Identification of carbon-centred radicals derived from linalyl hydroperoxide, a strong skin sensitizer: A possible route for protein modifications

Article abstract of DOI:10.1016/j.bmc.2005.04.004

Few studies are reported on the formation of reactive carbon-centred radical species from toxic xenobiotics. In this paper the formation of carbon radicals derived from the skin sensitizer linalyl hydroperoxide is described using radical trapping and EPR studies. Radical trapping used TMIO as scavenger agent and light, heat or TPP-Fe³⁺ as radical inducers. EPR spin trapping was based on the use of the parent alcohol, generating the same allyloxyl radical than the hydroperoxide by photolysis of the corresponding nitrite formed with t-BuONO, also playing the role of the spin trap. It is suggested that the generation of these carbon radical species could play an important role for the binding of the hydroperoxide with skin proteins to form antigenic structures, the first step of the skin sensitization mechanism.

Full text of DOI:10.1016/j.bmc.2005.04.004

Bioorganic & Medicinal Chemistry 13 (2005) 3977–3986
Identification of carbon-centred radicals derived from
linalyl hydroperoxide, a strong skin sensitizer: a possible route
for protein modifications
a
Michael Bezard, Elena Gim e´ nez-Arnau, Bernard Meurer, Loris Grossi,
a
b
c
a,
*
and Jean-Pierre Lepoittevin
a
Laboratoire de Dermatochimie, Universit e´ Louis Pasteur, CNRS-UMR 7123, Clinique Dermatologique CHU,
Place de l’H oˆ pital, 67091 Strasbourg, France
Institut Charles Sadron, CNRS-UPR 22, 6 rue Boussingault, 67083 Strasbourg, France
1
b
c
Dipartimento di Chimica Organica ‘A. Mangini’, Universit a` di Bologna, Viale Risorgimento 4, 40136 Bologna, Italy
Received 21 January 2005; revised 30 March 2005; accepted 5 April 2005
Available online 29 April 2005
Abstract—Few studies are reported on the formation of reactive carbon-centred radical species from toxic xenobiotics. In this paper
the formation of carbon radicals derived from the skin sensitizer linalyl hydroperoxide is described using radical trapping and EPR
3
+
studies. Radical trapping used TMIO as scavenger agent and light, heat or TPP–Fe as radical inducers. EPR spin trapping was
based on the use of the parent alcohol, generating the same allyloxyl radical than the hydroperoxide by photolysis of the correspond-
ing nitrite formed with t-BuONO, also playing the role of the spin trap. It is suggested that the generation of these carbon radical
species could play an important role for the binding of the hydroperoxide with skin proteins to form antigenic structures, the first
step of the skin sensitization mechanism.
ꢀ
2005 Elsevier Ltd. All rights reserved.
1
. Introduction
pathology is the reaction of an electrophilic function,
present on the allergen or hapten, with nucleophilic res-
2
,3
Allergic contact dermatitis (ACD) is a very common dis-
ease resulting of epidermal proteins being chemically
idues on proteins to form covalent bonds. However,
over the past few years, the radical mechanism, although
it has never been firmly demonstrated, has gained in-
creased interest in the discussion of the mechanism of
1
modified by haptens. The processing of such modified
proteins by Langerhans cells, the main antigen-present-
ing cells in the epidermis, generates altered peptides that
are subsequently presented, in association with major
histocompatibility complex molecules, to naive T-lym-
phocytes in the lymph nodes. The whole process results
in the selection and activation of T-lymphocyte subpopu-
lations with T-cell receptors specific of the chemical
4
hapten–protein binding for some specific haptens.
The high sensitizing potential of allylic hydroperoxides
derived from the autoxidation of terpenes is well known.
It was already demonstrated, 50 years ago that several
3
hydroperoxides derived from the autoxidation of D -car-
ene were responsible for the allergenic potential of tur-
1
modification. Haptens are usually a low molecular
5
weight molecules, lipophilic enough to penetrate the epi-
dermis through the Stratum corneum, and with a potent
chemical reactivity allowing the formation of a covalent
link with protein amino acid side chains. The classical
mechanism of antigen formation involved in the ACD
pentine. More recently, Karlberg and co-workers
have extensively reported on the allergenic potential of
allylic hydroperoxides derived from the autoxidation
6
of D-limonene, used as a fragrance material, or from
the autoxidation of abietic acid, the main resin acid of
7
,8
colophony.
Most of the studies reported in the literature on the
decomposition of allylic hydroperoxides concern those
derived from the oxidation of polyunsaturated fatty
acids. By studying the mechanisms of lipid oxidation it
Keywords: Skin sensitization; Allylic hydroperoxide; Carbon-centred
radicals; Radical trapping.
*
Corresponding author. Tel.: +33 3 88 35 06 64; fax: +33 3 88 14 04
7; e-mail: jplepoit@chimie.u-strasbg.fr
4
0
968-0896/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2005.04.004

3978
M. Bezard et al. / Bioorg. Med. Chem. 13 (2005) 3977–3986
has been shown that allylic hydroperoxides, when trea-
ted with a variety of radical inducers such as metal com-
plexes, form allyloxyl radicals by easy cleavage of the
peroxy bond. These unstable alkoxyl radicals readily
lead to the formation of carbon-centred radicals
through fragmentation, hydrogen abstraction or intra-
molecular cyclization that may react with molecular
the generation of these radical species could play an
important role for the binding of the hydroperoxide
with skin proteins to form antigenic structures, the first
step of the ACD mechanism.
2. Results
9
,10
oxygen and then decay to hydroxy derivatives.
Frag-
mentation of intermediate dioxetanes, presumably de-
rived from peroxyl radicals formed after cleavage of
the peroxy-hydrogen bond, has also been described.
2.1. Radical trapping experiments
1
1,12
The radical trapping technique that we have used relies
on the almost diffusion-controlled trapping of carbon-
centred radicals by stable nitroxides such as TMIO to
In the course of our studies on allergenic allylic hydro-
peroxides we have shown that such kind of carbon-cen-
tred radical reactions could be important for the binding
of those haptens containing hydroperoxide groups with
1
6
form alkoxyamine products (Scheme 1). The isoindo-
line-based aminoxyl TMIO is known to have a variety
of advantages over commercially available aminoxyls,
including the commonly used pyrrolidine and piperidine
1
3,14
skin proteins.
In a previous study, carried out on
model molecules potentially derived from the decompo-
sition of linalyl hydroperoxide, we have shown that
while hydroperoxide 1 was a very strong sensitizer, po-
tential rearrangement products such as epoxynerol 2,
epoxygeraniol 3, furan and pyran derivatives 4 and 5,
1
7
species. The fused aromatic moiety provides resistance
to the ring-opening reactions that are known decompo-
sition pathways for pyrrolidine and piperidine aminox-
yls. In addition, TMIO is thermally and chemically
stable in a wide range of chemical environments and
has a suitable UV absorbing chromophore that facili-
tates isolation of the adducts formed, whose structures
can be determined by use of classical spectroscopic
techniques.
1
5
respectively, were not sensitizing (Fig. 1). These results
ruled out a possible nucleophilic–electrophilic mecha-
nism involving epoxides as the actual sensitizers. How-
ever, the rearrangement of the allylic hydroperoxide is
highly probable and, since the rearranged derivatives
are not active, it may be supposed that a radical interme-
diate able to react with skin proteins is formed during
the process.
