Identification of radical species derived from allergenic 15-hydroperoxyabietic acid and insights into the behaviour of cyclic tertiary allylic hydroperoxides in Fe(II)/Fe(III) systems

Article abstract of DOI:10.1016/j.tet.2008.04.035

To understand the skin sensitization mechanism of 15-hydroperoxyabietic acid, the major allergen in colophony, we first examined the formation of potential reactive radicals derived from its reaction with light, heat and TPP-Fe³⁺. Trapping with 1,1,3,3-tetramethylisoindolin-2-yloxyl nitroxide confirmed the?formation of carbon-centred radicals derived from allyloxyl/allylperoxyl radicals as a consequence of the hydroperoxide scission. Particular interest was further given to the reactivity with Fe(II)/Fe(III) due to?the biological importance of haem containing enzymes. Using a monocyclic 15-hydroperoxyabietic acid-like compound as a model of allergenic allylic hydroperoxides, we evidenced, by the ESR spin-trapping technique, the competition between carbon and oxygen-centred radicals formed in the presence of Fe(II)/Fe(III) in organic/aqueous media. We complemented the study by showing the possibility of formation, via a radical mechanism induced by ferric chloride, of an adduct between the allylic hydroperoxide and N-acetyl-cysteine ethyl ester. The results gave new knowledge on the possible generation of highly reactive radicals that could lead to the formation of antigenic structures.

Full text of DOI:10.1016/j.tet.2008.04.035

Tetrahedron 64 (2008) 5680–5691
Contents lists available at ScienceDirect
Tetrahedron
journal homepage: www.elsevier.com/locate/tet
Identification of radical species derived from allergenic 15-hydroperoxyabietic
acid and insights into the behaviour of cyclic tertiary allylic hydroperoxides in
Fe(II)/Fe(III) systems
a,
a
b
a
*
´
Elena Gimenez-Arnau , Laure Haberkorn , Loris Grossi , Jean-Pierre Lepoittevin
a Institut de Chimie de Strasbourg (CNRS-ULP), Laboratoire de Dermatochimie, Clinique Dermatologique, CHU, 1 Place de l’Hoˆpital, 67091 Strasbourg, France
b
`
Dipartimento di Chimica Organica ‘A. Mangini’, Universita di Bologna, Viale Risorgimento 4, 40136 Bologna, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
To understand the skin sensitization mechanism of 15-hydroperoxyabietic acid, the major allergen in
colophony, we first examined the formation of potential reactive radicals derived from its reaction with
light, heat and TPP-Fe3þ. Trapping with 1,1,3,3-tetramethylisoindolin-2-yloxyl nitroxide confirmed
the formation of carbon-centred radicals derived from allyloxyl/allylperoxyl radicals as a consequence of
the hydroperoxide scission. Particular interest was further given to the reactivity with Fe(II)/Fe(III) due
to the biological importance of haem containing enzymes. Using a monocyclic 15-hydroperoxyabietic
acid-like compound as a model of allergenic allylic hydroperoxides, we evidenced, by the ESR spin-
trapping technique, the competition between carbon and oxygen-centred radicals formed in the pres-
ence of Fe(II)/Fe(III) in organic/aqueous media. We complemented the study by showing the possibility of
formation, via a radical mechanism induced by ferric chloride, of an adduct between the allylic hydro-
peroxide and N-acetyl-cysteine ethyl ester. The results gave new knowledge on the possible generation of
highly reactive radicals that could lead to the formation of antigenic structures.
Received 14 February 2008
Received in revised form 1 April 2008
Accepted 8 April 2008
Available online 11 April 2008
Keywords:
15-Hydroperoxyabietic acid
Allergenic allylic hydroperoxide
Radicals and radical reactions
Fe(II)/Fe(III) systems
Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction
structures, probably via the involvement of reactive radical in-
termediates that have gained increased importance in the discus-
The accepted mechanism for antigen formation in the allergic
contact dermatitis pathology is the generation of covalent bonds
between electrophilic chemical functions present on allergens,
low-molecular weight compounds penetrating the skin barrier, and
nucleophilic residues on skin proteins.1 The processing of such
hapten–protein complex by immunocompetent skin antigen-pre-
senting cells and the transmission of this information to T-cells in
the lymph nodes lead to erythema and oedema, the major clinical
signs of the pathology.2 Initially harmless molecules can be
converted into electrophilic derivatives and, therefore, possibly aller-
genic compounds, via metabolic processes mainly based on oxido-
reduction reactions,3 or non-enzymatically as for the reaction with
atmospheric oxygen.4 Terpenes, common compounds of natural
origin containing double bonds and, consequently, prone to oxi-
dation on air exposure, belong to this category of allergens. In fact,
many allylic hydroperoxides derived from autoxidation of terpenes
are known to be strong skin sensitizers.5–7 However, allylic hy-
droperoxides are not electrophiles and represent, therefore, an
alternative mode of reaction with proteins to form antigenic
sion of hapten–protein binding in the latest years.1
In the course of our studies on allergenic allylic hydroperoxides,
we have been interested in 15-hydroperoxyabietic acid 1 (Fig. 1),
identified as the major allergen in colophony, a naturally occurring
material obtained from coniferous trees and used in multiple
products due to its stickiness, emulsifying and isolating properties.8
15-Hydroperoxyabietic acid 1 is the most important oxidation
OOH
COOH
COOH
1
2
OOH
OOH
4
3
* Corresponding author. Tel.: þ33 3 88 35 06 64; fax: þ33 3 88 14 04 47.
Figure 1. Chemical structures of 15-hydroperoxyabietic acid 1, abietic acid 2 and
model compounds 3 and 4.
´
E-mail address: egimenez@chimie.u-strasbg.fr (E. Gimenez-Arnau).
0040-4020/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2008.04.035

E. Gime´nez-Arnau et al. / Tetrahedron 64 (2008) 5680–5691
5681
derivative in air of abietic acid 2, which is the principal diterpene
resinic acid constituting colophony.9
to the formation of antigenic structures, the first step of the allergic
contact dermatitis mechanism.
In the context of contact allergy and to understand the in-
teraction of 1-like compounds with proteins, we started our studies
by reporting the identification of potential reactive radicals derived
from two model compounds, 3 and 4, that were found to be also
strong sensitizers.10,11 Allylic hydroperoxides are known to form
allyloxyl and allylperoxyl radicals, which rapidly decompose via
intramolecular cyclization, fragmentation or hydrogen abstraction
to form carbon-centred radicals.12 Carbon-centred radicals derived
from the monocyclic model 3 were completely identified using
a combination of two spin-trapping techniques, the former leading
to neutral compounds and the second allowing the detection of
radical intermediates by ESR. For converting radical species in-
duced by light, heat or TPP-Fe3þ into neutral molecules, the stable
1,1,3,3-tetramethylisoindolin-2-yloxyl nitroxide 5 was used as trap.
For the ESR spin-trapping studies, the allylic alcohol precursor of
the hydroperoxide was used. In particular, the allyloxyl radical was
generated, in situ, by photolysis of the corresponding nitrite,
formed by reacting the parent allylic alcohol with t-BuONO, which
was also acting as the spin trap. The formation of epoxy alkyl
radicals via intramolecular cyclization of the parent allyloxyl radical
was proven, as well as the abstraction of hydrogens from different
allylic positions as evidenced by the trapping of the corresponding
allylic radicals.11 Similar studies carried out with the bicyclic model
4 and the epoxy alkyl radical derived from it showed the existence
of resonance hybrids of an allyl type radical due to the presence of
the conjugated double bond (Fig. 2).10
To complement our previous studies, we now report the po-
tentially reactive carbon-centred radicals that can be derived from
allergenic 15-hydroperoxyabietic acid 1, identified by radical trap-
ping by employing 5 and using light, heat or TPP-Fe3þ as radical
inducer. The structure of the isolated adducts was determined us-
ing one- and two-dimensional NMR. In parallel, being iron the most
abundant metal in humans, and for understanding the reactivity of
1 with TPP-Fe3þ, we studied more in detail the behaviour of such
cyclic tertiary allylic hydroperoxides in Fe(II)/Fe(III) systems by ESR
spectroscopy. But, since 1 was found to be unstable, and, therefore,
pure samples for ESR experiments could not be obtained, the model
compound 3 was used together with DMPO and DEPMPO as spin
traps. We confirmed the formation of carbon-centred radicals, as
well as the formation of allyloxyl/peroxyl radicals, the latter was
not evidenced by 5 since it was able to trap just carbon-centred
radicals. Finally, being cysteine the most labile amino acid in bi-
ological radical reactions, we studied the reaction of 3 with N-
acetyl-cysteine ethyl ester catalyzed by ferric chloride. An adduct
was formed by the reaction of the lateral chain of N-acetyl-cysteine
ethyl ester, through the thiyl radical, with a carbon-centred free
radical in an allylic position of 3. The evidence of carbon and oxy-
gen-centred radicals’ formation, which can be derived from aller-
genic allylic hydroperoxides, together with the possible
degradation of these compounds induced by the cysteine–ferric
chloride system, gave new insights into the in situ generation of
highly reactive radicals in the epidermis. These species could lead
2. Results and discussion
2.1. Radical trapping experiments with 15-
hydroperoxyabietic acid
It is well known that 1,1,3,3-tetramethylisoindolin-2-yloxyl
nitroxide 5 couples with carbon-centred radicals to give stable non-
radical alkoxyamine products, which can be characterized by using
a full range of spectroscopic techniques.13 In this study, we evalu-
ated the formation of carbon-centred radical species derived from
15-hydroperoxyabietic acid 1 by the aminoxyl radical trapping
technique using nitroxide 5 as a scavenger. The advantages of this
nitroxide are several: it exhibits a chemical inertness, uncommon
for most radicals, it is thermally stable, symmetrical (leading to
a simplified NMR spectrum of the products) and has a suitable UV
absorbing chromophore.
We carried out the experiments in a 1:1 (v/v) H2O/CH3CN
mixture. Non-polar organic solvents can mimic the reaction con-
ditions within, for example, a membrane, but the reactivity of allylic
hydroperoxides in aqueous solutions is also of interest, since it is
assumed that many biological processes proceed at aqueous/or-
ganic medium interface. Moreover, the skin could also be seen as
a semi-organic medium mixing hydrophilic and hydrophobic
properties. Radical formation was directly induced by thermal or
photochemical cleavage of 1 and by using the iron(III) porphyrin
complex TPP-Fe3þ. Reactions’ proceed was followed by TLC until
complete disappearance of 1. The reaction products (Scheme 1)
were isolated by column chromatography on silica gel and their
chemical structures identified by NMR (1H, 13C, DEPT, COSY, NOESY,
1H–13C HSQC and HMBC).
2.1.1. Structural analysis of compounds 6a, 6b, 7a, 7b and 8
The chemical structure of adducts resulting from radical trap-
ping by 5 was determined by two-dimensional (2D) NMR. 1H–13
C
HSQC experiments allowed the correlation of each proton to a car-
bon atom through 1J(C, H). Then, the skeleton of the molecules was
build thanks to the 2J(C, H) and 3J(C, H) long-range couplings ob-
served in the HMBC spectra. Finally, the relative stereochemistry of
the diverse substituents on the adducts was resolved using NOESY
data on 1H–1H correlations through space.
Analysis of the one-dimensional (1D) 1H and 13C NMR spectra of
isolated diastereomers 6a and 6b showed the presence of a new
carbonyl chemical function at C-13 instead of the 1 isopropyl
fragment. Also, the disappearance of the 1 vinyl proton at 5.42 ppm,
the new set of signals in the 6.8–7.2 ppm region corresponding to
the scavenger aromatic protons and the signals for a proton in
position
a to an alkoxyamine for each diastereomer at 4.10 and
4.39 ppm, respectively, were in agreement with the formation of an
adduct between 5 and a carbon-centred radical derived from 1 at C-
7. The relative stereochemistry of 6a, having H-7 (4.10 ppm) on an
equatorial position, was established by NOESY 1H–1H correlations
between H-7 and H-14 (6.20 ppm), and H-7 and the two H-6 (1.46
and 1.79 ppm). The absence of a NOESY cross-peak between H-7
(4.39 ppm) and H-14 (6.89 ppm) in adduct 6b, together with H-7
being correlated with H-5 (1.86 ppm), H-6eq (2.54 ppm) and H-9
(1.41 ppm), were in favour of H-7 being on an axial position (Fig. 3).
Mass spectrometric analysis of 6a and 6b showed an m/z value of
466 that supported the structures elucidated by the NMR analysis.
Diastereomers 7a and 7b were also isolated easily by column
chromatography on silica gel. As for 6a and 6b, the disappearance of
the 1 vinyl proton at 5.42 ppm was observed in the 1D 1H NMR
spectra, and new signals appeared instead for a proton in position
O
OR
OR
R = H, OH
O
Figure 2. Carbon-centred radicals derived from previously studied model compounds
3 and 4.
a
to an alkoxyamine for each diastereomer at 4.27 and 4.20 ppm,

