; Catalani, L. H.; Di
Mascio, P.; Gilbert, B. C.; Bechara, E. J. H. Chem. Res. Toxicol. 1997, 10,
1090-1096.
(15) Lee, S. H.; Blair, I. A. Chem. Res. Toxicol. 2000, 13, 698-702.
(16) Schneider, C.; Tallman, K. A.; Porter, N. A.; Brash, A. R. J. Biol. Chem.
2001, 276, 20831-20838.
(18) Chan, H. W.; Levett, G.; Matthew, J. A. Chem. Phys. Lipids 1979, 24,
245-256.
(19) Porter, N. A.; Wujek, J. S. J. Org. Chem. 1987, 52, 5085-5089.
(17) Minamoto, S.; Kanazawa, K.; Ashida, H.; Natake, M. Biochim. Biophys.
Acta 1988, 958, 199-204.
9
5700 J. AM. CHEM. SOC. VOL. 126, NO. 18, 2004

Oxidative Fragmentation of Hydroxy Octadecadienoates
A R T I C L E S
Figure 1. Autoxidation of hydroperoxides from linoleic acid at 37 °C and analyzed by LC-MS. (A) 13-HPODE (2), (B) 9-HPODE (3), (C) HODA (8)
generated from hydroperoxide autoxidation.
Scheme 3
peroxides such as hydroperoxyoctadecadienoates (HPODEs) 2
or 3, are mandatory for the allylic hydrogen atom abstraction
required to produce 6 from 4 (Scheme 1). To test this hypothesis,
we exposed pure 9-hydroxyoctadeca-10,12-dienoic acid (9-
HODE, 4) or 13-hydroxyoctadeca-9,11-dienoic acid (13-HODE,
9) to air in the absence or in the presence of 9,10,12,13-
tetradeuterated hydroperoxydiene 13-HPODE-d4 (2-d4). As
expected, the hydroxydienes 4 and 9 were unchanged upon
heating in air at 37 °C for 8 h. In contrast, both hydroxydienes
decomposed at rates similar to that of the hydroperoxydiene 2
when heated individually in air at 37 °C in the presence of
deuterium-labeled 2 (Figure 2).
Oxidative Fragmentation of Hydroxydiene 4 but Not 9
Generates γ-Hydroxyalkenal 8. These observations supported
the possibility that the hydroxydienes 4 and 9 can serve as
intermediates in the free radical-induced oxidative fragmentation
of linoleic acid. Detection of the fragmentation product 8
generated in the oxidative decomposition of the hydroxydiene
Figure 2. Decomposition of hydroxydiene 9-HODE (4) or 13-HODE (9)
in the presence of the hydroperoxyoctadecadienoate 13-HPODE-d4 (2-d4).
4 confirmed this presumption. By using tetradeuterated hydro-
peroxydiene 2-d4 in these experiments, LC-MS analysis allowed
a distinction between hydroxyalkenal 8-d3 generated from the
HODA (8) and the oxidation reaction product mixture were
derivatized with methoxylamine hydrochloride in pyridine.
HODA-methoxime ([M - H]-, m/z 256) produces a dominant
daughter ion m/z 197. The HPLC chromatogram of the HODA
methoxime derivative was monitored by ESI/MS/MS with
monitoring of the m/z 256 to m/z 197 mass transition. After
derivatization, the peak corresponding to HODA (8) in the
autoxidation reaction product mixture disappeared, and a new
hydroperoxide and hydroxyalkenal 8-d0 produced from the
hydroxydiene. The oxidative fragmentation product 8 was
generated upon air oxidation of both hydroxydiene 4 (in >5%
yield) and hydroperoxydiene 2-d4 but not from oxidation of
hydroxydiene 9 (Figure 3).
To further verify that HODA (8) was produced from 9-HODE
(4) in the presence of hydroperoxydiene 2-d4, both standard
9
J. AM. CHEM. SOC. VOL. 126, NO. 18, 2004 5701

A R T I C L E S
Sun and Salomon
Scheme 4
of the hydroxy hydroperoxide 14 derived from 9 is expected to
produce a different γ-hydroxyalkenal, 4-hydroxynon-2-enal
(HNE, 13 in Scheme 4). This cytotoxic product of lipid
oxidation in vivo20,21 forms protein adducts, causing the
inhibition of enzymes.22 Elevated levels of HNE and HNE-
derived protein modification have been implicated in
diseases.21,23-25
Generation of HNE (13) upon oxidation of 13-HODE (9) in
the presence of 13-HPODE-d4 (2-d4) was confirmed. To increase
the sensitivity of HNE (13) detection by ESI-MS, 13 was
derivatized with cyclohexanedione (CHD) and ammonium
acetate in aqueous acetic acid (Scheme 5).26 The ESI-MS
spectrum of the HNE CHD derivative 15 showed a predominant
pseudo parent [M - H2O + H]+ m/z 326 ion in addition to the
parent [M + H]+ m/z 344 ion. Tandem MS/MS analysis of the
dehydrated ion revealed m/z 308, in addition to m/z 216 which
was assigned previously.27 We propose that the HNE derivative
15 can be dehydrated at two different positions; one of them
leads to the m/z 216 ion, and the other leads to the m/z 308 ion
by loss of another molecule of water through electrocyclization
followed by dehydration.
Figure 3. Yield of HODA (8) from oxidative fragmentation of 9-HODE
(4) or 13-HODE (9) in the presence of 13-HPODE-d4 (2-d4).