Radical formation was chemically induced by use of
the
meso-tetraphenyl-porphyrin–iron(III)
chloride
3
+
complex (TPP–Fe ), and by direct thermal or photo-
chemical cleavage of 1. Anhydrous and degassed dichlo-
romethane was used as the reaction solvent. Structures
of all isolated and identified radical trapping adducts
are shown in Scheme 2.
During the last decades, much attention has been de-
noted to the endogenous or exogenous production of
radical reactive oxygen species because of their evident
link to the initiation and/or progression of many degen-
erative and chronic diseases. Research has been focussed
on the understanding of their mechanism of action and
on the development of compounds with powerful anti-
oxidant activity. However, few studies are reported in
the literature on the formation of reactive carbon-cen-
tred radical species from toxic xenobiotics that could
be responsible for the chemical modification of bioor-
ganic molecules. In this paper we report a complete
identification of the carbon-centred radicals derived
from the allergen linalyl hydroperoxide 1 using a radical
trapping technique with the stable nitroxide 1,1,3,3-
tetramethylisoindolin-2-yloxyl (TMIO) and electron
paramagnetic resonance (EPR) studies. We suggest that
The different yields obtained depending on the method
used to induce the radical generation are indicated in
Table 1. Structures of the isolated compounds were
1
13
established by a combination of H and C NMR data,
including H– C correlation (HSQC and HMBC) and
1
13
1
1
H– H correlation (COSY and NOESY) experiments.
2.1.1. Generation of radicals by reduction from TPP–
3
+
Fe . We have already reported, in a previous paper, our
first results on radical trapping experiments of 1 using
tritertiobutylphenol (TTBP) as trapping agent and
3
+
15
TPP–Fe as radical inducer. The major reaction tak-
ing place was the formation of a furan ring by intramo-
lecular reaction of an oxygen-centred radical with the
isoprenyl double bond, forming a tertiary carbon-cen-
tred radical. Compounds 2 and 3 were also isolated.
Work reported in the current paper aimed to compare
those results with radical trapping using TMIO. More-
2
+
over, transition metal ions and more precisely Fe /
Figure 1. Potential rearrangement compounds from linalyl
hydroperoxide.
Scheme 1. Formation of alkoxyamines by radical trapping with
TMIO.

M. Bezard et al. / Bioorg. Med. Chem. 13 (2005) 3977–3986
3979
gassed and anhydrous dichloromethane and in the pres-
ence of an excess of TMIO allowed the isolation of
compounds resulting from the rapid rearrangement of
unstable alkoxyl and peroxyl radical intermediates.
The reaction was followed as above. Once the hydroper-
oxide completely disappeared, the solvent was removed
under reduced pressure and the crude product was split
into two fractions (polar and nonpolar) by column chro-
matography on silica gel. The nonpolar fraction, contai-
ning the radical trapping adducts (6a/b, 7a/b and 8a/b),
was further purified by column chromatography on
silica gel and on aluminium oxide. The polar fraction,
containing compound 9, was purified by column chro-
matography on silica gel.
2.1.3. Generation of radicals by photochemical cleavage.
It has been reported in the literature that photochemical
cleavage of hydroperoxides is a quite facile method for
1
9
alkoxyl radical production. It presents the advantage
of being a cleaner method that avoids the presence of
metal contaminants. The formation of radical species
of linalyl hydroperoxide 1 was therefore also investi-
gated by irradiating the compound in the presence of
TMIO in degassed dichloromethane. Once there was
no further change in the reaction progress, the solvent
was removed under reduced pressure and products were
split into polar and nonpolar fractions by column chro-
matography on silica gel. Further purification of the
nonpolar fraction by column chromatography on silica
gel and on aluminium oxide afforded the radical trap-
ping adducts (6a/b, 7a/b and 8a/b). Compound 9 was iso-
lated from the polar fraction by column
chromatography on silica gel.
Scheme 2. Trapping of radicals derived from linalyl hydroperoxide 1.
Table 1. Adduct yields obtained depending on the radical inducer used
(
a
cis/trans ratio into brackets )
3+ b
TPP–Fe
c
d
hm
Compound
Heat
e
2
6
7
8
9
/3
9% (28/72)
48% (46/54)
15% (47/53)
1%
ND
ND
a/b
a/b
a/b
35% (43/57)
11% (86/14)
4% (38/62)
10%
40% (44/56)
9% (87/13)
8% (43/57)
11%
ND
a
b
c
1
Measured by H integration in the NMR spectra.
3+
2
.1.4. Mechanistic interpretations. Looking at the chem-
1
.1 equiv TPP–Fe , 4 equiv TMIO, 90 min.
Reflux, 2 equiv TMIO, 7 days.
ical structures of the diverse compounds obtained by
reaction of 1 in the presence of different radical initiators
and of TMIO, it could be easily seen that several car-
bon-centred radical intermediates were formed (Fig.
d
e
2 equiv TMIO, 3 weeks.
ND = not detected.
2
produced an alkoxyl radical 10 that underwent a rapid
). First of all, homolytic cleavage of the peroxy bond
3
+
Fe systems are suspected to be the most important
way of peroxides degradation in vivo. Therefore, the
use of TPP–Fe needed to be studied once more. The
2
0
intramolecular 5-exo cyclization with the isoprenyl
double bond, leading to the formation of a favourable
five-membered ring system and a rather stable carbon-
centred tertiary radical 11. This radical was then able
to react with TMIO to give diastereomers 6a/b. As
shown in Table 1, 6a/b were the major radical trapping
adducts obtained with yields ranging from 35% to 48%
3
+
3
+
reaction of 1 in the presence of TPP–Fe (1.1 equiv)
and an excess of TMIO was followed by thin layer chro-
matography until complete disappearance of the hydro-
peroxide. The solvent was then removed under reduced
pressure and the reaction compounds were split into two
fractions (polar and nonpolar) by column chromatogra-
phy on silica gel. The nonpolar fraction, containing the
radical trapping adducts (6a/b, 7a/b and 8a/b), was fur-
ther purified by HPLC on a C18 reversed phase column.
The polar fraction, containing compounds 2/3, was puri-
fied by column chromatography on silica gel with diffi-
culty due to the presence of multiple coloured
porphyrine derivatives.
2.1.2. Generation of radicals by thermal cleavage. It is
well known that peroxides undergo thermal cleavage
1
8
to a pair of alkoxyl radicals. However, thermolysis
of pure hydroperoxides for the production of alkoxyl
radicals is poorly documented. In the course of our
experiments, heating to reflux of a solution of 1 in de-
Figure 2. Radical intermediates derived from compound 1.

3980
M. Bezard et al. / Bioorg. Med. Chem. 13 (2005) 3977–3986
Å
depending on the radical initiation used. In all cases, the
less hindered trans diastereomer was slightly preferred
to the cis diastereomer. In order to validate the NMR
interpretations and to confirm the conformation estab-
lished for both adducts, the mixture 6a/b was treated
with an excess of zinc powder (10 equiv) and heated to
reflux in a 1:1 solution of acetic acid and water. Alcohol
–OH could subsequently be reduced liberating HO
Å
Å
(reaction 3). The hydroxyl (HO ) and alkoxyl (RO ) rad-
icals could then easily perform the hydrogen abstraction
affording peroxyl radicals ROO (reactions 2 and 4). In
the experimental conditions using thermal or photo-
chemical radical induction, reactions 2 and 4 may be
produced.
Å
4
was obtained as a previously characterized mixture of
1
5
diastereomers, having equivalent cis/trans ratios to
those of 6a/b. The chemical structures of 6a and 6b were
fully elucidated by NMR experiments on the diastereo-
meric mixture. An important NOE cross-peak between
H-5 of the furan ring and H-B of the ABX vinyl system
confirmed the trans conformation of adduct 6b (Fig. 3).
