Factors influencing the autoxidation of fatty acids: Effect of olefin geometry of the nonconjugated diene

Article abstract of DOI:10.1021/ja049104q

Autoxidations of cis,cis, cis,trans, and trans,trans nonconjugated octadecadienoates and pentadecadienes were carried out in the presence of α-tocopherol to investigate the effect of olefin geometry on this oxidation process and provide insight into the factors that influence the autoxidation of fatty acids. We have found that as the trans character of the diene increases, the amount of O₂ trapping at the central (bis-allylic) position of the pentadienyl radical also increases. In addition, the rate constant for β-fragmentation (k_β ≈ 10 ⁶ s^-1) of the bis-allylic peroxyl radical decreased on going from the cis,cis to the trans,trans diene. We have also found that for the cis,trans nonconjugated dienes, there is a preference for trapping of the pentadienyl radical by O₂ at the transoid end, generating the cis,trans conjugated hydroperoxide as the major product.

Full text of DOI:10.1021/ja049104q

Published on Web 07/13/2004
Factors Influencing the Autoxidation of Fatty Acids: Effect of
Olefin Geometry of the Nonconjugated Diene
Keri A. Tallman, Bill Roschek, Jr., and Ned A. Porter*
Contribution from the Department of Chemistry and Center in Molecular Toxicology,
Vanderbilt UniVersity, NashVille, Tennessee 37235
Received February 18, 2004; E-mail: n.porter@vanderbilt.edu
Abstract: Autoxidations of cis,cis, cis,trans, and trans,trans nonconjugated octadecadienoates and
pentadecadienes were carried out in the presence of R-tocopherol to investigate the effect of olefin geometry
on this oxidation process and provide insight into the factors that influence the autoxidation of fatty acids.
We have found that as the trans character of the diene increases, the amount of O2 trapping at the central
(bis-allylic) position of the pentadienyl radical also increases. In addition, the rate constant for â-fragmentation
(kâ ≈ 106 s-1) of the bis-allylic peroxyl radical decreased on going from the cis,cis to the trans,trans diene.
We have also found that for the cis,trans nonconjugated dienes, there is a preference for trapping of the
pentadienyl radical by O2 at the transoid end, generating the cis,trans conjugated hydroperoxide as the
major product.
Scheme 1. General Mechanism of Kinetically Controlled Linoleate
Oxidation
Introduction
The autoxidation of polyunsaturated fatty acids (PUFAs) and
the corresponding esters has been the focus of intense investiga-
tion due to its potential importance in biology.1 When PUFAs
and their esters are exposed to oxidative stress, the primary
products are hydroperoxides.2 This oxidation process is a free
radical chain reaction, which targets the nonconjugated diene
moieties of PUFAs (1). Oxidation is initiated by abstraction of
the bis-allylic hydrogen atom generating a pentadienyl radical
(2), which can undergo trapping at either terminus or at the
central (bis-allylic) position. For example, the pentadienyl radical
derived from methyl linoleate reacts with O2 at the 13, 11, or
9 positions (Scheme 1). A hydrogen atom donor subsequently
traps the peroxyl radical (3-5) generating a hydroperoxide (6-
8). The distribution of hydroperoxides formed is highly de-
pendent on the efficiency and concentration of the hydrogen
atom donor present. For example, when oxidations of linoleates
are carried out in the presence of low concentrations of
R-tocopherol (R-Toc, <10 mM), trans,cis conjugated hydro-
peroxides (6, 8) are the major products formed. However, in
the absence of antioxidant, â-fragmentation of the intermediate
peroxyl radicals leads to products having the more stable
trans,trans diene geometry.3
The formation of conjugated hydroperoxides, which result
from trapping of the pentadienyl radical at the terminal positions,
has been well documented. Although the formation of bis-allylic
hydroperoxides of cyclic dienes has been reported, products
arising from O2 trapping at the bis-allylic position of PUFAs
have eluded researchers until recently.4 When methyl linoleate
was oxidized in the presence of a high concentration of
R-tocopherol (∼0.1 M), the nonconjugated hydroperoxide (7)
was observed.4b We have previously reported that the bis-allylic
hydroperoxide is the major product formed when oxidations of
methyl linoleate were carried out in the presence of high
concentrations of R-tocopherol (0.05-1.8 M).5 The mechanism
shown in Scheme 1 was used as a basis for analysis of product
distribution in these oxidations. The high concentrations of
antioxidant were necessary in order to observe the bis-allylic
(1) (a) Steinburg, D.; Parhsarathy, S.; Carew, T. E.; Khoo, J. C.; Witztum, J.
L. New Engl. J. Med. 1989, 320, 915. (b) Marnett, L. J. Carcinogenesis
2000, 21, 361. (c) Sevanian, A.; Ursini, F. Free Radical Biol. Med. 2000,
29, 306. (d) Spiteller, G.; Spiteller, D.; Jira, W.; Kessling, U.; Dudda, A.;
Weisser, M.; Hecht, S.; Schwarz, C. Peroxide Chem. 2000, 179. (e)
Chisholm, G. M.; Steinburg, D. Free Radical Biol. Med. 2000, 28, 1815.
(2) (a) Chan, H. W.-S.; Levett, G. Lipids 1977, 12, 99. (b) Porter, N. A. Acc.
Chem. Res. 1986, 19, 262. (c) Porter, N. A.; Caldwell, S. E.; Mills, K. A.
Lipids 1995, 30, 277.
(3) Porter, N. A.; Wujek, D. G. J. Am. Chem. Soc. 1984, 106, 2626. A value
of 430 s-1 was reported in this publication, but we are currently
re-evaluating rate constants for fragmentation of the conjugated peroxyl
radicals for use as slow free radical clocks. These studies suggest a rate
constant for this fragmentation of 620 s-1. Gillmore, J. G.; Porter, N. A.
