Coordination (ag+) ion spray - Mass spectrometry of peroxidation products of cholesterol linoleate and cholesterol arachidonate: High-performance liquid chromatography - Mass spectrometry analysis of peroxide products from polyunsaturated lipid autoxidation

Article abstract of DOI:10.1021/ja001180f

Lipid peroxidation of polyunsaturated fatty acids and esters leads to a complex mixture of hydroperoxides and cyclic peroxides, some of which possess potent biological activity. These product mixtures contain dozens of diastereomers and regioisomers. The technique of coordination ion spray - mass spectrometry (CIS - MS), recently reported by Bayer and collaborators, proves to be a powerful tool for the analysis of complex peroxide mixtures. Silver ion forms readily detected Ag⁺ adducts of peroxides and hydroperoxides. These ions, observed at [M+ 107] and [M+ 109], undergo fragmentation typical of hydroperoxides, and cyclic peroxides. Thus, Hock fragmentation is observed from many of the silver ion adducts of hydroperoxides and cyclic peroxides undergo fragmentation to give aldehydes and epoxides. Silver ion coordination ion spray - mass spectrometry (Ag⁺CIS - MS) can be coupled to normal-phase high-performance liquid chromatography (HPLC) by postcolumn addition of AgBF₄, allowing the use of powerful techniques such as selected ion monitoring and selected reaction monitoring. This coupling permits, for the first time, the combination of powerful normal-phase separation techniques with detection methods that provide unambiguous structural information of complex peroxide compounds.

Full text of DOI:10.1021/ja001180f

8042
J. Am. Chem. Soc. 2000, 122, 8042-8055
Coordination (Ag+) Ion Spray-Mass Spectrometry of Peroxidation
Products of Cholesterol Linoleate and Cholesterol Arachidonate:
High-Performance Liquid Chromatography-Mass Spectrometry
Analysis of Peroxide Products from Polyunsaturated Lipid
Autoxidation
Christine M. Havrilla,†,§ David L. Hachey,‡ and Ned A. Porter*,†,‡,§
Contribution from the Department of Chemistry and Center in Molecular Toxicology,
Vanderbilt UniVersity, NashVille, Tennessee 37235, and the Department of Chemistry, Duke UniVersity,
Durham, North Carolina 27708
ReceiVed April 4, 2000
Abstract: Lipid peroxidation of polyunsaturated fatty acids and esters leads to a complex mixture of
hydroperoxides and cyclic peroxides, some of which possess potent biological activity. These product mixtures
contain dozens of diastereomers and regioisomers. The technique of coordination ion spray-mass spectrometry
(CIS-MS), recently reported by Bayer and collaborators, proves to be a powerful tool for the analysis of
complex peroxide mixtures. Silver ion forms readily detected Ag+ adducts of peroxides and hydroperoxides.
These ions, observed at [M+107] and [M+109], undergo fragmentation typical of hydroperoxides, and cyclic
peroxides. Thus, Hock fragmentation is observed from many of the silver ion adducts of hydroperoxides and
cyclic peroxides undergo fragmentation to give aldehydes and epoxides. Silver ion coordination ion spray-
mass spectrometry (Ag+CIS-MS) can be coupled to normal-phase high-performance liquid chromatography
(HPLC) by postcolumn addition of AgBF4, allowing the use of powerful techniques such as selected ion
monitoring and selected reaction monitoring. This coupling permits, for the first time, the combination of
powerful normal-phase separation techniques with detection methods that provide unambiguous structural
information of complex peroxide compounds.
The study of the free radical-mediated reaction of lipids with
oxide (R-OOH) products, and methods of analysis have been
reported for the corresponding alcohols (R-OH) formed by
reduction of the hydroperoxides.5 The identification of peroxide
products formed from lipids is tedious and time consuming,
usually requiring several chromatography steps or conversions
to known compounds.6
Analysis of free radical oxidation products from cholesteryl
arachidonate poses an even more formidable problem. Choles-
teryl arachidonate has sites for free radical oxidation on the
arachidonate chain at carbons 7, 10, and 13 as well as on the
molecular oxygen has been of interest for decades. In the recent
past, this process has received attention because of its probable
role in destructive biological processes. In living systems,
autoxidation of lipids has been implicated in DNA and protein
modification, radiation damage, aging, modification of mem-
brane structure, tumor initiation, and deposition of arterial
plaque.1
Low-density lipoprotein (LDL) is the major carrier of
cholesteryl esters in human blood plasma, and the free radical-
mediated modification of LDL may play an important role in
the development of atherosclerosis.2 For this reason, much effort
has been devoted to the study of lipoprotein oxidation and its
prevention over the past decade.3 Peroxidation of LDL neutral
lipids affords peroxides mainly from cholesteryl linoleate (Ch18:
2), and cholesteryl arachidonate (Ch 20:4), but the mixture of
products formed from these compounds is complex.4 Oxidation
of cholesteryl linoleate, 1, leads to several different hydroper-
(3) (a) Steinbrecher, H. P.; Lougheed, M.; Kwan, W. C.; Dirks, M. J.
Biol. Chem. 1989, 264, 15216-15223. (b) Morel, D. W.; DiCorleto, P. E.;
Chilsom, G. M. Arteriosclerosis 1984, 4, 357-364. (c) Steinbrecher, U.
P.; Parthasarathy, S.; Leake, D. S.; Witztum, J. L.; Steinberg, D. Proc. Natl.
Acad. Sci., U.S.A. 1984, 81, 3883-3887. (d) Ingold, K. U.; Bowry, V. W.;
Stocker, R.; Walling, C. Proc. Natl. Acad. Sci., U.S.A. 1993, 90, 45-49.
(e) Esterbauer, H.; Gebicki, J.; Puhl, H.; Ju¨rgens, G. Free Radical Biol.
Med. 1992, 13, 341-390. (f) Sato, K.; Niki, E.; Shimasaki, H. Arch.
Biochem. Biophys. 1990, 279, 402-405. (g) Bowry, V. W.; Stocker, R. J.
Am. Chem. Soc. 1993, 115, 6029-6044. (h) Bowry, V. W.; Stanley, K. K.;
Stocker, R. Proc. Natl. Acad. Sci., U.S.A. 1992, 89, 10316-10320. (i) Fu,
S.; Davies, M. J.; Stocker, R.; Dean, R. T. Biochem. J. 1998, 333, 519-
525. (j) Berliner, J. A.; Heinecke, J. W. Free Radical Biol. Med. 1996, 20,
707-727. (k) McIntyre, T. M.; Zimmerman, G. A.; Prescott, S. M. J Biol
Chem. 1999, 274, 25189-25192.
* To whom correspondence should be addressed at Department of
Chemistry, Vanderbilt University, Nashville, TN 37235. Telephone: 615-
343-2693; Fax: 615-343-5478; E-mail: n.porter@Vanderbilt.edu.
† Department of Chemistry, Vanderbilt University.
‡ Center in Molecular Toxicology, Vanderbilt University.
§ Duke University.
(4) For a complete discussion of cholesterol autoxidation, see Smith, L.
L. Cholesterol Autoxidation; Plenum Press: New York, 1981.
(1) (a) Nicholson, A. C.; Hajjar, D. P. Am. Scientist 1995, 83, 460-
467. (b) Feher, J.; Lengyel, G.; Blazovics, A. Scand. J. Gastroenterol. 1998,
288S, 38-46. (c) Welsch, C. W. Free Rad. Biol. Med. 1995, 18, 757-773.
(d) Knight, J. A. Ann. Clin. Lab. Sci. 1997, 27, 11-25.
(2) Steinberg, D.; Parhsarathy, S.; Carew, T. E.; Khoo, J. C.; Witztum,
J. L. N. Engl. J. Med. 1989, 320, 915-924.
(5) Kenar, J. A.; Havrilla, C. M.; Porter, N. A.; Guyton, J. R.; Brown,
S. A.; Klemp, K. R.; Selinger, E. Chem. Res. Tox. 1996, 9, 737-744.
(6) (a) Porter, N. A.; Logan, J.; Kontoyiannidou, V. J. Org. Chem. 1979,
44, 3177-3181. (b) Porter, N. A.; Wolf, R. A.; Yarboro, E. M.; Weenen,
H. Biochem. Biophys. Res. Comm. 1979, 89, 1058-1064.
10.1021/ja001180f CCC: $19.00 © 2000 American Chemical Society
Published on Web 08/03/2000