An LC-ESI-MS/MS method27 was employed to quantify HNE
(13) after derivatization based on monitoring unique transitions
between the mass-to-charge ratio of the pseudo parent ion
[M + H - 18]+ m/z 326 and the characteristic daughter ion
m/z 216 and its appropriate retention time. Before derivatization,
benzaldehyde was added as internal standard, and the amount
of HNE was calculated from a calibration curve (see Experi-
mental Procedures). As predicted, HNE was produced from 13-
HODE, but not from 9-HODE, upon autoxidation in the
presence of 13-HPODE-d4 (Figure 5). The yield of HNE (13)
peaked at about 4 h and is comparable with the yield of HODA
(8) from the autoxidation of 9-HODE (4) in the presence of
hydroperoxydiene-d4 (2-d4). This observation further confirmed
that 4 and 9 can serve as intermediates for production of HODA
(8) or HNE (13).
Figure 4. HPLC chromatogram of HODA and HODA methoxime
derivative. All chromatograms were monitored by LC-MS in the negative
ion mode with MRM of appropriate mass transitions as noted. (A) HODA-
d3 produced from autoxidation (2 h) of 13-HPODE-d4 in the presence of
9-HODE. (B) HODA-d0 produced from autoxidation (2 h) of 9-HODE in
the presence of 13-HPODE-d4. (C) Methoxime derivative of authentic
HODA-d0. (D) Methoxime derivative of HODA-d0 produced from autoxi-
dation (2 h) of 9-HODE in the presence of 13-HPODE-d4.
peak at the position of the HODA methoxime derivative
appeared (Figure 4).
The contrasting behavior of the hydroperoxydiene 13-HPODE
(2) and the hydroxydiene 13-HODE (9) is noteworthy. Autoxi-
dation of 2 generates HODA (8), but autoxidation of 9 does
not. Furthermore, the yield of 8 from the hydroperoxydiene 2
is only about one-half the yield of 8 from the regioisomeric
hydroxydiene 9-HODE (4). This is understandable if it is
presumed that: (i) the generation of 8 from the 13-hydroper-
oxydiene requires prior isomerization to the 9-hydroperoxydiene
9-HPODE (3), (ii) the formation of 8 through isomerization to
3 competes with alternative fragmentation reactions that are only
accessible to the regioisomeric hydroperoxydiene 13-HPODE
(2) or hydroxydiene 13-HODE (9), and (iii) while the regio-
isomeric hydroperoxydienes 2 and 3 readily interconvert, the
corresponding regioisomeric hydroxydienes 4 and 9 do not.
Oxidative Fragmentation of Hydroxydiene 9 Generates
γ-Hydroxynonenal (13). The failure of the oxidation of pure
13-HODE (9) to produce the γ-hydroxyalkenal 8 is understand-
able if C-C bond cleavage is presumed to require the
involvement of a hydroperoxy group or an alkoxyl radical on
one terminus of the scissile C-C bond. Oxidative fragmentation
Hydroxy- and Hydroperoxydienes Are Similarly Reactive
Precursors for γ-Hydroxyalkenals. The present study dem-
onstrates that hydroxydienes 4 and 9 are not stable end-products
(20) Benedetti, A.; Comporti, M.; Esterbauer, H. Biochim. Biophys. Acta 1980,
620, 281-296.
(21) Esterbauer, H.; Schaur, R. J.; Zollner, H. Free Radical Biol. Med. 1991,
11, 81-128.
(22) Benedetti, A.; Casini, A. F.; Ferrali, M.; Comporti, M. Biochem. Pharmacol.
1979, 28, 2909-2918.
(23) Hoff, H. F.; O’Neil, J.; Chisolm, G. M. d.; Cole, T. B.; Quehenberger, O.;
Esterbauer, H.; Jurgens, G. Arteriosclerosis 1989, 9, 538-549.
(24) Requena, J. R.; Fu, M. X.; Ahmed, M. U.; Jenkins, A. J.; Lyons, T. J.;
Baynes, J. W.; Thorpe, S. R. Biochem. J. 1997, 322, 317-325.
(25) Palinski, W.; Yla-Herttuala, S.; Rosenfeld, M. E.; Butler, S. W.; Socher,
S. A.; Parthasarathy, S.; Curtiss, L. K.; Witztum, J. L. Arteriosclerosis 1990,
10, 325-335.
(26) Matsuoka, M.; Imado, N.; Maki, T.; Banno, K.; Sato, T. Chromatographia
1996, 43, 501-506.
(27) O’Brien-Coker, I. C.; Perkins, G.; Mallet, A. I. Rapid Commun. Mass
Spectrom. 2001, 15, 920-928.
9
5702 J. AM. CHEM. SOC. VOL. 126, NO. 18, 2004

Oxidative Fragmentation of Hydroxy Octadecadienoates
A R T I C L E S
Scheme 5
dation of the hydroperoxydienes 13-HPODE (2) and 9-HPODE
(3) both afford HODA (8). However, this complexity is
undoubtedly the result of the interconversion of 2 and 3 under
the autoxidation reaction conditions. In contrast, only one
hydroxydiene, 9-HODE (4) and not 13-HODE (9), affords 8
upon autoxidation. Furthermore, only autoxidation of one
regioisomeric hydroxydiene, 13-HODE (9) and not 9-HODE
(4), affords HNE (13). Thus, in contrast with the chemistry of
the regioisomeric hydroperoxydienes, the immediate precursor
of the fragmentation products HODA (8) and HNE (13) is
unambiguously defined in the autoxidation reactions of the
regioisomeric hydroxydienes.