5682
E. Gime´nez-Arnau et al. / Tetrahedron 64 (2008) 5680–5691
OOH
1
COOH
H2O/CH3CN
1:1 (v/v)
N O
heat, hν or
TPP-Fe3+
5
16
OH
15
OH
N
12
17
12
O
O
11
11
20
17
13 OH
13
1
1
OH
14
14
2
2
3
9
9
10
10
8
8
3
8'
4
8'
1'
4
6
6
5
5
7
7
O
N
O
O
O
COOH
COOH
9'
COOH
COOH
9'
N
N
16
19
1'
15
18
7'a
5'
7'a
5'
3'
3'a
3'
3'a
10'
11'
10'
7'
6'
7'
6'
11'
6a
6b
7a
7b
4'
4'
16
17
OH
12
15
12
O
O
11
11
20
O
17
13
13
8
1
1
14
14
8'
2
3
2
3
9
6
9
6
O
OH
10
10
8
7
9'
O
N
7
1'
4
4
5
7'a
5'
5
3'
3'a
10'
7'
6'
COOH
COOH
COOH
18
COOH
15
16
19
11'
4'
8
9
10
11
Scheme 1. Trapping of carbon-centred radicals derived from 1.
respectively. This was in agreement with the formation of an ad-
duct with 5 at C-7. On another hand, the chemical shifts of carbon
atoms C-13 (141.6 ppm) and C-15 (81.9 ppm) of 1 did move to 73.8
and 74.2 ppm in the case of one diastereomer, and to 74.1 and
73.7 ppm in the case of the other one. These values were charac-
teristic of carbons bearing a tertiary alcohol. NOESY experiments
concluded the axial orientation of scavenger 5 in both 7a and 7b. In
the two structures, H-7 showed a cross-peak with the vinyl H-14
and the two H-6 protons. The relative stereochemistry dis-
tinguishing 7a from 7b was defined on C-13. Adduct 7a had the
isopropyl group on an axial position as protons of the methyl
groups C-16 (1.16 ppm) and C-17 (1.24 ppm) were correlated with
protons of the C-20 methyl group (0.69 ppm). Adduct 7b had the
isopropyl group on an equatorial position as protons of the methyl
groups C-16 (1.19 ppm) and C-17 (1.21 ppm) were correlated with
protons of a methyl group belonging to 5 and with the vinyl proton
H-14 (6.12 ppm) (Fig. 3). Mass spectrometric analysis of 7a and 7b
showed an m/z value of 526 that supported the structures eluci-
dated by the NMR analysis.
Complete NMR analysis of adduct 8 indicated the loss of the
isopropyl group of 1, its replacement by a carbonyl chemical func-
tion and the disappearance of all vinyl protons. Moreover, the
quaternary carbon C-8 of 1 at 134.7 ppm had also disappeared.
Again, new signals in the aromatic region corresponding to the
scavenger (6.89–7.09 ppm) and also a signal corresponding to
a proton in position a to an alkoxyamine (3.96 ppm) were observed.
A new tertiary carbon at 56.9 ppm and a new quaternary carbon at
58.6 ppm indicated the epoxidation of the C-7/C-8 double bond of
1. As a consequence, it was concluded that 5 could only be bound at
C-14. The relative stereochemistry of the substituents on positions
O
O
14
H
14
H
H
H
9
6
6
5
7
7
O
H
N
H
HOOC
H
HOOC
H
O
O
H
H
N
14
H
O
N
6b
6a
H
8
H
6
7
HOOC
O
H
OH
OH
H
H
H
15
H
H
13
13
OH
H
20
20
H
H
14
OH
14
8
H
6
6
7
7
H
H
HOOC
H
HOOC
H
O
O
N
N
7b
Figure 3. Main NOESY effects observed in compounds 6a, 6b, 7a, 7b and 8.
7a

E. Gime´nez-Arnau et al. / Tetrahedron 64 (2008) 5680–5691
5683
7, 8 and 14 was established thanks to the data furnished by the
NOESY experiments, being proton H-7 (3.31 ppm) correlated with
H-14 (3.96 ppm) and the two H-6 protons (1.61 and 1.88 ppm)
(Fig. 3). Mass spectrometric analysis of 8 showed an m/z value of
482 that supported the structure elucidated by the NMR analysis.
haem-hydroperoxide reactions for the last three decades, a clear
mechanistic consensus has not evolved on the homolytic or het-
erolytic O–O bond cleavage by iron porphyrin complexes. One of
the frequently used mechanistic tools to distinguish the homolytic/
heterolytic O–O bond cleavage of hydroperoxides is to analyze the
products formed in the epoxidation of olefins by iron porphyrin
complexes and hydroperoxides.15–17 It is commonly accepted that
a high yield of epoxidation indicates the formation of a high-valent
iron(IV) oxo porphyrin cation radical intermediate 12 via hetero-
lytic O–O bond cleavage (Scheme 2, pathway A), while the gener-
ation of a ferryl-oxo complex 13 via O–O homolysis of the
hydroperoxide bond affords low yield or none olefin epoxidation
(Scheme 2, pathway B). Traylor and co-workers suggested a differ-
ent mechanism for the product distribution of O–O bond homoly-
sis.15,18 It was proposed that the heterolytic O–O bond cleavage was
the starting process, followed by a fast side reaction between 12
and the hydroperoxide evidenced in the product distribution of
O–O bond homolysis (Scheme 2, pathway A followed by C). Further-
more, experiments conducted with oxidant mixtures let to con-
clude that even if 12 is reacting fast with ROOH, the low
epoxidation due to the homolytic O–O bond cleavage cannot be
excluded.
2.1.2. Structural analysis of compounds 9–11
Each compound was identified by comparison with the refer-
ence materials obtained from a control experiment carried out
without 5. Compound 9 was found to be an
a hydroxy epoxide. The
NMR spectra showed the loss of the C-13/C-14 double bond of the
starting material and there were no signals that could be assigned
to scavenger 5. However, a new tertiary carbon at 70.8 ppm, with
a related proton at 3.97 ppm, was pointed out. These chemical
shifts were typical of a carbon atom bearing a hydroxyl chemical
group and of a proton in geminal position to this hydroxyl. Also, the
chemical shifts of C-13 (141.6 ppm) and C-15 (81.9 ppm) of 1 did
move to 66.5 and 63.1 ppm, respectively, which were typical values
of quaternary carbon atoms belonging to an epoxide. The relative
stereochemistry of the substituents on positions 13 and 14 was
established thanks to the data furnished by the NOESY experi-
ments. The C-14 hydroxyl chemical group was on an axial position.
The epoxide was orientated to the
a
side of the molecule as protons
In the reaction between TPP-Fe3þ and 1, the epoxidation of the
double bonds was not observed. On the contrary, the presence of
both the peroxyl (ROOꢀ) and the alkoxyl (ROꢀ) radicals was neces-
sary to account for adducts 6a, 7a and 7b. Therefore, the O–O bond
cleavage had to take place and possibly through pathways A and C
(Scheme 2). The peroxyl radical ROOꢀ formed could rearrange via
intramolecular addition to the C-13/C-14 double bond to a carbon-
centred radical with a dioxetane structure, which is known to
cleave to get acetone and a carbonylic compound.19 The radical
formed at C-14, due to the presence of the allylic double bond,
coexisted with the resonance hybrid form containing the carbon-
centred radical at C-7, which when trapped by 5 led solely to the
diastereomer 6a (Scheme 3). On another hand, also the alkoxyl
of the methyl group at C-16 (1.09 ppm) were correlated with H-12eq
(1.82 ppm) and those of the methyl group at C-17 (1.26 ppm) with
H-14 (3.97 ppm). Finally, the structures of unsaturated ketone 10
and tertiary alcohol 11 were well known as these compounds were
characterized in previous work on the hemisynthesis of 1.14
2.1.3. Mechanistic interpretations
Reaction products were formed in different yields depending on
the method used to induce radical generation. As shown in Table 1,
in the presence of TPP-Fe3þ (1:1 (v/v) H2O/CH3CN solvent, room
temperature, 1 h) the decomposition of allylic hydroperoxide 1 was
very fast but only compounds 6a, 7a and 7b were isolated. In
contrast, both the thermal (1:1 (v/v) H2O/CH3CN solvent, reflux,
22 h) and the photochemical (1:1 (v/v) H2O/CH3CN solvent, UV–vis,
20 h) decomposition of 1 resulted slower, leading to several re-
action products and in the same percentage. The only difference
was the formation of compound 8 when the reaction was induced
photochemically. In general, the overall yield of trapped radicals
was very similar for both the thermal and the photochemical re-
actions. A complementary experiment in the absence of 5 was
carried out. Obviously, in this case, only compounds resulting from
the decay of carbon and oxygen-centred radicals were isolated.
Among them, 9 and 11 in a 3% and 5% yields, respectively. However,
the formation of other compounds in a very low yield, even if not
isolated, cannot be excluded.
radicals ROꢀ could undergo a cyclization and lead to an
a-oxir-
anylcarbinyl radical in resonance with other allyl radicals due to the
presence of the allylic double bond. In an aqueous medium, the
trapping of this radical by 5 and the following epoxide ring opening
can account for the formation of diastereomers 7a/7b (Scheme 3).
Under thermal or photochemical action, hydroperoxides can
cleave homolytically to produce ROꢀ and hydroxyl (HOꢀ) radicals.
The latter, via hydrogen abstraction from unreacted hydroperoxide
molecules, accounts for ROOꢀ formation,20 even if direct homolysis
of the O–H bond of the hydroperoxide group under heating or ir-
radiation conditions has also been reported.21,22 Alkoxyl ROꢀ and
peroxyl ROOꢀ radicals so-formed can account for the formation of
all the derivatives 6a/6b and 7a/7b (Scheme 3).
As models for the haem-containing enzymes, reactions of
iron(III) porphyrin complexes with oxidants, such as peroxyacids
and hydroperoxides, have been extensively studied with the in-
tention of elucidating the mechanisms of O–O bond activation and
oxygen atom transfer reactions. Despite intensive investigation on
H+
O
R
O
FeIII
FeIII
Porp
ROOH
Porp
Table 1
Heterolysis
A
Homolysis
B
Yields of the products formed in the radical trapping experiments with 1 according
to the radical inducer used
Radical inducer
Timea (h)
Products yield (%)
O
FeIV
OH
FeIV
C
6a
6b
7a
7b
8
9
10
11
ROH
Porp
Porp
RO
TPP-Fe3þ
Heat
1
22
5
7
4
d
d
3
2
1
d
2
6
d
d
d
d
3
d
5
3
d
10
17
10
d
3
5
12
13
olefin
no epoxide formation
ROOH ROO
Heatb
d
olefin
epoxide formation
h
n
20
4
3
1
6
5
3
a
The reactions were followed by TLC and stopped after complete disappearance
of 1.
Scheme 2. Heterolytic or homolytic mechanism for the O–O bond cleavage of hy-
droperoxides by iron(III) porphyrin complexes.
b
Control experiment was carried out in the absence of 5.