To be published.
(4) (a) Beckwith, A. L. J.; O’Shea, D. M.; Roberts, D. H. J. Am. Chem. Soc.
1986, 108, 6408. (b) Brash, A. R. Lipids 2000, 35, 947. (c) Hamberg, M.;
Su, C.; Oliw, E. J. Biol. Chem. 1998, 273, 1080.
(5) Tallman, K. A.; Pratt, D. A.; Porter, N. A. J. Am. Chem. Soc. 2001, 123,
11827.
9
9240
J. AM. CHEM. SOC. 2004, 126, 9240-9247
10.1021/ja049104q CCC: $27.50 © 2004 American Chemical Society

Factors Influencing the Autoxidation of Fatty Acids
A R T I C L E S
Scheme 2. Synthesis of Model Dienesa
product, since the peroxyl radical leading to the hydroperoxide
undergoes a very rapid â-fragmentation (106 s-1). In addition,
we have shown that this fragmentation serves as a useful radical
clock for antioxidant H-atom transfer to peroxyl radicals with
bimolecular rate constants of ca. 106 M-1 s-1
.
Herein we report more detailed studies on the formation of
bis-allylic peroxyl radicals and their subsequent â-fragmentation
in the oxidation of octadecadienoate and model diene systems.
In addition to studies involving methyl linoleate previously
reported, we have investigated the effect of olefin geometry on
the formation and reactivity of bis-allylic peroxyl radicals.
Although little is known about their reactivity, there is evidence
that geometric isomers of methyl linoleate (i.e., trans fatty acids)
can play a significant role in biological systems.6
a Reagents: (a) 1-heptyne, CuI, NaI, K2CO3, DMF; (b) Pd/BaSO4, H2,
EtOAc; (c) NH3/ether, Li0, (NH4)2SO4; (d) 1-heptyne, n-BuLi, DMPU, THF,
-78 °C.
Synthesis of Octadecadienoates and Model Dienes. Methyl
linoleate (9) and linoelaidate (11) are commercially available,
whereas the cis,trans octadecadienoates (10a, 10b) were syn-
thesized as previously reported. Although none of the model
dienes are commercially available, they could be synthesized
in a straightforward manner (Scheme 2). The cis,cis (12) and
trans, trans (14) dienes were synthesized from a common
intermediate diyne (15). The diyne (15) was synthesized by a
copper-promoted coupling of 1-heptyne with 1-chloro-2-octyne.7
Lindlar hydrogenation using the more reactive Pd/BaSO4
catalyst or a buffered Birch reduction8 of 15 yielded 12 or 14,
respectively. The unsymmetrical diene (13) was synthesized by
addition of 1-heptynyllithium to trans-1-bromo-2-octene,9 fol-
lowed by the same Lindlar hydrogenation. The synthesis of 16
via the copper coupling has been reported,10 but we found that
this procedure resulted in a mixture of regioisomers. Carrying
out the coupling of 1-heptynyllithium with the allylic bromide
in the presence of DMPU gave exclusively the desired regio-
isomer. All model dienes were purified by column chromatog-
raphy on silver-nitrate-impregnated silica gel (SNIS) to ensure
high isomeric purity.11
Results and Discussion
The effect of olefin geometry on the formation of the bis-
allylic hydroperoxides was studied using methyl octadecadi-
enoates having cis,cis (linoleate), cis,trans, and trans,trans
(linoelaidate) geometry (9-11). Studies were also carried out
using the analogous model diene compounds (12-14). Oxida-
Oxidation Products of Octadecadienoates and Model
Dienes. Oxidations of the model dienes and octadecadienoates
(0.2 M) were carried out in benzene in the presence of varying
concentrations of R-tocopherol (0.05-1.8 M) and were initiated
by 2,2′-azobis(4-methoxy-2,4-dimethylvaleronitrile) (MeOAM-
VN). The reaction vials were incubated at 37 °C for 4 h. The
relatively short reaction times allowed for low conversion so
that the R-tocopherol was not consumed to a significant extent
during any of the oxidations, ensuring pseudo-first-order condi-
tions. The octadecadienoate hydroperoxides were analyzed
directly by normal phase HPLC. The diene hydroperoxides from
12-14 were reduced to the corresponding alcohols with PPh3
and subsequently analyzed by GC. Under these oxidation
conditions, only the kinetic products were observed, as depicted
in Scheme 3.
The nonconjugated hydroperoxide formed in each oxidation
contained the same olefin geometry as the parent substrate. The
cis,cis and trans,trans substrates generated only the cis,trans and
trans,trans conjugated hydroperoxides, respectively, while a
mixture of the two conjugated hydroperoxides were generated
upon oxidation of the cis,trans substrates. The products of the
oxidations were identified by several methods. In the case of
tions of the model dienes enabled us to probe the generality of
the proposed mechanism (Scheme 1) by focusing only on the
reactivity of the nonconjugated diene moiety without potential
interference from any other functional group in the compound.
Although the ester present in the linoleate compounds would
appear to be too far removed from the reactive site to play any
role, oxidation of the model dienes addresses this issue. The
oxidation profiles for these three model compounds were
identical to the octadecadienoates, suggesting that the ester
indeed has no influence at the site of oxidation. In addition to
probing the generality of the oxidation mechanism, the use of
the octadecadienoates and model dienes offers the opportunity
to use complimentary methods of product analysis. The fatty
ester hydroperoxides were analyzed by HPLC, whereas the diene
oxidation products were analyzed by GC. These two methods
allowed us to verify the product distribution from each substrate.
Finally, the model dienes offer the advantage that the oxidation
products could be confirmed through straightforward indepen-
dent synthesis.