Figure 5. HNE production from autoxidation of HODEs and 13-HPODE-
d4 in air at 37 °C.
of lipid oxidation. In fact, under identical reaction conditions,
that is, in the presence of hydroperoxides that can initiate
autoxidation, the hydroxydienes and hydroperoxydienes are
consumed (Figure 2) and generate fragmentation products
(Figure 3) at similar rates. Thus, rather than being stable11
markers for lipid peroxidation, the hydroxydienes 4 and 9 are
reactive precursors to biologically active γ-hydroxyalkenals
HODA (8) and HNE (13). The increased levels of 9-HODE (4)
found in human low-density lipoprotein from atherosclerotic
patients as compared with individuals who have no cardiovas-
cular disease9 may have pathological significance because it can
serve as a precursor of HODA (8). The ester of HODA (8) with
2-lyso-phosphatidylcholine is a strong ligand for the scavenger
receptor CD-36 of macrophage cells.4 It promotes recognition
of oxidatively damaged low-density liproprotein, leading to
unregulated endocytosis.2 This avid uptake contributes to the
production of foam cells and atherosclerotic plaques. Because
hydroperoxydienes are rapidly reduced in vivo,6,7 hydroxydienes
are generally more abundant8 and, consequently, their oxidative
cleavage could be the primary route to γ-hydroxyalkenals. The
production of HNE from 13-HODE (9) may also be a
pathologically important process because the elevated levels of
HNE (13), produced through lipid oxidation in vivo, contribute
to various disease processes through covalent adduction with
proteins and consequent inhibition of enzymes.20-25
Oxidative Fragmentation of HODE-PC in Unilamellar
Vesicles. In vivo, hydroperoxy- and hydroxydienes are present
mainly esterified, for example, in phospholipids. Therefore, we
compared free radical-induced oxidative fragmentation of the
hydroperoxy- and hydroxyoctadecadienoate esters of 2-lyso-
phosphatidylcholine (PC), HPODE-PC (16) and HODE-PC (17),
respectively. Regioisomeric 13- and 9-hydroperoxyoctadecadi-
enoate esters 16a and 16b (4:1) are produced from 1-palmitoyl-
2-linoleoyl-sn-glycero-3-phosphocholine (LA-PC) by lipoxy-
genase-promoted autoxidation. Reduction with triphenylphosphine
provides the corresponding regioisomeric hydroxydienes 17a
and 17b. To evaluate the oxidative fragmentation of hydro-
peroxy and hydroxydienes in a model membrane under physi-
ologically relevant conditions, we compared the free radical-
induced oxidation of unilamellar vesicles comprised of HPODE-
PC (16) or HODE-PC (17) esters in pH 7.4 aqueous buffer at
37 °C. Myeloperoxidase (MPO) serves as an enzymatic catalyst
for initiation of lipid peroxidation in vivo,28 and we recently
showed that free radical-induced phospholipid peroxidation
promoted by the MPO/H2O2 system of leukocytes serves as one
Their similar reactivities toward autoxidation also suggest that
oxidative fragmentations of both the hydroxydienes and the
hydroperoxydienes proceed by identical mechanisms. Autoxi-
(28) Zhang, R.; Brennan, M. L.; Shen, Z.; MacPherson, J. C.; Schmitt, D.;
Molenda, C. E.; Hazen, S. L. J. Biol. Chem. 2002, 277, 46116-46122.
9
J. AM. CHEM. SOC. VOL. 126, NO. 18, 2004 5703

A R T I C L E S
Sun and Salomon
Scheme 6
for example, 16, are plausible intermediates in the production
of oxidatively truncated lipids, HNE (13) and HODA-PC (18),
in vivo where a constant flux of radicals and hydroperoxides is
present.
The Oxidative Fragmentations Do Not Require Peroxy-
cyclization to Dioxetane Intermediates. The mechanisms that
generate HODA (8) from 4 and HNE (13) from 9 remain to be
determined. However, the peroxyradical cyclization pathway
outlined in Scheme 2 is not feasible for oxidative fragmentation
of hydroxydienes. The presumption of identical mechanisms for
fragmentations of both the hydroxydienes and the hydroper-
oxydienes leads to the conclusion that the process involving
cleavage of a dioxetane intermediate, formed through perox-
yradical cyclization, is not a significant pathway for oxidative
fragmentations of hydroperoxydienes. A recent study discounted
the intermediacy of a similar dioxetane in the generation of
isoprostanes from cholesteryl-15-hydroperoxyeicosatetraenoate.30
The contrasting chemistry of 9-HODE (4) and 13-HODE (9)
summarized in Figures 3 and 5 supports our presumption that
C-C bond cleavage requires the involvement of a hydroperoxy
group or an alkoxyl radical on one terminus of the scissile C-C
bond involved in oxidative fragmentation. As suggested previ-
ously, the formation of 8 upon oxidation of LA (1) may involve
allylic oxygenation of 4 followed by â-scission of a hydroxy
alkoxyl radical derived from an intermediate hydroxy hydro-
peroxide 6 (Scheme 1).10 As expected for free radical oxidations,
the presence of 10 wt % of vitamin E, a lipophilic antioxidant,
inhibited the fragmentations of both hydroperoxydiene 2 and
hydroxydiene 4. Fragmentation of both 2 and 4 occurred
simultaneously, albeit more gradually, after a delay period during
which the antioxidant was consumed. However, these observa-
tions do not rule out the possibility that formation of γ-hy-
droxyalkenal 8 from 6 involves a polar Hock-cleavage (Scheme
6) analogous to that proposed recently to explain the formation
of HNE (13) through fragmentation of a dihydroperoxide
corresponding to the hydroxy hydroperoxide 14 in Scheme 4.16
Because the hydroxydienes 4 and 9 do not interconvert, in
contrast with the corresponding hydroperoxydienes 13-HPODE
(2) and 9-HPODE (3), the reaction product mixtures generated
from autoxidation of the hydroxydienes are less complex than
those generated from the HPODEs. Consequently, studies of
the 9-HODE (4) to HODA (8) and the 13-HODE (9) to HNE
(13) conversions are especially attractive for further studies
focusing on this fragmentation process.