Peroxide Products from Polyunsaturated Lipid Autoxidation
J. Am. Chem. Soc., Vol. 122, No. 33, 2000 8043
Several new ionization techniques were explored to provide
a more direct solution to the peroxide and hydroperoxide
analytical problem, and the method that we have found to be
most useful is coordination ion-spray-mass spectrometry (CIS-
MS), a technique reported recently by Bayer and collaborators
who have used this method to analyze unsaturated organic
compounds.8 This ionization method, when combined with
techniques available to modern tandem-mass spectrometers,
provides valuable structural information for several peroxides
and hydroperoxides. One particular advantage of CIS-MS is
that it can be coupled to normal-phase HPLC, whereas most
ionization methods are normally used in conjunction with
reversed-phase chromatography, in aqueous solvents.
A typical CIS-MS spectrum is shown in Figure 2 for
compound 3, the trans,cis-conjugated diene with peroxide
substitution on carbon 13 of the linoleate chain. The spectrum
results from a direct liquid injection (DLI) experiment of 3 in
1:1 hexane:2-propanol with 2 equivalents of AgBF4. The
dominant ions observed are the [M3+Ag]+ adducts at m/z )
787 and 789, formed from 107Ag and 109Ag isotopes that are
present in a ratio of ∼1:1. Collision Induced Dissociation (CID)
experiments on the complex of 3 and 107Ag+ at m/z ) 787 gave
fragment ions at m/z ) 769 [M3+Ag+-H2O]+ and m/z ) 687
(loss of hexanal). CID on the 109Ag+ peak gave corresponding
fragment ions at m/z ) 771 [M3+Ag+-H2O]+ and m/z ) 689.
Analysis of 4, the trans,trans-substituted geometric isomer of
3, gave essentially the same spectra as those observed for 3.
Analysis of the two hydroperoxides 5 and 6 having the
peroxide substitution at the 9 position on the linoleate chain
gave the [M5+Ag]+ ion doublet at m/z ) 787 and 789. CID
experiments on the Ag adducts of 5 (or 6) with 107Ag+ at m/z
) 787 gave fragment ions at m/z ) 769 [M5+Ag+-H2O]+ and
m/z ) 647. The spectra of the carbon-13-substituted hydrop-
eroxides (3 and 4) and the carbon-9-substituted isomers (5 and
6) are essentially identical with the exception that 3 and 4 give
CID fragments at 687/689 whereas 5 and 6 give fragments at
647/649.
Figure 1. Free radical oxidation products formed from cholesteryl
linoleate.
cholesterol backbone at, for example, carbon 7. These reactive
centers are indicated in structure 2.
One can quickly come to the conclusion that several dozen
peroxide- or hydroperoxide-containing compounds could be
formed in the oxidation of 2. Conventional techniques (product
isolation, spectroscopic identification, and conversion to and
comparison with known compounds) are unrealistic for the
analysis of oxidation products formed from complex biological
lipid sources. For that reason, we sought new methods that are
appropriate for peroxide and hydroperoxide analysis. Mass
spectrometric methods that provide structural information and
that are compatible with normal-phase high-performance liquid
chromatography (HPLC) would prove particularly useful for
examination of the oxidation products derived from cholesteryl
esters such as 1 and 2.
Results and Discussion
Cholesteryl Linoleate. We have previously developed a
method for analysis of the oxidation products of cholesteryl
linoleate5 and we have reported the use of this protocol for the
analysis of products formed in the azo-initiated peroxidation
of LDL. The four hydroperoxide products, 3-6, are shown in
Figure 1. Direct analysis of the hydroperoxides or the corre-
sponding alcohols, 7-10, by mass spectrometry (MS) is
complicated by the fact that most MS ionization methods result
in loss of water from the parent ion.7 Furthermore, the resultant
mass spectrum is a composite of the sterol and fatty acid
fragmentation pathways. Dehydration of secondary hydroper-
oxides such as 3-6 leads to ketones, which give fragment ions
that can provide structural information. Most of the useful MS
ionization techniques are not compatible with normal-phase
HPLC however, which is the method of choice for the separation
of complex mixtures of oxidation products from lipids such as
cholesteryl esters. Indeed, normal-phase HPLC separation of
the cholesteryl linoleate oxidation products 7-10 was possible,
and structures were assigned, after separation, by conversion
of the cholesteryl ester to the corresponding methyl esters and
comparison with samples previously characterized.5 The pro-
cedure was tedious and time consuming.
We suggest that this fragmentation pattern observed is due
to Hock fragmentation of the silver ion hydroperoxide complex.
Hock fragmentation (or rearrangement), shown in Figure 3, is
commonly observed in solution for lipid hydroperoxides such
as 3-6.9 This fragmentation is promoted by protic or Lewis
acids at moderate temperatures. The process catalyzed by silver
(8) (a) Bayer, E.; Gfrorer, P.; Rentel, C. Angew. Chem.-Int. Ed. 1999,
38, 992-995. (b) See also, Suma, K.; Raju, N. P., and Vairamani, M., Rapid
Comm. Mass Spectrom. 1997, 11, 1939-1944.
(9) (a) Hock, H. Angew. Chem. 1936, 49, 595. (b) For a review see,
Frimer, A. A. Chem. ReV. 1979, 79, 359-387. (c) Tanigawa, S.; Kajiware,
T.; Hatanaka, A. Phytochemistry 1984, 23, 2439-2444. (d) Gardner, H.
W.; Planter, R. D. Lipids 1984, 19, 294-298.
(7) (a) MacMillian, D. K.; Murphy, R. C. J. Am. Soc. Mass Spectrom.
1995, 6, 1190-1201. (b) See also, Nakamura, T.; Hall, L.; Murphy R. C.
In Eicosanoids and Other BioactiVe Lipids in Cancer, Inflammation, and
Radiation Injury, 4; Honn, K. V., Ed.; Advances in Experimental Medicine
and Biology; Kluwer Academic/Plenum: New York, 1999; Vol. 469, pp
539-545.

8044 J. Am. Chem. Soc., Vol. 122, No. 33, 2000
HaVrilla et al.
assigned to the CID fragments m/z ) 687/689 from hydroper-
oxides 3 and 4 and those assigned to the CID fragments m/z )
647/649 from hydroperoxides 5 and 6 are shown in Figure 3.
Coupling HPLC and Ag+ coordination ion spray-mass
spectrometry (HPLC-Ag+CIS-MS) was possible by postcol-
umn mixing of a 0.26 mM AgBF4 in 2-propanol with the
effluent from a normal-phase separation. Chromatography was
carried out with two Beckman Ultrasphere narrow bore silica
columns at 150 µL/min of mobile phase (0.35% 2-propanol in
hexane). The mixing of column effluent with the AgBF4
coordinating ion was done with a Upchurch PEEK high-pressure
mixing tee, with the silver solution added via syringe pump at
a rate of 75 µL/min. The tetrafluoroborate salt of silver is soluble
in most organic solvents making it ideal for use in this
application.
A typical chromatogram for the mixture of hydroperoxides
formed in the autoxidation of 1 is shown in Figure 4.10 Panel A
shows the chromatogram resulting from UV detection at 234
nm, and Panel B shows the total ion chromatogram of the Ag
adducts. The chromatograms shown as Panels C and D in Figure
4 are the result of operation of the spectrometer in the selected
reaction monitoring (SRM) mode. In SRM, a specific parent-
to-daughter mass conversion produced at a characteristic energy
in the collision cell is monitored. This mode clearly distinguishes
the carbon-9-substituted hydroperoxides (Panel C), from the
carbon-13-substituted compounds (Panel D). SRM essentially
utilizes the differences in Hock fragmentation of the hydro-
peroxides to differentiate the eluting regioisomers.
Comment should also be made about differences observed
between the chromatogram with UV detection (Figure 4, Panel
A) and detection by CIS-MS (Figure 4, Panel B). Two peaks
are observed between 24 and 25 min when CIS-MS detection
is used that are not observed in the chromatograms obtained
with UV detection at 234 nm. The first compound eluting in
this region of the chromatogram gives [M+Ag]+ ) 787 adducts
that undergo CID to give m/z ) 769 [M+Ag-H2O]+ and a
fragment at m/z ) 387. The m/z ) 387 fragment is assigned to
Ag+ + C18H32O2, the silver ion complex of unoxidized linoleic
acid. This fragment is observed in CID spectra of the parent
cholesteryl linoleate, with the side-chain fragmentation from the
cholesterol backbone apparently being an important dissociation
mode. This compound is clearly a hydroperoxide of cholesteryl
linoleate as evidenced by HPLC-CIS-MS, and it undergoes
dehydration in CID mode but has an intact linoleate chain. Based
on this evidence, we make the tentative structural assignment
to this compound as the B-ring hydroperoxides of cholesteryl
linoleate, 11R or 11â.4 Some cholesterol backbone oxidation
apparently competes with linoleate side-chain oxidation. The
compound eluting at 25 min is also clearly a hydroperoxide of
cholesteryl linoleate, but its CID spectrum is not instructive.
Figure 2. Direct injection silver ion coordination ion spray mass
spectrum of 3: (a) full scan; (b) blow up of [M+Ag}+ region; (c)
collision induced dissociation spectrum of m/z ) 787.
Cholesteryl Arachidonate. The free radical oxidation prod-
ucts from cholesteryl arachidonate have not been previously
characterized. Based on the products of oxidation formed from
other arachidonate esters, one anticipates formation of six
hydroperoxides, 12-17, analogous to the linoleate hydroper-
oxides. The acyclic hydroperoxides 12-17 are indeed formed
if oxidation of the ester is carried out in the presence of good
hydrogen atom donors, such as 1,4-cyclohexadiene. Structures
for two of the hydroperoxides, 12 and 14 [cholesteryl-5-
hydroperoxy eicosatetraenoate (Ch-5-HPETE) and cholesteryl-
9-hydroperoxy eicosatetraenoate, (Ch-9-HPETE), respectively]
Figure 3. Hock fragmentation of a diene hydroperoxide and silver
ion adducts observed for 3 and 5.
ion Lewis acid is shown in Figure 3, but a proton could promote
a similar rearrangment. Hock fragmentation in the Ag+-CID
experiment apparently occurs in the low-pressure collision cell
and it is promoted by energy deposition in the collision process.
It provides unambiguous information about the position of
peroxide substitution on the linoleate chain. The structures
(10) Autoxidation was initiated by 10% di-tert-butyl hyponitrite (DTBN),
see (a) Mendenhall, G. D. Tetrahedron Lett. 1983, 24, 451-454. (b) Traylor,
T. G.; Kiefer, H. Tetrahedron Lett. 1966, 49, 6163-6168.

