Rate constants for peroxidation of polyunsaturated fatty acids and sterols in solution and in liposomes

Article abstract of DOI:10.1021/ja9029076

Rate constants for autoxidation propagation of several unsaturated lipids in benzene solution at 37°C and in phosphatidylcholine liposomes were determined by a linoleate radical clock. This radical clock is based on competition between hydrogen atom abstraction by an intermediate peroxyl radical derived from linoleic acid that leads to a trans,cis-conjugated hydroxyoctadecadienoic product and β-fragmentation of the same peroxyl that gives the trans,trans-product hydroxyoctadecadienoic acid. Rate constants determined by this approach in solution relative to linoleic acid (k _p) 62 M^-1 s^-1) were: arachidonic acid (k _p = 197 ± 13 M^-1 s^-1), eicosapentaenoic acid (k_p = 249 ± 16 M^-1 s^-1), docosahexaenoic acid (k_p = 334 ± 37 M^-1 s ^-1), cholesterol (k_p= 11 ± 2 M^-1 s ^-1), and 7-dehydrocholesterol (k_p = 2260 ± 40 M^-1 s^-1). Free radical oxidations of multilamellar and unilamellar liposomes of various mixtures of glycerophosphatidylcholine molecular species were also carried out. In some experiments, cholesterol or 7-dehydrocholesterol was incorporated into the lipid mixture undergoing oxidation. A phosphatidylcholine bearing a linoleate ester at sn-2 was a component of each liposome peroxidation reaction and the ratio of trans,cis/trans,trans (t,c/t,t)-conjugated diene oxidation products formed from this phospholipid was determined for each oxidation reaction. This t,c/t,t-product ratio from linoleate was used to "clock" liposome constituents as hydrogen atom donors in the lipid bilayer. Application of this lipid bilayer radical clock gives relative autoxidation propagation rate constants of arachidonate (20:4), eicosapentaenoate (20:5), docosahexaenoate (22:6), and 7-dehydrocholesterol to be 115 ± 7, 145 ± 8, 172 ± 13, and 832 ± 86, respectively, a reactivity trend that parallels the one in solution. We also conclude from the liposome oxidations that linoleate peroxyl radicals at different positions on the eighteen-carbon chain (at C-9 and C-13) have different kinetic properties. This is in contrast to the results of solution oxidations of linoleate in which the C-9 and C-13 peroxyl radicals have similar reactivities. We suggest that peroxyl radical β-scission depends on solvent polarity and the polarity of the local environment of peroxyl radicals in liposomal oxidations depends on the position of the peroxyl radical on the 18-carbon chain.

Full text of DOI:10.1021/ja9029076

Published on Web 08/25/2009
Rate Constants for Peroxidation of Polyunsaturated Fatty
Acids and Sterols in Solution and in Liposomes
Libin Xu, Todd A. Davis,† and Ned A. Porter*
Department of Chemistry and Vanderbilt Institute of Chemical Biology, Vanderbilt UniVersity,
NashVille, Tennessee 37235
Received April 11, 2009; E-mail: n.porter@vanderbilt.edu
Abstract: Rate constants for autoxidation propagation of several unsaturated lipids in benzene solution at
37 °C and in phosphatidylcholine liposomes were determined by a linoleate radical clock. This radical
clock is based on competition between hydrogen atom abstraction by an intermediate peroxyl radical derived
from linoleic acid that leads to a trans,cis-conjugated hydroxyoctadecadienoic product and ꢀ-fragmentation
of the same peroxyl that gives the trans,trans-product hydroxyoctadecadienoic acid. Rate constants
determined by this approach in solution relative to linoleic acid (kp ) 62 M-1 s-1) were: arachidonic acid (kp
) 197 ( 13 M-1 s-1), eicosapentaenoic acid (kp ) 249 ( 16 M-1 s-1), docosahexaenoic acid (kp ) 334 (
37 M-1 s-1), cholesterol (kp) 11 ( 2 M-1 s-1), and 7-dehydrocholesterol (kp) 2260 ( 40 M-1 s-1). Free
radical oxidations of multilamellar and unilamellar liposomes of various mixtures of glycerophosphatidyl-
choline molecular species were also carried out. In some experiments, cholesterol or 7-dehydrocholesterol
was incorporated into the lipid mixture undergoing oxidation. A phosphatidylcholine bearing a linoleate
ester at sn-2 was a component of each liposome peroxidation reaction and the ratio of trans,cis/trans,trans
(t,c/t,t)-conjugated diene oxidation products formed from this phospholipid was determined for each oxidation
reaction. This t,c/t,t-product ratio from linoleate was used to “clock” liposome constituents as hydrogen
atom donors in the lipid bilayer. Application of this lipid bilayer radical clock gives relative autoxidation
propagation rate constants of arachidonate (20:4), eicosapentaenoate (20:5), docosahexaenoate (22:6),
and 7-dehydrocholesterol to be 115 ( 7, 145 ( 8, 172 ( 13, and 832 ( 86, respectively, a reactivity trend
that parallels the one in solution. We also conclude from the liposome oxidations that linoleate peroxyl
radicals at different positions on the eighteen-carbon chain (at C-9 and C-13) have different kinetic properties.
This is in contrast to the results of solution oxidations of linoleate in which the C-9 and C-13 peroxyl radicals
have similar reactivities. We suggest that peroxyl radical ꢀ-scission depends on solvent polarity and the
polarity of the local environment of peroxyl radicals in liposomal oxidations depends on the position of the
peroxyl radical on the 18-carbon chain.
Introduction
reaction with molecular oxygen. Cholesterol free radical oxida-
tion has been studied in great detail and some of its peroxidation
Autoxidation of polyunsaturated fatty acids and esters (PU-
FAs) and sterols, a process known as lipid peroxidation, has
attracted research attention over the last few decades due to its
involvement in the pathophysiology of common diseases like
atherosclerosis,1,2 asthma,3,4 neurodegenerative disorders such
as Alzheimer’s disease and Parkinson’s disease,5-8 and meta-
bolic disorders such as the Smith-Lemli-Opitz syndrome
(SLOS).9-12 PUFAs are important membrane constituents and
the highly unsaturated members of the group of linoleic (18:2,
ω-6), arachidonic (20:4, ω-6), eicosapentaenoic (20:5, ω-3)
and docosahexaenoic (22:6, ω-3) are extremely prone to
products have potent biological activities.13-15 Cholesterol is
also an important constituent of biological membranes and
7-dehydrocholesterol (7-DHC) is a biosynthetic precursor of
cholesterol as well as vitamin D3.16-18 7-DHC is present in
(5) Simonian, N. A.; Coyle, J. T. Annu. ReV. Pharmacol. Toxicol. 1996,
36, 83–106.