5684
E. Gime´nez-Arnau et al. / Tetrahedron 64 (2008) 5680–5691
O
O
O
6a/6b
R-OO
5
O
O
OO
COOH
COOH
5
O
O
O
COOH
COOH
COOH
1O2 and 5
or
5 and 1O2
8
R-O
O
O
O
H2O
7a/7b
5
COOH
COOH
COOH
HO
11
9
Scheme 3. Mechanistic interpretation for the generation of 6a/6b, 7a/7b, 8, 9 and 11.
As mentioned above, only 6a was isolated after 1-h reaction
between 1 and TPP-Fe3þ (complete disappearance of 1). However,
additional NMR follow-up of the reaction, in deuterated chloro-
form, showed the appearance of signals belonging to 6b concomi-
tantly to the decrease in the intensity of signals attributed to 6a and
reaching the same concentration in the reaction mixture after 2
days. We hypothesized that both keto diastereomers were in
equilibrium through an enol non-isolated intermediate as proved
by NMR experiments conducted in deuterated benzene, in which
the equilibrium between 6a and 6b was reached after 7 days. Fi-
nally, 1 resulted much less reactive when the process was induced
by heat or light and its decomposition took place in circa 20 h. This
allowed enough time for the 6a/6b equilibrium to be reached and
then the possibility to isolate both diastereomers. Concerning ad-
ducts 7a/7b, only the carbon-centred radical at C-7 was trapped by
5 and that is probably because of the steric hindrance of the scav-
enger and of the isopropyl group. The relative stereochemistry of
substituents on C-13 was defined by the epoxide ring opening and
allylic hydroperoxides to afford carbonylic compounds.25 Even if
this reaction, in general, occurs under acidic catalysis, several
studies describe the possible degradation of allylic hydroperoxides
in accordance with this mechanism by heating.26,27 Thus, the for-
mation of 10 in experiments carried out by heating or by photolysis
cannot be disregarded as it is obtained with significant yields. Fi-
nally, the tertiary alcohol 11 can be accounted for by hydrogen
abstraction of the corresponding alkoxyl radical ROꢀ.
2.2. ESR spin-trapping experiments with model compound 3
in the presence of Fe(II)/Fe(III)
Combined with ESR spectroscopy, the spin-trapping technique,
nowadays considered as the best method for the detection of short-
lived radicals, has been extensively used for the detection of free
radicals generated in biological milieu.28 This method is based on
the reaction of a spin trap with a radical to form a more stable
radical adduct that can be detected by ESR. Cyclic nitrones, espe-
cially pyrroline N-oxides, exhibit very interesting properties as spin
traps. Among them, 5-diethoxy-phosphoryl-5-methyl-1-pyrroline
N-oxide (DEPMPO) has gained wider acceptance because of the
very characteristic ESR spectra that are observed upon reaction
with free radicals. The hyperfine coupling constants (hfc) afford
valuable information on the radical trapped and today DEPMPO is
recognized as the most efficient spin trap for superoxide in vivo.
Indeed, the HOO–DEPMPOꢀ adduct results much more persistent
compared to the analogous HOO–DMPOꢀ adduct obtained with 5,5-
dimethyl-1-pyrroline N-oxide (DMPO).29 Moreover, the DEPMPO
allows an easy distinction between oxygen-centred adducts (per-
oxyl/alkoxyl) and carbon-centred adducts.30 Thus, to get more in-
formation on the reactivity of allergenic cyclic tertiary allylic
hydroperoxides with iron systems, experiments were conducted
using the stable model compound 3 and DEPMPO, and the results
compared with those obtained using DMPO.
not by the cyclization of the alkoxyl radical. Indeed, only the
oxiranylcarbinyl radical resulting from the cyclization of the alkoxyl
radical on the
-side of the molecule was trapped by HOꢀ, a less
a-
a
bulky group able to trap rapidly the carbon-centred radical at C-14
accounting for the formation of 9 (Scheme 3).
Compound 8 was detectable only when the parent peroxyl
radical was formed via photolysis. In particular, to account for the
carbonyl group at C-13 the decomposition of an intermediate
dioxetane as described for 6a/6b has to be invoked and the
resulting radical at C-14 trapped by 5. The formation of the C-7
a/C-
8
a
epoxide could be explained by photo-oxidation of the double
bond in the presence of singlet dioxygen (Scheme 3). The mecha-
nism of this reaction, even if not completely well understood, has
been described on the basis of the formation and evolution of in-
termediates of the perepoxide kind.23,24
A possible mechanism explaining the formation of
a,b-un-
saturated ketone 10 would involve the heterolytic fragmentation of
the hydroperoxide combined with a migration of the vinyl group,
which is known as the Hoch rearrangement, a facile reaction of
Figure 4 shows the ESR spectra obtained from the reaction of 3
with Fe(II) in the presence of DEPMPO or DMPO, in CH3CN or in
a 1:1 (v/v) H2O/CH3CN mixture. Using DEPMPO in a CH3CN solution