Peroxide Products from Polyunsaturated Lipid Autoxidation
J. Am. Chem. Soc., Vol. 122, No. 33, 2000 8045
Figure 4. Chromatograms of cholesteryl linoleate oxidation mixture: (a) UV detection at 234 nm; (b) HPLC-CIS-MS in selected ion monitoring
mode for m/z ) 787 and 789; (c) HPLC-CIS-MS in selected reaction monitoring mode for m/z ) 787 f 647; (d) HPLC-CIS-MS in selected
reaction monitoring mode for m/z ) 787 f 687.
are shown, and analogous compounds are also formed as
follows: 13 (Ch-8-HPETE), 15 (Ch-11-HPETE), 16 (Ch-12-
HPETE), and 17 (Ch-15-HPETE). HPLC analysis of the
hydroperoxides 12-17 on silica separates all but 13 and 14,
which are only partially resolved. Reduction of the hydroper-
oxides to the corresponding alcohols and preparative HPLC
provided pure compounds that were characterized by proton
nuclear magnetic resonance spectroscopy (1H NMR) and by
conversion to the hydroxy eicosatetraenoate (HETE) methyl
esters by reaction with sodium methoxide in methanol. The
HETE methyl esters have been previously characterized, and
comparison of the compounds derived from 12-17 with the
known standards permitted assignments of structures to the
hydroperoxides.
derived from Hock fragmentation of the arachidonate. Hock
fragmentation of Ch-15-HPETE gives a fragment at m/z ) 343
(Panel C). The 11- and 12-hydroperoxides of cholesteryl
arachidonate give identical Hock fragmentation ions, and Panel
D shows the sum of signals detected from the SRM for m/z )
303. Ch-8-HPETE and Ch-9-HPETE both give the same Hock
fragmentation product and Panel C shows the SRM for this ion
at m/z ) 283, and the corresponding fragmentation of the Ch-
5-HPETE gives m/z ) 327.
Peroxyl Radical Cyclization. Hydroperoxide products 7-12
are major products of free radical oxidation of 2 only in the
presence of good hydrogen atom donors, such as 1,4-cyclo-
hexadiene. The product mixture becomes much more complex
when cyclohexadiene is not used. Peroxyl radicals derived from
2 can undergo peroxyl radical cyclization in competition with
hydrogen atom abstraction, and cyclization can lead to a host
of different products.11 Although cyclization products have not
been previously reported from cholesteryl arachidonate, one
expects them to form based on studies carried out and reported
on the corresponding free acids and methyl esters.12
HPLC-Ag+CIS-MS of the mixture of 12-17, formed by
autoxidation of 2, provided support for the assigned structures.
The chromatogram is presented in Figure 5, with UV chroma-
tography detection shown in Panel A. SRM analysis of the
parent 107Ag+ adduct at m/z ) 811 gives fragments that identify
the position of substitution on the arachidonate chain. All of
the SRM fragments result from cleavage of the cholesterol fatty
acid link with characteristic fragments, shown in Figure 5,
Scheme 1 outlines a mechanism that is useful in understand-
ing the mixture of products that can form in the oxidation of

8046 J. Am. Chem. Soc., Vol. 122, No. 33, 2000
HaVrilla et al.
Figure 5. Chromatograms of cholesteryl arachidonate oxidation mixture formed by co-oxidation with cyclohexadiene: (a) UV detection at 234
nm; (b) HPLC-CIS-MS total ion current; (c) HPLC-CIS-MS in selected reaction monitoring mode for m/z ) 787 f 343; (d) HPLC-CIS-MS
in selected reaction monitoring mode for m/z ) 787 f 303; (e) HPLC-CIS-MS in selected reaction monitoring mode for m/z ) 787 f 327.
Scheme 1
different products 7-12. Only one of these peroxyl radicals,
18, is shown in the Scheme, but similar chemistry derives from
the other radicals. Radical 18 can undergo hydrogen atom
abstraction to give the corresponding hydroperoxide or 5-exo
cyclization of the radical can occur giving 19. Two competing
pathways exist for reactions of the carbon radical 19. Cyclization
onto the conjugated diene leads ultimately to isoprostane
structures, whereas addition of oxygen, which occurs presumably
at the diffusion-controlled rate, gives a new peroxyl radical that
can abstract hydrogen or cyclize again to form a second 1,2-
dioxolane. The addition of oxygen to radical 19 would give
(11) (a) Porter, N. A., Caldwell, S. E., Mills, K. A. Lipids 1995, 30,
277-290. (b) Porter, N. A. Acc. Chem. Res. 1986, 19, 262-268. (c)
O’Connor, D. E.; Mihelech, E. D.; and Coleman, M. C. J. Am. Chem. Soc.
1983, 106, 3577-3584. (d) Porter, N. A.; Funk, M. O.; Gilmore, D.; Isaac,
R.; Nixon, J. J. Am. Chem. Soc. 1976, 98, 6000-6005. (e) Funk, M. O.;
Isaac, R.; Porter, N. A. J. Am. Chem. Soc. 1975, 97, 1281-1282. (f) Porter,
N. A.; Funk, M. O. J. Org. Chem. 1975, 40, 3614-3615. (g) Pryor, W. A.;
Stanley, J. P. J. Org. Chem. 1975, 40, 3615-3617.
cholesteryl arachidonate. Six different peroxyl radicals are
formed in oxidation of arachidonate, corresponding to the six
(12) Khan, J. A.; Porter, N. A., Angew. Chem. 1982, 94, 220-221.

Peroxide Products from Polyunsaturated Lipid Autoxidation
J. Am. Chem. Soc., Vol. 122, No. 33, 2000 8047
Scheme 2
temperature source of t-butoxy radicals, led to a complex product
mixture of more polar peroxidic compounds. HPLC analysis
of this product mixture gave a chromatogram that is displayed
in Figure 6. The chromatogram shown is in SIM detection mode
at m/z ) 875 or 877, although essentially the same trace is
obtained by UV detection (at 234 mm) or by SRM mode m/z
) 875 f 591. Several fractions were isolated by semipreparative
chromatography of this product mixture. Material sufficient for
analysis by NMR spectroscopy and direct-injection MS was
obtained from five of the fractions identified as B, C, E, F/G,
and I in Figure 6.
In Figure 7 is presented the Ag+CIS-MS spectrum obtained
by direct injection of fraction B of the chromatogram shown in
Figure 6. The dominant ions observed are the [M21+Ag]+ ion
at m/z ) 875 and 877 (see insert B), which is consistent with
the addition of four oxygen atoms (+64) to the starting
hydroperoxide, 17. The CID experiments on the complex of
107Ag+ at m/z ) 875 gave fragment ions at m/z ) 709, 651,
637, 591, and 475, as shown in Figure 7C. Similar CID spectra
were obtained from each of the fractions isolated. The data from
the CID experiments suggest that the compounds eluting in
fractions B, C, E, F/G, and I are diastereomers having the serial
cyclic structure, 21 (Figure 8). The molecular ion of 21 + Ag+
is 875/877, and the fragments observed in the CID spectra are
consistent with structure 21. Decomposition of peroxides
initiated by heat, light, protic acid, Lewis acid, or reducing metal
usually results in fragmentation of bonds â to peroxides and
hydroperoxides.14 The fragment ions observed in the CID spectra
and shown in Figure 8 are all the result of such fragmentations
in 21. The fragments may provide unique markers for the serial
cyclic structure, with â fragmentation leading to aldehyde and
epoxide functional groups at multiple sites along the chain being
the dominant CID pathway. In fact, chromatograms essentially
identical to the one shown in Figure 6 can be obtained by
operation in the SRM mode, with reaction of the m+107Ag+
complex (m/z ) 875) to fragments at m/z ) 709, 651, 637,
591, 475 being used for detection. Multiple-ion SRM confirms
the structure because each of these ions is observed in the CID
spectrum of 21.
Scheme 3
members of the class of compounds identified in Scheme 1 as
monocyclic and serial cyclic products. Note that the monocyclic
compound contains three stereogenic centers, giving rise to eight
possible stereoisomers (four pairs of enantiomers), whereas the
five stereogenic centers in the serial cyclic structure leads to
16 pairs of enantiomers.
To simplify the mixture of higher oxidation products of
cholesteryl arachidonate with the hope of isolating and char-
acterizing pure compounds, we sought to generate one specific
peroxyl radical capable of cyclization. The approach, which has
been used in analogous systems,9c-f is outlined in Scheme 2.
The hydroperoxide 17, which can be prepared in large quantities
by a chemo-enzymatic synthesis, may be used as a source of
peroxyl radical 20. Hydroperoxide hydrogens are readily
abstracted by alkoxyl radicals, and hydrogen atom exchange
between hydroperoxides is also a facile process. Thus, treatment
of 17 with radical initiators that provide alkoxyl radicals starts
a chain process in which a peroxyl radical is generated from
17, and this peroxyl radical undergoes fragmentation and re-
addition of oxygen, giving peroxyl radical 20.12 Note that radical
20 has a conjugated diene with trans,trans geometry, whereas
peroxyl 18 shown in Scheme 1, has trans,cis diene geometry.
This diene geometry is the only difference between the radical
formed directly in the oxidation of cholesteryl arachidonate (18)
and a radical generated indirectly (20) via a primary oxidation
product 17. In an oxidation mixture derived from cholesteryl
arachidonate, one would likely see products derived from both
18 and 20. Thus, products containing either trans,cis or trans,-
trans diene geometry are expected.
1
Analysis of fraction B by H NMR showed the presence of
five vinylic Hs, one being the cholesterol C-7 vinylic H observed
at 5.38 δ, and the other four observed in a pattern consistent
with a trans,trans-conjugated diene. In addition to conjugated
diene, other protons were observed between 3.8 and 4.8δ, a
region of the spectrum where protons R to peroxides and
hydroperoxides are observed, and between 2.5 and 3.2δ, where
protons substituted at C-4 of a 1,2-dioxolane are observed.
Above 2δ, the complex spectrum typical of steroid derivatives
made simple analysis impossible. The spectra of all other
fractions (C, F/G, and I) showed evidence of trans,trans-
conjugated diene structure and other features similar to those
observed for fraction B. Most of the fractions appeared to
1
contain one major component, but analysis of the H NMR
spectra obtained was not straightforward because most of the
fractions were apparently not completely homogeneous. At-
tempts to obtain 13C NMR spectra were abandoned because of
the long acquisition times required. The NMR experiments defy
simple analysis, but they are also supportive of the structures
suggested on the basis of MS data.
The cholesteryl ester hydroperoxide 17 was prepared in
quantity by the sequence outlined in Scheme 3. Dussault’s
peroxyketal group13 was used to protect 15-hydroperoxyeico-
satetraenoic acid (15-HPETE) and the coupling to cholesterol
was achieved with DCC. Deprotection of the peroxyketal with
acetic acid/THF/water gave 17. Treatment of 0.15 M 17 in
benzene with di-tert-butyl hyponitrite (DTBN; 0.10 eq), a room-
(14) (a) Adam, W.; Duran, N. J. Am. Chem. Soc. 1977, 99, 2729-2734.
(b) Adam, W.; Duran, N. J. Org. Chem. 1973, 38, 1434-1436. (c)
Bloodworth, A. J.; Baker, D. S. Chem. Commun. 1981, 11, 547-549. (d)
Bloodworth, A. J.; Eggelte, H. J. Tetrahedron Lett. 1984, 25, 1525-1528.
(e) Balci, M.; Suetbeyaz, Y. Tetrahedron Lett. 1983, 24, 311-314. (f)
Yoshida, M.; Miura, M.; Nojima, M.; Kusabayashi, S. J. Am.Chem. Soc.
1983, 105, 6279-6285.
(13) Dussault, P.; Sahli, A. J. Org. Chem. 1992, 57, 1009-1012.