(6) Sayre, L. M.; Zelasko, D. A.; Harris, P. L. R.; Perry, G.; Salomon,
R. G.; Smith, M. A. J. Neurochem. 1997, 68, 2092–2097.
(7) Fahn, S.; Cohen, G. Ann. Neurol. 1992, 32, 804–812.
(8) Yoritaka, A.; Hattori, N.; Uchida, K.; Tanaka, M.; Stadtman, E. R.;
Mizuno, Y. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 2696–2701.
(9) Gaoua, W.; Chevy, F.; Roux, C.; Wolf, C. J. Lipid. Res. 1999, 40,
456–463.
† Current address: Department of Chemistry, Campus Box 8023, Idaho
State University, Pocatello, ID 83209.
(10) Richards, M. J.; Nagel, B. A.; Fliesler, S. J. Exp. Eye. Res. 2006, 82,
538–541.
(1) Esterbauer, H.; Gebicki, J.; Puhl, H.; Jurgens, G. Free Radical Biol.
Med. 1992, 13, 341–390.
(11) Valencia, A.; Rajadurai, A.; Carle, A. B.; Kochevar, I. E. Free Radical
Biol. Med. 2006, 41, 1704–1718.
(2) Berliner, J. A.; Heinecke, J. W. Free Radical Biol. Med. 1996, 20,
707–727.
(12) Valencia, A.; Kochevar, I. E. Free Radical Biol. Med. 2006, 40, 641–
650.
(3) Antczak, A.; Nowak, D.; Shariati, B.; Krol, M.; Piasecka, G.;
Kurmanowska, Z. Eur. Respir. J. 1997, 10, 1235–1241.
(4) Montuschi, P.; Corradi, M.; Ciabattoni, G.; Nightingale, J.; Kharitonov,
S. A.; Barnes, P. J. Am. J. Respir. Crit. Care. Med. 1999, 160, 216–
220.
(13) Smith, L. L. Cholesterol Autoxidation; Plenum Press: New York, 1981.
(14) Murphy, R. C.; Johnson, K. M. J. Biol. Chem. 2008, 283, 15521–
15525.
(15) Girotti, A. W. J. Lipid Res. 1998, 39, 1529–1542.
9
10.1021/ja9029076 CCC: $40.75  2009 American Chemical Society
J. AM. CHEM. SOC. 2009, 131, 13037–13044 13037

A R T I C L E S
Xu et al.
Scheme 1
relatively high concentrations in skin where it is exposed to
exogenous radical sources and oxygen, and it is of additional
interest because of syndromes in which high levels of this sterol
are associated with severe developmental defects, SLOS being
one example. Oxidative stress and peroxidation have been
associated with SLOS in several studies and the oxidative
instability of 7-DHC has been implied in more than one
instance.11,12
+3b)], can be related to the concentrations and rate constants
of all hydrogen atom donors present in the reaction mixture.
Experiments to determine important fragmentation rate constants
and oxygen partitioning parameters for processes shown in
Scheme 1 have been carried out for reactions in benzene at 37 °C
and this permits the formulation of a simple kinetic expression
for peroxidation under these standard conditions, relating the
trans,cis/trans,trans-product ratio and kinetic rate constants, as
shown in eq 1.19,24 The boundary limit of eq 1 as kpi [Ri -
H]approaches zero corresponds to oxidation conditions in which
poor H-atom donors are present or good donors are present at
very low concentrations. Under these conditions, the trans,cis/
trans,trans product ratio approaches 0.16 and the reaction is
effectively under thermodynamic control. Under conditions in
which good H-atom donors such as R-tocopherol are present,
the reaction is under kinetic control and the trans,cis isomers
dominate the product mixture.
The slow step in free radical chain oxidation is the step in
which a hydrogen atom is transferred from an organic reactant,
R-H, to an intermediate peroxyl radical, R-OO•. The overall
rate of oxidation depends directly on the rate constant for this
propagation step, kp, the concentration [R-H], and the rate
constants for radical initiation and radical-radical termination.
We have recently made use of a peroxyl radical clock to
determine propagation rate constants for organic compounds
undergoing free radical chain oxidation. The radical clock is
based on the unimolecular ꢀ-fragmentation rate constants of
peroxyl radicals derived from linoleic acid (Scheme 1) and these
clocks have been applied in the measurement of a number of
propagation rate constants for autoxidation in solution.19 Two
trans,cis-conjugated diene peroxyl radicals, 1a and1b, are
formed in linoleate oxidation and trapping of these radicals by
hydrogen atom transfer with a rate of kp[R-H][ROO•] gives
rise to trans,cis-conjugated diene oxidation products, 2a and
2b.20-23 ꢀ-Fragmentation of the peroxyl radicals 1a and1b,
leading to trans,trans-conjugated diene products, competes with
hydrogen atom transfer and this competition can be used to
“clock” the bimolecular atom-transfer reaction.19
kip[Ri - H]
214 s-1
(2a + 2b)
(3a + 3b)
trans, cis
trans, trans
)
)
+ 0.16
∑
i)1-n
(1)
Peroxidation of phospholipids in lipid bilayers follows the
same kinetic rate law as in homogeneous systems.25-31 How-
ever, generally applicable radical clocks in membranes have
not been developed due to the complexity of autoxidation
reactions in heterogeneous systems and the difficulty in defining
necessary parameters. Barclay and co-workers determined kinh
Based upon the mechanism shown in Scheme 1, the ratio of
trans,cis/trans,trans-products from linoleate, [(2a + 2b)/(3a
(24) Davis, T. A.; Gao, L.; Yin, H.; Morrow, J.; Porter, N. A. J. Am. Chem.
Soc. 2006, 128, 14897–14904.
(16) Kresge, N.; Simoni, R. D.; Hill, R. L. J. Biol. Chem. 2005, 280, e7.
(17) Javitt, N. B. Steroids 2008, 73, 149–57.
(25) Antunes, F.; Barclay, L. R. C.; Vinqvist, M. R.; Pinto, R. E. Int.
J. Chem. Kinet. 1998, 30, 753–767.
(18) Herman, G. E. Hum. Mol. Genet. 2003, 12 (Spec No 1), R75–88.