Feasibility of reaction 2 could have been verified by
determining the content of linalool in the crude prod-
ucts. Unfortunately, it was not possible to isolate linal-
ool from the excess of TMIO. The addition of the
peroxyl radical on the isoprenyl double bond was not
surprising neither, even if peroxyl radicals are much
more known for their aptitude to abstract hydrogen
atoms from activated positions. Indeed, peroxyl radical
cyclization by addition on double bonds is a property
that has been used for the synthesis of cyclic peroxides
In parallel, the peroxyl–hydrogen bond of 1 could be
also cleaved homolytically affording the peroxyl radical
1
2 (Fig. 2).
2
3
with success.
This one could then rearrange, via an intramolecular 6-
exo addition to the isoprenyl double bond, leading to
the tertiary stable carbon-centred radical 13, which re-
acted with TMIO to afford peroxides 7a/b as a nonsepa-
rable mixture of two diastereomers. Yields, ranging
from 9% to 15%, were approximately 3-fold lower than
those for 6a/b, but not insignificant. It is interesting to
note the differences observed in the diastereomeric ratios
obtained, depending on the experimental conditions
Finally, adducts 8a/b were those obtained to a lower ex-
3
+
tent. Only 1% when 1 was in the presence of TPP–Fe ,
4% by thermal decomposition and 8% when exposed to
irradiation. Alkoxyl radical 10 underwent a 3-exo cycli-
zation with the vinyl double bond to the a-oxiranylcar-
binyl radical 14. This species was trapped by TMIO
giving the nonseparable mixture of diastereomers 8a/b.
3
+
The reduction from TPP–Fe afforded also the mixture
of a-hydroxyepoxides 2/3 (28/72) with a yield of 9%.
These compounds, most probably, were the result of a
process taking place in the metal coordination sphere.
This may explain why 2/3 were not detected during the
thermal and photochemical treatment of 1, the a-oxira-
3
+
(
Table 1). In the presence of TPP–Fe , the reaction
was very fast and the trans diastereomer was slightly
preferred. Thermal and photochemical reactions were
slower and the cis diastereomer, thermodynamically fa-
voured, was considerably preferred. In our previous
studies on the sensitizing potential of model molecules
potentially derived from the decomposition of linalyl
hydroperoxide 1 we considered pyran derivative 5
4
+
nylcarbinyl radicals formed being free from TPP–Fe
–OH molecules they were easily trapped by TMIO.
(
Fig. 1) but not the possibility of obtaining a cyclic per-
In order to elucidate the chemical structures of 8a/b
NMR experiments were performed at low temperature.
At room temperature, due to the high flexibility of 8a/b
compared to 6a/6b and 7a/b, dynamic processes associ-
ated to the 1,1,3,3-tetramethylisoindolin moiety, free
to rotate rapidly, prevented good NMR resolution.
The four methyl groups of the spin trap were undistin-
oxide by intramolecular rearrangement of a peroxyl rad-
ical. However, the formation of peroxyl radicals derived
from compound 1 is not surprising.
tions can be suggested for explanation (Scheme 3). The
peroxy compound being considered essentially as an oxi-
dant, it may produce RO and TPP–Fe –OH in the
2
1,22
Several reac-
Å
4+
3
+
4+
1
presence of TPP–Fe (reaction 1). Also, TPP–Fe
guished in the H NMR spectra and even not observed
in the C NMR spectra (Fig. 4).
1
3
Less rapid rotation of the 1,1,3,3-tetramethylisoindolin
moiety was achieved at lower temperatures. As a conse-
quence, the four methyl groups became nonequivalent
as they did not have the same time-averaged chemical
environment, and could therefore be identified (Fig. 4).
1
The 8a/b diastereomeric ratios were calculated by H
integration of the methyl groups at carbon C-3 (Fig.
Figure 3. NOE observed in adduct 6b confirming the trans
conformation.
4
A).
The a,b-unsaturated ketone 9 was also isolated from the
thermal and photochemical experiments with a yield of
1
0% and 11%, respectively. A mechanism that could ac-
count for its formation, starting from radical intermedi-
ate 13 and based on the known aptitude of cyclic
peroxides to be homolytically cleaved by the action of
2
4
heat or light to form oxygenated biradicals, is shown
in Scheme 4.
Scheme 3. Formation of peroxyl radicals.

M. Bezard et al. / Bioorg. Med. Chem. 13 (2005) 3977–3986
3981
Scheme 5. Spin trapping of alkyl radicals derived from linalool 15.
The experiments were conducted under continuous flow
1–2 mL/h) of acetonitrile solutions of 15 in the presence
(
of an excess of t-BuONO. The solutions were photo-
lyzed with a laser tuned for UV light (334–364 nm) di-
rectly in the EPR spectrometer cavity thermostated at
Figure 4. Identification of methyl groups in the mixture of diastereo-
1 13
mers 8a/b by H (A) and C NMR (B) at different temperatures.
240 K. The analysis of all EPR spectra allowed the iden-
tification of nitroxide radicals derived from the trapping
of tertiary carbon-centred radicals (Scheme 5). Nitrox-
ide 16 was formed by t-BuONO trapping of the five-
membered ring system and stable carbon-centred
tertiary radical 11, formed via 5-exo cyclization of allyl-
oxyl radical 10. The hyperfine splitting constant of the
nitrogen atom (a = 25.00 G) was characteristic of an
N
Å
t
R–N(O )–O Bu nitroxide type (25–28 G) as well as the
2
5
g value (2.0052–2.0053). This radical was observed
when using more concentrated linalool solutions, at
low flow rate, under a full power laser irradiation (Fig.
5
A). Nitroxides derived from the coupling of species
generated in situ could also be observed, as 17. The loss
of the tert-butoxide group from 16 afforded a nitroso
derivative, which acted as a spin trap of carbon-centred
Scheme 4. Mechanism suggested for the obtention of 9.
2
.2. EPR experiments
radical 11 to give nitroxide 17. Its a hyperfine splitting
N
constant value (14.60 G) was indeed comparable to those
0
Å
In order to get more detailed information on the evolu-
tion of radical intermediate 10, a spin trapping EPR
technique described in preliminary communications
was employed. The technique is based on the known
ability of alkyl nitrites, in particular of tert-butyl nitrite
of R–N(O )–R systems, as well as the g value of 2.0057.
Nitroxide radical 17 was observed under the same con-
ditions as 16. However, it was the only radical observed
when less concentrated solutions of linalool were used,
at low power laser irradiation, but higher flow condi-
tions (Fig. 5B). EPR studies confirmed thus the forma-
tion of the carbon-centred radical 11, via the favoured
5-exo cyclization of the parent allyloxyl radical. How-
ever, spin adducts formed by trapping of radicals of
the oxiranylcarbinyl type could not be identified, as no
b-proton hyperfine couplings were observed.
2
5
(
t-BuONO), to act as a spin trap for short-lived alkyl
2
6
radicals. More precisely, the allyloxyl radical 10 can
be generated by photolysis of the corresponding nitrite
prepared via the facile exchange reaction between the
parent alcohol and t-BuONO. The alkyl radicals formed
by rapid decay of the allyloxyl radical are then trapped
by the excess of t-BuONO in the reaction medium to
give stable nitroxides detectable by EPR. This method
affords the same allyloxyl radical obtained by homolytic
cleavage of the O–O bond of the hydroperoxide chemi-
cal function. Moreover, hydroperoxides are often unsta-
ble species difficult to handle. Linalool 15, the stable
tertiary alcohol precursor of linalyl hydroperoxide 1,
was therefore used to carry out these studies.