E. Gime´nez-Arnau et al. / Tetrahedron 64 (2008) 5680–5691
5685
P(O)(OEt)2
spectrometry and tandem mass spectrometry, further studies
characterized the previously assigned tert-butylperoxyl and
cumylperoxyl DMPO adducts, formed from the reaction of Fe(II)
with tert-butyl or cumene hydroperoxide, as the result of the
trapping of the tert-butyloxyl and cumyloxyl radicals, re-
spectively.33,34 In accordance with the described controversial as-
signment of peroxyl and alkoxyl radical adducts, with both spin
traps, it was necessary to consider the possibility of having the
concurrent presence of both types of oxygen-centred radicals. Ac-
tually, from the mechanistic point of view, a Fenton-like reaction
with Fe(II) will at first produce ROꢀ radicals, which easily can ab-
stract the hydrogen of the hydroperoxide group (ROO–H) to lead to
ROOꢀ radicals. These radicals are known to undergo bimolecular
self-reactions to form a tetroxide intermediate, ROOOOR, which
decomposes irreversibly to give alkoxyl ROꢀ radicals and dioxygen
(Scheme 4).31
Spin-trapping with
DEPMPO
Spin-trapping with
DMPO
N
O
N
O
A
25 G
B
When experiments with DEPMPO were conducted in H2O/
CH3CN (1:1 (v/v)) as solvent, the ESR spectrum (Fig. 4C) showed the
presence of an oxygen-centred adduct, besides a species charac-
terized by hfc (aN¼14.81 G, aH¼21.72 G and aP¼47.33 G; g¼2.0059)
typical of carbon-centred radical adducts.35 The formation of
a carbon-centred radical was also confirmed when DMPO was used,
resulting in the formation of a carbon-centred DMPO adduct
(aN¼15.51 G and aH¼22.26 G).36 As carbon-centred radicals were
not trapped when CH3CN was used as solvent, it might be expected
that oxygen-centred radicals are trapped before rearranging. On
the other hand, in H2O/CH3CN as solvent, the spin-trapping process
should result slower, thus allowing enough time for the rear-
rangement to take place, and then account for the formation of
carbon-centred radicals.
C
D
Allyloxyl radicals have been shown to undergo a rapid cycliza-
tion to give a-oxiranylcarbinyl radicals, as well as a b-scission to
Figure 4. ESR spectra obtained from the reaction of 3 with Fe(II) in the presence of
DEPMPO or DMPO, in CH3CN or in a 1:1 (v/v) H2O/CH3CN mixture. ESR settings: mi-
crowave power 1 mW, microwave frequency 9 GHz, modulation frequency 100 KHz,
modulation amplitude 1 G, scan width 100 G, conversion time 25 ms, receiver gain
1ꢁ105. (A) A solution in CH3CN of compound 3 (60 mM) and the spin trap (90 mM) was
reacted with FeSO4 heptahydrate (2 mg). (B) Computer simulation of spectra of panel
A. The best fit was obtained with coupling constants aN¼13.17 G, aH¼11.14 G and
aP¼45.24 G (g¼2.0064) for RO–DEPMPOꢀ or ROO–DEPMPOꢀ, and aN¼14.15 G,
aH¼10.60 G and aH¼1.49 G (g¼2.0063) for RO–DMPOꢀ or ROO–DMPOꢀ. (C) A 1:1 (v/v)
H2O/CH3CN solution of compound 3 (60 mM) and the spin trap (90 mM) was reacted
with FeSO4 heptahydrate (2 mg). (D) Computer simulation of spectra of panel C.
yield alkyl radicals and conjugated ketones or aldehydes. In the
case of tertiary allyloxyl radicals, as the one derived from 3, it has
been shown that the ratio between the amount of oxiranylcarbinyl
and alkyl radicals formed is 6 to 1, confirming that the formation of
the oxiranylcarbinyl radical is the fastest process.37 On another
hand, we evidenced in previous studies that other hydrogens,
mainly in allylic positions, can also undergo abstraction.11 Finally,
we have proved for 1 (Scheme 3), as previously for 3,11 that peroxyl
radicals can rearrange, via intramolecular addition to a double
bond, leading to a carbon-centred radical with a dioxetane struc-
ture that can cleave to get a carbonylic compound. Therefore, the
possible formation of several carbon-centred radicals in the re-
action medium must be considered (Scheme 4).
A
composite spectrum was obtained when using DEPMPO with coupling
constants aN¼13.81 G, aH¼11.92 G and aP¼47.79 G (g¼2.0064) for RO–DEPMPOꢀ or
ROO–DEPMPOꢀ (6), and aN¼14.81 G, aH¼21.72 G and aP¼47.33 G (g¼2.0059) for
a DEPMPO-carbon-centred adduct (O). For DMPO the best fit was obtained with
coupling constants aN¼15.51 G, aH¼22.26 G (g¼2.0059) for a DMPO-carbon-centred
adduct.
ESR spin-trapping experiments were also carried out with Fe(III)
derivatives, such as FeCl3, and due to the ubiquitous nature of haem
complexes also with Fe(III)-porphyrins TPP-Fe3þ and haemin.
When 3 was reacted with FeCl3, in the presence of DEPMPO in
CH3CN or in a 1:1 (v/v) H2O/CH3CN solvent, the ESR spectra showed
that an oxygen-centred radical was the main trapped species in
organic solvents, while in an organic/aqueous mixture also carbon-
centred radicals could be trapped (Fig. 5). As shown in Scheme 2,
peroxyl ROOꢀ and alkoxyl ROꢀ radicals could be obtained from the
controversial homolytic/heterolytic O–O bond cleavage by iron
porphyrin complexes. These oxygen-centred radicals can further
evolve to form carbon-centred radicals as indicated in Scheme 4.
Thus, ESR spin-trapping experiments confirmed the formation
of carbon-centred radicals and evidenced the formation of allyloxyl
and peroxyl radicals even if cannot be trapped by 5.
(Fig. 4A), the main trapped radical was characterized by hfc values
(aN¼13.17 G, aH¼11.14 G and aP¼45.24 G; g¼2.0064) that let to
hypothesize the presence of an oxygen-centred radical, in partic-
ular, the adduct of the superoxide to the DEPMPO.29,30 As ROO–
DEPMPOꢀ adducts are in general characterized due to their close
similarity to that of the HOO–DEPMPOꢀ adduct, the trapping of an
ROOꢀ radical could be suggested. However, similar values of hfc are
reported for adducts to DEPMPO of alkoxyl radicals, such as tert-
butyloxyl and radicals derived from peroxidized lipids, and then
the possibility of trapping of an ROꢀ radical cannot be excluded.30
This was confirmed in experiments conducted in the presence of
DMPO, which let to detect a radical species characterized by hfc
(aN¼14.15 G, aH¼10.60 G and aH¼1.49 G; g¼2.0063) that were in
good agreement with those reported for tertiary peroxyl radicals,
such as those derived from tert-butyl or cumene hydroperoxide,
trapped by DMPO.31 Later, the spin-trapping literature reported
that many DMPO peroxyl radical adducts have virtually the same
hfc as alkoxyl radical adducts, raising the issue of correct assign-
ment to peroxyl radical adducts.32 Based on 17O-labelling, but also
on liquid chromatography/ESR, electrospray-ionization mass
2.3. Addition of N-acetyl-cysteine ethyl ester to model
compound 3 catalyzed by ferric chloride
Analyses of amino acid residues from enzymes, and from other
proteins treated with lipid hydroperoxides, have demonstrated that

5686
E. Gime´nez-Arnau et al. / Tetrahedron 64 (2008) 5680–5691
OO
O
P(O)(OEt)2
P(O)(OEt)2
N
O
N
O
DEPMPO
DEPMPO
OH
O O
O
OO
Fe2+
Fe3+
O
DEPMPO
O
3
3
HO
P(O)(OEt)2
R-O
R-OO
carbon-centred
radical
N
O
dimerization
O2
O O O O
OH
(EtO)2(O)P
OH
OH
O
O
O
N
DEPMPO
OH
P(O)(OEt)2
N
O
O N
carbon-centred radicals
P(O)(OEt)2
Scheme 4. Potential nitroxide radicals that could be obtained by DEPMPO spin-trapping of oxygenated and/or carbon-centred radicals derived from compound 3.
cysteine is one of the most labile residues. Thus, we tried to trap
radical intermediates issued from allylic hydroperoxide 3 using
a cysteine derivative as trapping agent. Similar experiments had
been already conducted with fatty acids containing a hydroperox-
ide group and were suggested as possible models for the in-
teractions of allylic hydroperoxides with proteins.38 More precisely,
several products resulted from the reaction of linoleic acid hydro-
peroxide with cysteine catalyzed by iron ion (FeCl3), some of which
were the result of a thio-bond formation between the thiyl radical
derived from the lateral chain of cysteine and a carbon-centred
radical derived from the hydroperoxide.39
context and also to avoid the Schiff base that could be formed be-
tween the free amino group and carbonylic compounds produced
in the hydroperoxide decomposition. A catalytic amount of FeCl3
was employed to induce the radical formation. In fact, it has been
reported that certain reductants, such as cysteine, induce an in-
crease of the rate of hyroperoxide decomposition in the presence of
Fe(III), via reduction of Fe(III) to Fe(II).40 A redox reaction between
the reductant and the hydroperoxide can be thus catalyzed by
Fe(II)/Fe(III) ions. Reactions were carried out in argon atmosphere
to minimize the competing reaction with oxygen and assuring
higher yields of adducts. It is well known that the reaction of fatty
acids carrying a hydroperoxide group in their structure with cys-
teine/FeCl3 in the presence of oxygen results in the formation of
numerous oxygenated fatty acids, as well as cystine and oxides of
cysteine.38,40 This has also been described for the Fe(II)/Fe(III)-
mediated chemistry of artemisinin and synthetic 1,2,4-trioxanes
with glutathione.41
N-Acetyl-cysteine ethyl ester was selected as the reactant be-
cause of its resemblance with cysteine when located in a protein
Date: 11.12.1907 Time: 17:01
C:\...\sp1506~1.spc
A
Hydroperoxide 3 and N-acetyl-cysteine ethyl ester (2 equiv) in
a deaerated 1:1 (v/v) H2O/CH3CN mixture were treated with a cat-
alytic amount of FeCl3 (0.1 equiv), at room temperature and under
argon atmosphere. The reaction, followed by TLC, let to detect the
formation of new products after 90 min. However, the starting
hydroperoxide was still present in the reaction mixture after 4 days.
As the reaction did not evolve after that time, it was stopped and
the solvent removed under reduced pressure. The reaction mixture
was separated by column chromatography, but only one fraction
containing a non-separable mixture of two new products could be
isolated, whose structures were determined using one- and two-
dimensional NMR techniques, especially 1H, 13C, DEPT, COSY,
NOESY and 1H–13C HSQC and HMBC experiments (Scheme 5).
An adduct originated from 3 and N-acetyl-cysteine ethyl ester
was identified from the NMR data as 1-S-(N-acetyl-cysteine ethyl
ester)-2-(10-hydroxy-10-methyl-ethyl) cyclohex-2-ene 14. The
HSQC spectrum showed a proton at 6.31 ppm directly correlated to
a carbon atom at 123.4 ppm, indicating the existence of the vinylic
group (]CH), and also a proton at 4.04 ppm correlated to a carbon
atom at 41.2 ppm, chemical shifts characteristic of the –CH in po-
25 G
B
50
3275
3300
3325
3350
3375
3400
3
[G]
Figure 5. ESR spectra of DEPMPO radical adducts from the reaction of 3 with Fe(III).
ESR settings: microwave power 1 mW, microwave frequency 9 GHz, modulation fre-
quency 100 KHz, modulation amplitude 1 G, scan width 100 G, conversion time 25 ms,
receiver gain 1ꢁ105. (A) A solution in CH3CN of compound 3 (60 mM) and the spin trap
(90 mM) was reacted with FeCl3 (2 mg). In the computer simulation the best fit was
obtained with coupling constants aN¼13.17 G, aH¼11.14 G and aP¼45.24 G (g¼2.0064)
for RO–DEPMPOꢀ or ROO–DEPMPOꢀ. (B) A 1:1 (v/v) H2O/CH3CN solution of compound
3 (60 mM) and the spin trap (90 mM) was reacted with FeCl3 (2 mg). A composite
spectrum was obtained with coupling constants aN¼13.81 G, aH¼11.92 G and
sition
firmed the two quaternary carbons C-2 at 139.6 ppm and C-10 at
75.1 ppm (C–OH). All these positions were corroborated by 1H–13
a
to the cysteine thiyl group. The 13C NMR spectrum con-
aP¼47.79 G (g¼2.006) for RO–DEPMPOꢀ or ROO–DEPMPOꢀ
(
D
), and aN¼14.81 G,
aH¼21.72 G and aP¼47.33 G (g¼2.0059) for a DEPMPO-carbon-centred adduct (O).
C

