Synthesis of allylic hydroperoxides and EPR spin-trapping studies on the formation of radicals in iron systems as potential initiators of the sensitizing pathway

Article abstract of DOI:10.1021/jo200948x

Many terpenes used as fragrance compounds autoxidize when exposed to air, forming allylic hydroperoxides that have the potential to be skin contact allergens. To trigger the immunotoxicity process that characterizes contact allergy, these hydroperoxides a

Full text of DOI:10.1021/jo200948x

ARTICLE
pubs.acs.org/joc
Synthesis of Allylic Hydroperoxides and EPR Spin-Trapping Studies
on the Formation of Radicals in Iron Systems as Potential
Initiators of the Sensitizing Pathway
Dany Kao,† Alain Chaintreau,‡ Jean-Pierre Lepoittevin,† and Elena Gimꢀenez-Arnau*,†
†Laboratoire de Dermatochimie, Institut de Chimie de Strasbourg (UMR 7177), Universitꢀe de Strasbourg, 4 Rue Blaise Pascal,
67081 Strasbourg, France
‡Corporate R&D Division, Firmenich SA, 1 Route des Jeunes, 1211 Geneva 8, Switzerland
S Supporting Information
b
ABSTRACT: Many terpenes used as fragrance compounds
autoxidize when exposed to air, forming allylic hydroperoxides
that have the potential to be skin contact allergens. To trigger
the immunotoxicity process that characterizes contact allergy,
these hydroperoxides are supposed to bind covalently to
proteins in the skin via radical pathways. We investigated the
formation of reactive radical intermediates from 7-hydroperoxy-
3,7-dimethylocta-1,5-dien-3-ol and 2-hydroperoxylimonene,
responsible for the sensitizing potential acquired by autoxidized
linalool and limonene. Both compounds were synthesized
through new short and reproducible synthetic pathways. The
hydroperoxide decomposition catalyzed by Fe(II)/Fe(III) redox systems, playing a key role in degradating peroxides in vivo, was
examined by spin-trapping-EPR spectroscopy. Alkoxyl and carbon-centered free radicals derived from the hydroperoxides were
successfully trapped by the spin-trap 5,5-dimethyl-1-pyrroline N-oxide, whereas peroxyl radicals were characterized by spin-trapping
studies with 5-diethoxyphosphoryl-5-methyl-1-pyrroline N-oxide. Using liquid chromatography combined with mass spectrometry,
we demonstrated the formation of adducts, via radical mechanisms induced by Fe(II)/Fe(III), between the hydroperoxides and N-
acetylhistidine methyl ester, a model amino acid that is prone to radical reactions. Free radicals derived from these hydroperoxides
can thus induce amino acid chemical modifications via radical mechanisms. The study of these mechanisms will help to understand
the sensitizing potential of hydroperoxides.
Chart 1. Structures of Linalool 1, R-(+)-Limonene 2,
7-Hydroperoxy-3,7-dimethylocta-1,5-dien-3-ol 3, and
2-Hydroperoxylimonene 4
’ INTRODUCTION
Terpenes, principally mono- and sesquiterpenes, are common
fragrance compounds of natural origin used in the production of fine
perfumes, but also in a large variety of fragranced cosmetic, house-
hold, and industrial products. Some terpenes are known to be
converted by autoxidation into allylic hydroperoxides which have
the potential to cause skin contact allergy.1 This is the case for the
acyclic terpene alcohol linalool 1 and for the cyclic terpene hydro-
carbon R-(+)-limonene 2 (Chart 1). Both occur in many essential
oils, such as those of lavender and citruses, respectively. Linalool and
R-(+)-limonene are among the most used fragrance compounds in
scented products. A study on the quantitative chemical composition
of a variety of deodorants (vapo- and aerosol spray, roll-on),
purchased at retail outlets in five European countries, showed that
97% contained linalool.2 R-(+)-Limonene was the most detected
fragrance substance when analyzing the chemical composition of 59
domestic and occupational products intended for hand exposure
and was present in 71% of 300 cosmetic products examined.3,4
Linalool and R-(+)-limonene are not themselves sensitizing
compounds, but they do autoxidize to allylic hydroperoxides that
have been shown to be potent sensitizers in the oxidation mixtures
formed upon air exposure. In particular, 7-hydroperoxy-3,7-di-
methylocta-1,5-dien-3-ol 3 and 2-hydroperoxylimonene 4 have
been identified as the primary forms responsible for the sensitizing
potential of autoxidized linalool 1 and limonene 2, respectively
(Chart 1).5À7
Allergic contact dermatitis (ACD) is an immunologically
based disease resulting from skin sensitization to a chemical.
Received: May 10, 2011
Published: June 07, 2011
r
2011 American Chemical Society
6188
dx.doi.org/10.1021/jo200948x J. Org. Chem. 2011, 76, 6188–6200
|