Figure 6. MPO-promoted autoxidation of HPODE-PC or HODE-PC in
unilamellar liposomes.
mechanism for the generation of isolevuglandins in vivo.29 To
provide free radicals that can initiate autoxidation, we used the
-
MPO-H2O2-NO2 system and generated a constant flux of
the H2O2 through the glucose oxidase-promoted oxidation of
glucose. The consumption of starting dienes and production of
the γ-hydroxyalkenal phospholipid HODA-PC (18) through
oxidative fragmentation of HPODE-PC (16) or HODE-PC (17)
were monitored by LC-ESI-MS/MS (Figure 6).
Both HPODE-PC (16) and HODE-PC (17) were readily
consumed under these physiomimetic autoxidation reaction
conditions. Oxidative fragmentation of both 16 and 17 produced
the biologically active truncated phospholipid HODA-PC (18).
The yield of γ-hydroxyalkenal 18 was 3-fold higher from the
hydroxydiene 17 than from the hydroperoxydiene 16. HNE (13)
is also produced in the MPO-promoted autoxidation of 16 and
17. Yields reached a maximum within 5 h and were about 5%
from the hydroxydiene 17 and about 3% from the hydroper-
oxydiene 16. Therefore, hydroxydienes, for example, 17, that
are generally more abundant in vivo than hydroperoxydienes,
Experimental Procedures
General. High performance liquid chromatography (HPLC) purifica-
tion was performed with HPLC grade solvents using a Waters M600A
solvent delivery system and a Waters U6K injector or a Waters 717
(29) Poliakov, E.; Brennan, M. L.; Macpherson, J.; Zhang, R.; Sha, W.; Narine,
(30) Yin, H.; Havrilla, C. M.; Gao, L.; Morrow, J. D.; Porter, N. A. J. Biol.
Chem. 2003, 278, 16720-16725.
L.; Salomon, R. G.; Hazen, S. L. FASEB J. 2003, 17, 2209-2220.
9
5704 J. AM. CHEM. SOC. VOL. 126, NO. 18, 2004

Oxidative Fragmentation of Hydroxy Octadecadienoates
A R T I C L E S
Table 1. Optimized Parameters for the Mass Spectrometer
Table 2. HPLC Gradients for the Purification of Hydroperoxidiene
acids
CHD derivs
PFB derivs
time (min)
flow (mL/min)
methanol (%)
water (%)
ion mode
capillary (kV)
cone (V)
hex 1 (V)
aperture (V)
hex 2 (V)
LM 1 resolution
HM 1 resolution
ion energy 1
LM 2 resolution
HM 2 resolution
ion energy 2
negative
positive
positive
0
20
25
30
3
3
3
3
70
100
100
70
30
0
0
3
30
0
3
30
10
0
0.5
19
19
1
4
30
10
0
0.5
15
15
1
30
0
0.5
10
10
1
15
15
2
9-Hydroperoxyoctadeca-10,12-dienoic Acid (3, 9-HPODE, Cay-
man). ESI-MS/MS analysis yielded the following diagnostic frag-
ments: m/z 311 [M - H]-; m/z 293; m/z 197; m/z 185; m/z 125 (see
Supporting Information Figure S1B).
13-Hydroxyoctadeca-9,11-dienoic acid (9, 13-HODE, Cayman)
was purified by the same HPLC system as 13-HPODE-d4. ESI-MS/
MS analysis yielded the following diagnostic fragments: m/z 295 [M
- H]-; m/z 278; m/z 195; m/z 112 (see Supporting Information Figure
S2A).
9-Hydroxyoctadeca-10,12-dienoic acid (4, 9-HODE, Cayman) was
purified by the same HPLC system as 13-HPODE-d4. ESI-MS/MS
analysis yielded the following diagnostic fragments: m/z 295 [M -
H]-; m/z 278; m/z 171; m/z 125 (see Supporting Information Figure
S2B).
13-Hydroperoxy-9,10,12,13-tetradeuterio-octadeca-9,11-dienoic
acid (13-HPODE-d4, 2-d4) was prepared by incubating 9Z,12Z-
octadecadienoic-9,10,12,13-d4 acid (1-d4, Cayman) with lipoxygenase
(Sigma, type V)31 and purified with a RP-HPLC column (Phenomenex
LUNA, 10.0 × 250 mm) using methanol/water linear gradients (Table
2). Products were monitored at 234 nm. The retention time of the
product is about 12 min. ESI-MS/MS analysis yielded the following
diagnostic fragments: m/z 315 [M - H]-; m/z 297; m/z 227; m/z 198;
m/z 114 (see Supporting Information Figure S3A).