(19) Roschek, B.; Tallman, K. A.; Rector, C. L.; Gillmore, J. G.; Pratt,
D. A.; Punta, C.; Porter, N. A. J. Org. Chem. 2006, 71, 3527–3532.
(20) Porter, N. A.; Weber, B. A.; Weenen, H.; Khan, J. A. J. Am. Chem.
Soc. 1980, 102, 5597–5601.
(26) Barclay, L. R. C.; Baskin, K. A.; Kong, D.; Locke, S. J. Can. J. Chem.
1987, 65, 2541–2550.
(27) Barclay, L. R. C. Can. J. Chem. 1993, 71, 1–16.
(28) Barclay, L. R. C.; Ingold, K. U. J. Am. Chem. Soc. 1980, 102, 7792–
7794.
(21) Porter, N. A.; Lehman, L. S.; Weber, B. A.; Smith, K. J. J. Am. Chem.
Soc. 1981, 103, 6447–6455.
(29) Barclay, L. R. C.; Ingold, K. U. J. Am. Chem. Soc. 1981, 103, 6478–
6485.
(22) Tallman, K. A.; Roschek, B.; Porter, N. A. J. Am. Chem. Soc. 2004,
126, 9240–9247.
(30) Barclay, L. R. C.; Locke, S. J.; Macneil, J. M.; Vankessel, J.; Burton,
G. W.; Ingold, K. U. J. Am. Chem. Soc. 1984, 106, 2479–2481.
(31) Barclay, L. R. C.; Vinqvist, M. R.; Antunes, F.; Pinto, R. E. J. Am.
Chem. Soc. 1997, 119, 5764–5765.
(23) Pajunen, T. I.; Johansson, M. P.; Hase, T.; Hopia, A. Lipids 2008, 43,
599–610.
9
13038 J. AM. CHEM. SOC. VOL. 131, NO. 36, 2009

Membrane Lipid Peroxidation
A R T I C L E S
Table 1. Autoxidation Propagation Rate Constants Determined by
the Linoleic Radical Clocka
fatty acid or sterol, R-H
kp in solution at 37 °C (M-1s-1
)
linoleic (18:2)
62 (ref 34)
197 ( 13
249 ( 16
334 ( 37
11 ( 2
arachidonic (20:4)
eicosapentaenoic (20:5)
docosahexaenoic (22:6)
cholesterol
7-dehydrocholesterol
2260 ( 40
a errors are 2σ
of R-tocopherol in model membranes by applying the linoleate
peroxyl radical clock, but this approach has not been widely
applied.31,32
We report here the propagation rate constants of PUFAs and
sterols in solution and also an expanded study of the linoleate
peroxyl radical clock in lipid bilayers such as those found in
biomembranes with a goal being to predict and/or interpret
peroxidation reactions in ViVo in a more substantive way.33 One
of our aims is to develop a clocking equation for reactions in
lipid bilayers where the ratio of trans,cis/trans,trans-products
from linoleate is expressed as a function of kp, the propagation
rate constants of hydrogen atom donors and the bilayer mole
fraction of those donors.
sn-glycero-3-phosphatidylcholine (PLPC) was used as the
linoleate oxidation substrate in all experiments and it was diluted
by mixing with non- or poorly oxidizable PCs to determine the
effect of linoleate mole fraction in the bilayer on the trans,cis/
trans,trans-product ratio.20,33 The peroxidation chain reactions
were initiated with a catalytic amount of the water-soluble azo
initiator 2,2′- azobis(2-(2-imidazolin-2-yl) propane) dihydro-
chloride (AIPH) at 37 °C and were quenched at an appropriate
time to avoid overoxidation of linoleate (in all cases <5%).
Conjugated diene products were analyzed by HPLC-UV after
reduction with PPh3 and hydrolysis of the phosphatidylcholine
under basic conditions.
The trans,cis/trans,trans-product ratio, [(2a + 2b)/(3a +3b)],
for oxidation of PLPC liposomes is presented in Figure 1.
Dilution experiments of PLPC with different nonoxidizable
phospholipids are presented in the Figure in different colors.
The data in the figure show that the ratio of trans,cis/trans,trans
Results
Propagation Rate Constants for Autoxidation of
Important Lipids in Solution. We first sought to determine
propagation rates for the important fatty acids and sterols in
homogeneous solution by the use of the linoleate clock.19 The
experiments require determining the trans,cis/trans,trans-product
ratio from linoleate as a function of the concentration of other
oxidizable substrates. In a typical experiment, 0.2 to 0.4 M
solutions of linoleic acid or methyl linoleate in the presence of
0.1 to 0.6 M of the compound under study (e.g., arachidonic
acid etc.) in benzene or chlorobenzene at 37 °C were oxidized
for 30 to 90 min in reactions initiated by 2,2′-azobis(4-methoxy-
2,4-dimethylvaleronitrile) (MeOAMVN, < 0.5 mol %). Less than
a percent of linoleate or substrate were consumed in a typical
experiment. The linoleate trans,cis/trans,trans-product ratio was
determined by HPLC/MS or HPLC/UV at various substrate
concentrations and kp for the substrate was then determined by
the application of Equation 1 (Figure S1-S3, Supporting
Information). Typically, duplicate analyses of triplicate experi-
ments were carried out for six or seven concentrations of
substrate between 0.1 and 0.6 M. Table 1 gives the propagation
rate constants determined by this approach for five substrates.
Note that all of the rate constants are based upon the propagation
rate constant for autoxidation of linoleic acid, a rate constant
that has been determined by the rotating sector method.25,34 We
find that rate constants determined using either linoleic acid or
methyl linoleate are within experimental error under the
conditions described.
(2a + 2b)
(3a + 3b)
trans, cis
trans, trans
)
) a
ki ni
+ b
R-H
(2)
∑
p
i)1-n
products from linoleate depends on the mole fraction of linoleate
in the lipid bilayer. In eq 2, niR-H is the mole fraction of any
H-atom donor in the liposome, kip is the propagation rate constant
for a given H-atom donor in the liposome,35 b is a constant
that is essentially the thermodynamic product ratio in a given
lipid bilayer and a is a constant having units of time. Equation
2 is based on the analogy with eq 1, and the assumption that it
is the bilayer concentration of H-atom donors and not the
solution molarity that effectively controls the linoleate product
ratio. The bilayer mole fraction can, of course, be converted to
a bilayer concentration by making assumptions about bilayer
density and lipid average molecular weight, but for purposes
of our analysis the mole fraction analysis seems more straight-
forward and just as instructive. Statistical analyses suggest that
the various series of data shown in Figure 1 are not significantly
Calibration of the Linoleate Radical Clock in Liposomes.