8048 J. Am. Chem. Soc., Vol. 122, No. 33, 2000
HaVrilla et al.
Figure 6. Chromatograms of product mixture formed from Ch-15-HPETE rearragement selected ion monitoring of m/z ) 875.
The serial cyclic compounds with the structure 21 result from
5-exo cyclization of a peroxyl radical substituted at C-11 onto
the ∆-8 double bond of the arachidonate (see Scheme 2). A
different set of serial cyclic products is expected starting from
a peroxyl radical substituted at C-9 that is formed by H atom
abstraction at C-7 and oxygen addition at C-9. To access
members of this serial cyclic class, we isolated milligram
quantities of the 5-hydroperoxy cholesteryl arachidonate, 12,
from a product mixture of a cholesteryl arachidonate oxidation
in the presence of cyclohexadiene.
Compound 12 was treated with the free radical initiator
(DTBN) under conditions identical to those used to generate
the serial cyclic structures 21 from 17. This procedure led to a
product mixture that was chromatographed on silica with
detection by UV (234 nm) as well as by Ag+CIS-MS in the
SIM (m/z ) 875-877) and SRM modes. The chromatograms
obtained were essentially identical for the different detection
methods, and an illustrative HPLC chromatogram is presented
in Figure 9. In Figure 9, UV detection of the chromatogram is
shown in Panel A, and Panel B shows the SRM of the reaction
for ion m/z ) 875 f 657. Indeed, the CID of this set of serial
cyclic products, 22, is much simpler than the CID obtained from
serial cyclic products 21 shown in Figure 7. The fragment ion
23 shown in Panel B of Figure 9 is the dominant fragment.
CID showed other ions consistent with the structure 22, but
they were formed to a much lesser extent than was 23.
Autoxidation of cholesteryl arachidonate in benzene with
DTBN initiator gave a mixture that was chromatographed under
conditions similar to those used to analyze the serial cyclic
products 21 and 22. An illustrative chromatogram obtained from
the autoxidation mixture is shown in Figure 10. Panel A of
Figure 10 shows UV detection of this mixture at 234 nm, Panel
B is an SRM chromatogram that detects the formation of serial
cyclic stereoisomers of 21 by detecting the CID reaction of m/z
875 f 591, and Panel C is an SRM that detects stereoisomers
of structure 22 by detecting the CID reaction of m/z 875 f
657. It is noteworthy that most of the fractions observed by
UV detection at 234 nm are also observed by the SRM method.
The MS detection method apparently is detecting the principal
compounds in the product mixture that have conjugated diene
absorption at 234 nm. SRM chromatograms detecting fragments
other than m/z 657 or 591 but that are observed in the CID of
21 or 22 were similar to those shown in Panels B and C. Thus,
SRM of the conversion m/z 875 f 709 gave a chromatogram
essentially identical to that shown in Panel B. This result can
be understood based on the fact that m/z 709 and 591 are both
observed in the CID spectrum of 21 and either fragment detects
elution of a stereoisomer of 21.
As can be seen in Figure 10, the mixture of products with
fragmentation indicative of 21 is more complex than the
chromatogram of 21 containing only the trans,trans diene
functionality (compare Panel B of Figure 10 and Figure 6).
Similarly, the mixture of products with fragmentation consistent
with structure 22 is more complex than a mixture of 22
containing only trans,trans diene functionality (compare Panel
C of Figure 10 and Figure 9). As noted, the mixture obtained
from direct oxidation (Figure 10) contains both trans,cis and
trans,trans dienes, and we have assigned trans,trans compounds
in the Panels B and C of Figure 10 by comparison with the
chromatograms shown in Figures 6 and 9. The other compounds
eluting in the chromatograms that have essentially the same
fragmentation patterns are assigned the structure of the corre-
sponding trans,cis isomers of the serial cyclic products. Thus,
in Panel B, the pair of peaks identified as “B” and “C” (trans,-
trans-conjugated diene), based on comparison with the chro-
matogram shown in Figure 6, are immediately preceded by two
peaks that we tentatively assign as the corresponding trans,cis
isomers. We hesitate to provide more than this tentative
identification at this time because to do so would require product
isolation and a more classical structure proof.
The studies just outlined provide a basis for analysis of an
autoxidation mixture derived from cholesteryl arachidonate. The
product mixture will likely include diastereomers with the
general structures 21 and 22. Another set of serial cyclic
products is also expected, as shown in Scheme 1. Thus, products
analogous to 21 and 22 with trans,cis-conjugated diene stere-
ochemistry are expected in a direct autoxidation of 2. The trans,-
cis products are formed from the first-formed set of peroxyl
radicals (see Scheme 1), whereas the trans,trans products form
after â fragmentation-re-addition, as illustrated in Scheme 2.

Peroxide Products from Polyunsaturated Lipid Autoxidation
J. Am. Chem. Soc., Vol. 122, No. 33, 2000 8049
Figure 7. Direct injection silver ion coordination ion spray mass spectrum of fraction b from the chromatogram shown in Figure 6: (a) full scan;
(b) blow up of [M+Ag]+ region; (c) collision induced dissociation spectrum of m/z ) 875.
Monocyclic Peroxides and Isoprostanes.15 Product mixtures
derived from Ch-15-HPETE containing 21 and from Ch-5-
HPETE containing 22 showed only small amounts of com-
pounds with molecular weights [(M + Ag+) ) 843/845]
consistent with monocyclic peroxide or isoprostane structures
(see Scheme 1). Consider the experiments in which Ch-15-
HPETE, 17, served as precursor for the serial cyclic compounds
21, as shown in Scheme 4. The hydroperoxide 17 was con-
verted to the peroxyl radical by H atom transfer and a mono-
cyclic compound, 24, would be formed from 5-exo cycliza-
tion and oxygen addition. Compound 24 is, itself, also a
hydroperoxide, which means that 24 can serve as a hydrogen-
atom donor and generate a peroxyl radical that can cyclize
again. Thus, even if 24 is formed in the reaction, it can be
converted subsequently to 21 by further hydrogen-atom ex-
change, cyclization, oxygen addition, and hydrogen-atom
transfer.
(15) Morrow, J. D.; Minton, T. A.; Mukundan, C. R.; Campbell, M. D.;
Zacker, W. E.; Daniel, V. C.; Badr, K. R.; Blair, I. A.; Roberts, J. L., Jr. J.
Biol. Chem. 1994, 269, 4317-4326 and references cited.