3. Discussion
The results presented in this paper, coming from a com-
bination of radical trapping and EPR studies, confirmed
the formation of reactive carbon-centred radicals from

3982
M. Bezard et al. / Bioorg. Med. Chem. 13 (2005) 3977–3986
In order to confirm the participation of such kind of
radicals in the reaction of 1 with proteins, studies were
started on its reactivity towards a synthetic model pep-
tide. We used an analogue of the N-terminal chain of
the globine protein: H N-Val-Leu-Ser-Pro-Ala-Asp-
2
Lys-Thr-Asn-Trp-Gly-His-Glu-Tyr-Arg-Met-Phe-Gln-
Ile-Gly-CO H. This peptide was reacted by irradiation
2
1
3
with compound 1 C labelled at the isoprenyl double
bond, which is involved in the formation of carbon-cen-
tred radicals 11 and 13. The peptide was then analyzed
1
3
by C NMR in order to identify peaks corresponding
1
3
to potential adducts. New C signals were observed that
could correspond, up to the chemical shifts, to the
formation of C–O bonds between the two entities.
1
3
1
3
This suggested that 11 or 13, now C-centred radicals,
could react with peptide chemical functions containing
the oxygen heteroatom. These preliminary studies need
further development and are a preface for future analysis.
However, they reinforced the hypothesis that the in situ
generation of such reactive radicals in the epidermis
could lead to the formation of antigenic structures
explaining the sensitizing potential of 1.
4. Conclusion
Few studies are reported in the literature on the forma-
tion of reactive carbon-centred radical species from
toxic xenobiotics. We have demonstrated the formation
of carbon radicals from the skin sensitizer linalyl hydro-
peroxide by using radical trapping and EPR studies. The
generation of these carbon radical species could play an
important role for the binding of the hydroperoxide
with skin proteins to form antigenic structures, the first
step of the skin sensitization mechanism. Moreover,
these radical intermediates could be responsible for the
modification of amino acids that are not usually consid-
ered in the mechanism of ACD, going further into the
mechanism of hapten–protein interactions.
Figure 5. A) EPR spectrum obtained in the photolysis of an
acetonitrile solution of 15, 104 mM and t-BuONO, 224 mM, at
2
40 K (flow rate 1 mL/h; laser at 50 mW power), showing radicals 16
d) and 17 (m). (B) EPR spectrum obtained in the photolysis of an
acetonitrile solution of 15, 50 mM, and t-BuONO, 110 mM, at 240 K
flow rate 2 mL/h; laser at 5 mW power), showing radical 17 (m).
(
(
linalyl hydroperoxide 1. The easy cleavage of the hydro-
peroxide chemical function afforded unstable oxygen-
ated radicals that readily rearranged to form stable
carbon radicals. An intramolecular 5-exo cyclization of
the alkoxyl radical 10 with the isoprenyl double bond
led to the formation of the furan derivative 11 (35–
5. Experimental
Caution: Skin contact with hydroperoxides must be
avoided. Since these compounds are skin sensitizing sub-
stances, they must be handled with care.
4
8% yield). Radical 10 could also rearrange via a 3-
exo cyclization with the vinyl double bond to form the
a-oxiranylcarbinyl radical 14 (4–10% yield). Finally,
the peroxyl radical 12 followed an intramolecular 6-
exo addition to the isoprenyl double bond to form the
peroxide 13 (9–15%). The results presented in this paper
were in accordance with our previous studies on the
allergenic activity of compounds suggested as a result
of the linalyl hydroperoxide decomposition (Fig. 1).
However, in the present study we did not observe the
pyran derivative 5 derived from the 6-endo ring closure
of intermediate 10, neither trapped by TMIO or by
Elemental microanalyses were performed by the micro-
analytical laboratory of the CRM (Centre de Recherche
sur les Macromol e´ cules), Strasbourg, France. IR spectra
were recorded on a Perkin–Elmer 1600 Series Fourier
ꢀ1
Transform spectrophotometer (cm
scale) and refer
1
to thin films of liquids (neat). H NMR spectra were re-
corded at 200 MHz on a Bruker AC-200 spectrometer
and at 500 MHz on a Bruker ARX-500 spectrometer.
The chemical shifts (d) are reported in parts per million
(ppm) and are indirectly referenced to tetramethylsilane
via the solvent signal (CDCl , 7.26 ppm). Multiplicities
3
Å
HO , but the formation of the peroxide 13 as well as
are denoted as s (singlet), d (doublet), t (triplet),
sept (septuplet), m (multiplet) and br (broad).
1
3
of compound 9. We cannot therefore exclude the pres-
ence of these intermediates in vivo and their implication
in the allergenic activity of linalyl hydroperoxide should
be also evaluated.
C
NMR spectra were recorded at 50 MHz on a Bruker
AC-200 spectrometer and at 125 MHz on a Bruker
1
3
ARX-500 spectrometer. C chemical shifts (d) are

M. Bezard et al. / Bioorg. Med. Chem. 13 (2005) 3977–3986
3983
ꢁ
reported in ppm and are indirectly referenced to tetra-
France). Dichloromethane was dried over Sicapent
1
3
methylsilane via the C isotope signal of the solvent,
used as internal reference (CDCl , 77.03 ppm; C D ,
(Merck, Nogent-sur-Marne, France) before distillation.
Hexane (99%) was obtained from SDS (Peypin, France).
Aqueous solutions were prepared with deionized water.
All other chemicals were obtained from Sigma Aldrich
(St Quentin de Fallavier, France).
3
6
6
1
28.60 ppm). The different types of carbon in the struc-
tures (quaternary, methine, methylene and methyl) were
identified by the DEPT-135 technique. The two-dimen-
1
1
sional phase-sensitive H– H COSY (correlation spec-
troscopy) spectra were collected using a standard pulse
sequence with a 254 ms acquisition time, a sweep
5.1. Radical trapping experiments
width of 12.25 ppm, 512 complete t data points, and
1
5.1.1. Generation of radicals by reduction from TPP–
Fe . To a solution of linalyl hydroperoxide 1 (200 mg,
3
+
8
scans per increment. The two-dimensional phase-sensi-
1
1
tive H– H NOESY (nuclear Overhauser effect spectros-
copy) were collected using a 800 ms mixing time and a
1.17 mmol) in anhydrous dichloromethane (25 mL) were
added TMIO (894 mg, 4.70 mmol) and TPP–Fe
3
+
2
5
54 ms acquisition time, a sweep width of 12.25 ppm,
12 complete t data points, and 16 scans per increment.
(910 mg, 1.29 mmol). The reaction mixture was stirred
at room temperature until complete disappearance of 1
followed by TLC (90 min). The mixture was then con-
centrated under reduced pressure, taken up with the
minimum amount of dichloromethane, and firstly puri-
fied by column chromatography on silica gel (diethyl
ether/hexane, 16/84; followed by ethyl acetate/hexane,
35/65) to fractionate the crude product into a nonpolar
fraction of compounds and a mixture of 2/3 as a colour-
less oil (19 mg, 9% yield). The nonpolar fraction was
purified by HPLC (C18 reversed-phase; water/acetoni-
trile, 30/70) to give the different adducts as yellowish
oils, which were further purified by column chromato-
graphy on aluminium oxide (hexane) to obtain the mix-
tures of diastereomers 6a/b (195 mg, 48% yield), 7a/b
(65 mg, 15% yield) and 8a/b (4 mg, 1% yield) as colour-
less oils.