(7) (a) Ivanov, I. V.; Groza, N. V.; Romanov, S. G.; Ku¨hn, H.; Myagkova, G.
I. Synthesis 2000, 691. (b) Dasse, O.; Mahadevan, A.; Han, L.; Martin, B.
R.; Di Marzo, V.; Razdan, R. K. Tetrahedron 2000, 56, 9195. (c) Jeffery,
T.; Gueugnot, S.; Linstrumelle, G. Tetrahedron Lett. 1992, 33, 5757.
(8) Brandsma, L.; Nieuwenhuizen, W. F.; Zwikker, J. W.; Ma¨eorg, U. Eur. J.
Org. Chem. 1999, 775.
(6) (a) Sargis, R. M.; Subbaiah, P. V. Biochemistry 2003, 42, 11533. (b) Katan,
M. B.; Zock, P. L.; Mensink, R. P. Annu. ReV. Nutr. 1995, 15, 473. (c)
Mensink, R. P.; Katan, M. B. New Engl. J. Med. 1990, 323, 439.
(9) Chun, J.; Li, G.; Byun, H.-S.; Bittman, R. J. Org. Chem. 2002, 67, 2600.
(10) Capon, R. J.; Barrow, R. A. J. Org. Chem. 1998, 63, 75.
(11) Williams, C. M.; Mander, L. N. Tetrahedron 2001, 57, 425.
9
J. AM. CHEM. SOC. VOL. 126, NO. 30, 2004 9241

A R T I C L E S
Tallman et al.
Scheme 3. Oxidation Products of Octadecadienoates (9-11) and
Dienes (12-14)
Figure 1. HPLC chromatograms of methyl linoleate oxidation mixture.
(A) HPLC-UV detection at 234 nm, (B) HPLC-UV detection at 207 nm,
(C) HPLC-MS (Ag CIS), intact hydroperoxides detected at 433.2 (M +
Ag). HPLC-MS analyses were carried out using a Beckman Ultrasphere
narrowbore (0.20 × 25 cm2) silica column with 0.5% iPrOH/hexanes
(0.15 mL/min) and postcolumn addition of AgBF4/iPrOH (0.3 mM, 0.075
mL/min) as described in ref 13a.
the octadecadienoates (9-11), the product mixtures were
analyzed by HPLC-MS utilizing silver coordination ion spray
mass spectrometry (Ag-CIS-MS).12 This method has proven to
be very useful in identifying intact hydroperoxides, as well as
distinguishing the regioisomers present in the oxidation mix-
ture.13 Coordination of silver to the hydroperoxides provides a
mild method of generating a positively charged species that is
detected by MS (Figure 1).
Scheme 4. Hock Fragments of Linoleate Hydroperoxide Silver
Adducts
In addition, the silver promotes Hock fragmentation14 of the
hydroperoxides in the MS to generate two aldehydes (Scheme
4). For example, the 13- and 9-hydroperoxides (18, 19) fragment
to generate aldehydes having a mass of 333 and 293, respec-
tively. The 11-hydroperoxide (17) undergoes fragmentation on
either side of the hydroperoxide generating a mixture of two
aldehydes (m/z 319 and 307). The characteristic fragments that
arise from each hydroperoxide are diagnostic and allow the
assignment of each regioisomer present in the oxidation
mixture.15 It should be noted that only the aldehyde fragment
containing the ester functionality (shown in Scheme 4) was
detected by MS. In addition to HPLC-MS analysis, the
Scheme 5. Synthesis of Diene Alcoholsa
1
hydroperoxides were isolated and analyzed by H NMR to
confirm the double-bond geometry.
The products formed from oxidation of the model dienes 12-
14 were identified and verified by independent synthesis of the
alcohols. The cis,cis bis-allylic alcohol (25) was synthesized
by addition of 1-heptynyllithium to ethyl formate, followed by
(12) (a) Bayer, E.; Gfrorer, P.; Rentel, C. Angew. Chem., Int. Ed. 1999, 38,
992. (b) Suma, K.; Raju, N. P.; Vairamani, M. Rapid Commun. Mass
Spectrom. 1997, 11, 1939.
(13) (a) Havrilla, C. M.; Hachey, D. L.; Porter, N. A. J. Am. Chem. Soc. 2000,
122, 8042. (b) Seal, J. R.; Havrilla, C. M.; Porter, N. A.; Hachey, D. L. J.
Am. Soc. Mass Spectrom. 2003, 14, 872.
a Reagents: (a) 1-heptyne, n-BuLi, THF, -78 °C; (b) Pd/CaCO3/Pb, H2,
EtOAc; (c) LiAlH4, THF, reflux; (d) H11C5MgBr, THF; (e) MsCl, pyr; (f)
(PhSe)2, LiAlH4, THF; (g) m-CPBA, THF, -78 °C; LDA, hexanal, -78
°C; i-Pr2NH, reflux.
(14) Hock fragmentation is commonly observed in solution for lipid hydro-
peroxides and is promoted by protic or Lewis acids. (a) Hock, H. Angew.
Chem. 1936, 49, 595. (b) Frimer, A. A. Chem. ReV. 1979, 79, 359. (c)
Tanigawa, S.; Kajiware, T.; Hatanaka, A. Phytochemistry 1984, 23, 2439.
(d) Gardner, H. W.; Planter, R. D. Lipids 1984, 119, 294.
Lindlar hydrogenation with the mild Pd/CaCO3/Pb catalyst
(Scheme 5).16 Addition of 1-heptyne to trans-2-octenal yielded
31. This intermediate was converted to the cis,trans (27) and
(15) Regardless of the olefin geometry in the hydroperoxide, the Hock
fragmentation of each regioisomer was virtually identical. Olefin geometry
1
was confirmed by H NMR analysis and comparison to published spectra
for similar compounds (cholesteryl linoleate). Kenar, J. A.; Havrilla, C.