Liposomes of multilamellar vesicles (MLV) from phosphatidyl
cholines were the colloids studied in all experiments unless
otherwise noted. Oxidation of unilamellar vesicles (ULV) gave
similar results to experiments carried out in MLV (Table 2; also
see Figure S4, Supporting Information). 1-Palmitoyl-2-linoleoyl-
(32) Culbertson, S. M.; Antunes, F.; Havrilla, C. M.; Milne, G. L.; Porter,
N. A. Chem. Res. Toxicol. 2002, 15, 870–876.
(35) Mole fraction is unitless but we define a unit n for calculation
convenience. Therefore, the unit of the propagation rate constant (kip)
in mole fraction will be n-1s-1, while the unit of the pseudo-first order
(33) Weenen, H.; Porter, N. A. J. Am. Chem. Soc. 1982, 104, 5216–5221.
(34) Howard, J. A.; Ingold, K. U. Can. J. Chem. 1967, 45, 793–802.
rate constant (kip nRi -H) will be s-1
.
9
J. AM. CHEM. SOC. VOL. 131, NO. 36, 2009 13039

A R T I C L E S
Xu et al.
Table 2. Analysis of Linoleate Product Ratio 2a/3b or 2b/3a ) a ΣkpnR-H + b (eq 2) for Liposomal Oxidations and Co-Oxidationsa
d
R-H
seriesb
slope
intercept
∆
∆
kp(n-1s-1
)
slope
intercept
PLPC
MLVs
PLPC
ULVs
PLPC
PAPC (20:4)
PLPC
PEPC (20:5)
PLPC
PDPC (22:6)
PLPC
1
2
1
2
1
2
1
2
1
2
1
2
1
2
0.76 ( 0.06
0.64 ( 0.07
0.79 ( 0.11
0.76 ( 0.09
2.79 ( 0.22
2.14 ( 0.17
3.47 ( 0.30
2.93 ( 0.27
4.04 ( 0.43
3.39 ( 0.58
28 ( 3
0.34 ( 0.02
0.25 ( 0.02
0.40 ( 0.04
0.30 ( 0.03
0.40 ( 0.05
0.27 ( 0.04
0.36 ( 0.07
0.26 ( 0.06
0.39 ( 0.09
0.32 ( 0.12
0.59 ( 0.14
0.53 ( 0.15
0.17 ( 0.01
0.17 ( 0.01
0.12 ( 0.13
0.03 ( 0.20
0.65 ( 0.39
0.54 ( 0.57
0.65 ( 1.01
18 ( 6
0.09 ( 0.04
0.10 ( 0.07
0.13 ( 0.09
0.10 ( 0.13
0.07 ( 0.21
0.06 ( 0.29
0 ( 0.02
35 (ref 25)
115 ( 7
145 ( 8
172 ( 13
832 ( 86
7-DHC
MeLin
in solution
10 ( 3
0.27 ( 0.02
0.25 ( 0.01
0.02 ( 0.03
c
a Errors are 2σ. b Series 1: 2a/3a, series 2: 2b/3b. c Data for experiments in solution are presented in molar units, not mole fraction. d Mole fraction
propagation rate constants obtained from (2a + 2b)/(3a +3b) vs nR-H (see Figure 2).
Figure 1. Values [(2a + 2b)/(3a + 3b)] from oxidation of mixtures of
Figure 2. Values of [(2a + 2b)/(3a + 3b)] from oxidation of mixtures of
PLPC and nonoxidizable PCs in lipid bilayers at 37 °C vs linoleate mole
PLPC/POPC and oxidizable PCs in lipid bilayers at 37 °C vs oxidizable
fraction in the bilayer. Different mixtures are shown in different colors:
lipid mole fraction in the bilayer. Different mixtures are shown in different
PLPC/POPC (black), PLPC/DMPC (red), PLPC/DOPC (blue), and PLPC/
colors: PLPC/PAPC (black), PLPC/PEPC (red), PLPC/PDPC (green), and
POPC/DPPC (green). The line shows the fit all of the data in the Figure to
PLPC/7-DHC (blue). The lines show the fit to eq 2. POPC was used as a
eq 2.
nonoxidizable bilayer diluent.
different (p > 0.05). Therefore, all data in Figure 1 were fit to
The data obtained from oxidation of PLPC with nonoxidizable
substrates can be plotted such that the chemistry associated with
eq 2, which gave a ) 0.022 ( 0.002 s and b ) 0.28 ( 0.02.25
Co-Oxidation of Linoleate Liposomes with Other
Oxidizable Substrates. To obtain the propagation rate constants
of other oxidizable lipids compared to linoleate, liposomes
containing a fixed mole fraction of PLPC and variable amounts
of other oxidizable phospholipid and sterol substrates were
oxidized at 37 °C in reactions initiated with the water-soluble
initiator AIPH. Phospholipid mixtures generally contained nLA
) 0.1-0.4 and varaiable amounts of another oxidizable phos-
pholipid, for example, PAPC, and a nonoxidizable phospholipid,
POPC. Analysis by HPLC/MS rather than HPLC/UV was
required for these oxidations since products arise from both
PLPC and the other oxidizable lipids but only the linoleate
products are used in the clocking experiments. Data from the
experiments with mixtures of PLPC, POPC and several co-
oxidants are presented in Figure 2, and propagation rate
constants of various hydrogen donors in the liposome can be
calculated from the slopes according to eq 2 (see Table 2).
Cholesterol acts, in effect, as an inert diluent when it is a
constituent of MLVs, a result that is not surprising based upon
its relative reactivity as an H-atom donor compared to linoleate
(Figure S5, Supporting Information).
peroxyl radicals 1a and 1b are differentiated. In solution, these
two regioisomeric peroxyl radicals follow identical reaction
paths at virtually the same rates, the result of this being that
peroxidation of linoleic acid or linoleate esters in solution gives
rise to essentially a symmetrical distribution of products, that
is, [2a] ) [2b] and [3a] ) [3b].