E. Gime´nez-Arnau et al. / Tetrahedron 64 (2008) 5680–5691
5687
O
OOH
O
H
N
O
S
O
O
HN
H
H
O
H
O
SH
3
3
O
H
O
Fe2+
Fe3+
FeCl3 (cat.)
H2O/CH3CN 1:1 (v/v)
+ R-SH
H2O
S
O
H
H
O
H
1
H
S
S
N
H
N
OH
H
2'
O
5
H
6
O
1'
1
O
3
OH
3'
4
N
H
2
1'
2
H
H
O
O
H
H
H
O
2'
R-SH
O
3'
14
HO
15
H
N
S
Scheme 5. Ferric chloride catalyzed reaction of 3 with N-acetyl-cysteine ethyl ester.
The products identified are indicated together with the main NOE effects observed
allowing their structural elucidation.
O
O
H
O
HO
14
long-range correlations existing between H-3 at 6.31 ppm, C-1 at
41.2 ppm and C-5 at 20.8 ppm, and between the protons of the
methyl groups at 1.32 ppm, C-10 at 75.1 ppm and C-2 at 139.6 ppm.
But more decisive for the structural elucidation of 14 was that the
C-1 at 41.2 ppm was correlated via long-range coupling with pro-
tons at 2.92 and 3.07 ppm, both belonging to the methylene group
(33.7 ppm) of the lateral chain of cysteine. Also, protons at 2.92 and
Scheme 6. A possible mechanistic pathway for the formation of adduct 14.
trapped by dioxygen, could give a peroxyl radical that after di-
merization and loss of dioxygen led to a C-1 alkoxyl radical.
Opening of the epoxide and hydrogen abstraction afforded com-
pound 15.
3.07 ppm were long-range correlated with the a-carbon position of
the amino acid at 51.7 ppm. These results were supported by NOE
experiments as proton at 4.04 ppm (H-1) exhibited important NOE
cross-peaks with protons of the methylene group at 2.92 and
3.07 ppm (Scheme 5). Mechanistically, a redox reaction between N-
acetyl-cysteine ethyl ester and allylic hydroperoxide 3, catalyzed by
FeCl3, was conducted. Cysteine is known to reduce Fe(III) and hy-
droperoxides are known to oxidize Fe(II). Under reduction of Fe(III)
to Fe(II), thiyl radicals were generated from N-acetyl-cysteine ethyl
ester. The produced Fe(II) can then generate ROꢀ radicals through
a Fenton-like reaction. Easily, ROꢀ radicals could perform an allylic
hydrogen abstraction from position 1 as the abstraction of hydro-
gens from different allylic positions was evidenced previously by
3. Conclusion
To complement our previous studies on 15-hydroperoxyabietic
acid 1-like compounds 3 and 4, we now confirm, using radical
trapping experiments, the formation of carbon-centred radicals
derived from the allergenic bis-allylic tertiary hydroperoxide 1.
Even if the trapping of these radicals by 1,1,3,3-tetramethyli-
soindolin-2-yloxyl nitroxide 5 does not prove that they are able to
form adducts with skin proteins in vivo, their possible generation in
the epidermis could be responsible of the formation of antigenic
structures that are able to trigger the immunological mechanism of
allergic contact dermatitis. As Fe(II)/Fe(III) redox cycles are com-
mon in biological media, we initially wanted to test the reactivity of
15-hydroperoxyabietic acid 1 in the presence of Fe(II)/Fe(III), but its
instability let it difficult to prepare a pure sample of 1 for ESR
experiments. Thus, alternatively, we used the monocyclic allylic hy-
droperoxide 3 as a model. We confirmed the formation of carbon-
centred radicals in the presence of Fe(II)/Fe(III) redox systems and
also the formation of oxygen-centred allyloxyl and peroxyl radicals
even if not trapped by 5. Therefore, carbon and oxygen-centred
radicals must both be considered as possible haptens of allylic hy-
droperoxides in allergic contact dermatitis. Finally, and to the best
of our knowledge, we report here for the first time the formation of
an adduct between an allergenic allylic hydroperoxide and cysteine
catalyzed by iron. Identification of this adduct provided evidence
on the participation of oxygen-centred allyloxyl radicals, which
could abstract hydrogen from an allylic position and coupled with
N-acetyl-cysteine ethyl ester through a carbon–sulfur bond. These
results provided interesting data at the molecular level for this class
of allergenic compounds as well as useful insights into the forma-
tion of antigenic structures via radical mechanisms.
the ESR spin trapping of the corresponding allylic radicals.11
A
radical termination reaction between the carbon-centred radical
so-formed and the cysteine thiyl radical to afford adduct 14 was
unlikely because the concentrations of the transient species were
low and the chances of one radical encountering another were
small. A suggestion for a mechanistic pathway explaining the for-
mation of 14 could be based on a more probable concerted mech-
anism (Scheme 6). That would imply that a thiyl radical could
abstract an allylic hydrogen42 and concomitantly a second cysteinyl
thiyl radical, formed via an hydrogen abstraction by the alkoxyl
radical from a neutral thiol molecule (2 equiv used in the reaction),
was adding to the intermediate.
The NMR data of the fraction containing adduct 14 showed also
the presence of the non-separable 2-R-(10-hydroxy-10-methyl-
ethyl)-1-R-cyclohexanol 15. The quaternary carbon of the vinyl
group of 3 had been replaced by a –CH at 1.38/53.7 ppm bearing the
10-hydroxy-10-methyl-ethyl substituent. In parallel, the vinyl ]CH
carbon had been replaced by a –CH at 3.65/73.3 ppm, chemical
shifts that are characteristic of this carbon atom bearing an hydroxy
group as substituent. The stereochemistry of the substituents was
established on the basis of NOE cross-peaks existing between H-2
at 1.38 ppm, H-1 at 3.65 ppm, axial H-4 at 0.85 ppm and protons of
the methyl group at 1.19 ppm. Proton H-1 also displayed an NOE
cross-peak with axial H-4. From the mechanistic point of view, the
4. Experimental
Caution. Skin contact with allylic hydroperoxides must be
avoided. Since these compounds are skin sensitizing substances,
they must be handled with care.
easily formed alkoxyl radical underwent a rapid cyclization to an
a-
oxiranylcarbinyl radical. This C-1 carbon-centred radical species,

5688
E. Gime´nez-Arnau et al. / Tetrahedron 64 (2008) 5680–5691
4.1. Chemicals and reagents
ESP 380 spectrometer equipped with an X-band microwave bridge.
Standard spectrometer settings: microwave power 1 mW, micro-
wave frequency 9 GHz, modulation frequency 100 KHz, modulation
amplitude 1 G, scan width 100 G, conversion time 25 ms and re-
ceiver gain 1ꢁ105. Hyperfine splitting assignments were obtained
by means of computer simulation using the Bruker WINEPR Sim-
Fonia software (version 1.25; Rheinstetten, Germany). For the ESR
spin-trapping experiments in iron systems, special glassware
equipment was designed where the hydroperoxide and the spin
trap in solution were initially in a distinct compartment (round-
bottom container) to the one of the iron reagent (capillary tube).
Both compartments were interlinked in order to allow sub-
sequently mix of the reagents. An ESR spectrum of the reaction