M.; Porter, N. A.; Guyton, J. R.; Brown, S. A.; Klemp, K. F.; Selinger, E.
Chem. Res. Toxicol. 1996, 9, 737.
(16) Guo, C.; Lu, X. Synlett 1992, 405.
9
9242 J. AM. CHEM. SOC. VOL. 126, NO. 30, 2004

Factors Influencing the Autoxidation of Fatty Acids
A R T I C L E S
Figure 2. % Oxidation of octadecadienoates and dienes in the presence of R-tocopherol. (A) Octadecadienoates (blue diamonds) 9, (red triangles) 10a,
(green squares) 11. (B) Dienes (blue diamonds) 12, (red triangles) 13, (blue squares) 14. Oxidations were carried out in benzene with octadecadienoate or
diene (0.1 M), MeOAMVN (0.01 M), and R-tocopherol (0.05-1.8 M) at 37 °C for 4 h. The oxidations were stopped by the addition of BHT, and the diene
hydroperoxides reduced with PPh3. The octadecadienoate and diene products were analyzed by HPLC and GC, respectively, as described in the Experimental
Section.
Figure 3. Oxidation profile of octadecadienoates and dienes, ratio of bis-allylic/conjugated products versus [R-tocopherol]: (A) Octadecadienoates (blue
diamonds) 9, (red squares) 10a, (red triangles) 10b, (green circles) 11. (B) Dienes (blue diamonds) 12, (red triangles) 13, (green circles) 14. Oxidations were
carried out in benzene with octadecadienoate or diene (0.1 M), MeOAMVN (0.01 M), and R-tocopherol (0.05-1.8 M) at 37 °C for 4 h. The oxidations were
stopped by the addition of BHT, and the diene hydroperoxides reduced with PPh3. The octadecadienoate and diene products were analyzed by HPLC and
GC, respectively, as described in the Experimental Section.
trans,trans (29) alcohols by Lindlar hydrogenation and LiAlH4
reduction, respectively. The trans,trans conjugated alcohol (28)
was easily prepared in one step by addition of pentylmagnesium
bromide to trans,trans-2,4-decadienal. Synthesis of the cis,trans
conjugated alcohol was achieved through an interesting cascade
of reactions.17 First, cis-3-nonenol was converted to the selenide
(32) by displacement of the mesylate. Oxidation of the selenide
to the selenoxide with m-CPBA increases the acidity of the
R-protons. Hence, deprotonation followed by addition of hexanal
generates the â-phenylseleno alcohol, which undergoes a
selenoxide elimination to yield the desired compound.
Extent of Oxidation. Oxidations of the octadecadienoates
and model dienes were carried out side-by-side, and the amount
of oxidation products were determined by HPLC and GC,
respectively. GC detection was standardized using the authentic
synthetic alcohols (25-29). To standardize HPLC detection,
the hydroperoxides were first isolated by semipreparative HPLC.
The concentration of each solution was determined by 1H NMR
relative to an internal standard and subsequently standardized
by HPLC-UV at 207 nm.
the concentration of R-tocopherol increases, the amount of
oxidation also increases, indicative of the ability of the R-
tocopheroxyl radical to mediate oxidation. However, at the
highest concentrations, the amount of oxidation decreased. At
higher concentrations of antioxidant, peroxyl radicals are more
effectively scavenged, interrupting chain propagation. It is also
clear from the plot that olefin geometry influences the extent
of oxidation. The cis,cis dienes (9, 12) undergo at least twice
as much oxidation as the trans,trans dienes (11, 14), with the
cis,trans dienes (10a, 10b, 13) falling between these two. This
is consistent with the BDEs that have been calculated for these
types of compounds. The bis-allylic C-H BDEs for cis,cis,
cis,trans, and trans,trans 2,5-heptadiene have been calculated
as 72.7, 73.1, and 73.5 kcal/mol, respectively.19 Since the
trans,trans diene has the strongest C-H bond, it is less prone
to oxidation than the other compounds. This is also consistent
with reports that trans fatty acids undergo less oxidation than
their cis counterparts in LDL oxidations.6a
Oxygen Partitioning and Peroxyl Radical Fragmentation.
The plot of the ratio of bis-allylic to conjugated products versus
the concentration of R-tocopherol is shown in Figure 3. The
profile for the product distribution is essentially the same for
the octadecadienoates and dienes. This indicates that the ester
functionality does not influence the site of oxidation. In addition,
regardless of the substrate or the method of analysis, the
oxidation profiles are nearly identical.
The oxidation profile somewhat demonstrates the delicate
balance between the prooxidant and antioxidant properties of
R-tocopherol (Figure 2).18 This prooxidant behavior is more
clearly observed in oxidations of the dienes (Figure 2B). As
(17) Baldwin, J. E.; Reed, N. V.; Thomas, E. J. Tetrahedron 1981, 37, 263.
(18) (a) Upston, J. M.; Terentis, A. C.; Stocker, R. FASEB 1999, 13, 977. (b)
Bowry, V. W.; Ingold, K. U. Acc. Chem. Res. 1999, 32, 27. (c) Culbertson,
S. M.; Vinqvist, M. R.; Barclay, L. R. C.; Porter, N. A. J. Am. Chem. Soc.
2001, 123, 8951.
(19) Pratt, D. A.; Mills, J. H.; Porter, N. A. J. Am. Chem. Soc. 2003, 125, 5801.
9
J. AM. CHEM. SOC. VOL. 126, NO. 30, 2004 9243

A R T I C L E S
Tallman et al.