It is possible that a nonsymmetrical product distribution may
be obtained for peroxidation reactions in liposomes since the
peroxyl radicals at different chain positions experience different
local environments in the lipid bilayer. We therefore considered
differentiation of the two peroxyls according to the radical clock
framework presented in Scheme 1. In the clock mechanism,
peroxyl radical 1a is trapped by hydrogen atom donors to give
product 2a, but fragmentation of 1a leads to the trans,trans-
product 3a. The clock product ratios that provide differential
information about peroxyl radicals 1a and 1b formed from
oxidation in a lipid bilayer are therefore [2a]/[3a] and [2b]/
[3b]. Analysis of the data plotted in Figure 1 in two plots of
linoleate mole fraction vs [2a]/[3a] or [2b]/[3b] leads to Figure
3. It seems clear from the Figure that the 9-peroxyl radical 1a
gives rise to more trans,cis- or less trans,trans-product than
9
13040 J. AM. CHEM. SOC. VOL. 131, NO. 36, 2009

Membrane Lipid Peroxidation
A R T I C L E S
Linoleate has one such center; arachidonate, 3; eicosapen-
taenoate, 4 and docosahexaenoate, 5 and the relative propagation
rate constants determined for these compounds are in a ratio of
1, 3.2, 4.0, and 5.4. Cosgrove, Church and Pryor reported that
the oxidizability ratios of linoleate, arachidonate and docosa-
hexaenoate were 1, 2.8 and 5.0, in good agreement with the
clock values of kp reported here.36
The rate constant determined by the clock method for
propagation of cholesterol autoxidation, 11 M-1 s-1, falls in
line with those determined for other cycloalkenes; cyclohexene
and cyclopentene both having rate constants of about 6 M-1
s-1.37 The propagation rate constants for these cycloalkenes are
considerably larger than those measured for simple acyclic
olefins such as methyl oleate, which has a value of 0.8 M-1
s-1.38 The increased reactivity of cycloalkenes as H-atom donors
compared to acyclic alkenes is undoubtedly due to conforma-
tional effects that orient the reactive C-H bond so that
maximum orbital overlap between the alkene π bond and the
developing radical center is possible in the transition state.
The propagation rate constant determined for autoxidation
of 7-DHC, 2260 M-1 s-1, is surprisingly large. 1,4-Cyclohexa-
diene, considered to be an excellent H-atom donor that is
frequently used for mechanistic studies, has a kp of only 265
M-1 s-1and arachidonic acid, which is normally considered to
be a highly oxidizable lipid is over an order of magnitude less
reactive than 7-DHC, Table 1.19 Indeed, the propagation rate
constant for 7-DHC makes this substrate competitive for peroxyl
radicals with some hindered phenols that are used as antioxi-
Figure 3. Values of 9-trans,cis/13-trans,trans(red) and 13-trans,cis/9-
trans,trans(blue) vs linoleate mole fraction in the bilayer for oxidation of
PLPC in liposomes with various nonoxidizable substrates as shown in Figure
1.
does the corresponding 13-peroxyl radical, 1b (Figure S10-S11,
Supporting Information). Statistical analysis of the data leads
to the conclusion that the intercepts for the two plots are
significantly different (P < 0.0001) while the slopes of the two
plots do not differ (P > 0.05). Data from the experiment with
mixtures of PLPC, PAPC and POPC are presented according
to the analysis of [2a]/[3a] and [2b]/[3b] in Figure 4A and data
from a similar set of experiments with 7-dehydrocholesterol is
presented in Figure 4B. Analogous plots of [2a]/[3a] and [2b]/
[3b] for co-oxidants PEPC (Figure S6) and PDPC (Figure S7)
are presented in Supporting Information. Analysis of data from
a series of oxidations of linoleic acid in homogeneous solution
is also presented in Supporting Information (Figure S9).
dants. 2,6-Di-tert-butylphenol, for example, has a rate constant
19
for H-atom transfer to peroxyl radicals of 3100 M-1 s-1
.
Molecular mechanics calculations indicate that the torsion angles
C(7)-C(8)-C(9)-H(9) and C(7)-C(8)-C(14)-H(14) are
close to 90°, suggesting that H(9) and H(14) are aligned for
removal from the open R face of the steroid so that maximum
delocalization stabilization is experienced in the transition state
for H-atom transfer, see Figure 5.
Discussion
Propagation Rate Constants of PUFAs and 7-DHC in
Lipid Bilayers. The data presented in Figure 1 shows that the
ratio of linoleate products [(2a + 2b)/(3a +3b)] depends directly
on the mole fraction of linoleate, nLA, in lipid multilamellar
vesicles. These experiments were used to calibrate the constants
a and b in eq 2 since the rate constant for autoxidation
propagation in these aggregates, kLpA, was determined to be 16.6
M-1 s-1 by Antunes et al.25 This molar rate constant can be
converted to a mole fraction propagation rate constant, 35 n-1s-1,
based upon the authors’ assumption that the average molecular
weight of phospholipid in their MLVs was ∼800 and the density
approximately 0.8.25,28,39 In this way, the data in Figure 1 was
fit to eq 2 with kpLA ) 35 n-1s-1, which leads to a ) 0.022 (
0.002 s and b ) 0.28 ( 0.02.
Propagation Rate Constants of PUFAs and Sterols in
Solution. The rate of oxygen and substrate consumption in
radical chain oxidation depends on the propagation and termina-
tion rate constants kp and kt,, as well as the rate of radical
initiation, Ri, Equation 3.34,36 Oxidizability is the inherent
propensity to undergo oxidation, defined as kp/(2kt)1/2, and this
parameter can be determined by measurement of oxygen
consumption at a known rate of initiation. The determination
of kp has typically involved a set of experiments requiring
measurement both of oxidizability and kt, the latter usually by
means of a spectroscopic probe. The radical clock method,
described in Scheme 1 and eq 1, permits an independent and
straightforward determination of kp based on the competition
of unimolecular ꢀ-fragmentation of a linoleate peroxyl radical
and hydrogen atom transfer to that radical.19
Having determined a and b from the experiments in which
linoleate is the only oxidizable lipid, Equation 2 can then be
used to provide rate constants for other H-atom donors in the
lipid bilayer. The data for arachidonate, eicosapentaenoate,
docosahexaenoate and 7-DHC are plotted in Figure 2 and the
-d[O2]/dt ) {kp/(2kt)1/2}[R - H]Ri1/2
(3)
Table 1 shows the results obtained for application of the clock
method to determine kp for a number of important polyunsatu-
rated fatty acids and sterols in solution. The propagation rate
constant for fatty acid oxidation depends, as expected, on the
number of oxidizable bis-allylic -CH2- centers in the molecule.
(37) Korcek, S.; Chenier, J. H. B.; Howard, J. A.; Ingold, K. U. Can.
J. Chem. 1972, 50, 2285–2897.
(38) Porter, N. A.; Mills, K. A.; Carter, R. L. J. Am. Chem. Soc. 1994,
116, 6690–6696.