1
1
13
The two-dimensional H– C HSQC (heteronuclear sin-
1
13
gle-quantum correlation) spectra and H– C HMBC
heteronuclear multiple-bond correlation) spectra were
collected using a standard pulse sequence, with a
(
2
54 ms acquisition time, 512 complete t data points,
1
6 (HSQC) or 32 (HMBC) scans per increment, and a
1
spectral width in the f dimension of 209.23 ppm. The
1
data were processed using XWINNMR software. The
HPLC analyses were performed on a Waters Associates
apparatus equipped with a model M680 automated gra-
dient controller and two M510 pumps. The solutions
were analyzed and compounds separated on reverse-
phase columns (C18, Nucleosil 5 lm, Interchrom,
4
.6 mm; C18, Rsil 5 lm, Interchrom, 10 mm), and the
effluent absorption was monitored by a M686 tunable
absorbance detector at 250 nm. All radical trapping
experiments were conducted in flame-dried glassware
under an inert atmosphere of argon. The generation of
radicals by photochemical cleavage was achieved using
a Philips 150 W tungsten filament lamp. EPR spectra
were recorded on a Bruker ESP 300 spectrometer
equipped with a X-band microwave bridge, 100 KHz
modulation, and a variable temperature apparatus Bru-
ker BVT 2000. Irradiation of the solutions inside the
EPR cavity was performed with a laser Spectra Physics
Stabilite 2018-RM tuned for UV light (334–364 nm).
Hyperfine splitting assignments were obtained by means
of computer simulation using WINEPR SimFonia.
5.1.2. Generation of radicals by thermal cleavage. Linalyl
hydroperoxide 1 (200 mg, 1.17 mmol) was dissolved in
anhydrous dichloromethane (25 mL), together with
TMIO (447 mg, 2.35 mmol). The solution was heated
to reflux and the reaction followed by TLC until com-
plete disappearance of the hydroperoxide. After 7 days
stirring under reflux, the solution was cooled to room
temperature and concentrated under reduced pressure.
The crude product was fractionated into polar and non-
polar compounds by silica gel column chromatography
(diethyl ether/hexane, 8/92; followed by ethyl acetate/
hexane, 40/60). Further purification of the nonpolar
fraction was carried out by column chromatography
on silica gel (diethyl ether/hexane, 5/95) to give the dif-
ferent adducts as yellowish oils, which were afterwards
purified by column chromatography on aluminium
oxide (hexane) to obtain the mixtures of diastereomers
6a/6b (140 mg, 35% yield), 7a/b (46 mg, 11% yield) and
8a/b (15 mg, 4% yield) as colourless oils. The polar frac-
tion was purified by column chromatography on silica
gel (ethyl acetate/hexane, 20/80) and was found to con-
tain compound 9 (20 mg, 10% yield) as a yellow oil.
TLC analyses were performed on 0.25 mm Merck silica
gel plates (60F254). After elution, the TLC plates were
inspected under ultraviolet light (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 sulfu-
ric acid (5 mL), followed by heating. Compounds were
purified by column chromatography on silica gel (Merck
6
standardized). Linalyl hydroperoxide 1 and TMIO were
0, 230–400 mesh) or on aluminium oxide (Merck 90,
5.1.3. Generation of radicals by photochemical cleavage.
Linalyl hydroperoxide 1 (200 mg, 1.17 mmol) was dis-
solved in anhydrous dichloromethane (25 mL). TMIO
(447 mg, 2.35 mmol) was added to the solution and
the stirred reaction mixture was irradiated for 3 weeks
with a 150 W tungsten filament lamp at a distance of
30 cm. The evolution of the reaction was followed by
TLC until complete disappearance of the hydroperox-
ide. The mixture was concentrated under reduced
1
5,27
synthesized as previously reported in the literature.
Linalool 15 was purchased from Sigma–Aldrich (St
3
+
Quentin Fallavier, France), TPP–Fe and t-BuONO
were purchased from Fluka (St Quentin Fallavier,
France), and all used without further purification. Aceto-
nitrile (99.8%, spectrophotometric grade), diethyl ether
(
(
99.8%), ethyl acetate (99.8%) and dichloromethane
99.5%) were purchased from Carlo Erba (Val de Reuil,

3984
M. Bezard et al. / Bioorg. Med. Chem. 13 (2005) 3977–3986
pressure and fractionated into polar and nonpolar com-
pounds by silica gel column chromatography (diethyl
ether/hexane, 8/92; followed by ethyl acetate/hexane,
(A part of an ABX system, J = 10.8, 1.5 Hz, 1H,
–CH@CH–C–O–), 5.22 (B part of an ABX system,
J = 17.5, 1.5 Hz, 1H, –CH@CH–C–O–), 6.02 (X part
4
0/60). Further purification of the nonpolar fraction
of an ABX system, J = 17.5, 10.8 Hz, 1H, CH @CH–
2
0
0
00
was carried out by column chromatography on silica
gel (diethyl ether/hexane, 5/95) to give the different ad-
ducts as yellowish oils, which were afterwards purified
by column chromatography on aluminium oxide (hex-
ane) to obtain the mixtures of diastereomers 6a/b
C–O–), 7.03–7.16 (m, 2H, H-4 , H-7 ), 7.17–7.30 (m,
00
00
13
2H, H-5 , H-6 ); C NMR (C D , 50 MHz): d 23.6,
6
6
24.2, 26.0, 26.1, 26.2, 27.6, 30.6 (2C), 38.0, 68.6 (2C),
80.5, 82.7, 84.7, 111.1, 122.0 (2C), 127.5 (2C), 145.2,
145.9, 146.0.
(
(
160 mg, 40% yield), 7a/b (38 mg, 9% yield) and 8a/b
32 mg, 8% yield) as colourless oils. The polar fraction
0 0 00 00 00 00
5.2.2.2. trans-2-Methyl-5-[1 -methyl-1 -(1 ,1 ,3 ,3 -
tetramethyl-1 ,3 -dihydroisoindol-2 -yloxy)ethyl]-2-vinyl-
00
00
00
was purified by column chromatography on silica gel
1
(
(
ethyl acetate/hexane, 20/80) to obtain compound 9
22 mg, 11% yield) as a yellow oil.
tetrahydrofuran (6b). Major isomer (56% yield, H inte-
1
gration); H NMR (CDCl , 200 MHz): d 1.29 (s, 3H,
3
–
CH ), 1.34 (s, 6H, 2·–CH ), 1.35 (s, 3H, –CH ), 1.37
3
3
3
00 00
(s, 3H, –CH ), 1.51 (s, 3H, CH -1 or CH -3 ), 1.53 (s,
3 3 3
5
5
.2. Characterization data for radical trapping adducts
00 00
H, CH -1 or CH -3 ), 1.60–2.10 (m, 4H, –CH–CH
3 3 2
3
.2.1. Mixture of diastereomers 2 and 3. Anal. Calcd for
–CH –), 3.97 (t, J = 6.9 Hz, 1H, –CH–CH –CH –),
2 2 2
C H O (170.25): C, 70.55; H, 10.66. Found: C, 70.28;
1
5.02 (A part of an ABX system, J = 10.6, 1.7 Hz, 1H,
–CH@CH–C–O–), 5.23 (B part of an ABX system,
J=17.2, 1.7 Hz, 1H, –CH@CH–C–O–), 5.91 (X part of
0
18
2
H, 10.71. IR (neat) mmax 3420 (O–H), 2969, 2866 (C–H),
ꢀ
1
1
454, 1380 (C–H), 1035 (C–O) cm
.