mixture was registered directly by placing the capillary tube in the
spectrometer cavity.
15-Hydroperoxyabietic acid 1 and 1-(1-hydroperoxy-1-methyl-
ethyl) cyclohexene 3 were synthesized using procedures we had
reported previously.14,43 The free-radical trapping agent 1,1,3,3-
tetramethylisoindolin-2-yloxyl nitroxide 5 was prepared using
a method described in the literature from N-benzylphthalimide.13
meso-Tetraphenylporphyrin iron(III) chloride (TPP-Fe3þ) was pur-
chased from Fluka (St. Quentin Fallavier, France). Iron(III) chloride
(FeCl3), iron(II) sulfate heptahydrate (FeSO4$7H2O) and polyvinyl
alcohol were purchased from Aldrich (St. Quentin Fallavier, France).
Haemin and 5,5-dimethyl-1-pyrroline N-oxide (DMPO) were acquired
from Sigma (St. Quentin Fallavier, France). 5-Diethoxy-phosphoryl-
5-methyl-1-pyrroline N-oxide (DEPMPO) was synthesized in two
steps from 2-methyl-1-pyrroline according to the method de-
scribed by Stolze et al.30 N-Acetyl-cysteine ethyl ester was syn-
thesized from N-acetyl-cysteine, purchased from Sigma (St.
Quentin Fallavier, France), by a classical esterification method using
absolute ethanol and thionyl chloride.44 Spectrophotometric grade
acetonitrile (99.8%) was purchased from Carlo Erba (Val de Reuil,
France) and used as-delivered. Aqueous solutions were prepared
with deionized water. All other chemicals were purchased from
Aldrich (St. Quentin Fallavier, France) and used without further
purification. All reactions were conducted in flame-dried glass-
ware, under inert atmosphere of argon, and with previously dea-
erated solvents. Thin layer chromatography (TLC) was carried out
on 0.25 mm Merck silica gel plates (60 F254). After elution, the TLC
plates were visualized with UV light at 254 nm or sprayed with
a solution of phosphomolybdic acid in ethanol (5% w/v) or with
a solution of m-anisaldehyde (0.5 mL) and p-anisaldehyde (0.5 mL)
in glacial acetic acid (100 mL), ethanol (85 mL) and concentrated
sulfuric acid (5 mL) followed by heating. Compounds were purified
by column chromatography on silica gel (Merck 60, 230–400
mesh).
4.3. Radical trapping experiments with 15-
hydroperoxyabietic acid
4.3.1. Generation of radicals by reaction with TPP-Fe3þ
To a solution of hydroperoxide 1 (173 mg, 0.52 mmol) in a 1:1
(v/v) mixture of H2O/CH3CN (100 mL) were added 5 (197 mg,
1.03 mmol) and TPP-Fe3þ (400 mg, 0.57 mmol). The reaction mix-
ture was stirred at room temperature for 1 h (complete disap-
pearance of 1 as shown by TLC, hexane/AcOEt 7:3). The mixture was
filtered and the solid obtained was washed with CH3CN (20 mL). All
filtrates together were concentrated under reduced pressure and
taken up with the minimum amount of CH2Cl2 (40 mL). The solu-
tion was then dried over magnesium sulfate, filtered and concen-
trated under reduced pressure. The crude product was purified at
first by column chromatography on silica gel (hexane/AcOEt 7:3) in
order to eliminate the excess of TPP-Fe3þ. Further column chro-
matography on silica gel (eluent gradient from hexane/AcOEt 8:2 to
AcOEt) allowed the purification, as white solids, of compounds 6a
(16.9 mg, 36.2
m
mol, 7% yield), 7a (4.5 mg, 8.5
mmol, 2% yield) and
4.2. Instrumentation
7b (3.6 mg, 6.8
mmol, 2% yield).
1H and 13C NMR spectra were recorded on Bruker Avance 300
and Bruker Avance 500 spectrometers. Chemical shifts are reported
in parts per million (d) and are indirectly referenced to Me4Si via
4.3.2. Generation of radicals by thermal cleavage
Compound 1 (123 mg, 0.37 mmol) was dissolved in a 1:1 (v/v)
mixture of H2O/CH3CN (100 mL) together with nitroxide 5 (175 mg,
0.92 mmol). The solution was heated to reflux and the reaction
followed by TLC (hexane/AcOEt 7:3) until complete disappearance
of the hydroperoxide. After 22 h stirring under reflux, the solution
was allowed to cool down to room temperature and concentrated
under reduced pressure. The residue obtained was taken up with
the minimum amount of CH2Cl2 (40 mL). The solution was then
dried over magnesium sulfate, filtered and concentrated under
reduced pressure. The crude product was purified by silica gel
column chromatography (eluent gradient from hexane/AcOEt 8:2
the solvent signals (CDCl3, 7.26 ppm for 1H and 77.0 ppm for 13C;
benzene-d6, 7.15 ppm for 1H and 128.6 ppm for 13C). Multiplicities
are indicated by s (singlet), d (doublet), t (triplet) and m (multiplet).
1H coupling constants (J) were obtained by first-order analysis and
are reported in hertz. The different types of carbon in the structures
have been identified by the DEPT-135 technique. The two-di-
mensional phase-sensitive 1H–1H COSY (correlation spectroscopy)
spectra were collected using a standard pulse sequence with
254 ms acquisition time, sweep width 12.25 ppm, 512 complete t1
data points and 8 scans per increment. The two-dimensional
phase-sensitive 1H–1H NOESY (nuclear Overhauser effect spec-
troscopy) spectra were collected using 800 ms mixing time and
254 ms acquisition time, sweep width of 12.25 ppm, 512 complete
t1 data points, and 16 scans per increment. The two-dimensional
1H–13C HSQC (heteronuclear single-quantum correlation) spectra
to AcOEt) to give, as white solids, compounds 6a (5.9 mg,12.6
4% yield), 6b (5.5 mg, 10.4 mol, 3% yield), 7a (2 mg, 3.8 mol, 1%
yield), 7b (9.6 mg, 20.6 mol, 6% yield), 9 (5.6 mg, 16.7 mol, 5%
mol, 10% yield) and 11 (3.4 mg, 10.6 mol,
mmol,
m
m
m
m
yield), 10 (10 mg, 36.1
m
m
3% yield).
and 1H–13
C
HMBC (heteronuclear multiple-bond correlation)
4.3.3. Generation of radicals by photochemical cleavage
spectra were collected using a standard pulse sequence, with
254 ms acquisition time, 512 complete t1 data points, 16 (HSQC) or
32 (HMBC) scans per increment and a spectral width in the f1 di-
mension of 209.2 ppm. The data were processed using the NMR
notebook software (version 2.5; NMRtec S.A.S., Illkirch-Graffen-
staden, France). Fast atom bombardment mass spectrometry was
carried out on an Autospect spectrometer (Micromass Inc., United
Kingdom), in a 3-nitrobenzyl alcohol matrix and using caesium
cations for the bombardment in positive and negative ion modes.
Electron spin resonance (ESR) spectra were recorded on a Bruker
Compound 1 (132 mg, 0.39 mmol) was dissolved in a 1:1 (v/v)
mixture of H2O/CH3CN (100 mL). Nitroxide 5 (188 mg, 0.98 mmol)
was added to the solution and the stirred reaction mixture was
irradiated for 20 h with a 150 W tungsten filament lamp at a dis-
tance of 30 cm. The evolution of the reaction was followed by TLC
(hexane/AcOEt 7:3) until complete disappearance of the hydro-
peroxide. The mixture was concentrated under reduced pressure
and the residue obtained was taken up with the minimum amount
of CH2Cl2 (40 mL). The solution was then dried over magnesium
sulfate, filtered and concentrated under reduced pressure. The