(39) The pseudo first-order rate constant for a given hydrogen donor at a
given concentration can be expressed as kMp [R-H] ) kpnnR-H, where
kMp is the propagation rate constant in units of M-1s-1 and knp is the
constant in units of n-1s-1. Thus, knp can be expressed as a function of
kMp as follows: kpn ) (kMp [R-H])/(nR-H).
(36) Cosgrove, J. P.; Church, D. F.; Pryor, W. A. Lipids 1987, 22, 299–
304.
9
J. AM. CHEM. SOC. VOL. 131, NO. 36, 2009 13041

A R T I C L E S
Xu et al.
Figure 4. Values of 9-trans,cis/13-trans,trans (red circle) and 13-trans,cis/9-trans,trans (blue square) vs in A, arachidonate mole fraction in the bilayer for
oxidation of PLPC in liposomes containing PAPC as an oxidizable cosubstrate and nonoxidizable POPC and in B, 7-dehydrocholesterol (7-DHC) mole
fraction in the bilayer for oxidation of PLPC in liposomes containing 7-DHC as an oxidizable cosubstrate and nonoxidizable POPC.
peroxyl radicals is known to decrease in polar solvents and being
closer to the aqueous interface, the 9-trans,cis-peroxyl may well
have a slower rate of fragmentation (kꢀ) than the 13-trans,cis-
species (kꢀ).40 Recently, Lucarini and co-workers reported that
peroxyl radicals can be stabilized by hydrogen bonding, which
is another factor that could result in a lower kꢀ for the 9-trans,cis-
peroxyl radical.41 Radical ꢀ-fragmentation is the basis of the
radical clock competition and any factor that would differentially
Figure 5. 7-Dehydrocholesterol low-energy conformation (MM2) showing
reactive H-9 and H-14 and the B-ring conjugated diene frame.
affect the fragmentation rates of 1a and 1b would differentiate
the 2a/3a and 2b/3b product ratios.
The slopes of the two series are roughly parallel in Figure 3,
but the intercept of the plot in the Figure is statistically different
values for kp determined by Equation 2 are presented in the
last column in Table 2. The relative reactivities of 18:2, 20:4,
20:5, 22:6 and 7-DHC in vesicles determined in this way are
1:3.3:4.1:4.9:24 while in solution the relative reactivities are
1:3.2:4.0:5.4:62. The fatty acids have remarkably similar
reactivity in solution and in the lipid bilayer while 7-DHC
appears to be somewhat less reactive as an H-atom donor in
vesicles as compared to isotropic media.
for the two series (p < 0.0001). This intercept is dependent on
processes that do not involve H-atom transfer-ꢀ-fragmentation
of the peroxyls, and the data presented in Figure 3 suggest that
these processes occur at different rates for the two radicals, 1a
and 1b. On the other hand, the slope difference between 2a/3a
and 2b/3b can be given by eq 4 (Figure 4), where (kꢀ) is the
ꢀ-fragmentation rate constant of 1a and (kꢀ) is that of 1b, kp is
Differential Kinetic Behavior of the Regioisomeric 9- and
13-Peroxyl Radicals in Lipid Bilayers. The data plotted in Figure
3 suggest that the two trans,cis-peroxyl radicals 1a and 2a (see
Scheme 1) have distinctly different reaction dynamics in
oxidation reactions carried out in vesicles. The consequences
of the kinetic differences between these two radicals are
expressed in the ratios of products 2a/3a (series 1 in Figure 3)
and 2b/3b (series 2). Significant differences between the two
radicals were also observed in the series of reactions in which
arachidonate or 7-DHC was the co-oxidant in the clocking
experiments (Figure 4A and B), while the differences were
smaller for eicosapentaeonate and they were negligible for the
docosahexaenoate clocking experiments (see Supporting Infor-
mation). The data shown in Figures 3 and 4 are in stark contrast
to the results obtained in isotropic media. The radicals 1a and
2a undergo chemistry in benzene solution, for example, that is
virtually identical. There is no difference in the distribution of
products derived from these radicals in reactions carried out in
solution (Figure S9, Supporting Information).
xthe propagation rate constant to 9-peroxyl for a given H-atom
donor and kp′ is that to 13-peroxyl, and a′ is a constant that is
determined by the
kp
kꢀ
kꢀp
kꢀꢀ
Slope Difference ) a′
-
(4)
(
)
partitioning of oxygen to different sites on the pentadienyl
radical.19 Thus, when hydrogen donors are equivalent to 9- or
13-peroxyls, the slope difference would be proportional to the
propagation rate constant.
Studies of the co-oxidations of PLPC and 7-DHC, presented
in Figure 4b, give results in striking contrast to those for
oxidations of PLPC alone or PLPC/PAPC co-oxidations. For
these experiments, the slope of the 2a/3a (series 1) and 2b/3b
(series 2) are substantially different. This indicates that for these
co-oxidations, 7-DHC is a better H-atom donor to radical 1a
than it is to radical 2a (kp > kp′). The results suggest that in the
local bilayer environment, the reactive hydrogen atoms at C-9
and C-14 of 7-DHC are better positioned for transfer to the
9-peroxyl radical than to the 13-peroxyl. In addition, no
statistically significant slope differences were observed for
We speculate that differentiation of peroxyl radicals 1a and
1b in lipid bilayers is due to the nonisotropic nature of these
colloids. The 9-trans,cis-peroxyl radical occupies a position in
the bilayer that is closer to the aqueous interface and this radical
is therefore exposed to a more polar environment than is the
13-trans,cis-peroxyl. The rate constant for ꢀ-fragmentation of
(40) Jha, M.; Pratt, D. A. Chem. Commun. 2008, 1252–1254.
(41) Mugnaini, V.; Lucarini, M. Org. Lett. 2007, 9, 2725–2728.
9
13042 J. AM. CHEM. SOC. VOL. 131, NO. 36, 2009

Membrane Lipid Peroxidation
A R T I C L E S
PLPC/PEPC or PLPC/PDPC co-oxidation experiments. (Figure
S6-S7, Supporting Information).
and some have been shown to have interesting and diverse
biological activities.14,17,49 Because of this, we are in the process
of defining the reaction pathways and describing the products
of 7-DHC peroxidation. The results of those studies will be
reported in due course.