an ABX system, J = 17.2, 10.6 Hz, 1H, CH @CH–C
2
00
00
5
.2.1.1. 2,3-Epoxynerol (2). Minor isomer (28% yield,
–O–), 7.03–7.16 (m, 2H, H-4 , H-7 ), 7.17–7.30 (m,
00
1
1
00
13
H integration); H NMR (CDCl , 200 MHz): d 1.32 (s,
2H, H-5 , H-6 ); C NMR (C D , 50 MHz): d 23.4,
3
6 6
3
1
H, CH –C–O–), 1.37–1.74 (m, 2H, –CH –CH –C–O–),
.60 (br s, 3H, CH –C@CH–), 1.67 (br s, 3H, CH –
24.3, 26.2 (2C), 27.2, 27.3, 30.6 (2C), 37.4, 68.6 (2C),
80.4, 83.0, 85.1, 111.2, 122.0 (2C), 127.5 (2C), 144.5,
145.9 (2C).
3
2
2
3
3
C@CH–), 1.96–2.17 (m, 2H, –CH –CH –C–O–), 2.18–
2
2
2
.30 (m, 1H, –OH), 2.95 (X part of an ABX system,
J = 6.9, 4.4 Hz, 1H, –O–CH–CH –OH), 3.63 (B part
5.2.3. Mixture of diastereomers 7a and 7b. Anal. Calcd
for C H NO (359.51): C, 73.50; H, 9.25. Found: C,
2
of an ABX system, J = 12.1, 6.9 Hz, 1H, –O–CH–CH–
OH), 3.80 (A part of an ABX system, J = 12.1, 4.4 Hz,
1
2
2
33
3
73.10; H, 9.58. IR (neat) mmax 3082 (Ar C–H), 2973,
2857 (C–H), 1486, 1461, 1375, 1363 (C–H), 1144, 1018
H, –O–CH–CH–OH), 5.08 (tsept, J = 7.1, 1.5 Hz, 1H,
13
ꢀ1
(
2
CH ) C@CH–); C NMR (CDCl , 50 MHz): d 17.6,
(C–O) cm
.
3
2
3
2.2, 24.2, 25.6, 33.1, 61.2, 61.5, 64.2, 123.3, 132.5.
0 0 00 00 00 00
.2.3.1. cis-6-Methyl-3-[1 -methyl-1 -(1 ,1 ,3 ,3 -tetra-
methyl-1 ,3 -dihydroisoindol-2 -yloxy)ethyl]-6-vinyl-1, 2-
5
00
00
00
5
.2.1.2. 2,3-Epoxygeraniol (3). Major isomer (72%
1
1
1
yield, H integration); H NMR (CDCl , 200 MHz): d
1
dioxane (7a). Major isomer (87% yield, H integration);
H
3
1
.28 (s, 3H, CH –C–O–), 1.36–1.75 (m, 2H, –CH –
NMR (CDCl3, 200 MHz):
d
1.21 (s, 3H,
3
2
CH –C–O–), 1.59 (br s, 3H, CH –C@CH–), 1.67 (br s,
–CH ), 1.25 (s, 3H, –CH ), 1.29 (s, 3H, –CH ), 1.32 (s,
3 3 3
3H, –CH ), 1.33 (s, 3H, –CH ), 1.48 (s, 3H, –CH ),
3 3 3
1.49 (s, 3H, –CH ), 1.65–2.12 (m, 4H, –CH–CH –
3 2
2
3
3
O–), 2.19–2.41 (m, 1H, –OH), 2.96 (X part of an ABX
H, CH –C@CH–), 1.94–2.16 (m, 2H, –CH –CH –C–
3
2
2
system, J = 6.6, 4.2 Hz, 1H, –O–CH–CH –OH), 3.65
2
CH –), 4.08 (dd, J = 10.7, 2.1 Hz, 1H, –CH–CH –
2 2
(
B part of an ABX system, J = 12.1, 6.6 Hz, 1H, –O–
CH –), 5.18 (A part of an ABX system, J = 10.8,
2
CH–CH–OH), 3.81 (A part of an ABX system,
J = 12.1, 4.2 Hz, 1H, –O–CH–CH–OH), 5.06 (tsept,
1.2 Hz, 1H, –CH@CH–C–O–), 5.19 (B part of an
ABX system, J = 18.0, 1.2 Hz, 1H, –CH@CH–C–O–),
6.14 (X part of an ABX system, J = 18.0, 10.8 Hz, 1H,
1
3
J = 7.1, 1.5 Hz, 1H, (CH ) C@CH–); C NMR (CDCl3,
3
2
00
00
5
0 MHz): d 16.7, 17.6, 23.7, 25.6, 38.5, 61.2, 61.4, 63.0,
CH @CH–C–O–), 7.01–7.15 (m, 2H, H-4 , H-7 ),
2
00 00
7.16–7.33 (m, 2H, H-5 , H-6 ); C NMR (CDCl ,
3
13
1
23.3, 132.1.
5
33.4, 68.2, 68.4, 79.4, 80.9, 86.7, 114.0, 121.7 (2C),
127.1 (2C), 141.9, 145.4, 145.6.
0 MHz): d 21.1, 23.4, 24.1, 26.5, 25.8, 25.9, 30.2, 30.5,
5.2.2. Mixture of diastereomers 6a and 6b. Anal. Calcd
for C H NO (343.51): C, 76.92; H, 9.68. Found: C,
2
2
33
2
7
2
1
6.92; H, 9.89. IR (neat) mmax 3076 (Ar C–H), 2975,
870 (C–H), 1644 (C@C), 1486, 1362 (C–H), 1154,
0
0
00 00 00 00
5.2.3.2. trans-6-Methyl-3-[1 -methyl-1 -(1 ,1 ,3 ,3 -
tetramethyl-1 ,3 -dihydroisoindol-2 -yloxy)ethyl]-6-vinyl-
ꢀ
1
00 00 00
1
,2-dioxane (7b). Minor isomer (13% yield, H integra-
tion); H NMR (CDCl , 200 MHz): d 1.32 (s, 3H,
033 (C–O) cm
.