E. Gime´nez-Arnau et al. / Tetrahedron 64 (2008) 5680–5691
5689
crude product was purified by silica gel column chromatography
(eluent gradient from hexane/AcOEt 8:2 to AcOEt) to give, as white
12), 1.57 (m, 1H, H-3ax), 1.60 (m, 1H, H-1ax), 1.61 (s, 3H, H-110), 1.62
(m, 1H, H-6eq), 1.62 (s, 3H, H-100), 1.73 (m, 1H, H-11eq), 1.83 (m, 1H,
H-11ax), 1.90 (m, 1H, H-6ax), 1.93 (m, 1H, H-3eq), 2.33 (d, J¼7.8 Hz,
1H, H-9), 2.65 (dd, J¼13.0 and 2.6 Hz, 1H, H-5), 4.27 (m, 1H, H-7),
6.17 (d, J¼1.4 Hz, 1H, H-14), 6.91 (m, 1H, H-70), 7.02 (m, 1H, H-40),
solids, compounds 6a (5.9 mg, 12.6
m
m
m
mol, 4% yield), 6b (5.5 mg,
mol, 1% yield), 7b (9.6 mg,
mol, 3% yield), 9 (5.6 mg,
mol, 5% yield), 10 (10 mg, 36.1 mol, 10% yield) and 11
mol, 3% yield).
10.4
20.6
16.7
m
m
m
mol, 3% yield), 7a (2 mg, 3.8
mol, 6% yield), 8 (4.4 mg, 9.1
m
7.11 (m, 2H, H-50 and H-60); 13C NMR (benzene-d6, 125 MHz):
d 15.7
(3.4 mg, 10.6
m
(C-20), 16.6 (C-19), 17.7 (C-11), 18.4 (C-2), 24.7 (C-17), 24.9 (C-16),
25.4 (C-90), 25.6 (C-110), 29.4 (C-12), 30.3 (C-80), 30.7 (C-100), 31.7
(C-6), 36.9 (C-3), 38.6 (C-1), 39.9 (C-10), 43.5 (C-5), 45.5 (C-9), 47.4
(C-4), 67.7 (C-10), 68.2 (C-30), 73.8 (C-13), 74.2 (C-15), 85.6 (C-7),
121.7 (C-40),122.0 (C-70),127.5 (C-50 and C-60), 128.3 (C-14),141.8 (C-
8), 145.0 (C-30a and C-70a), 184.4 (C-18); FABMS [MþH]: 526.4,
FABMS [MꢂH]: 524.3.
4.3.4. Radical trapping experiment in the absence of 5
Allylic hydroperoxide 1 (96 mg, 0.29 mmol) was dissolved in
a 1:1 (v/v) mixture of H2O/CH3CN (100 mL). The solution was
heated to reflux. The reaction was followed by TLC (hexane/AcOEt
7:3) until complete disappearance of the hydroperoxide. After 5 h
stirring under reflux, the solution was allowed to cool down to
room temperature and concentrated under reduced pressure. The
residue obtained was taken up with the minimum amount of
CH2Cl2 (30 mL). The solution was then dried over magnesium sul-
fate, filtered and concentrated under reduced pressure. The crude
product was purified by silica gel column chromatography (eluent
gradient from hexane/AcOEt 6:4 to AcOEt) to give, as white solids,
4.4.4. 7
13 ,15-dihydroxy-abietan-8(14)-enoic acid (7b)
1H NMR (benzene-d6, 400 MHz):
0.69 (s, 3H, H-20), 1.01 (m,
a
-(10,10,30,30-Tetramethyl-10,30-dihydroisoindol-20-yloxyl)-
b
d
1H, H-1eq), 1.19 (s, 3H, H-16), 1.21 (s, 3H, H-17), 1.23 (s, 3H, H-19),
1.33 (m, 2H, H-2), 1.40 (m, 1H, H-12eq), 1.46 (s, 3H, H-110), 1.50 (m,
1H, H-1ax), 1.50 (s, 3H, H-80), 1.55 (m, 1H, H-11ax), 1.57 (m, 1H, H-
3
ax), 1.59 (m, 1H, H-6eq), 1.60 (s, 3H, H-90), 1.66 (m, 1H, H-12ax), 1.67
compounds 9 (3.1 mg, 9.2
mmol, 3% yield), 10 (14 mg, 50.6
mmol,17%
yield) and 11 (4.3 mg, 13.5
m
mol, 5% yield).
(m, 1H, H-11eq), 1.67 (s, 3H, H-100), 1.92 (m, 1H, H-6ax), 1.97 (m, 1H,
H-3eq), 2.33 (t, J¼8.1 Hz,1H, H-9), 2.61 (dd, J¼13.1 and 2.3 Hz,1H, H-
5), 4.20 (m, 1H, H-7), 6.12 (s, 1H, H-14), 6.92 (m, 1H, H-70), 7.04 (m,
1H, H-40), 7.10 (m, 2H, H-50 and H-60); 13C NMR (benzene-d6,
4.4. Characterization of radical trapping adducts derived
from 15-hydroperoxyabietic acid
100 MHz): d 14.2 (C-20), 16.8 (C-19), 17.5 (C-11), 18.2 (C-2), 24.9 (C-
4.4.1. 7
oxo-deisopropyl-abietan-8(14)-enoic acid (6a)
1H NMR (benzene-d6, 500 MHz):
0.50 (s, 3H, H-17), 0.81 (m,
a
-(1010,30,30-Tetramethyl-10,30-dihydroisoindol-20-yloxyl)-13-
17), 25.2 (C-16), 25.8 (C-90 and C-110), 29.2 (C-12), 30.6 (C-80), 31.0
(C-6), 31.5 (C-100), 36.9 (C-3), 38.0 (C-1), 38.2 (C-10), 43.4 (C-5), 47.1
(C-9), 47.2 (C-4), 67.9 (C-10), 68.3 (C-30), 73.7 (C-15), 74.1 (C-13), 84.5
(C-7), 121.8 (C-40), 121.9 (C-70), 127.7 (C-50 and C-60), 129.9 (C-14),
142.0 (C-8), 145.7 (C-30a and C-70a), 184.5 (C-18); FABMS [MþH]:
526.4.
d
1H, H-1eq), 1.14 (s, 3H, H-16), 1.23 (m, 2H, H-2), 1.30 (m, 1H, H-1ax),
1.31 (s, 3H, H-80), 1.34 (s, 3H, H-90), 1.46 (m, 2H, H-6eq and H-11eq),
1.52 (m, 1H, H-3ax), 1.54 (s, 3H, H-100), 1.56 (s, 3H, H-110), 1.58 (m,
1H, H-11ax), 1.79 (m, 1H, H-6ax), 1.83 (m, 1H, H-3eq), 2.16 (m, 1H, H-
12eq), 2.33 (m, 1H, H-9), 2.35 (m, 1H, H-12ax), 2.50 (m, 1H, H-5), 4.10
(m, 1H, H-7), 6.20 (d, J¼1.9 Hz, 1H, H-14), 6.85 (dd, J¼7.5 and 1.0 Hz,
1H, H-70), 7.03 (dd, J¼7.6 and 1.0 Hz, 1H, H-40), 7.07 (m, 1H, H-60),
4.4.5. 14
,8 -epoxy-13-oxo-deisopropyl-abietic acid (8)
1H NMR (benzene-d6, 500 MHz):
0.63 (m, 1H, H-1eq), 0.73 (s,
a
-(10,10,30,30-Tetramethyl-10,30-dihydroisoindol-20-yloxyl)-
7a a
d
7.12 (m, 1H, H-50); 13C NMR (benzene-d6, 125 MHz):
d
15.2 (C-17),
3H, H-17), 1.18 (s, 3H, H-16), 1.22 (m, 2H, H-2), 1.34 (m, 1H, H-1ax),
1.35 (m, 1H, H-11eq), 1.37 (s, 3H, H-100), 1.39 (s, 3H, H-80), 1.47 (m,
1H, H-9), 1.50 (s, 3H, H-90), 1.53 (m, 1H, H-3ax), 1.55 (s, 3H, H-110),
1.61 (m, 1H, H-6eq), 1.74 (m, 1H, H-3eq), 1.88 (m, 1H, H-6ax), 2.00 (m,
1H, H-11ax), 2.10 (m, 1H, H-12ax), 2.13 (m, 1H, H-5), 2.55 (m, 1H, H-
12eq), 3.31 (m, 1H, H-7), 3.96 (m, 1H, H-14), 6.89 (m, 2H, H-40 and H-
70), 7.09 (m, 2H, H-50 and H-60); 13C NMR (benzene-d6, 125 MHz):
16.8 (C-16), 18.2 (C-2), 20.3 (C-11), 25.6 (C-90), 25.8 (C-110), 30.2 (C-
80), 30.6 (C-6), 30.9 (C-100), 36.5 (C-3), 36.8 (C-12), 38.4 (C-1), 39.1
(C-10), 43.2 (C-5), 47.2 (C-4), 48.1 (C-9), 67.4 (C-10), 68.3 (C-30), 83.9
(C-7), 121.6 (C-40), 122.0 (C-70), 127.6 (C-60), 128.2 (C-50), 130.1 (C-
14), 145.0 (C-30a and C-70a), 158.4 (C-8), 184.3 (C-15), 197.8 (C-13);
FABMS [MþH]: 466.4.
d
15.7 (C-17), 17.6 (C-16), 17.8 (C-11), 18.1 (C-2), 24.6 (C-6), 25.9 (C-
4.4.2. 7
13-oxo-deisopropyl-abietan-8(14)-enoic acid (6b)
1H NMR (benzene-d6, 500 MHz):
0.54 (s, 3H, H-17), 0.64 (m,
b
-(10,10,30,30-Tetramethyl-10,30-dihydroisoindol-20-yloxyl)-
90), 25.7 (C-110), 29.8 (C-80), 30.6 (C-100), 34.1 (C-10), 37.1 (C-3), 37.8
(C-12), 38.4 (C-1), 40.4 (C-5), 46.3 (C-4), 51.1 (C-9), 56.9 (C-7), 58.6
(C-8), 67.6 (C-10), 68.7 (C-30), 91.5 (C-14), 121.8 (C-40 and C-70), 127.5
(C-50 and C-60), 144.6 (C-70a), 144.9 (C-30a), 183.3 (C-15), 205.0 (C-
13); FABMS [MþH]: 482.2, FABMS [MꢂH]: 480.2.
d
1H, H-1eq), 1.17 (s, 3H, H-16), 1.22 (m, 2H, H-2), 1.24 (m, 1H, H-1ax),
1.36 (s, 3H, H-90), 1.41 (m, 1H, H-9),1.42 (m, 2H, H-11),1.45 (s, 3H, H-
110), 1.50 (m, 1H, H-3eq), 1.54 (m, 1H, H-6ax), 1.56 (s, 6H, H-80 and H-
100), 1.64 (m, 1H, H-3ax), 1.86 (m, 1H, H-5), 2.05 (m, 1H, H-12eq), 2.33
(m,1H, H-12ax), 2.54 (m,1H, H-6eq), 4.39 (m,1H, H-7), 6.89 (s,1H, H-
14), 6.90 (m, 1H, H-70), 7.02 (m, 1H, H-40), 7.09 (m, 1H, H-60), 7.18 (m,
4.4.6. 13
a
,15
a
-Epoxy-14
a
-hydroxy-abietic acid (9)
1H NMR (benzene-d6, 500 MHz):
d
0.73 (s, 3H, H-20), 0.85 (m,
1H, H-1eq), 1.09 (s, 3H, H-16), 1.24 (s, 3H, H-19), 1.25 (m, 1H, H-11ax),
1.26 (s, 3H, H-17), 1.34 (m, 2H, H-2), 1.46 (m, 1H, H-12ax), 1.55 (m,
1H, H-1ax), 1.60 (m, 1H, H-3ax), 1.67 (m, 1H, H-11eq), 1.74 (m, 1H, H-
9), 1.82 (m, 1H, H-12eq), 1.83 (m, 1H, H-3eq), 1.96 (m, 2H, H-6), 2.22
(m, 1H, H-5), 3.97 (m, 1H, H-14), 5.79 (m, 1H, H-7); 13C NMR
1H, H-50); 13C NMR (benzene-d6, 125 MHz):
d 15.6 (C-17), 17.1 (C-
16),18.2 (C-2), 20.3 (C-11), 25.6 (C-90), 25.8 (C-110), 30.0 (C-80 and C-
100), 32.4 (C-6), 36.5 (C-12), 36.8 (C-3), 38.0 (C-1), 39.0 (C-10), 45.7
(C-5), 47.4 (C-4), 49.3 (C-9), 67.3 (C-10), 68.2 (C-30), 82.5 (C-7), 122.0
(C-40), 122.2 (C-70), 124.1 (C-14), 128.0 (C-60), 128.1 (C-50), 145.0 (C-
30a and C-70a), 160.7 (C-8), 184.2 (C-15), 197.2 (C-13); FABMS
[MþH]: 466.3, FABMS [MꢂH]: 464.3.
(benzene-d6, 125 MHz):
d 14.2 (C-20), 16.9 (C-19), 18.3 (C-2), 19.4
(C-11), 20.3 (C-17), 22.2 (C-16), 25.7 (C-6), 26.7 (C-12), 35.1 (C-10),
37.5 (C-3), 38.2 (C-1), 45.2 (C-5), 46.5 (C-4), 50.5 (C-9), 63.1 (C-15),
66.5 (C-13), 70.8 (C-14), 123.8 (C-7), 137.5 (C-8), 183.5 (C-18);
FABMS [Mꢂ(OH)]: 317.2.
4.4.3. 7
13 ,15-dihydroxy-abietan-8(14)-enoic acid (7a)
1H NMR (benzene-d6, 500 MHz):
0.69 (s, 3H, H-20), 1.00 (m,
a
-(10,10,30,30-Tetramethyl-10,30-dihydroisoindol-20-yloxyl)-
a
d
4.4.7. 13-Oxo-deisopropyl-abietan-8(14)-enoic acid (10)
1H, H-1eq), 1.16 (s, 3H, H-16), 1.20 (s, 3H, H-19), 1.24 (s, 3H, H-17),
1.35 (m, 2H, H-2), 1.42 (s, 3H, H-90), 1.45 (s, 3H, H-80), 1.53 (m, 2H, H-
1H NMR (CDCl3, 200 MHz):
d
0.84 (s, 3H, H-17),1.23 (s, 3H, H-16),
1.41–2.53 (m, 16H, H-1, H-2, H-3, H-5, H-6, H-7, H-9, H-11 and H-