Table 1. Values for O2 Partitioning to the Bis-Allylic Position (R)
and the Rate Constant for Subsequent â-Fragmentation (kâ) of the
Bis-Allylic Peroxyl Radical
consistent with ESR and theoretical predictions. We note that
ESR21 and theory19 (Table 2) indicate that the pentadienyl radical
formed from the trans,trans diene precursor has higher unpaired
spin (38% from the ESR data) at the bis-allylic carbon than
does the radical formed from the cis,cis diene (36%). For this
pair of reactants, unpaired electron spin density on the inter-
mediate pentadienyl radical parallels trapping by O2 at the bis-
allylic position and a higher value for R (0.55) is observed for
the trans,trans diene oxidations compared to the cis,cis precursor
(0.43). Likewise, ESR experiments suggest that the pentadienyl
from the trans,cis precursors has 37% unpaired spin at the bis-
allylic position and the R values obtained (0.47) for these
substrates (10a, 10b, and 13) are consistent with unpaired spin
density being important in controlling the kinetic product
distribution. The experimentally determined values for R are
somewhat higher than the spin density ratios that have been
calculated for model compounds and that are observed by ESR,
although experiment and theory show the same trends. The
calculations were carried out using 2,5-heptadiene as the model
for octadecadienoates, and experiments to be published indicate
that there is a significant substituent effect on the partitioning
of oxygen to pentadienyl radicals.22,23
substrate
R
kâ (106 s-1
)
9
0.44 ((0.01)
0.48 ((0.01)
0.47 ((0.01)
0.55 ((0.02)
0.43 ((0.01)
0.48 ((0.01)
0.55 ((0.01)
2.32 ((0.09)
2.74 ((0.18)
2.84 ((0.14)
2.18 ((0.17)
2.36 ((0.10)
2.69 ((0.11)
2.02 ((0.09)
10a
10b
11
12
13
14
The data clearly shows that, for all compounds studied, the
amount of bis-allylic hydroperoxide increases with increasing
R-tocopherol concentration at the expense of the conjugated
products until a maximum limit is reached. This observation is
consistent with the proposed mechanism for linoleate shown in
Scheme 1.5 In this mechanism, O2 is partitioned among the three
positions of the pentadienyl radical as follows: terminal )
1 - R/2, bis-allylic ) R, and terminal ) 1 - R/2. The peroxyl
radicals subsequently abstract a hydrogen atom from R-to-
copherol to generate the hydroperoxides. Under the oxidation
conditions used in these studies, the rate constant for â-frag-
mentation of the conjugated peroxyl radical is slow (620 s-1
)
There is also a rather small influence of olefin geometry on
the propensity for â-fragmentation of nonconjugated peroxyl
radicals. The BDEs for the bis-allylic 4-peroxyl(s) derived from
cis,cis, cis,trans, and trans,trans 2,5-heptadiene have been
calculated to be 7.4, 7.9, and 8.4 kcal/mol, respectively.19 Since
the trans,trans peroxyl radical has a stronger C-OO• bond than
the analogous cis,cis substrate, kâ is expected to be slower for
the trans,trans peroxyl. However, the cis,trans substrates do not
follow this trend. In fact these compounds have the highest
observed kâ of the substrates studied.
relative to trapping by R-tocopherol and formation of the
thermodynamically favored trans,trans conjugated hydroperox-
ides is not observed.
In contrast to the conjugated peroxyl radicals, the noncon-
jugated peroxyl radical undergoes a rapid â-fragmentation (kâ)
to regenerate the pentadienyl radical (see Scheme 1). Competing
with this â-fragmentation is hydrogen atom transfer (kinh) to
the peroxyl radical by R-tocopherol with a rate constant of
3.5 × 106 M-1 s-1 20
.
As the concentration of R-tocopherol
increases, â-fragmentation becomes negligible and a limit is
reached. At this limit, the product distribution reflects the O2
partition to the three positions of the pentadienyl radical.
From Figure 3 it can be seen that olefin geometry influences
the partitioning of O2 across the pentadienyl radical. The fraction
of bis-allylic products arising from the oxidation of trans,trans
compounds reaches a higher limit than the other compounds.
This indicates that as the trans character of the pentadienyl
radical increases, so does O2 addition at the bis-allylic position.
Analysis of the mechanism leads to a kinetic expression (eq 1),
Terminal Trapping of Pentadienyl Radical. The results for
oxidation of the octadecadienoates with both alkenes having
the same geometry (cis,cis or trans,trans) showed that O2
trapping at the terminal positions of the pentadienyl radicals
(C9 and C13) was identical (Figure 4A, data for 11 not shown).
The yield of the 13- and 9-conjugated hydroperoxide was always
equal for oxidations of 9 and 11. This is not surprising, since
one expects the unpaired electron spin density to be identical
at the terminal pentadienyl positions for these substrates.
Unpaired spin density for the radicals derived from cis,trans
dienes is not equal at the pentadienyl terminal positions, as
indicated by theory and ESR spectroscopy.19,21 Indeed, spin
appears to be slightly greater at the cisoid end of the pentadienyl
compared to the transoid position for these unsymmetrical
radicals (see Table 2). If spin density controls the site of initial
addition of oxygen to delocalized radicals, the major conjugated
product would be expected to be the trans,trans conjugated diene
hydroperoxide, as shown in Scheme 6.
which describes the product ratio as a function of R, kâ, kinh
,
and [R-tocopherol].
kinh[R-Toc]
[bis-allylic]
[conjugated]
R
1 - R
)
(1)
(
)
kinh[R-Toc] + kâ
Nonlinear least-squares analysis of the data using eq 1 allowed
us to determine values for R and kâ for the various substrates
(Table 1). The data clearly shows that olefin geometry has an
effect on the O2 partitioning (R) and on fragmentation (kâ) of
the bis-allylic peroxyl radical (Table 1).