1
0
0
0
00 00 00
1
5.2.2.1. cis-2-Methyl-5-[1 -methyl-1 -(1 ,1 ,3 ,3 -tetra-
methyl-1 ,3 -dihydroisoindol-2 -yloxy)ethyl]-2-vinyltetra-
3
00
00
00
–CH ), 1.34 (s, 6H, 2·–CH ), 1.36 (s, 3H, –CH ), 1.49
3
3
3
1
hydrofuran (6a). Minor isomer (44% yield, H integra-
(s, 6H, 2·–CH ), 1.51 (s, 3H, –CH ), 1.65–2.12 (m,
3
3
1
tion); H NMR (CDCl , 200 MHz): d 1.26 (s, 3H,
4H, –CH–CH –CH –), 4.07 (dd, J = 10.7, 2.1 Hz, 1H,
2 2
3
–
CH ), 1.32 (s, 6H, 2·–CH ), 1.34 (s, 3H, –CH ), 1.35
–CH–CH –CH –), 5.15 (A part of an ABX system,
2 2
3
3
3
00
00
(
3
s, 3H, –CH ), 1.50 (s, 3H, CH -1 or CH -3 ), 1.52 (s,
J = 10.8, 1.1 Hz, 1H, –CH@CH–C–O–), 5.26 (B part
of an ABX system, J = 17.7, 1.1 Hz, 1H, –CH@CH–
C–O–), 5.83 (X part of an ABX system, J = 17.7,
3
3
3
00 00
H, CH -1 or CH -3 ), 1.70–2.10 (m, 4H, –CH–CH –
3 3 2
CH –), 4.04 (t, J = 7.0 Hz, 1H, –CH–CH –CH –), 4.99
2
2
2

M. Bezard et al. / Bioorg. Med. Chem. 13 (2005) 3977–3986
3985
0
0
0
1
H-7 ), 7.167.33 (m, 2H, H-5 , H-6 ); C NMR (CDCl ,
0.8 Hz, 1H, CH @CH–C–O–), 7.017.15 (m, 2H, H-4 ,
or CH -3 ), 1.61 (s, 3H, CH –C@CH–), 1.68 (s, 3H,
2
3
3
0
0
00
00 13
CH –C@CH–), 2.06–2.20 (m, 2H, –CH –CH –C–O–),
3
3
2
2
50 MHz): d 20.6, 21.0, 23.5, 24.2, 25.9, 26.0, 30.3, 30.6,
32.8, 68.3, 68.5, 79.4, 79.8, 86.9, 114.7, 121.7 (2C),
127.1 (2C), 141.4, 145.3, 145.6.
3.10–3.20 (m, 1H, –O–CH–CH –), 3.91–4.19 (m, 2H,
–O–CH–CH –O–N–), 5.02–5.14 (m, 1H, (CH ) @CH–),
2
2
3 2
0
0
7.10–7.20 (m, 2H, H-4 , H-7 ), 7.22–7.36 (m, 2H,
0
0
13
H-5 , H-6 ); C NMR (CDCl , 50 MHz, 300 K): d
3
5.2.4. Mixture of diastereomers 8a and 8b. Anal. Calcd
for C H NO (343.51): C, 76.92; H, 9.68. Found: C,
17.0, 17.7, 23.7, 25.7, 38.5, 60.6, 60.7, 67.4 (2C),
76.0, 121.5 (2C), 123.6, 127.2 (2C), 132.0, 145.0 (2C);
C NMR (CDCl , 125 MHz, 213 K): d 16.8, 17.7,
3
23.4, 25.9, 24.8, 25.0, 29.7, 29.9, 38.1, 60.9 (2C), 67.1
(2C), 75.7, 121.5 (2C), 122.9, 127.2 (2C), 132.4, 144.2
2
2
33
2
1
3
7
6.91; H, 9.83. IR (neat) mmax 3047 (Ar C–H), 2972,
859 (C–H), 1484, 1455, 1376, 1361 (C–H), 1143 (C–
2
O) cm
ꢀ
1
.
(
2C).
0 0 0 0
.2.4.1. cis-2,3-Epoxy-3,7-dimethyl-1-(1 ,1 ,3 ,3 -tetra-
5
methyl-1 ,3 -dihydroisoindol-2 -yloxy)-6-octene
0
0
0
(8a).
Minor isomer (43% yield, H integration); H NMR
5.2.5. 6-Hydroxy-2,6-dimethylocta-1,7-dien-3-one (9).
CAS registry number [56698-54-5]; IR (neat) mmax 3462
(O–H), 2960, 2859 (C–H), 1676 (C@O), 1631 (C@C),
1453, 1373 (C–H), 1087 (C–O) cm ; H NMR (CDCl ,
200 MHz): d 1.30 (s, 3H, CH –C–OH), 1.73–1.99 (m,
1
1
(
1
1
CDCl , 500 MHz, 300 K): d 1.37 (s, 3H, CH –C–O–),
3 3
ꢀ
1 1
.45–1.73 (m, 2H, –CH –CH –C–O–), 1.34–1.58 (m,
0
2
2
3
0
2H, 2 · CH -1 , 2 · CH -3 ), 1.62 (s, 3H, CH –
3
3
3
3
C@CH–), 1.69 (s, 3H, CH –C@CH–), 2.06–2.20 (m,
5H, –CH –CH –CO–, CH @C–CH ), 2.70–2.86 (m,
2H, –CH –CH –CO–), 5.06 (A part of an ABX system,
2 2
3
2
2
2
3
2
J = 6.3, 4.8 Hz, 1H, –O–CH–CH –), 3.99 (B part of an
H, –CH –CH –C–O–), 3.12 (X part of an ABX system,
2
2
J = 10.7, 1.4 Hz, 1H, –CH@CH–C–OH), 5.23 (B part of
an ABX system, J = 17.3, 1.4 Hz, 1H, –CH@CH–C–
2
ABX system, J = 11.4, 6.3 Hz, 1H, –O–CH–CH–O–N–),
.09 (A part of an ABX system, J = 11.4, 4.8 Hz, 1H,
O–CH–CH–O–N–), 5.11 (tsept, J = 6.9, 1.4 Hz, 1H,
4
–
OH), 5.72–5.78 (m, 1H, –CH@C–CH ), 5.96–6.01 (m,
3
1H, –CH@C–CH ), 5.85 (X part of an ABX system,
3
0
0
0
13
(
J = 7.6, 7.4, 1.1 Hz, 2H, H-4 , H-7 ), 7.24 (XX part of
CH ) @CH–), 7.11 (AA part of an AA XX system,
J = 17.3, 10.7 Hz, 1H, CH @CH–C–OH); C NMR
3
2
2
0
0
0
(CDCl , 50 MHz): d 17.7, 28.7, 32.3, 35.7, 72.7, 112.2,
3
0
0
0
an AA XX system, J = 7.6, 1.1, 0.6 Hz, 2H, H-5 , H-
); H NMR (CDCl , 500 MHz, 213 K): d 1.38 (s,
124.7, 144.5 (2C), 202.6.
0
1
6
3
3
H, CH –C–O–), 1.40–1.70 (m, 2H, –CH –CH –C–O–
0
5.3. EPR experiments
3
2
2
0
0
), 1.38 (s, 3H, CH -1 or CH -3 ), 1.39 (s, 3H, CH -1
0
or CH -3 ), 1.51 (s, 3H, CH -1 or CH -3 ), 1.52 (s,
3
3
3
0
0
An oxygen free acetonitrile solution containing linalool
15 (104 mM) and t-BuONO in excess (224 mM) was
continuously flowed (flow rate 1 mL/h) through a flat
quartz cell inside the EPR cavity and directly irradiated
with a laser Spectra Physics Stabilite 2018-RM, tuned
for UV light (334–364 nm), at 240 K. The solutions were
3
3
3
0
0
3
H, CH -1 or CH -3 ), 1.59 (s, 3H, CH –C@CH–),
3
3
3
1
CH –C–O–), 3.10–3.20 (m, 1H, –O–CH–CH –), 3.91–
.67 (s, 3H, CH –C@CH–), 2.06–2.20 (m, 2H, –CH –
3
2
2
2
4
(
7
.19 (m, 2H, –O–CH–CH –O–N–), 5.02–5.14 (m, 1H,
2
0
0
CH ) @CH–), 7.10–7.20 (m, 2H, H-4 , H-7 ), 7.22–
3
2
0
0
13
.36 (m, 2H, H-5 , H-6 ); C NMR (CDCl , 50 MHz,
deaerated prior to use by purging with N -gas for
2
3
3
00 K): d 17.7, 22.2, 24.2, 25.7, 33.4, 60.2, 62.1, 67.4
2C), 75.7, 121.5 (2C), 123.6, 127.2 (2C), 132.2, 145.0
30 min.