19
19
2
15
15
2
autosampler. The eluants were monitored using an ISCO V4 UV-vis
detector or a Waters 2996 photodiode array detector. Linoleic acid-d4
(9,10,12,13-tetradeutero linoleic acid), 9(S)-hydroxyoctadeca-10Z,12E-
dienoic acid (9-HODE), 9(S)-hydroperoxyoctadeca-10Z,12E-dienoic
acid (9-HPODE), 13(S)-hydroxyoctadeca-9Z,11E-dienoic acid (13-
HODE), and 13(S)-hydroperoxyoctadeca-9Z,11E-dienoic acid (13-
HPODE) were from Cayman Chemical Co. (Ann Arbor, MI), 1-palm-
itoyl-2-linoleoyl-sn-glycero-3-phosphocholine (LA-PC) was from Avanti
Polar Lipids (Alabaster, AL), catalase (bovine liver, thymol-free) and
glucose oxidase (grade II) were from Boehringer Mannheim (India-
napolis, IN), and myeloperoxidase (human leukocytes) was from Sigma
(St. Louis, MO).
Liquid Chromatography/Mass Spectrometry. LC/ESI/MS/MS
analysis of autoxidation reaction product mixtures was performed on
a Quattro Ultima (Micromass, Wythenshawe, UK) connected to a
Waters 2790 solvent delivery system with an auto-injector. The source
temperature was maintained at 100 °C, and the desolvation temperature
was kept at 200 °C. The drying gas (N2) was maintained at ca. 450
L/h, and the cone flow gas was kept at ca. 50 L/h. The multiplier was
set at an absolute value of 500. Optimized parameters for acids,
cyclohexanedione (CHD) derivatives of HNE, and pentafluorobenzyl
(PFB) derivatives of acids are presented in Table 1. MS scans at m/z
20-400 were obtained for standard compounds.
Argon was used as collision gas at a pressure of 5 psi for MS/MS
analysis. For MS/MS analysis, the collision energy was optimized for
each compound. For multiple reaction monitoring (MRM) experiments,
the optimum collision energy (giving the strongest signal) was
determined for each m/z ion pair. For acidic compounds, the mass
spectrometer was operated in the negative ion mode. For CHD-
derivatized HNE and PFB derivatives, it was operated in the positive
ion mode.
Online chromatographic separation was achieved using a 150 × 2.0
mm i.d. Prodigy ODS-2, 5 µ column (Phenomenex, UK), with a binary
solvent (water and methanol) gradient. The solvents were supplemented
with 0.2% formic acid, whenever the mass spectrometer was operated
in positive mode. For acidic compounds, the gradient started with 100%
water and rose to 100% methanol linearly in 15 min, and elution was
continued for 5 min with 100% methanol. The gradient was then
reversed to 100% water in 0.5 min, and then held for 9.5 min at 100%
water. For CHD-derivatized HNE, the gradient started with 70%
methanol and rose to 100% methanol linearly in 10 min, and was then
held for 5 min with 100% methanol. The gradient was then reversed
to 70% methanol in 0.2 min, and then held for 4.8 min at 70% methanol.
For dipentafluorobenzyl ester compounds, the gradient started with 85%
methanol and rose to 88% methanol linearly in 12 min, and then rose
to 100% methanol linearly over 3 min, which was held for 12 min.
The gradient was then reversed to 85% methanol in 4 min, and held
for 10 min. The solvents were delivered at 200 µL/min.
9-(2-Oxanyloxy)-11-(3,3-dimethyl-2,4-dioxolanyl)undec-10-eno-
ic acid was used as internal standard and was prepared as described
previously.32 ESI-MS/MS analysis yielded the following diagnostic
fragments: m/z 384 [M - H]-; m/z 281; m/z 223; m/z 101 (see
Supporting Information Figure S3B).
9-Hydroxy-12-oxo-10-dodecenoic acid (HODA, 8) was synthesized
according to the published protocol.32 ESI-MS/MS analysis yielded the
following diagnostic fragments: m/z 227 [M - H]-; m/z 209; m/z 163;
m/z 84 (see Supporting Information Figure S4A).
9-Hydroxy-12-oxo-9,10,12-trideuterio-10-dodecenoic acid (HODA-
d3, 8-d3) was prepared by the vitamin C-catalyzed oxidation of 13-
HPODE-d4. 13-HPODE-d4 (500 µg) was dissolved in 50 mM pH 7.4
phosphate buffer with vitamin C (2 mM).33 The mixture was heated at
37 °C under air for 2 h. Catalyst was added to the mixture followed by
BHT. The mixture was then acidified to pH 3 with 0.2 M HCl. The
products were extracted and then fractionated on a RP-HPLC column
(Phenomenex LUNA, 1.0 × 25 cm) with a methanol/water gradient at
0.8 mL/min. The fractions containing HODA-d3, that is, with retention
time similar to HODA-d0, were combined and concentrated by rotary
evaporation. ESI-MS/MS analysis yielded the following diagnostic
fragments: m/z 230 [M - H]; m/z 212; m/z 166, m/z 86 (see Supporting
Information Figure S4B).
9-Hydroxy-12-methoxyimino-dodec-10-enoic Acid (HODA) Meth-
oxime. Synthetic HODA (100 µg) in chloroform was transferred to a
3.7-mL vial, and the solvent was evaporated with a stream of dry N2.
The freshly distilled pyridine (200 µL) containing 2 wt % of methoxime
hydrochloride was added.34 The mixture was sealed under nitrogen and
kept at room temperature overnight with stirring. Pyridine was then
evaporated with a stream of nitrogen, and the residue was dissolved in
0.5 mL of water (pH 3). The solution was extracted with ethyl ether (3
× 0.5 mL). The combined organic layer was washed with water (pH
13-Hydroperoxyoctadeca-9,11-dienoic Acid (2, 13-HPODE, Cay-
man). ESI-MS/MS analysis yielded the following diagnostic frag-
ments: m/z 311 [M - H]-; m/z 293; m/z 223; m/z 113 (see Supporting
Information Figure S1A).