8050 J. Am. Chem. Soc., Vol. 122, No. 33, 2000
HaVrilla et al.
is evidenced from UV detection. It should also be noted that
there is a fraction observed in Panels B and C of Figure 11 at
15 min that does not have significant absorbance in the UV
spectrum at 234. The compound eluting in this fraction may be
an endoperoxide precursor to an isoprostane based on the MS
and UV evidence. Assignment of structure to fractions based
on the information obtained from only HPLC-MS and UV is
risky, and further experiments are clearly in order to confirm
or deny these structural assignments.
Conclusions. Ag+CIS-MS provides a powerful new tool for
the study of lipid peroxidation. The technique relies on the
interaction of the soft Lewis acid, Ag+, with soft Lewis base
centers such as carbon-carbon double bonds. Highly oxidizable
lipids generally have polyunsaturation and the products of lipid
peroxidation are also highly unsaturated, making the technique
appropriate to the problem. In addition, the silver ion complexes
of hydroperoxides and polyperoxides provide useful structural
information based on understandable fragmentation patterns.
The analysis of cholesteryl linoleate hydroperoxides, Ch-
HPETEs and complex peroxide mixtures derived from rear-
rangements of Ch-5- and Ch-15-HPETEs, as well as products
formed by direct oxidation of Ch-20:4 illustrate that complex
mixtures of diverse peroxides can be studied. Coupling Ag+-
CIS-MS to HPLC makes possible the use of the powerful tools
of SIM and SRM, which can dissect complex mixtures into
constituent molecular classes. The SRM chromatograms shown
in Figure 10, for example, provide detailed information about
the serial cyclic structures that are present in the product mixture,
whereas Figure 11 is an analysis of the same reaction mixture
but with focus on monocyclic or isoprostane structures. The
two analyses serve to deconvolute the chromatogram of the
complex mixture into chromatograms of constituent classes of
peroxide compounds.
This report is intended to demonstrate the potential utlility
of this method and it is not intended to be exhaustive in its
scope. Extensive studies of oxidation product mixtures of diverse
lipid classes are ongoing and will be reported in due course.
Figure 8. Structure of serial cyclic product 21 and fragment ions
formed by collision induced dissociation of the [M+Ag]+ adduct.
The experiments that lead to formation of the serial cyclic
compounds from the isolated hydroperoxide precursors were
biased toward the isolation of serial cyclic compounds. This
fate is not, however, expected for monocyclic peroxides formed
from precursor peroxyl radicals substituted at the 8 or 12
position of the arachidonate chain. Thus, monocyclic peroxide
25, derived from Ch-12-HPETE, cannot form a serial cyclic
product by subsequent reactions. There is some expectation,
therefore, that monocyclic peroxides might be formed in the
autoxidation of Ch-20:4 because compounds such as 25 are
possible products.
In Figure 11 is presented the chromatogram obtained from
analysis of the mixture formed by autoxidation of Ch-20:4.
Detection of products formed was by UV at 234 nm (Panel A)
and by SIM at m/z ) 843/845 (Panel B). Serial cyclic
compounds are detected in HPLC-Ag+CIS-MS at m/z ) 875
and 877, monocyclic compounds and isoprostanes at m/z ) 843
and 845, and Ch-HPETEs at m/z ) 811 and 813. A recon-
structed ion current (RIC) chromatogram of ions at m/z ) 875/
877, 843/845, and 811/813 is presented in Panel C. It seems
clear from the chromatogram shown in Panel B that monocyclic
peroxides or isoprostane endoperoxides are present in the
reaction mixture. Based on the fact that monocyclic compounds
have a conjugated diene whereas isoprostanes do not (see
Scheme 1) and by comparison of Panels A and B, the fractions
identified with an “m” are suggested to be monocyclic peroxides.
This assignment is based on the observation that these fractions
have m/z ions at 843/845 and also have a conjugated diene, as
Experimental Section
General Methods. Reactions involving hydroperoxides were moni-
tored by TLC using a stain of 1.5 g of N,N′-dimethyl-p-phenylenedi-
amine dihydrochloride/25 mL of H2O/125 mL of MeOH/1 mL of acetic
acid. Hydroperoxides yield an immediate pink color,4 whereas protected
hydroperoxides exhibit a pink color after mild charring. TLC was carried
out using 0.2-mm layer thickness, Si-coated aluminum columns (EM
Scientific) that were visualized by UV254, phosphomolybdic acid char,
or the peroxide stain. In general, hydroperoxides were stored as dilute
solutions with 1 mol % butylated hydroxytoluene (BHT) in either
hexanes or benzene at -78 °C, and they were never exposed to
temperatures >40 °C. All HPLC solvents were filtered through
Whatman Nylon membrane filters (0.45-µm pore size) prior to use.
Chemicals. All lipids were purchased from Nu Chek Prep (Elysian,
MI) and were of the highest purity (>99+%). Soybean lipoxidase (Type
I-B, EC. 1.13.11.12) was purchased from Sigma Chemical Company
(St. Louis, MO). Sodium hyponitrite was purchased from Michigan
Technological Institute (Houghton, MI). Chelex-100 Resin was pur-
chased from Bio-Rad Laboratories (Hercules, CA) as the 50-100 mesh,
sodium form. Organic solvents, such as 2-propanol, hexane, benzene,
and ethyl acetate, were HPLC quality and purchased from either
Mallinckrodt, Inc. (St. Louis, MO) or Fisher Chemical (Phillipsburg,
NJ). Hexanes for LC-MS were purchased from Burdick & Jackson
(Muskegon, MI). All other reagents were purchased from Aldrich
Chemical Company (Milwaukee, WI) and used without further purifica-
tion. The free radical initiator di-tert-butyl hyponitrite (DTBN) was
synthesized prior to use.10 Borate buffer (pH 9.0, 0.2 M) was prepared
and treated with Chelex-100 resin (10 g/L buffer) for at least 24 h to
remove transition metal contaminants.

Peroxide Products from Polyunsaturated Lipid Autoxidation
J. Am. Chem. Soc., Vol. 122, No. 33, 2000 8051
Figure 9. Chromatograms of product mixture formed from Ch-5-HPETE rearrangement: (a) UV detection at 234 nm; (b) selected reaction monitoring
of m/z ) 875 f 657.
Figure 10. Chromatograms of cholesteryl arachidonate oxidation mixture: (a) UV detection at 234 nm; (b) HPLC-CIS-MS in selected ion
monitoring mode for m/z ) 875 f 591; (c) HPLC-CIS-MS in selected reaction monitoring mode for m/z ) 875 f 657.
1
Instruments. H NMR spectra were recorded on either a Varian
INOVA 400 MHz spectrometer or a Bruker DRX-400 spectrometer in
CDCl3. Chemical shifts are reported in ppm (δ) with respect to the
residual H signal in CDCl3 (δ ) 7.26 ppm) and coupling constants (J
values) are given in Hz. Multiplicities are indicated by s (singlet), d
(doublet), t (triplet), and m (multiplet). Immediately prior to NMR
sample preparation, CDCl3 was passed through a plug of basic alumina
to remove adventitious HCl.
Analytical HPLC was conducted on a Waters model 600 HPLC
instrument with a Hewlett-Packard 1050 Multiwavelength detector and
a Hewlett-Packard 3396 Series III integrator. For cholesteryl linoleate
and cholesteryl arachidonate hydroperoxide analysis, the HPLC was
equipped with two tandem Beckman Ultrasphere 5-µm silica columns
(4.6 mm × 25 cm), and 0.5% 2-propanol in hexane at a delivery rate
of 1.0 mL/min was used. For analytical analysis of the more polar cyclic
peroxides, HPLC was conducted on a Waters model 600E pump with