(
(
1
3
2C); C NMR (CDCl , 125 MHz, 213 K): d 17.7,
3
2
6
1
2.1, 23.9, 25.8, 24.8, 25.0, 29.7, 29.9, 33.0, 60.5, 61.9,
7.0 (2C), 75.2, 121.5 (2C), 122.9, 127.2 (2C), 132.5,
44.4 (2C).
Acknowledgements
Financial support from the Ministere de lÕEducation
Nationale (France) is gratefully acknowledged for M.B.
0 0 0 0
trans-2,3-Epoxy-3,7-dimethyl-1-(1 ,1 ,3 ,3 -
tetramethyl-1 ,3 -dihydroisoindol-2 -yloxy)-6-octene (8b).
5
.2.4.2.
0
0
0
1
1
Major isomer (57% yield, H integration); H NMR
(
References and notes
CDCl , 500 MHz, 300 K): d 1.33 (s, 3H, CH –C–O–),
3
3
1
1
.45–1.73 (m, 2H, –CH –CH –C–O–), 1.34–1.58 (m,
2 2
1
. Rustemeyer, T.; van Hoogstraten, I. M. W.; von Blom-
berg, B. M. E.; Scheper, R. J. In Textbook of Contact
Dermatitis; Rycroft, R. J. G., Menn e´ , T., Frosch, P. J.,
Lepoittevin, J.-P., Eds.; Springer: Berlin, Heidelberg,
2001, pp 13–58.
0
0
2H, 2 · CH -1 , 2 · CH -3 ), 1.63 (s, 3H, CH –
3
3
3
C@CH–), 1.69 (s, 3H, CH –C@CH–), 2.06–2.20 (m,
3
2
J = 6.3, 4.8 Hz, 1H, –O–CH–CH –), 4.02 (B part of an
H, –CH –CH –C–O–), 3.11 (X part of an ABX system,
2 2
2
2
. Roberts, D. W.; Lepoittevin, J.-P. In Allergic Contact
Dermatitis: The Molecular Basis; Lepoittevin, J.-P., Bas-
ketter, D. A., Goossens, A., Karlberg, A.-T., Eds.;
Springer: Berlin, Heidelberg, 1998, pp 81–111.
. Lepoittevin, J.-P. In Textbook of Contact Dermatitis;
Rycroft, R. J. G., Menn e´ , T., Frosch, P. J., Lepoittevin,
J.-P., Eds.; Springer: Berlin, Heidelberg, 2001, pp 59–89.
ABX system, J = 11.3, 6.3 Hz, 1H, –O–CH–CH–O–N–
)
, 4.12 (A part of an ABX system, J = 11.3, 4.8 Hz,
1
1
H, –O–CH–CH–O–N–), 5.12 (tsept, J = 7.1, 1.4 Hz,
0
0
0
H, (CH ) @CH–), 7.11 (AA part of an AA XX sys-
3
2
3
0
0
0
tem, J = 7.6, 7.4, 1.1 Hz, 2H, H-4 , H-7 ), 7.24 (XX part
0
0
0
of an AA XX system, J = 7.6, 1.1, 0.6 Hz, 2H, H-5 , H-
0
1
6
); H NMR (CDCl , 500 MHz, 213 K): d 1.32 (s, 3H,
3
4. Smith, C. K.; Hotchkiss, S. A. M. In Allergic Contact
Dermatitis: Chemical and Metabolic Mechanisms; Smith,
C. K., Hotchkiss, S. A. M., Eds.; Taylor and Francis:
London, 2001, pp 119–205.
CH –C–O–), 1.40–1.70 (m, 2H, –CH –CH –C–O–), 1.38
3
2
2
0 0 0
s, 3H, CH -1 or CH -3 ), 1.39 (s, 3H, CH -1 or CH -
3 3 3 3
(
0
0
0
0
3
), 1.51 (s, 3H, CH -1 or CH -3 ), 1.52 (s, 3H, CH -1
3
3
3

3986
M. Bezard et al. / Bioorg. Med. Chem. 13 (2005) 3977–3986
5
6
. Hellerstrom, S.; Thyresson, N.; Widmark, G. Dermato- 16. Beckwith, A. L. J.; Bowry, V. W.; Ingold, K. U. J. Am.
logica 1957, 115, 277.
Chem. Soc. 1992, 114, 4983.
17. Griffiths, P. G.; Rizzardo, E.; Solomon, D. H. Tetrahedron
Lett. 1982, 23, 1309.
. Matura, M.; Goossens, A.; Bordalo, O.; Garcia-Bravo, B.;
Magnusson, K.; Wrangsjo, K.; Karlberg, A.-T. Contact
Dermatitis 2003, 49, 15.
18. Dussault, P. In Active Oxygen in Chemistry; Foote, C. S.,
Valentine, J. S., Greenberg, A., Liebman, J. F., Eds.;
Blackie Academic and Professional, Chapman and Hall:
New York, 1995, pp 141–203.
7
8
9
. Karlberg, A.-T.; Gafvert, E. Contact Dermatitis 1996, 35,
201.
. Karlberg, A.-T.; Basketter, D.; Goossens, A.; Lepoittevin,
J.-P. Contact Dermatitis 1999, 40, 183.
. Wilcox, A. L.; Marnett, L. J. Chem. Res. Toxicol. 1993, 6,
19. Schieberle, P.; Trebert, Y.; Firl, J.; Grosch, W. Chem.
Phys. Lipids 1985, 37, 99.
4
13.
20. Baldwin, J. E.; Thomas, R. C.; Kruse, L. I.; Silberman, L.
J. Org. Chem. 1977, 42, 3846.
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. Boukouvalas, J.; Pouliot, R.; Fr e´ chette, Y. Tetrahedron
Lett. 1995, 36, 4167.
24. Careless, H. A. J.; Atkins, R.; Fekarurhobo, G. K.
Tetrahedron Lett. 1985, 26, 803.
25. Grossi, L.; Strazzari, S.; Gilbert, B. C.; Whitwood, A. C.
J. Org. Chem. 1998, 63, 8366.
26. Grossi, L. Tetrahedron 1997, 53, 3205.
27. Griffiths, P. G.; Moad, G.; Rizzardo, E.; Solomon, D. H.
Aust. J. Chem. 1983, 36, 397.
1
1
1
0. Loidl-Stahlhofen, A.; Hannemann, K.; Spiteller, G. Bio-
chim. Biophys. Acta 1994, 1213, 140.
1. Kaur, K.; Salomon, R. G.; OÕNeil, J.; Hoff, H. F. Chem.
Res. Toxicol. 1997, 10, 1387.
2. Timmins, G. S.; dos Santos, R. E.; Whitwood, A. C.;
Catalani, L. H.; di Mascio, P.; Gilbert, B. C.; Bechara, E.
J. H. Chem. Res. Toxicol. 1997, 10, 1090.
3. Mutterer, V.; Gim e´ nez-Arnau, E.; Karlberg, A.-T.; Lep-
oittevin, J.-P. Chem. Res. Toxicol. 2000, 13, 1028.
4. Gim e´ nez-Arnau, E.; Haberkorn, L.; Grossi, L.; Lepoitte-
vin, J.-P. Tetrahedron 2002, 58, 5535.
1
1
1
5. Bezard, M.; Karlberg, A.-T.; Montelius, J.; Lepoittevin,
J.-P. Chem. Res. Toxicol. 1997, 10, 987.