(31) Funk, M. O.; Isaac, R.; Porter, N. A. Lipids 1975, 11, 113-117.
(32) Deng, Y. H.; Salomon, R. G. J. Org. Chem. 1998, 63, 7789-7794.
(33) Lee, S. H.; Oe, T.; Blair, I. A. Science 2001, 292, 2083-2086.
(34) Gardner, H. W. In AdVances in Lipid Methodology-Four; Christie, W. W.,
Ed.; Oily Press: Dundee, Scotland, 1997; pp 1-43.
9
J. AM. CHEM. SOC. VOL. 126, NO. 18, 2004 5705

A R T I C L E S
Sun and Salomon
3). Solvent was evaporated with a stream of nitrogen, and the residue
was analyzed with ESI-MS/MS, producing the following fragments:
m/z 256 [M - H]-; m/z 197 (see Supporting Information Figure S5).
1-Palmitoyl-2-(13-hydroperoxyoctadeca-9Z,10E-dienoyl)-sn-glyc-
ero-3-phosphatidylcholine (16a) and 1-Palmitoyl-2-(9-hydroper-
oxyoctadeca-10Z,12E-dienoyl)-sn-glycero-3-phosphatidylcholine (16b).
An 81:19 regioisomeric mixture of hydroperoxydienes 16a and 16b
was prepared by enzyme-catalyzed oxidation of 1-palmitoyl-2-linoleoyl-
sn-glycero-3-phosphatidylcholine (LA-PC) by a modification of the
method reported previously.35 LA-PC (20 mg, 0.026 mmol, Avanti Polar
Lipids, AL) was suspended in sodium borate buffer (26 mL, pH 9.0,
0.1 M, pretreated with Chelex 100) and was emulsified with vigorous
magnetic stir for 5 min at room temperature. Sodium cholate (112 mg)
was then added. After the suspension became transparent, 26 mg of
soybean lipoxygenase (Fluka) was added. The reaction was terminated
after 1.5 h by the addition of 2:1 CHCl3/MeOH (25 mL) and
centrifugation. The organic layer was collected, and the aqueous layer
was further extracted with chloroform (3 × 25 mL). The combined
organic extracts were concentrated, and the product was purified on a
silica gel column (CHCl3/MeOH/H2O: 65/24/4, TLC: Rf ) 0.4) to
give the title hydroperoxydienoyl phospholipids (15 mg, 73%). The
hydroperoxy functional group was detected on TLC plates with
ammonium thiocyanate-ferrous sulfate solution spray. The ratio of
regioisomeric hydroperoxydienes was determined after reduction to the
Figure 7. LC/ESI/MS/MS analysis of 9-HODE and 13-HODE. For
calibration curves and experimental details, see Quantification of Acids
section.
125 mg of CHD, 5 g of ammonium acetate, and 2.5 mL of acetic acid
were dissolved in deionized water and diluted to 50 mL. The pale yellow
solution was heated at 60 °C for 1 h, cooled with an ice bath, and
purified by passing through a C18 SPE cartridge (1 mL), which was
preconditioned with 5 mL of water and 5 mL of methanol. The elute
was collected and stored at 4 °C.27
1
corresponding hydroxydienes (vide infra). H NMR (CDCl3): δ 6.54
(dd, J ) 14.7, 11 Hz, 1H), 6.03 (dd, J ) 10.8, 10.8, 1H), 5.61 (dd, J
) 17.4, 7.8 Hz, 1H), 5.46 (dd, J ) 17.4, 8.4 Hz, 1H), 5.24 (m, 1H),
4.3-4.4 (4H), 4.17 (dd, J )12.2, 7.4 Hz, 1H), 3.9-4.1 (2H), 3.77 (bm,
2H), 3.35 (bs, 9H), 2.0-2.4 (6H), 1.5-1.7 (4H), 1.1-1.45 (42H), 0.85-
0.95 (6H). HRMS (ESI-MS): m/z 791.0 (MH+) calcd for C42H81NO10P
found 791.0; m/z 813.0 calcd for C42H80NO10PNa+ found 813.0 (see
Supporting Information Figure S6A).
Derivatization of 4-Hydroxy-2-nonenal and Benzaldehyde with
CHD/NH4OAc. Stock solutions of HNE and benzaldehyde (100 ng/
µL) were prepared. First, 10 µL of benzaldehyde solution was mixed
with 10 µL, 50 µL, 100 µL, 500 µL, and 1000 µL of HNE solution.
The solvent was removed with a steam of nitrogen, and the residue
was redissolved in 100 µL of methanol. CHD reagent (1 mL) was
added, and the mixture was incubated at 60 °C for 1 h. The resulting
solution was cooled, and then passed through a C18 SPE cartridge,
which had been washed with 5 mL of methanol and 5 mL of water
sequentially. The cartridge was washed with 5% acetonitrile in water
(9 mL), and the product was washed off with acetonitrile containing
5% water (9 mL). The solvent was removed with a stream of nitrogen.