5690
E. Gime´nez-Arnau et al. / Tetrahedron 64 (2008) 5680–5691
12), 5.88 (br s, 1H, –COOH); 1C NMR (CDCl3, 50 MHz):
d
15.6 (C-17),
25.7 (C-6), 27.2 (C-4), 29.6 (C-20 and C-30), 33.7 (–S–CH2–), 41.2 (C-
1), 51.7 (–NH–CH–), 61.6 (–CH2–CH3), 75.1 (C-10), 123.4 (C-3), 139.6
(C-2), 170.0 (–NH–CO–), 170.9 (–CO–O–).
16.8 (C-16), 17.9 (C-2), 20.4 (C-11), 24.1 (C-6), 34.5 (C-10), 35.2 (C-3),
36.7 (C-1), 36.9 (C-7), 38.3 (C-12), 47.1 (C-4), 47.9 (C-9), 51.7 (C-5),
126.3 (C-14), 165.0 (C-8), 184.3 (C-15), 199.9 (C-13); CAS registry
number [5708-85-0].
4.6.2. 2-R-(10-Hydroxy-10-methyl-ethyl)-1-R-cyclohexanol (15)
1H NMR (CDCl3, 500 MHz):
d
0.85 (m, 1H, H-4ax), 1.12 (m, 1H, H-
4.4.8. 15-Hydroxyabietic acid (11)
5
eq), 1.17 (s, 3H, H-20), 1.19 (s, 3H, H-30), 1.24 (m, 1H, H-3ax), 1.30 (m,
1H NMR (CDCl3, 200 MHz):
d
0.83 (s, 3H, H-20), 1.18–2.36 (m,
1H, H-6eq), 1.38 (m, 1H, H-2), 1.63 (m, 1H, H-5ax), 1.67 (m, 1H, H-4eq),
1.70 (m, 1H, H-3eq), 1.95 (m, 1H, H-6ax), 3.65 (m, 1H, H-1); 13C NMR
14H, H-1, H-2, H-3, H-5, H-6, H-9, H-11 and H-12), 1.26 (s, 3H, H-19),
1.33 (s, 3H, H-16 or H-17), 1.35 (s, 3H, H-16 or H-17), 5.48 (m, 1H, H-
(CDCl3, 100 MHz):
d
23.7 (C-20), 24.8 (C-3), 25.9 (C-5), 27.6 (C-4),
7), 6.07 (s, 1H, H-14); 1C NMR (CDCl3, 50 MHz):
d
14.0 (C-20), 16.7
29.97 (C-30), 35.8 (C-6), 53.7 (C-2), 73.3 (C-1), 74.2 (C-10).
(C-19), 18.1 (C-2), 22.6 (C-11), 25.7 (C-12), 25.8 (C-6), 28.6 (C-16),
28.7 (C-17), 34.4 (C-10), 37.2 (C-3), 38.3 (C-1), 44.8 (C-9), 46.3 (C-4),
50.7 (C-5), 72.9 (C-15), 122.5 (C-14), 122.9 (C-7), 135.0 (C-8), 144.5
(C-13), 184.3 (C-18); CAS registry number [101821-23-2].
Acknowledgements
`
We thank the Ministere de l’Education Nationale, France, for
financial support through a fellowship to L.H. We also gratefully
acknowledge Bernard Meurer for the excellent technical assistance
in the ESR studies.
4.5. ESR spin-trapping experiments with model
compound 3 in the presence of Fe(II)/Fe(III)
Compound 3 (5 mg, 0.032 mmol) and the spin trap, DMPO or
References and notes
DEPMPO (0.05 mmol), were dissolved in 500
mL of deaerated
CH3CN or in 500 L of a deaerated 1:1 (v/v) H2O/CH3CN mixture.
m
´
1. Lepoittevin, J.-P. Contact Dermatitis; Frosch, P. J., Menne, T., Lepoittevin, J.-P.,
Eds.; Springer: Berlin, Heidelberg, New York, NY, 2006; pp 45–68.
The solution was placed in the round-bottom compartment of
a glassware equipment specially designed for the experiments,
whilst in another compartment, a capillary tube, the Fe(II) or Fe(III)
derivatives (0.035 mmol). The compartments were interlinked to
allow the subsequent mix of the reagents. In order to avoid a fast
mix of the hydroperoxide when poured over the iron derivative,
a layer of polyvinyl alcohol was used as a filter. The whole system
was degassed by a vacuum line using the standard freeze–thaw
technique, procedure repeated three times. Then, the solution
containing 3 and the spin trap was poured into the capillary tube
containing the layer of polyvinyl alcohol and the iron reagent
underneath. An ESR spectrum of the reaction mixture was regis-
tered directly by placing the capillary tube in the spectrometer
cavity.
2. Rustemeyer, T.; van Hoogstraten, I. M. W.; von Blomberg, B. M. E.; Scheper, R. J.
´
Contact Dermatitis; Frosch, P. J., Menne, T., Lepoittevin, J.-P., Eds.; Springer:
Berlin, Heidelberg, New York, NY, 2006; pp 11–43.
3. Smith, C. K.; Hotchkiss, S. A. M. Allergic Contact Dermatitis: Chemical and Met-
abolic Mechanisms; Smith, C. K., Hotchkiss, S. A. M., Eds.; Taylor and Francis:
London, 2001; pp 87–117.
4. Karlberg, A.-T.; Basketter, D.; Goossens, A.; Lepoittevin, J.-P. Contact Derm. 1999,
40, 183.
5. Matura, M.; Sko¨ld, M.; Bo¨rje, A.; Andersen, K. E.; Bruze, M.; Frosch, P. J.;
Goossens, A.; Johansen, J. D.; Svedman, C.; White, I. R.; Karlberg, A.-T. Contact
Derm. 2005, 52, 320.
6. Sko¨ld, M.; Bo¨rje, A.; Harambasic, E.; Karlberg, A.-T. Chem. Res. Toxicol. 2004, 17,
1697.
´
7. Matura, M.; Goossens, A.; Bordalo, O.; Garcıa-Bravo, B.; Magnusson, K.;
Wrangsjo, K.; Karlberg, A.-T. Contact Derm. 2003, 49, 15.
8. Karlberg, A.-T.; Bohlinder, K.; Boman, A.; Hacksell, U.; Hermansson, J.; Jacobson,
S.; Nilsson, J. L. J. Pharm. Pharmacol. 1988, 40, 42.
9. Joye, N. M.; Lawrence, R. V. J. Chem. Eng. Data 1967, 12, 279.
´
10. Mutterer, V.; Gimenez-Arnau, E.; Karlberg, A.-T.; Lepoittevin, J.-P. Chem. Res.
Toxicol. 2000, 13, 1028.
4.6. Addition of N-acetyl-cysteine ethyl ester to the model
compound 3 catalyzed by ferric chloride
´
11. Gimenez-Arnau, E.; Haberkorn, L.; Grossi, L.; Lepoittevin, J.-P. Tetrahedron 2002,
58, 5535.
12. Dussault, P. Active Oxygen in Chemistry; Foote, C. S., Valentine, J. S., Greenberg, A.,
Liebman, J. F., Eds.; Blackie Academic and Professional: London, Glasgow, New
York, NY, 1995; pp 141–203.
13. Griffiths, P. G.; Moad, G.; Rizzardo, E.; Solomon, D. H. Aust. J. Chem. 1983, 36,
397.
Allylic hydroperoxide 3 (300 mg, 1.92 mmol) was dissolved in
a deaerated 1:1 (v/v) mixture of H2O/CH3CN (100 mL). N-Acetyl-
cysteine ethyl ester (734 mg, 3.84 mmol) and FeCl3 (31 mg,
0.20 mmol) were added to the solution. The reaction mixture
continuously stirred for 90 min, at room temperature, let to iden-
tify, by TLC (pentane/AcOEt 4:6), new products. As the starting
hydroperoxide was still present in the reaction mixture after 4 days,
and the reaction did not evolve after that time, the solvent was
removed under reduced pressure. The crude product obtained was
fractionated by column chromatography on silica gel (pentane/
AcOEt 4:6) into three fractions, which were analyzed by NMR
(1H, 13C, DEPT, COSY, NOESY, HSQC and HMBC). Only the fra-
ction obtained at an average TLC retention factor of 0.45 (pentane/
AcOEt 4:6) was shown to contain the non-separable compounds 14
and 15.
´
14. Haberkorn, L.; Mutterer, V.; Gimenez-Arnau, E.; Lepoittevin, J.-P. Synlett 2001,
1723.
15. Traylor, T. G.; Kim, C.; Fann, W.-P.; Perrin, C. L. Tetrahedron 1998, 54, 7977.
16. Almarsson, O.; Bruice, T. C. J. Am. Chem. Soc. 1995, 117, 4533.
17. Labeque, R.; Marnett, L. J. J. Am. Chem. Soc. 1989, 111, 6621.
18. Traylor, T. G.; Tsuchiya, S.; Byun, Y.-S.; Kim, C. J. Am. Chem. Soc. 1993, 115,
2775.
19. Frimer, A. A. Chem. Rev. 1979, 79, 359.
20. Walling, C. Active Oxygen in Chemistry; Foote, C. S., Valentine, J. S., Greenberg, A.,
Liebman, J. F., Eds.; Blackie Academic and Professional: London, Glasgow,
New York, NY, 1995; pp 24–65.
21. Roe, A. N.; McPhail, A. T.; Porter, N. A. J. Am. Chem. Soc. 1983, 105, 1199.
22. Porter, N. A.; Zuraw, P. J. J. Org. Chem. 1984, 49, 1345.
23. Shimizu, N.; Bartlett, P. D. J. Am. Chem. Soc. 1976, 98, 4193.
24. Jefford, C. W.; Boschung, A. F. Helv. Chim. Acta 1977, 60, 2673.
25. Hock, H.; Kropf, H. Angew. Chem. 1957, 69, 313.
26. Turner, J. A.; Hertz, W. J. Org. Chem. 1977, 42, 1657.
27. Frimer, A. A.; Rot, D.; Sprecher, M. Tetrahedron Lett. 1977, 1927.
28. Janzen, E. G. Acc. Chem. Res. 1971, 4, 31.
29. Frejaville, C.; Karoui, H.; Tuccio, B.; Le Moigne, F.; Culcasi, M.; Pietri, S.; Laur-
icella, R.; Tordo, P. J. Med. Chem. 1995, 38, 258.
30. Stolze, K.; Udilova, N.; Nohl, H. Free Radical Biol. Med. 2000, 29, 1005.
31. Chamulitrat, W.; Takahashi, N.; Mason, R. P. J. Biol. Chem. 1989, 264,
7889.
32. Hanna, P. M.; Chamulitrat, W.; Mason, R. P. Arch. Biochem. Biophys. 1992, 296,
640.
4.6.1. 1-S-(N-Acetyl-cysteine ethyl ester)-2-(10-hydroxy-10-methyl-
ethyl) cyclohex-2-ene (14)
1H NMR (CDCl3, 500 MHz):
d
1.12 (m, 1H, H-5eq), 1.13 (m, 1H, H-
6
ax), 1.26 (m, 3H, –CH2–CH3), 1.32 (m, 6H, H-20 and H-30), 1.59 (m,
1H, H-6eq), 1.70 (m, 1H, H-4ax), 1.85 (m, 1H, H-5ax), 1.97 (m, 1H,
H-4eq), 2.01 (s, 3H, –CO–CH3), 2.92 (m, 1H, –S–CH2–), 3.07 (m,
1H, –S–CH2–), 4.04 (m, 1H, H-1), 4.18 (m, 2H, –CH2–CH3), 4.75 (m,
1H, –NH–CH–), 6.31 (m, 1H, H-3), 6.40 (s, 1H, –NH–); 13C NMR
33. Dikalov, S. I.; Mason, R. P. Free Radical Biol. Med. 1999, 27, 864.
34. Guo, Q.; Qian, S. Y.; Mason, R. P. J. Am. Soc. Mass Spectrom. 2003, 14, 862.
35. Khramtsov, V.; Berliner, L. J.; Clanton, T. L. Magn. Reson. Med. 1999, 42, 228.
(CDCl3, 100 MHz):
d 14.11 (–CH2–CH3), 20.8 (C-5), 22.9 (–CO–CH3),

E. Gime´nez-Arnau et al. / Tetrahedron 64 (2008) 5680–5691
5691
36. Sinha, B. K. J. Biol. Chem. 1983, 258, 796.
37. Grossi, L.; Strazzari, S.; Gilbert, B. C.; Whitwood, A. C. J. Org. Chem. 1998, 63,
40. Gardner, H. W.; Jursinic, P. A. Biochim. Biophys. Acta 1981, 665, 100.
41. Singh, C.; Gupta, N.; Tiwari, P. Tetrahedron Lett. 2005, 46, 4551.
8366.
42. Lunazzi, L.; Placucci, G.; Grossi, L. J. Chem. Soc., Chem. Commun. 1979,
533.
43. Lepoittevin, J.-P.; Karlberg, A.-T. Chem. Res. Toxicol. 1994, 7, 130.
44. Atlas, D.; Melamed, E.; Ofen, D. U.S. Patent 5,874,468, 1999.
38. Gardner, H. W.; Plattner, R. D.; Weisleder, D. Biochim. Biophys. Acta 1985,
834, 65.
39. Gardner, H. W.; Weisleder, D.; Kleiman, R. Lipids 1976, 11, 127.