Different ꢀ-fragmentation rate constants and propagation rate
constants to 9- and 13-peroxyl radicals also lead to differences
in the total amounts of 9- and 13-hydroperoxides (Figure S10,
Supporting Information). As indicated in Figure S10, more
9-products than 13-products were consistently observed in the
co-oxidation experiments of PLPC with nonoxidizable PCs,
PAPC, and 7-DHC, while they are about the same for oxidation
of linoleate in solution. Barclay and co-workers reported a
difference between 9- and 13-product formation in a study of
linoleate oxidation in micelles, more 13-product being formed
than 9.42 The different regioselectivities observed in bilayer and
micellar oxidations of linoleate are not readily explained but
the source of the selectivity must result from different exposures
of intermediate peroxyl radicals to the aqueous medium in the
micelle and bilayer environments.
Experimental Section
General Procedure for Calibration Experiments and
Co-Oxidation with Cholesterol in Liposomes. PLPC and nonre-
active PCs or cholesterol (in chloroform; [PC]final ) 50-80 mM)
were added to a 2 mL vial. The solution was dried by a flow of N2,
and was kept under vacuum for 10 min. PBS buffer (190 µL, 50
mM, pH ) 7.4) was then added and the mixture was sonicated for
10-20 s. The resulting milky suspension was incubated at 37 °C
for 10 min, which is followed by sonication for an additional 10-20
s. After addition of AIPH (10 µL, 0.02 M in PBS buffer), the milky
suspension was kept at 37 °C for 2 h. The oxidation reaction was
then quenched by adding BHT (100 µL, 0.1 M in ethanol) and
PPh3 (100 µL, 0.1 M in ethanol/CH2Cl2 ) 4/1). The mixture was
vortexed, and was then kept at room temperature for 30 min. Upon
addition of ethanol (500 µL) and LiOH (500 µL, 3 M), the resulting
mixture was stirred at 37 °C for 1 h, followed by acidification with
HCl (600 µL, 3 M). After extraction with hexanes (2 mL × 3), the
organic layers were combined and dried over MgSO4. The upper
clear solution was collected after centrifugation, and was blown
dry by a flow of N2. The residue remaining was dissolved in
benzene (300 + 100 µL) and was further stabilized by addition of
BHT in benzene (50 µL, 0.1 M). Samples were analyzed on normal
phase HPLC-UV (solvent, hexane/1.4% 2-propanol/0.1% acetic
acid; column 5 µm silica, 4.6 mm × 25 cm column, 1.0 mL/min;
detection UV 234 nm).
General Procedure for Clocking Experiments of PUFAs
and 7-DHC in Liposomes. One co-oxidation experiment of PLPC/
POPC/PAPC with nAA ) 0.1125 in the liposome is used as an
illustration: Chloroform solutions of PLPC (1.7 µmol), POPC (4.5
µmol) and PAPC (1.8 µmol) were added to a 2-mL vial and the
solution was dried by a flow of N2, after which it was kept under
vacuum for 10 min. Liposomes from this mixture were then
prepared by a procedure similar to that described in the previous
section. Thus, 95 µL of PBS buffer (50 mM, pH ) 7.4) was added
and the mixture was sonicated for an additional 10-20 s. The
resulting milky suspension was incubated at 37 °C for 10 min,
which was followed by another 10-20 s of sonication. After adding
the initiator AIPH (5 µL, 0.02 M in PBS buffer), the milky
suspension was kept at 37 °C for 1 h, and the reaction was quenched
by addition of BHT (100 µL, 0.1 M in ethanol) and PPh3 (150 µL,
0.1 M in ethanol/CH2Cl2 ) 4/1). The mixture was kept at room
temperature for 30 min and it was then hydrolyzed and worked up
in a procedure as described in the preceding section. Samples were
analyzed on normal phase HPLC-APCI-MS/MS (hexane/1.4%
2-propanol/0.1% acetic acid, 4.6 mm × 25 cm column, 5 µ, 1.0
mL/min). Selective reaction monitoring (SRM) was employed for
detection (m/z 295 f 195 for 13-substituted products 2b and 3a,
Implications
of
the
Exceptional
Reactivity
of
7-Dehydrocholesterol in Biology. The unusual reactivity of
7-DHC toward peroxyl radicals in solution and in bilayer
aggregates calls attention to the importance of this steroid in
biology. 7-DHC is the immediate biosynthetic precursor of
cholesterol,16,18,43,44 with the enzyme 7-dehydrocholesterol
reductase catalyzing the reduction of the C7dC8 double bond
of 7-DHC. The Smith-Lemli-Opitz syndrome (SLOS) is a
recessive disorder caused by mutations of the DHCR7 gene.
Impaired reductase activity for individuals having this syndrome
leads to a build-up of 7-DHC and lower than normal levels of
cholesterol in tissues and plasma.18,43,45-47 The experiments that
we report here show that 7-DHC is not just oxidatively unstable,
but it is one of the best chain-carrying molecules that has been
evaluated in free radical oxidation. This property of 7-DHC
should serve as an alert for the possibility of oxidative stress
whenever elevated levels of the compound are found, SLOS
being such an instance.48 Clearly, the compound should be added
to the list of lipids that are susceptible to peroxidation since
our studies show that 7-DHC is greater than ten times more
reactive than arachidonic acid, a polyunsaturated fatty acid that
is normally thought of as being particularly prone to peroxida-
tion. Products of 7-DHC peroxidation are also of interest because
of their potential as markers of oxidative stress in ViVo.
Furthermore, there are a host of oxysterol natural products17,48-63
(42) Barclay, L. R. C.; Crowe, E.; Edwards, C. D. Lipids 1997, 32, 237–
245.
(43) Haas, D.; Kelley, R. I.; Hoffmann, G. F. Neuropediatrics 2001, 32,
113–22.
(44) Waterham, H. R. Clin. Genet. 2002, 61, 393–403.
(45) Opitz, J. M.; Gilbert-Barness, E.; Ackerman, J.; Lowichik, A. Pediatr.
Pathol. Mol. Med. 2002, 21, 153–81.
(46) Kelley, R. I. AdV. Pediatr. 2000, 47, 1–53.
(47) Kelley, R. I.; Hennekam, R. C. J. Med. Genet. 2000, 37, 321–35.
(48) Gaoua, W.; Chevy, F.; Roux, C.; Wolf, C. J. Lipid Res. 1999, 40,
456–63.
(56) Mansoor, T. A.; Hong, J.; Lee, C. O.; Bae, S. J.; Im, K. S.; Jung,
J. H. J. Nat. Prod. 2005, 68, 331–336.
(49) Javitt, N. B. Curr. Opin. Lipidol. 2007, 18, 283–8.