8052 J. Am. Chem. Soc., Vol. 122, No. 33, 2000
HaVrilla et al.
µL/min. Column effluent was passed through an Applied Biosystems
785A programmable absorbance UV detector, with detection at 234
nm. An Upchurch PEEK high-pressure mixing tee was connected next
in series for the postcolumn addition of the silver salts. The silver
tertafluoroborate (AgBF4) solution (0.25 mM in 2-propanol) was added
via a Harvard Apparatus (Cambridge, MA) syringe pump at a flow
rate of 75 µL/min. A section of PEEK tubing (1.04 m, 0.25 mm i.d.)
allowed time for the complexation of the silver to the lipid while
delivering effluent to the mass spectrometer. A Rheodyne 7725 injector
was fitted with a 100-µL PEEK loop for 20-50 µL sample injections.
Autoxidation of Cholesteryl Esters for Lipid Hydroperoxide
Analysis. Oxidation of Cholesteryl Linoleate. A round-bottomed flask
was charged with cholesteryl linoleate (300 mg, 0.462 mmol) and was
diluted to 0.20 M with 1,4-cyclohexadiene (0.876 mL, 9.26 mmol) and
dry benzene (1.43 mL). DTBN (2 mg) was added to the solution and
the sealed flask was heated to 37 °C under an oxygen atmosphere.
After 24 h, TLC indicated the formation of peroxidic products.
Butylated hydroxytoluene (BHT, ∼2 mg) was added to the reaction.
Analytical HPLC (tandem Si columns, 0.5% 2-propanol in hexanes, λ
) 234 nm) indicated the formation of four major components in the
mixture: 3, tR ) 15.0 min; 4, tR ) 17.5 min; 5, tR ) 20.1 min; and 6,
tR ) 20.5 min. Semipreparative HPLC (0.66% 2-propanol in hexane)
was used to separate the components. Compounds 5 and 6 were isolated
as a mixture.
Scheme 4
a Waters 717plus Autosampler and a Waters 996 Photodiode array
detector. The autosampler and PDA detector were controlled by
Millenium chromatography software (Waters Corp., Milford, MA). A
single Beckman Ultrasphere 5-µm silica column (4.6 mm × 25 cm)
with a flow rate of 1.0 mL/min of 1.0% 2-propanol in hexanes was
utilized for analytical cyclic lipid hydroperoxide analysis. Semiprepara-
tive HPLC was conducted on a Waters model 600E HPLC instrument,
with a Waters model 481 variable wavelength detector operating at
234 nm and with output to a Fisher Record-All Series 5000 strip chart
recorder. A Dynamax-60 Å silica 83-121-C column (21.4 mm × 25
cm × 8 µm particle) purchased from Rainin Instrument Company
(Woburn, MA), 0.66% 2-propanol in hexane, and a solvent delivery
rate of 10 mL/min were used for the preparative scale separations of
cholesteryl arachidonate hydroperoxides (Ch-HPETEs). Semipreparative
isolation of the more polar cyclic peroxides was accomplished using
the same column with 1.1% 2-propanol in hexane at a flow of 10 mL/
min.
Oxidation of Cholesteryl Arachidonate. In a round-bottomed flask,
cholesteryl arachidonate (400 mg, 0.594 mmol) was dissolved in 1,4-
cyclohexadiene (1.6 mL, 17.0 mmol) and dry benzene (2.7 mL) to give
a solution 0.15 M in lipid and 4.0 M in 1,4-cyclohexadiene. DTBN (3
mg) was added and the reaction was stirred under oxygen at 37 °C for
24 h, after which time BHT (2 mg) was added. Analytical HPLC
(tandem Si columns, 0.5% 2-propanol in hexanes, λ ) 234 nm)
indicated the formation of six major fractions: A, tR ) 11.44 min; B,
tR ) 12.14 min; C, tR ) 13.12 min; D.1, tR ) 16.97 min; D.2, tR
)
17.29 min; and E, tR ) 22.57 min. Preparative HPLC (0.66% 2-propanol
in hexane) was used to separate the components. Compounds D.1 and
D.2 were isolated as a mixture.
Mass Spectrometry. CIS-MS was performed using a Finnigan TSQ-
7000 (San Jose, CA) triple quadrupole mass spectrometer equipped
with a standard API-1 electrospray ionization source outfitted with a
100-µm deactivated fused Si capillary. Data acquisition and spectral
analysis were conducted with ICIS software, version 8.3.2, on a Digital
Equipment Corp. Alpha Station 200 4/166. Data collected for selected
reaction monitoring (SRM) experiments was also processed using
Xcalibur, version 1 (Finnigan, San Jose, CA). Nitrogen gas served both
as the sheath and auxiliary gas, and argon served as the collision gas.
The electrospray needle was maintained at 4.6 kV, and the heated
capillary temperature was 200 °C. The tube lens and capillary voltages
were optimized to maximize ion current for electrospray, with the
optimals determined to be 80 and 20 V, respectively, for cholesteryl
ester analysis. For CID experiments, the collision gas pressure was
typically 2.30-2.56 mTorr. To obtain fragmentation information of
each compound, the dependence of offset-voltage and relative ion
current (RIC) was studied. The collision energy offset was varied from
10 to 40 eV depending on the compound being analyzed. Positive ions
were detected scanning from 100 to 1000 amu in both the parent and
daughter ion scans, with a scan duration of 1 s. Profile data were
recorded for 1 min (∼60 scans) and averaged for analysis.
Identification of Cholesteryl Arachidonate Oxidation Products.
A mixture of cholesteryl arachidonate hydroperoxides (Ch-HPETEs)
was prepared as already described. Preparative-scale HPLC isolation
of the five major fractions was achieved with a Dynamax-60 Å Si 83-
121-C column (21.4 mm × 25 cm × 8 µm particle) in isocratic mode,
with 0.66% 2-propanol in hexane as solvent at a delivery rate of 10
mL/min. Each of the eluting hydroperoxides was collected into a round-
bottomed flask containing BHT (3 mg) and an excess of triphenylphos-
phine (50 mg) for in situ reduction of the hydroperoxides to the
corresponding alcohols. Each of the fractions was concentrated and
rechromatographed under the same solvent conditions to give pure
1
cholesteryl arachidonate alcohols, which were examined by H NMR
and by conversion to their corresponding methyl esters, which were
previously fully characterized.6
Cholesteryl 15-hydroxy-5(Z),8(Z),11(Z),13(E)-eicosatetraenoate
(fraction A). Analytical HPLC (0.5% IPA in hexanes, λ ) 234 nm):
tR 11.44 min (-OOH), tR 15.36 min (-OH); 1H NMR (400 MHz,
CDCl3): δ 6.50 (dd, J ) 15.1, 11.2 Hz, 1H), 5.98 (t, J ) 10.9, 10.9
Hz, 1H), 5.68 (dd, J ) 15.1, 6.7 Hz, 1H), 5.38 (m, 6H), 4.59 (m, 1H),
4.15 (m, 1H), 3.46 (bs, 1H), 2.94 (m, 2H), 2.81 (m, 2H), 2.30-0.83
(m, 57H), 0.66 (s, 3H).
Cholesteryl 12-hydroxy-5(Z),8(Z),10(E),14(Z)-eicosatetraenoate
(fraction B). Analytical HPLC (0.5% IPA in hexanes, λ ) 234 nm):
tR 12.14 min (-OOH), tR 14.87 min (-OH); 1H NMR (400 MHz,
CDCl3): δ 6.54 (dd, J ) 15.2, 11.3 Hz, 1H), 5.97 (t, J ) 11.1, 11.1
Hz, 1H), 5.71 (dd, J ) 15.2, 6.3 Hz, 1H), 5.55 (m, 1H), 5.37 (m, 5H),
4.59 (m, 1H), 4.20 (m, 1H), 3.49 (bs, 1H), 2.91 (m, 2H), 2.34-0.83
(59H), 0.65 (s, 3H).
Cholesteryl 11-hydroxy-5(Z),8(Z),12(E),14(Z)-eicosatetraenoate
(fraction C). Analytical HPLC (0.5% IPA in hexanes, λ ) 234 nm):
tR 13.12 min (-OOH); tR 17.76 min (-OH); 1H NMR (400 MHz,
CDCl3): δ 6.51 (dd, J ) 15.3, 11.0 Hz, 1H), 5.95 (t, J ) 10.5, 10.5
Hz, 1H), 5.67 (dd, J ) 15.1, 6.3 Hz, 1H), 5.45 (m, 6H), 4.60 (m, 1H),
4.21 (m, 1H), 3.56 (bs, 1H), 2.78 (m, 2H), 2.37-0.82 (59H), 0.66 (s,
3H).
Samples were introduced either by direct liquid infusion or HPLC.
For direct liquid injection, stock solutions of the lipids (100 ng/µL in
1% 2-propanol in hexane) were prepared and mixed 1:1 with silver
tetrafluoroborate (51.4 ng/µL in 2-propanol). Samples were introduced
to the ESI source with a syringe pump at a rate of 10 µL/min. For
HPLC sample introduction, a Hewlett-Packard 1090 HPLC system was
used. The auxiliary gas flow rate was increased to between 5 and 10
units to assist in desolvation of the samples. For cholesteryl linoleate
and cholesteryl arachidonate hydroperoxide analysis, normal-phase
HPLC sample introduction was carried out using two tandem Beckman
Ultrasphere narrowbore 5-µm silica columns (2.0 mm × 25 cm)
operated in isocratic mode with 0.35% 2-propanol in hexanes. For
analysis of cyclic peroxide mixtures, sample introduction was carried
out using a single Beckman Ultrasphere narrowbore 5-µm silica column
(2.0 mm × 25 cm) operated in isocratic mode with 1.0% 2-propanol
in hexanes. The flow rate for both modes of chromatography was 150