As the trans character of the pentadienyl radical increases,
the addition of O2 at the bis-allylic position (R) increases,
A previously reported study of the oxidations of 10a and 10b
in the absence of phenolic antioxidants showed, however, that
twice as much trans,cis conjugated diene product was formed
(21) (a) Bascetta, E.; Gunstone, F. D.; Scrimgeour, C. M.; Walton, J. C. J. Chem.
Soc., Chem. Commun. 1982, 110. (b) Bascetta, E.; Gunstone, F. D.; Walton,
J. C. J. Chem. Soc., Perkin Trans. 2 1983, 603.
(22) In comparing 6,9-pentadecadiene, 2,5-undecadiene, and 2,5-heptadiene, we
found that O2 trapping at the bis-allylic position (R) decreases in this order.
(23) The stability of peroxyl radicals has been shown to be influenced by
substituent effects due to hyperconjugation. (a) Pratt, D. A.; Porter, N. A.
Org. Lett. 2003, 5, 387. (b) Kranenburg, M.; Ciriano, M. V.; Cherkasov,
A.; Mulder, P. J. Phys. Chem. A 2000, 104, 915.
(20) The rate constant for R-tocopherol at 37 °C was derived from an Arrhenius
plot of known rate constants: 30 °C (3.2 × 106 M-1 s-1) (a) Burton, G.
W.; Doba, T.; Gabe, E. J.; Hughes, L.; Lee, F. L.; Prasad, L.; Ingold, K.
U. J. Am. Chem. Soc. 1985, 107, 7053. (b) Burton, G. W.; Hughes, L.;
Ingold, K. U. J. Am. Chem. Soc. 1983, 105, 5950. 50 °C (4.1 × 106 M-1
s-1) (c) Pratt, D. A.; DiLabio, G. A.; Brigati, G.; Pedulli, G. F.; Valgimigli,
L. J. Am. Chem. Soc. 2001, 123, 4625.
9
9244 J. AM. CHEM. SOC. VOL. 126, NO. 30, 2004

Factors Influencing the Autoxidation of Fatty Acids
A R T I C L E S
Table 2. Unpaired Spin as Determined by ESR and Theory on Isomeric Pentadienyl Radicals
a R and R′) methyl; see ref 21b. b R and R′) methyl; see ref 19. The values given here were the result of UB3LYP/6-311+G(2d,2p). c This work,
product distributions at the kinetic limit. Average distribution from diene substrates having identical diene geometry: 9 and 12; 10a, 10b, and 13; and 11
and 14.
Figure 4. Mole fractions of oxidation products. (A) Oxidation of 9, (red diamonds) 13-cis,trans conjugated, (black triangles) 9-cis,trans conjugated, (blue
circles) 11-cis,cis. (B) Oxidation of 10a, (red diamonds) 13-cis,trans conjugated, (black triangles) 9-trans,trans conjugated, (blue circles) 11-cis,trans. Oxidations
were carried out in benzene with 9 or 10a (0.1 M), MeOAMVN (0.01 M), and R-tocopherol (0.05-1.8 M) at 37 °C for 4 h. The oxidations were stopped
by the addition of BHT. The products were analyzed by HPLC as described in the Experimental Section.
Scheme 6. Terminal Trapping of Pentadienyl Radical
Table 3. Kinetically Controlled O2 Trapping of Unsymmetrical
Pentadienyl Radicals
substrate
âa
R
γb
as compared to the trans,trans product. Our data for O2 trapping
(Table 2) also indicate that the trans,cis conjugated diene product
is formed in preference to the trans,trans compound at the kinetic
limit. Theory and ESR experiments cannot account for the
observed distribution of products based upon unpaired spin for
these “unsymmetrical” systems. Because of this discrepancy,
we analyzed the distribution of products formed from 10a, 10b,
and 13 as a function of antioxidant concentration by a more
complex kinetic model. In Figure 4 the distribution of the three
products formed from oxidation of an isomerically symmetrical
diene 9 (cis,cis) and one of the isomerically unsymmetrical
dienes, 10a, is presented. The analysis of products from the
other two unsymmetrical dienes, 10b and 13, gives results that
are essentially identical to those of 10a.
The data from Figure 4B are analyzed under the assumption
that formation of each conjugated hydroperoxide is a separate
pathway leading to eqs 2 and 3. These equations are essentially
the same as eq 1, but now it is not assumed that O2 is partitioned
equally at the terminal positions of the pentadienyl radical. The
fraction of trapping at the bis-allylic position is still denoted as
R, whereas trapping at the transoid and cisoid ends of the
pentadienyl radical are denoted as â and γ, respectively (Scheme
6). Instead of collectively referring to terminal trapping as 1 -
R, the two termini are treated separately (i.e., â + γ ) 1 - R).
Analysis of the data using these equations, in conjunction with
eq 1, gives the ratio of O2 trapping at the three positions of the
pentadienyl radical (Table 3). The data presented in Table 3
show the results of this analysis for 10a, 10b, and 13.
10a
10b
13
0.35
0.36
0.32
0.48
0.47
0.48
0.16
0.17
0.21
a Refers to trapping at transoid end of pentadienyl radical to generate
cis,trans conjugated product. b Refers to trapping at cisoid end to generate
the trans,trans conjugated product.
The best-fit parameters for the analysis shown in Figure 4B
for 10a and similar analyses for 10b and 13 indicate that â,
representing oxygen addition at the transoid end of a pentadi-
enyl, is almost twice the value of γ, oxygen addition at the cisoid
end of the radical. This analysis is consistent with the earlier
experimental results and substantiates the conclusion that the
unpaired spin density apparently does not control the partitioning
of oxygen in this geometrically unsymmetrical pentadienyl
radical.