Before LC-MS/MS analysis, the residue was dissolved in methanol
containing 25% water, and diluted to 0.01 ng/µL benzaldehyde, of
which 0.25 ng was injected into the LC-MS for quantification. ESI-
MS/MS analysis of the HNE-CHD derivative produced characteristic
fragments: m/z 326 ([MH - 18]+); m/z 308 ([MH - 18 - 18]+); m/z
216. ESI-MS/MS analysis of the benzaldehyde-CHD derivative pro-
duced characteristic fragments: m/z 294 ([MH]+); m/z 216 (see
Supporting Information Figure S7).
Autoxidation of HODEs with and without the Presence of the
Hydroperoxide 13-HPODE-d4. Into five 1-mL glass vials was placed
100 µL of 9-HODE or 13-HODE (50 µg/mL, in methanol). Solvent
was evaporated with a stream of nitrogen. The vials were incubated at
37 °C for 1, 2, 4, and 8 h and were stored at -80 °C immediately after
incubation. Into six 1-mL glass vials were placed 100 µL of 9-HODE
or 13-HODE (50 µg/mL, in methanol) and 100 µL of 13-HPODE-d4
(100 µg/mL, in methanol). Solvent was evaporated with nitrogen. The
vials were incubated at 37 °C for 1, 2, 4, 8, and 18 h and were stored
at -80 °C immediately after incubation. To each vial was added 100
µL of methanol with 20 ng internal standard before analysis. The
solution (50 µL) was injected onto a RP-HPLC column (Phenomenex
Prodigy, 2.0 × 150 mm) using methanol/water gradients at 200 µL/
min. The eluant was monitored by an ESI-MS/MS (Quattro Ultima) in
the negative ion mode. To confirm the presence of HODA in the
oxidation product mixture, a product mixture obtained from the mixture
of 10 µg of 9-HODE and 20 µg of 13-HPODE-d4 after 2 h autoxidation
was derivatized with methoxylamine hydrochloride and then analyzed
following the same procedure as used for characterizing authentic
HODA-methoxime (Figure S5). To quantify HNE, to each vial
1-Palmitoyl-2-(13-hydroxyoctadeca-9Z,10E-dienoyl)-sn-glycero-
3-phosphatidylcholine (17a) and 1-Palmitoyl-2-(9-hydroxyoctadeca-
10E,12Z-dienoyl)-sn-glycero-3-phosphatidylcholine (17b). The cor-
responding hydroperoxides (5 mg, 0.0068 mmol) in CHCl3 (5 mL) were
reduced with Ph3P (16 mg, 0.06 mmol). The mixture was stirred for
10 min and then concentrated. The residue was purified on a silica gel
column (65/24/4 CHCl3/MeOH/H2O, TLC Rf ) 0.4) to give the title
hydroxydienoyl phospholipids (4 mg, 82%). 1H NMR (CDCl3): δ 6.41
(dd, J ) 15, 11.4 Hz, 1H), 5.9 (dd, J ) 10.8, 10.8 Hz, 1H), 5.60 (dd,
J ) 15, 6.6 Hz, 1H), 5.34 (dd, J ) 18, 7.8 Hz, 1H), 5.14 (m, 1H),
4.3-4.4 (3H), 4.0-4.1 (4H), 3.94 (2H), 3.89 (bm, 2H), 3.33 (bs, 9H),
2.1-2.3 (4H), 2.05-2.15 (2H), 1.5-1.7 (4H), 1.1-1.45 (42H), 0.85-
0.95 (6H). HRMS (ESI-MS): m/z 775.0 (MH+) calcd for C42H81NO9P
found 775.0; m/z 797.0 calcd for C42H80NO9PNa+ found 797.0 (see
Supporting Information Figure S6B). To determine the ratio of the
9-hydroxy to 13-hydroxy regioisomers, the mixture (20 µg) was
hydrolyzed with 0.2 N NaOH (1 mL) at 45 °C for 1 h. The resulting
solution was acidified to pH 4 and extracted with chloroform (4 × 1
mL). The ratio of 13-hydroxyoctadeca-9Z,10E-dienoic acid to 9-hy-
droxyoctadeca-10E,12Z-dienoic acid was determined to be 81:19,
respectively, by quantifying both isomers using LC-MS (Figure 7).
Autoxidation of 13-HPODE and 9-HPODE. Five 1 mL glass vials
containing 10 µg of 13-HPODE or 9-HPODE in each were heated at
37 °C in an incubator. At 0, 1, 2, 4, and 8 h, a vial was removed from
the oven and stored at -80 °C. Before analysis, 100 µL of internal
standard solution (0.2 µg/mL) in methanol was added, and 50 µL of
the mixture was injected into the LC-MS/MS system. HPLC conditions
and MS conditions are the same as those in other experiments for other
acidic compounds. The amounts of 13-HPODE, 9-HPODE, and HODA
in the autoxidation reaction mixture were monitored simultaneously.
Preparation of Cyclohexanedione (CHD) Reagent. Cyclohex-
anedione (CHD) was freshly recrystallized from ethyl acetate. First,
(35) Brash, A. R.; Ingram, C. D.; Harris, T. M. Biochemistry 1987, 26, 5465-
5471.
9
5706 J. AM. CHEM. SOC. VOL. 126, NO. 18, 2004

Oxidative Fragmentation of Hydroxy Octadecadienoates
A R T I C L E S
Figure 8. (A) Chromatograms of HNE and benzaldehyde after derivati-
zation by CHD reagent and (B) calibration curve for HNE.
containing 5 µg of 9-HODE (or 13-HODE) and 10 µg of 13-HPODE-
d4 was added 10 ng benzaldehyde in 100 µL of methanol, and the
resulting mixture was derivatized with CHD reagent as described for
HNE derivatization (see above).