Peroxide Products from Polyunsaturated Lipid Autoxidation
J. Am. Chem. Soc., Vol. 122, No. 33, 2000 8053
Figure 11. Chromatograms of cholesteryl arachidonate oxidation mixture: (a) UV detection at 234 nm; (b) HPLC-CIS-MS in selected ion
monitoring mode for m/z ) 843 and 845; (c) HPLC-CIS-MS, reconstructed ion current (RIC) for ions 811 and 813 (monohydroperoxides), 843
and 845 (monocyclic peroxides or isoprostanes), and 875 and 877 (serial cyclic peroxides).
Cholesteryl 9-hydroxy-5(Z),7(E),11(Z),14(Z)-eicosatetraenoate (frac-
tion D.1). Analytical HPLC (0.5% IPA in hexanes, λ ) 234 nm): tR
16.97 min (-OOH); tR 22.89 min (-OH); 1H NMR (400 MHz,
CDCl3): δ 6.48 (dd, J ) 15.1, 11.2 Hz, 1H), 6.00 (t, J ) 10.8, 10.8
Hz, 1H), 5.69 (dd, J ) 15.5, 6.6 Hz, 1H), 5.53 (m, 1H), 5.36 (m, 5H),
4.60 (m, 1H), 4.20 (m, 1H), 3.53 (bs, 1H), 2.79 (m, 2H), 2.37-0.83
(59H), 0.65 (3H).
Cholesteryl 8-hydroxy-5(Z),9(E),11(Z),14(Z)-eicosatetraenoate (frac-
tion D.2). Analytical HPLC (0.5% IPA in hexanes, λ ) 234 nm): tR
17.29 min (-OOH), tR 25.80 min (-OH); 1H NMR (400 MHz,
CDCl3): δ 6.57 (dd, J ) 15.1, 11.2 Hz, 1H), 5.97 (t, J ) 10.9, 10.9
Hz, 1H), 5.69 (dd, J ) 15.1, 6.6 Hz, 1H), 5.41 (m, 6H), 4.60 (m, 1H),
4.21 (m, 1H), 3.50 (bs, 1H), 2.91 (m, 2H), 2.34-0.83 (59H), 0.66 (s,
3H).
Cholesteryl 5-hydroxy-6(E),8(Z),11(Z),14(Z)-eicosatetraenoate (frac-
tion E). Analytical HPLC (0.5% IPA in hexanes, λ ) 234 nm): tR
22.57 min (-OOH), tR 41.31 min (-OH);1H NMR (400 MHz,
CDCl3): δ 6.51 (dd, J ) 15.1, 11.2 Hz, 1H), 5.97 (t, J ) 10.9, 10.9
Hz, 1H), 5.67 (dd, J ) 15.4, 6.6 Hz, 1H), 5.36 (m, 6H), 4.60 (m, 1H),
4.17 (m, 1H), 3.52 (bs, 1H), 2.94 (m, 2H), 2.80 (m, 2H), 2.32-0.82
(57H), 0.66 (s, 3H).
Conversion of Cholesteryl Esters to Methyl Esters. Methyl
arachidonate hydroperoxides were prepared as described5,6 by autoxi-
dation, analogous to that described for cholesteryl arachidonate. The
mixture of methyl arachidonate hydroperoxides was treated with an
excess of triphenylphosphine to generate the corresponding alcohols
for comparison purposes. Each isolated cholesteryl arachidonate alcohol
was dissolved in benzene (0.5 mL) and converted to the methyl ester
by treatment with an excess of NaOMe in methanol for 2 h. The
reactions were worked up by the addition of deionized water (1.0 mL)
and glacial acetic acid (100 µL), followed by the extraction of the
aqueous layer with hexane (2 × 5 mL). The combined organic layers
were dried, concentrated, and analyzed by analytical HPLC equipped
with two tandem Beckman Ultrasphere 5µm silica columns (4.6 mm
× 25 cm) in isocratic mode with 0.6% 2-propanol in hexanes containing
0.1% acetic acid at a flow rate of 3.0 mL/min with UV detection at
234 nm. Co-injection of each transesterified fraction with the methyl
arachidonate alcohol mixture confirmed assignment of the elution order.
The elution order of the transesterified fractions was confirmed and
correlated with the original cholesteryl arachidonate compounds: 16,
tR 11.187 min; 17, tR 11.675 min; 15, tR 13.052 min; 14, tR 17.541
min; 13, tR 17.81 min; 12, tR 28.598 min.
arachidonate hydroperoxide fractions were isolated by semipreparative
HPLC (0.66% IPA in hexanes, 10.0 mL/min) and analyzed by direct
liquid infusion of the sample. Briefly, each isolated lipid fraction was
concentrated to dryness and a lipid solution (100 ng/µL) was prepared.
This lipid solution was mixed 1:1 with AgBF4 in 2-propanol (51.4 ng/
µL) to give a solution of ∼50 ng/µL lipid and 25 ng/µL Ag+. On a
molar basis, the silver-to-lipid concentration is approximately 2:1. The
lipid + Ag+ solution is introduced to the ESI source via a syringe pump
at a rate of 10 µL/min. Parent ion scans were collected over the mass
range 100-1000 amu for a scan duration of 1 s. Mass spectra were
typically collected in profile mode for 1 min (60 scans) and averaged.
LC-MS-MS experiments were carried out with a argon gas pressure
of 2.56 mTorr. Collision offset voltages were optimized with respect
to each compound and ranged from 10 to 40 eV for cholesteryl linoleate
and cholesteryl arachidonate hydroperoxides.
HPLC-MS Analysis (LC-CIS-MS). For on-line LC-MS sepa-
ration and analysis of the cholesteryl ester hydroperoxides, the peroxidic
oxidation products were isolated from the unoxidized lipid by semi-
preparative chromatography (0.5% 2-propanol in hexanes, 10 mL/min).
The oxidized fractions for cholesteryl linoleate (10-25 min) and
cholesteryl arachidonate (10-35 min) were collected, concentrated, and
analyzed by LC-CIS-MS. A stock solution of the oxidized fraction (1
mg/mL in mobile phase) was prepared, and typically 2-20 µg of the
mixture was injected per analysis. Offset voltages for SRM experiments
were as determined by optimization in the DLI experiments.
Chemo-enzymatic Synthesis of Cholesteryl 15(S)-hydroperoxy-
5(Z),8(Z),11(Z),13(E)-eicosatetraenoate, 17. 15(S)-Hydroperoxy-
5(Z),8(Z),11(Z),13(E)-eicosatetraenoic Acid. The following reaction
was prepared on a 1.0 L scale. The reaction sequence was repeated six
times for a total of 1 g of arachidonic acid substrate. An ethanolic
solution of arachidonic acid (167 mg, 7 mL) was added in one aliquot
to 1000 mL of boric acid buffer (pH 9.0, 0.2 M) that was saturated
with oxygen. To the slightly cloudy solution was added a solution of
soybean lipoxidase (3.1 mg/mL, 2.5 mL) in boric acid buffer. Oxygen
was slowly bubbled into the solution for 10 min. The reaction mixture
was acidified to pH 4 by the addition of HCl and extracted with EtO2
(3 × 250 mL). The ether extracts were dried and concentrated in vacuo
to a pale yellow oil. The resulting crude hydroperoxide was immediately
carried forward without purification or characterization.
Methyl 15(S)-hydroperoxy-5(Z),8(Z),11(Z),13(E)-eicosatetraenoate.
The crude hydroperoxide was dissolved in EtO2 (10 mL) and treated
with an excess of an Etheral solution of freshly prepared diazomethane.
The reaction was allowed to stir at 0 °C for 30 min. Excess
diazomethane was removed by gently bubbling with argon until the
CIS-MS Analysis of Cholesteryl Ester Hydroperoxides. Direct
Liquid Injection Analysis (DLI). Individual cholesteryl linoleate and