Dioxygen-Radical Complexes as Intermediates in Chain
Oxidation. The results for oxidation of the octadecadienoates
have been analyzed here assuming that a simple â-fragmentation
of peroxyl radicals is the critical process that interconverts the
bis-allylic and conjugated peroxyl radicals. There is, however,
a rich experimental history for peroxyl rearrangements that
9
J. AM. CHEM. SOC. VOL. 126, NO. 30, 2004 9245

A R T I C L E S
Tallman et al.
Scheme 7. Mechanisms for Peroxyl Rearrangements
should not be ignored when considering product distributions
in autoxidations of unsaturated substrates.24 The use of isoto-
pically labeled hydroperoxides or molecular oxygen in studies
of hydroperoxide rearrangements has led to the conclusion that
the rearrangement of conjugated diene hydroperoxides proceeds
via a â-fragmentation (see Scheme 7A).2a,25 The rearrangement
of simple allylic hydroperoxides, on the other hand, is more
complex. A recent computational study has led to the proposal
of an allyl-triplet dioxygen complex 33 (CX), as an intermediate
in the rearrangement; see Scheme 7B.26 Oxygen labeling and
stereochemical studies of such rearrangements are consistent
with this proposal.27,28 Calculations suggest that oxygen and an
allyl radical are held in the complex 33 by dispersion forces
and that “The observed high degree of stereoselectivity of the
rearrangement results from the formation of the CX complex,
which prevents the diffusion of the allyl radical and triplet
dioxygen partners and maintains the stereocontrol along the
fragmentation-recombination processes.”26
as 13 could rearrange to the conjugated peroxyls through two
distinct isomeric radical-dioxygen complexes, 34 and 35. The
complex 34 derives from rearrangement across the trans alkene,
while 35 is formed from rearrangement across the cis double
bond. The calculated low-energy conformation of 33 is a typical
five-membered ring envelope structure, and based upon that
structure, one anticipates that the formation of 34 would be
favored over that of 35.29 Substitution in 34 is diequatorial on
the ring, while 35 must have one axial substituent. Thus, product
distribution in the oxidation may be biased by the preferential
formation of 34 over 35, the consequence of which is the
formation of the trans,cis hydroperoxide. Experiments on
rearrangement of the bis-allylic hydroperoxides are currently
underway in our laboratory to support or deny the intermediacy
of radical-dioxygen complexes in these rearrangements. We
believe such species may well be intermediates in the first
addition of oxygen to unsaturated carbon radicals in chain
oxidation.
We speculate that such radical-dioxygen triplet complexes
may intervene in the chemistry of the bis-allylic peroxyl radicals
that are intermediates in this study. The bis-allylic peroxyl
radical derived from oxidation of an unsymmetrical diene such
Conclusion
The autoxidations of cis,cis, cis,trans, and trans,trans non-
conjugated dienes and octadecadienoates give rise to three
kinetically controlled hydroperoxides. Formation of the bis-
allylic peroxyl radical and its subsequent fragmentation depend
on the geometry of the alkene precursor and consequently the
pentadienyl radical intermediate. Significant unpaired electron
spin density is present at the central carbon of pentadienyls,
and the bis-allylic hydroperoxide product that arises from
addition at this position is the major kinetic product for each of
the systems studied. Unpaired spin density is not the only factor
that determines the position of oxygen addition to a delocalized
radical, however, and we speculate that radical-triplet dioxygen
complexes may be intermediates in the formation and rear-
rangement of delocalized radicals. Rearrangement of the bis-
(24) (a) Schenck, G. D.; Neumuller, O. A.; Eisfeld, K. C. Angew. Chem. 1958,
595. (b) Brill, W. F. J. Am. Chem. Soc. 1965, 87, 3286. (c) Beckwith, A.
L.; Davies, A. G.; Davison, I. G. E.; Maccoll, A.; Mruzek, M. H. J. Chem.
Soc., Chem. Commun. 1988, 475. (d) Avila, D. V.; Davies, A. G.; Davison,
I. G. E. J. Chem. Soc., Perkin Trans 2 1988, 1847. (e) Davies, A. G.; Kinart,
W. J. J. Chem. Res. 1989, 22. (f) Dang, H.; Davies, A. G.; Davison, I. G.
E.; Schiesser, C. H. J. Org. Chem. 1990, 55, 5085. (g) Porter, N. A.; Kaplan,
J. K.; Dussault, P. H. J. Am. Chem. Soc. 1990, 112, 1266.
(25) Chan, H. W.; Levett, G.; Matthew, J. A. Chem. Phys. Lipids 1979, 24,
245.
(26) Olivella, S.; Sole´, A. J. Am. Chem. Soc. 2003, 125, 10641.
(27) (a) Mills, K. A.; Caldwell, S. E.; Dubay, G. R.; Porter, N. A. J. Am. Chem.
Soc. 1992, 114, 9689. (b) Porter, N. A.; Mills, K. A.; Carter, R. L. J. Am.
Chem. Soc. 1994, 116, 6690. (c) Porter, N. A.; Mills, K. A.; Caldwell, S.
E.; Dubay, G. R. J. Am. Chem. Soc. 1994, 116, 6697. (d) Lowe, J. R.;
Porter, N. A. J. Am. Chem. Soc. 1997, 119, 11534.
(28) Previous studies have ruled out the intermediacy of the cyclic 1,2-
dioxolanyl-4-yl radical. (a) Brill, W. F. J. Chem. Soc., Perkin Trans. 2
1984, 621. (b) Porter, N.; Zuraw, P. J. Chem. Soc., Chem. Commun. 1985,
1472. (c) Porter, N. A.; Wujek, J. S. J. Org. Chem. 1987, 52, 5085. (d)
Boyd, S. L.; Boyd, R. J.; Shi, Z.; Barclay, R. C.; Porter, N. A. J. Am.