(50) To¨ro¨csik, D.; Szanto, A.; Nagy, L. Mol. Aspects Med. 2009, electronic
ahead of print.
(57) Mansoor, T. A.; Lee, Y. M.; Hong, J.; Lee, C.-O.; Im, K. S.; Jung,
J. H. J. Nat. Prod. 2006, 69, 131–134.
(51) Akihisa, T.; Franzblau, S. G.; Tokuda, H.; Tagata, M.; Ukiya, M.;
Matsuzawa, T.; Metori, K.; Kimura, Y.; Suzuki, T.; Yasukawa, K.
Biol. Pharm. Bull. 2005, 28, 1117–1119.
(58) Yaoita, Y.; Amemiya, K.; Ohnuma, H.; Furumura, K.; Masaki, A.;
Matsuki, T.; Kikuchi, M. Chem. Pharm. Bull. 1998, 46, 944–950.
(59) Yaoita, Y.; Endo, M.; Tani, Y.; Machida, K.; Amemiya, K.; Furumura,
K.; Kikuchi, M. Chem. Pharm. Bull. 1999, 47, 847–851.
(60) Yaoita, Y.; Matsuki, K.; Iijima, T.; Nakano, S.; Kakuda, R.; Machida,
K.; Kikuchi, M. Chem. Pharm. Bull. 2001, 49, 589–594.
(61) Yaoita, Y.; Yoshihara, Y.; Kakuda, R.; Machida, K.; Kikuchi, M.
Chem. Pharm. Bull. 2002, 50, 551–553.
(62) Yue, J. M.; Chen, S. N.; Lin, Z. W.; Sun, H. D. Phytochemistry 2001,
56, 801–806.
(63) Albro, P. W.; Bilski, P.; Corbett, J. T.; Schroeder, J. L.; Chignell,
C. F. Photochem. Photobiol. 1997, 66, 316–325.
(52) Akihisa, T.; Nakamura, Y.; Tagata, M.; Tokuda, H.; Yasukawa, K.;
Uchiyama, E.; Suzuki, T.; Kimura, Y. Chem. BiodiVersity 2007, 4,
224–231.
(53) Ishizuka, T.; Yaoita, Y.; Kikuchi, M. Chem. Pharm. Bull. 1997, 45,
1756–1760.
(54) Kawagishi, H.; Katsumi, R.; Sazawa, T.; Mizuno, T.; Hagiwara, T.;
Nakamura, T. Phytochemistry 1988, 27, 2777–2779.
(55) Luo, X.; Li, F.; Shinde, P. B.; Hong, J.; Lee, C.-O.; Im, K. S.; Jung,
J. H. J. Nat. Prod. 2006, 69, 1760–1768.
9
J. AM. CHEM. SOC. VOL. 131, NO. 36, 2009 13043

A R T I C L E S
Xu et al.
and m/z 295 f 171 for 9-substituted products 2a and 3b; the
collision energy was 21 eV; see Supporting Information). The
elution order of the four products is 13-t,c (2b), 13-t,t (3a), 9-t,c
(2a), and 9-t,t (3b).2 Response factors relative to the 13-t,c-product
were determined before, during, and after each series of HPLC
analyses using a standard mixture with known ratios of four
products.
For a set of experiments with a given PUFA, arachidonate in
this example, the mole fraction of linoleic acid (nLA, from PLPC)
was kept at the same value. the mole fraction of arachidonic acid
(nAA, from PAPC) was varied, and the mole fraction of oleic acid
(nOA, from POPC) was adjusted to compensate for the change in
ethylvaleronitrile); HPLC, high performance liquid chromatography;
MS, mass spectrometry; MLV, multilamellar vesicles; ULV,
unilamellar vesicles; PLPC, 1-Palmitoyl-2-linoleoyl-sn-glycero-3-
phosphatidylcholine; POPC, 1-palmitoyl-2-oleoyl-sn-glycero-3-
phosphotidylcholine; DMPC, 1,2-dimyristoyl-sn-glycero-3-phos-
photidylcholine;
phosphotidylcholine;
DOPC,
DPPC,
1,2-dioleoyl-sn-glycero-3-
1,2-dipalmitoyl-sn-glycero-3-
phosphotidylcholine; PAPC, 1-Palmitoyl-2-arachidonyl-sn-glycero-
3-phosphatidylcholine; PEPC, 1-Palmitoyl-2-eicosapentaenoyl-sn-
glycero-3-phosphatidylcholine;PDPC,1-Palmitoyl-2-docosahexaenoyl-
sn-glycero-3-phosphatidylcholine; AIPH, 2,2′-azobis(2-(2-imidazolin-
2-yl) propane) dihydrochloride; MeLin, methyl linoleate; SLOS,
Smith-Lemli-Opitz syndrome; PC, phosphotidylcholine; BHT,
butylated hydroxytoluene; APCI, atmospheric pressure chemical
ionization; AA, arachidonic acid; EPA, eicosapentaenoic acid;
DHA, docosahexaenoic acid.
nAA
.
Clocking experiments on other fatty esters (PEPC and PDPC)
and 7-DHC was carried out with a similar procedure to that
described here for PAPC. In the case of 7-DHC, the oxidation
reactions were carried out for 30 min to avoid overoxidation of
7-DHC due to its high oxidizability.
Oxidation of PLPC/POPC in Liposomes Made of ULV. MLV
liposomes were prepared in the same way as described the
Calibration Experiments. The MLV liposomes were then passed
through a LiposoFast Basic extruder (Avestin, Inc., Ottawa, Canada)
19 times to give ULV liposomes as a transparent emulsion. After
addition of the initiator AIPH (10 µL, 0.02 M in PBS buffer), the
reaction mixture was kept at 37 °C for 1 h. The oxidation reaction
was worked up and analyzed in a manner similar to the Calibration
Experiments.
Acknowledgment. We thank Dr. Heidi Chen from VICC
Cancer Biostatistics Center for statistical analyses. This work was
supported by National Institutes of Health Grant ES013125 and
the National Science Foundation (CHE 0717067).
Supporting Information Available: General methods and
materials, clocking experiments in solutions, figures, typical
chromatogram, etc. This material is available free of charge via
the Internet at http://pubs.acs.org.
Abbreviations: PUFA, polyunsaturated fatty acid; 7-DHC,
7-dehydrocholesterol; MeOAMVN, 2,2′-azobis(4-methoxy-2,4-dim-
JA9029076
9
13044 J. AM. CHEM. SOC. VOL. 131, NO. 36, 2009