8054 J. Am. Chem. Soc., Vol. 122, No. 33, 2000
HaVrilla et al.
solution was colorless. BHT (∼3 mg) was added and the ether removed
in vacuo. The resulting bright yellow oil was purified by flash
chromatography on silica gel (20% Et2O in hexanes) yielding 0.9736
g of a pale yellow oil (80.5%): TLC: Rf 0.21 (25% Et2O in hexanes);
1H NMR (400 MHz, CDCl3): δ 8.42 (s, 1H), 6.51 (dd, J ) 15.2, 11.1
Hz, 1H), 5.94 (t, J ) 10.9, 10.9 Hz, 1H), 5.55 (dd, J ) 15.2, 8.0 Hz,
1H), 5.32 (m, 5H), 4.30 (m, 1H), 3.59 (s, 3H), 2.89 (m, 2H), 2.73 (m,
2H), 2.25 (m, 2H), 2.03 (m, 2H), 1.61 (m, 3H), 1.43-1.16 (8H), 0.81
(t, 3H); 13C NMR (100 MHz, CDCl3): δ 174.26, 133.52, 132.28,
130.62, 128.78, 128.47, 127.69, 127.21, 86.21, 60.41, 51.41, 33.20,
32.40, 31.56, 26.33, 25.95, 25.43, 24.77, 24.51, 22.31, 13.81; FTIR
(neat, NaCl plates): 3417.7 (s), 1717.26 (s) cm-1; CIS-MS (DLI,
AgBF4): m/z 457.0/459.1[M + Ag]+.
25.43, 25.08, 24.93, 24.86, 24.67, 24.26, 23.80, 23.02, 22.83, 22.80,
22.54, 22.50, 21.00, 19.30, 18.69. 14.03, 11.83; FTIR (neat, NaCl
plates): 1732.56 (s) cm-1; CIS-MS (DLI, AgBF4): m/z 883.6/885.6
[M + Ag]+.
Cholesteryl 15(S)-hydroperoxy-5(Z),8(Z),11(Z),13(E)-eicosatet-
raenoate, 17. The protected cholesteryl arachidonate hydroperoxide
(0.7519 g, 0.967 mmol) was dissolved in THF (9 mL). BHT (∼3 mg)
was added to the solution to prevent hydroperoxide rearrangement.
Glacial acetic acid (5 mL) and deionized water (3 mL) were added
with stirring resulting in a cloudy white solution, which was stirred
overnight under argon. The reaction was worked up by drying with
anhydrous NaSO4, followed by removal of the acetic acid by azeotrope
with cyclohexane. The crude material was purified by flash column
chromatography (5% EtOAc in hexanes), resulting in the recovery of
0.0321 g of the starting protected hydroperoxide and 0.5183 g of the
desired hydroperoxide (76.0%): TLC: Rf 0.38 (15% EtOAc in
hexanes); 1H NMR (400 MHz, CDCl3): δ 8.33 (s, 1H), 6.49 (dd, J )
15.2, 11.1 Hz, 1H), 5.91 (t, J ) 10.9, 10.9 Hz, 1H), 5.51 (dd, J )
15.2, 8.0 Hz, 1H), 5.30 (m, 6H), 4.52 (m, 1H), 4.27 (q, 14.0, 6.7 Hz,
1H), 2.86 (m, 2H), 2.71 (m, 2H), 2.19 (m, 4H), 2.03-0.75 (53H), 0.57
(s, 3H); 13C (100 MHz, CDCl3): δ 173.26, 139.50, 132.22, 130.90,
129.10, 129.03, 128.70, 128.55, 128.22, 127.73, 127.27, 122.58, 86.41,
73.89, 56.60, 56.06, 49.93, 42.23, 39.65, 39.45, 38.05, 36.90, 36.50,
36.12, 35.73, 33.97, 32.54, 31.82, 31.77, 31.71, 28.16, 27.94, 27.72,
26.50, 26.12, 25.58, 24.92, 24.78, 24.21, 23.77, 22.76, 22.50, 20.96,
19.24, 18.651, 13.97, 11.79; CIS-MS (DLI, AgBF4): m/z 811.6/813.6
[M + Ag]+.
Conversion of 15-OOH Cholesteryl Arachidonate (Ch-15-
HPETE) to More Polar Compounds. A solution of Ch-15-HPETE
(17, 0.2858 g, 0.405 mmol) was dissolved in dry benzene (2.7 mL) for
a final lipid concentration of 0.15 M. DTBN (7.2 mg, 0.041 mmol, 10
mol %) was added, and the lipid mixture was stirred at 37 °C in an oil
bath under an atmosphere of oxygen for 16-18 h. BHT (∼2 mg) was
added and the reaction was stored at -78 °C until HPLC and MS
analysis. The reaction was repeated for a total of 0.5781 g of Ch-15-
HPETE converted to more polar products. The analytical traces of each
individual conversion were identical to one another, so the mixtures
were pooled. The polar compounds were isolated by semipreparative
HPLC (1.0% 2-propanol in hexanes, 10 mL/min). Typically, 15-20
mg of the oxidized material was separated per injection to prevent
column overloading. Several major UV active (λ ) 234 nm) fractions
were collected (B, C, E, F/G, I). Minor fractions (A, D, H, J) were
also collected, but were typically mixtures of 2-4 compounds. The
major fractions were analyzed by 1H NMR, two-dimensional correlation
spectroscopy (2D COSY), and CIS-MS.
Direct Liquid Injection Analysis (DLI) and HPLC-MS Analysis
of More Polar Cyclic Peroxides. Individual isolated fractions (B, C,
E, F/G, I) were analyzed by direct liquid infusion of the sample as
described for the cholesteryl linoleate hydroperoxides. For on-line LC-
CIS-MS analysis of cyclic peroxide mixtures, sample introduction was
carried out using a single Beckman Ultrasphere narrowbore 5-µm silica
column (2.0 mm × 25 cm) operated in the isocratic mode with 1.0%
2-propanol in hexanes at a flow rate of 150 µL/min. A stock solution
of the Ch-15-HPETE conversion mixture (1 mg/mL in mobile phase)
was prepared in mobile phase and typically 20-40 µL (20-40 µg lipid)
of the solution was injected per analysis. Offset voltages for SRM
experiments, usually 20 eV were set as determined by optimization in
the DLI experiments
Conversion of 5-OOH Cholesteryl Arachidonate (Ch-5-HPETE)
to More Polar Compounds. Ch-5-HPETE (12) was isolated from a
cholesteryl arachidonate autoxidation with 1,4-cyclohexadiene, as
already described. The isolated hydroperoxide (6.8 mg, 9.64 × 10-3
mmol) was dissolved in benzene (195 µL). DTBN was added as a
benzene solution (65 µL of 2.7 mg/mL, 0.176 mg, 1 × 10-3 mmol, 10
mol % DTBN) for a final lipid concentration of 0.0375 M. The reaction
vial was flushed with oxygen, sealed, and stirred at 37 °C for 17 h.
BHT (∼2 mg) was added and the reaction was stored at -78 °C until
analysis. LC/CIS-MS was performed as described for the above cyclic
product mixtures.
Methyl 15(S)-[(1-methoxy-1-methylethyl)dioxy]-5(Z),8(Z),11(Z),-
13(E)-eicosatetraenoate. Methyl arachidonate hydroperoxide (0.8538
g, 2.44 mmol) was dissolved in dry CH2Cl2 (14 mL). To this solution,
2-methoxypropene (360 µL, 0.263 g, 3.65 mmol) and pyridinium-p-
toluenesulfonate (∼2 mg) were added. The reaction was stirred
overnight at room temperature under argon. The reaction was added
to a separatory funnel with CH2Cl2 (50 mL). The organic layer was
washed with NaHCO3 (1 × 20 mL) and saturated NaCl (1 × 20 mL),
dried with MgSO4, and concentrated to dryness to yield 1.22 g of a
1
yellow oil (100%): TLC: Rf 0.35 (25% Et2O in hexanes); H NMR
(400mHz, CDCl3): δ 6.47 (dd, J ) 15.2, 11.1 Hz, 1H), 5.98 (t, J )
11.2, 11.2 Hz, 1H), 5.61 (dd, J ) 15.2, 7.9 Hz, 1H), 5.35 (m, 5H),
4.37 (m, 1H), 3.63 (s, 3H), 3.25 (s, 3H), 2.92 (m, 2H), 2.77 (m, 2H),
2.28 (t, 2H), 2.07 (q, 2H), 1.66 (m, 3H), 1.43-1.22 (13H), 0.84 (t,
3H); 13C NMR (100 MHz, CDCl3): δ 173.95, 133.60, 130.08, 128.98,
128.65, 128.53, 128.17, 127.56, 127.40, 104.51, 84.56, 51.41, 49.20,
33.35, 33.03, 31.73, 26.49, 26.06, 25.56, 25.05, 24.71, 22.98, 22.79,
22.47, 13.98; FTIR (neat, NaCl plate) 1738.12 cm-1; CIS-MS (DLI,
AgBF4): m/z 529.1/531.1 [M + Ag]+.
15(S)-[(1-Methoxy-1-methylethyl)dioxy]-5(Z),8(Z),11(Z),13(E)-
eicosatetraenoic Acid. The protected methyl arachidonate hydroper-
oxide (1.0358 g, 2.4510 mmol) was dissolved in tetrahydrofuran (THF,
10 mL) and LiOH (4.0 M, 10 mL). The biphasic solution was vigorously
stirred at room temperature overnight. The reaction was poured into a
separatory funnel containing 1 M HCl (60 mL) and EtOAc (50 mL).
The acidified layer was extracted with EtOAc (3 × 50 mL). The
combined organic layers were washed with brine (1 × 25 mL), dried
with anhydrous MgSO4, and concentrated to 1.0228 g of a bright yellow
oil (100% yield): TLC: Rf 0.09 (25% Et2O in hexanes), Rf 0.27 (50%
Et2O in hexanes); 1H NMR (400 MHz, CDCl3): δ 9.91 (bs, 1H), 6.48
(dd, J ) 15.2, 11.1 Hz, 1H), 5.98 (t, J ) 11.2, 11.2 Hz, 1H), 5.61 (dd,
J ) 15.2, 7.9 Hz, 1H), 5.36 (m, 5H), 4.38 (q, 1H), 3.26 (s, 3H), 2.93
(t, 2H), 2.78 (t, 2H), 2.33 (t, 2H), 2.10 (q, 2H), 1.67 (m, 3H), 1.44-
1.20 (13H), 0.85 (t, 3H); 13C NMR (100 MHz, CDCl3): δ 179.49,
133.51, 130.08, 128.81, 128.46, 128.16, 127.57, 127.44, 104.57, 84.54,
49.18, 33.30, 33.02, 31.72, 26.39, 26.06, 25.57, 25.02, 24.43, 22.94,
22.78, 22.46, 13.98; FTIR (neat, NaCl plates): 2923.21 (s), 1693.61
(s); CIS-MS (DLI, AgBF4): m/z 515.1/517.1 [M + Ag+].
Cholesteryl 15(S)-[(1-methoxy-1-methylethyl)dioxy]-5(Z),8(Z),-
11(Z),13(E)-eicosatetraenoate. The protected arachidonic acid hydro-
peroxide (0.6023 g, 1.47 mmol) was dissolved in dry CH2Cl2 (20 mL).
dimethylaminopyridine (0.3978 g, 3.26 mmol) and cholesterol (0.6932
g, 1.792 mmol) were added and stirred for 5 min. dicyclohexylcarbo-
diimide (0.6744 g, 3.26 mmol) was added in one portion, and the
reaction was stirred under argon for 24 h. TLC indicated the formation
of a less polar fraction, which stained pink with the hydroperoxide
stain after heating. After three flash columns (5% EtOAc in hexanes),
the protected cholesteryl ester hydroperoxide was obtained as 0.7519
g of a thick, colorless oil (65.7%): TLC: Rf 0.33 (10% EtOAc in
hexanes), Rf 0.51 (15% EtOAc in hexanes); 1H NMR (400 MHz,
CDCl3): δ 6.48 (dd, J ) 15.4, 11.1 Hz, 1H), 5.99 (t, J ) 11.2, 11.2
Hz, 1H), 5.62 (dd, J ) 15.5, 8.0 Hz, 1H), 5.36 (m, 6H), 4.59 (m, 1H),
4.39 (q, 14.0, 6.7 Hz, 1H), 3.27 (s, 3H), 2.94 (m, 2H), 2.79 (m, 2H),
2.27 (m, 4H), 2.11-0.83 (59H), 0.65 (s, 3H); 13C (100 MHz, CDCl3):
δ 172.95, 139.62, 133.62, 130.11, 129.14, 128.61,128.59, 128.21,
127.59, 127.44, 122.60, 104.54, 84.59, 73.77, 56.66, 56.10, 49.99, 49.24,
42.28, 39.70, 39.49, 38.13, 36.97, 36.57, 36.16, 35.77, 34.90, 34. 02,
33.06, 31.88, 31.83, 31.77, 28.21, 27.99, 27.79, 26.54, 26.10, 25.59,
Autoxidation of Cholesteryl Arachidonate without 1,4-Cyclo-
hexadiene. Cholesteryl arachidonate (0.400 g, 0.594 mmol) was

Peroxide Products from Polyunsaturated Lipid Autoxidation
J. Am. Chem. Soc., Vol. 122, No. 33, 2000 8055
dissolved in dry benzene (4.0 mL) for a final lipid concentration of
0.15 M. Oxidation was initiated by the addition of DTBN (∼3 mg).
The reaction was placed into a 37 °C oil bath under air for 48 h.
Reaction was inhibited by the addition of BHT (∼2 mg). A portion of
the reaction mixture (1 mL) was removed and concentrated to dryness.
The residue was diluted with CHCl3 (100 µL) and applied to a Burdick
& Jackson Inert SPE System (silica, 500 mg) prewashed with hexane
(10 mL) and equilibrated with 1.5% MTBE in hexane (10 mL). The
unoxidized cholesteryl arachidonate was eluted with 1.5% MTBE in
hexanes (3 × 5 mL). The oxidized cholesteryl arachidonate material
was eluted from the column with 5% methanol in MTBE (3 × 5 mL).
The oxidized fractions were combined and concentrated to yield 61.2
mg oxidized material. A stock solution of the oxidized material was
prepared in benzene (2 mL) for a concentration of 30.6 mg/mL.
Subsequent dilutions of this stock solution were used for MS analysis.
Analytical HPLC analysis (single Si column, 1.0% 2-propanol in
hexane) indicated no difference in the product distribution of polar
products after removal of the unoxidized lipid. The complex oxidation
product mixture was analyzed by LC-CIS-MS as described for the
cyclic peroxide analysis. Typically, 20-60 µg of the lipid oxidation
mixture was injected for SIM and SRM analysis.
Acknowledgment. We thank Lisa Manier of the Vanderbilt
University Mass Spectrometry Research Center for her as-
sistance in MS measurements. Support by USPHS Grants
HL1792, P30 ES00267, and NSF CHE 9996188 is gratefully
acknowledged. C.M.H. also acknowledges funding from the
toxicology training grant NIEHS T32ES07031 (Duke University).
JA001180F