Chem. Soc. 1993, 115, 687.
(29) It has been shown that for allylperoxyl rearrangements, the trans allyl
radical-dioxygen reaction is faster than the cis allyl radical-dioxygen reaction
(see ref 27c).
9
9246 J. AM. CHEM. SOC. VOL. 126, NO. 30, 2004

Factors Influencing the Autoxidation of Fatty Acids
A R T I C L E S
products. The initiator, 2,2′-azobis(4-methoxy-2,4-dimethylvaleronitrile)
(MeOAMVN), was obtained from Wako and dried under high vacuum
for 2 h. R-Tocopherol was purchased from Aldrich and purified by
flash chromatography (10% EtOAc/hexanes on silica), protected from
light and oxygen. It is crucial that the R-tocopherol be purified prior
to use and dried overnight under high vacuum. Benzene was passed
through a column of alumina.
allylic peroxyls to conjugated peroxyls occurs with a rate
constant between 2.2 and 2.8 × 106 s-1. Regardless of the
mechanism by which this rearrangement takes place, this
rearrangement can be used as a peroxyl radical clock.5 Although
the previous studies were carried out using methyl linoleate in
the clock reactions, any of the substrates discussed here could
be used depending on the preference for HPLC or GC analysis.
This offers a very straightforward method for determining the
rate constant of H-atom transfer to a peroxyl in the kinetic range
General Procedure for Oxidations. Stock solutions of each
octadecadienoate or diene (1.5-1.7 M), MeOAMVN (0.1 M), and
R-tocopherol (1.0 M) were made up in benzene. For samples containing
high concentrations, neat R-tocopherol (2.2 M) was used. Samples were
made up in autosampler vials with a total reaction volume of 100 µL.
It was important to add the solutions in the following order to avoid
premature oxidation: R-tocopherol (0.05-1.8 M), octadecadienoate or
diene (0.15-0.17 M), MeOAMVN (0.01 M) and diluted to 100 µL
with benzene. The sealed samples were then incubated at 37 °C for
4 h.
of 105 to 106 M-1 s-1 5,30
Although the formation of bis-allylic
.
hydroperoxides only can be observed at concentrations of
R-tocopherol that are not normally encountered in vivo, the
formation of these products may become important in confined
systems, such as an enzyme active site4b or a lipid particle having
excellent H-atom donors present.20c,30,31
Experimental Section
After 4 h, the oxidation was stopped by the addition of BHT (50 µL
of 1.0 M solution in hexanes), followed by the addition of the internal
standard (for octadecadienoates, 5 mM benzyl alcohol; for dienes, 5
mM tetradecane). The octadecadienoate samples were diluted to 1.0
mL with hexanes and analyzed by HPLC as their hydroperoxides. The
diene samples were reduced with PPh3 (50 µL of 1.0 M solution/
hexanes) and analyzed by GC.
1
General Methods. H and 13C NMR spectra were collected on a
300 or 400 MHz NMR, using Bruker software. HPLC analyses were
carried out with a Waters 600 liquid chromatograph interfaced to a
Waters 996 PDA detector. Linoleate hydroperoxides were separated
on a Beckman Ultrasphere silica column (0.46 × 25 cm2) using 0.5%
iPrOH/hexanes at 1.0 mL/min and detected at 207 nm. GC analyses
were carried out with a Hewlett-Packard 6890 gas chromotagraph
equipped with a DB-5 (30 m × 0.32 mm × 0.25 mm) fused silica
column from J&W Scientific. The diene alcohols were separated using
a temperature program of 100-180 °C @ 5 °C/min, 180-280 °C @
20 °C/min (10 min). HPLC-MS analyses were carried out as previously
described13a using a Beckman Ultrasphere narrowbore silica column
(0.20 × 25 cm2) with 0.5% iPrOH/hexanes (0.15 mL/min) and
postcolumn addition of AgBF4/iPrOH (0.3 mM, 0.075 mL/min).
Methyl linoleate and linoelaidate were purchased from Nu-Chek Prep
and chromatographed on silica (10% EtOAc/hexanes) prior to use. All
other compounds were synthesized (see Supporting Information) and
freshly chromatographed on silica (10a and 10b with 10% EtOAc/
hexanes, 12-14 with hexanes) prior to use to remove any oxidation
Acknowledgment. We thank Derek A. Pratt and Alan R.
Brash for helpful discussions. N.A.P. thanks NIH (HL-17921),
NIEHS (P30ES00267), and the National Science Foundation
for support of this work. K.A.T. and B.R. acknowledge support
from the Center in Molecular Toxicology, Vanderbilt University
(T32ES07028). K.A.T. also acknowledges support from the NIH
(F32HL68439-01).
Supporting Information Available: Experimental procedures
for the synthesis of all compounds, HPLC and GC standardiza-
tion of octadecadienoates and dienes, and plots of mole fraction
of oxidation products of 10b, 11, and the dienes (12-14). This
material is available free of charge via the Internet at
http://pubs.acs.org.
(30) Wijtmans, M.; Pratt, D. A.; Valgimigli, L.; DiLabio, G. A.; Pedulli, G. F.;
Porter, N. A. Angew. Chem., Int. Ed. 2003, 42, 4370.
(31) Valgimigli, L.; Brigati, G.; Pedulli, G. F.; DiLabio, G. A.; Mastragostino,
M.; Arbizzani, C.; Pratt, D. A. Chem.sEur. J. 2003, 9, 4997.
JA049104Q
9
J. AM. CHEM. SOC. VOL. 126, NO. 30, 2004 9247