Quantification of HNE as the CHD Derivative 15. The CHD
derivative 15 of HNE and the analogous derivative of benzaldehyde
were identified individually by reverse phase HPLC coupled to tandem
mass spectrometry in the positive ion mode using multiple reaction
monitoring (MRM). The mass transition m/z 396 to 216 was monitored
for the HNE-CHD derivative, and m/z 294 to 216 was monitored for
the benzaldehyde-CHD derivative (Figure 8A). The amount of ben-
zaldehyde injected into LC-MS was 0.25 ng. A standard curve (Figure
8B) was generated by preparing samples containing a fixed amount of
internal standard (benzaldehyde) and various amounts of HNE, and
plotting peak area ratio versus amount of HNE. Each point in the figure
is the mean of triplicate determinations.
Quantification of Acids. 13-HPODE-d4, 13-HPODE, 13-HODE,
9-HODE, HODA, and HODA-d3 were quantified by LC-ESI-MS/MS
with MRM. For 13-HPODE-d4, the mass transition m/z 315.4 to 114.6
was monitored. For 13-HPODE, the mass transition m/z 311.4 to 112.6
was monitored. For 13-HODE, the mass transition m/z 295 to 195 was
monitored. For 9-HODE, the mass transition m/z 295 to 171 was
monitored. For HODA, the mass transition m/z 227.3 to 84 was
monitored. For HODA-d3, the mass transition m/z 230.3 to 86 was
monitored. Addition of 9-(2-oxanyloxy)-11-(3,3-dimethyl-2,4-dioxola-
nyl)undec-10-enoic acid as internal standard to the sample prior to
analysis permitted for adjustment for ionization variations for each
injection. For internal standard, the mass transition m/z 382.9 to 100.6
was monitored. Calibration curves were built by injecting various
amounts of 13-HPODE, 13-HODE, 9-HODE, HODA, and 10 ng
internal standard into the LC/MS/MS (Figure 9).
Figure 9. LC/ESI/MS/MS analysis of lipid oxidation products. (A) Each
compound was identified individually by reverse phase HPLC coupled to
tandem mass spectrometry in the negative ion mode. Multiple reaction
monitoring analysis for individual compounds was performed by monitoring
the characteristic parent to daughter ion transitions noted. (B) Calibration
curves generated by incorporating a fixed amount of internal standard and
varying levels of each analyte, and plotting peak area ratio versus nanograms
for each analyte.
Briefly, vesicles of HPODE-PC or HODE-PC (0.13 µmol/mL), MPO
(60 nM), sodium nitrite (50 µM), glucose (100 µg/mL), and glucose
oxidase (20 ng/mL) were incubated in sodium phosphate buffer (50
mM, pH 7.4) supplemented with DTPA (200 µM) at 37 °C. For analysis
of the reaction mixture, 1,2-dimyristoyl-sn-glycero-3-phosphatidylcho-
line (DMPC, 7.4 nmol/mL) was added as an internal standard, and
lipids were extracted immediately by the method of Bligh and Dyer.37
The extracts were dried over nitrogen and stored at -80 °C for less
than 24 h before analysis. The extracts were analyzed by LC-ESI-MS/
MS. HODA-PC, HPODE-PC, and HODE-PC were quantified by
monitoring the mass transition from the parent ion (MH+) to a common
fragment (m/z 184) as described previously.4 To determine the
production of HNE, an aliquot (60 µg) in 100 µL of methanol was
treated with benzaldehyde (10 ng) in methanol (20 µL) followed by
addition of CHD reagent (1 mL) as described above.
Acknowledgment. We thank the National Institutes of Health
for support of this research by grants GM21249 and HL53315.
We thank Lian Shan for guiding our LC-ESI-MS/MS analyses.
All mass spectrometry experiments were performed in the
Cleveland Clinic Foundation Mass Spectrometry Core Facility.
Myeloperoxidase-Promoted Oxidation of HPODE-PC and HODE-
PC. Vesicles comprised of the 13-hydroperoxy regioisomer 16a (81%)
and the 9-hydroperoxy regioisomer 16b (19%) of HPODE-PC or the
corresponding regioisomers 17a and 17b of HODE-PC were prepared
by extrusion as described previously.4 Autoxidation promoted by
myeloperoxidase (MPO) was performed as described previously.36
Supporting Information Available: Negative ion ESI-MS/
MS spectra of 13-HPODE (2), 13-HPODE-d4 (2-d4), 9-HPODE
(36) Gu, X.; Sun, M.; Gugiu, B.; Hazen, S.; Crabb, J. W.; Salomon, R. G. J.
Org. Chem. 2003, 68, 3749-3761.
(37) Bligh, E. G.; Dyer, W. J. Can. J. Biochem. Physiol. 1959, 37, 911-917.
9
J. AM. CHEM. SOC. VOL. 126, NO. 18, 2004 5707

A R T I C L E S
Sun and Salomon
(3), 9-HODE (4), HODA-d0 (8), HODA-d3 (8-d3), 13-HODE
(9), and HODA methoxime. Positive ion ESI-MS/MS spectra
of HPODE-PC (16), HODE-PC (17), and CHD derivatives of
(16) and HODE-PC (17). This material is available free of
charge via the Internet at http://pubs.acs.org.
1
JA038756W
HNE (13) and benzaldehyde. H NMR spectra of HPODE-PC
9
5708 J. AM. CHEM. SOC. VOL. 126, NO. 18, 2004