The Journal of Organic Chemistry
ARTICLE
The sensitization process starts by the chemical modification of
epidermal proteins by the allergen or hapten. Processing of the
haptenÀprotein complex by immunocompetent skin antigen-
presenting cells ensures the presentation of the altered peptides
Scheme 1. Synthesis of 3
€
to naive T-lymphocytes in the lymph nodes. T-lymphocyte
subpopulations with T-cell receptors specific for the chemical
modification are then selected and will be activated following a
second contact with the allergen. Individuals sensitized to the
chemical will thus be predisposed to develop ACD, characterized
by the appearance of erythema and edema, after a new skin
exposure to the same chemical.8,9 The established mechanism for
the initial haptenÀprotein complex formation is the reaction of
an electrophilic function present on the allergen with nucleo-
philic residues on proteins. Instead, allergenic allylic hydroper-
oxides are believed to react through mechanisms involving
radical intermediates.10À12
Many studies have been reported recently on the identification
and possible involvement in ACD of radicals derived from allylic
hydroperoxides using chemical traps.10À12 Indeed, homolytic
cleavage of the OÀO and OÀH bonds of hydroperoxides, having
bond dissociation energies that are relatively weak (around 42
and 88 kcal molÀ1, respectively13), is an easy process that affords
alkoxyl and peroxyl radicals, which rapidly can lead via intramo-
lecular cyclization, fragmentation, or hydrogen abstraction to the
formation of carbon-centered radicals. Today, there is a real
belief that this kind of oxygen- and carbon-centered radicals
could be important for the binding of those haptens containing
hydroperoxide groups with skin proteins. Previous studies have
already examined radical formation from allylic hydroperoxides
found in oxidation mixtures of limonene, including 4.14,15
Chemical trapping experiments were performed using tetraphe-
nylporphyrin-Fe3+ (TPP-Fe3+) as initiator, and 1,1,3,3-tetra-
methylisoindolin-2-yloxyl (TMIO) known to trap exclusively
carbon-centered radicals by forming alkoxyamine products.
Studies were complemented by preliminary EPR experiments
using photolysis for radical initiation. Herein we report a
complete identification of radicals issued from the allergenic
acyclic tertiary allylic hydroperoxide 3, derived from linalool,
compared to the cyclic and secondary allylic hydroperoxide 4,
derived from limonene, by using the spin-trapping technique
combined with EPR spectroscopy, specifically employed for
the identification of transient radicals in chemical and bio-
logical systems. 5,5-Dimethyl-1-pyrroline N-oxide (DMPO) and
5-diethoxyphosphoryl-5-methyl-1-pyrroline N-oxide (DEPMPO)
were used as spin traps in organic/aqueous media, and Fe(II)/
Fe(III) redox systems were used as initiators of the radical reac-
tions, as they are described to play a key role in the degradation of
peroxides in vivo.16 Further reactivity studies of 3 and 4 toward
N-acetylhistidine methyl ester in iron systems, by liquid chro-
matography electrospray ionization tandem mass spectrometry
(LC-ESI-MS/MS), showed the formation of adducts via radical
pathways, suggesting that these radicals could play an important
role for the binding of the hydroperoxides to skin proteins to
form antigenic structures.
Scheme 2. The Schenck Reaction
oxidation mixtures. Even though compound 3 has already been
isolated from linalool autoxidation mixtures, it has never been
previously prepared by synthesis to our knowledge.17 On the
other hand, the synthesis of compound 4 has been reported in
the literature but with moderate overall yields (20À30%) that we
sought to improve.14
a. Allylic Hydroperoxide 3. The synthetic route developed for
the generation of 3 is illustrated in Scheme 1. We targeted the
synthesis of 3 by subjecting linalool 1 to the ene or ‘Schenck
reaction’, allowing the preparation of allylic hydroperoxides from
alkenes containing allylic hydrogens via singlet oxygen (1O2)
reaction (Scheme 2).
Concentrating on the factors that control the stereoselectivity and
regioselectivity of the 1O2Àalkene ene reaction, Alberti and Orfa-
nopoulos demonstrated that in the case of trisubstituted alkenes,
including a geminal dimethyl substituent, the allylic hydrogens next
to the large alkyl substituent are in general more reactive than those
next to the small group.18 Therefore, the ‘Schenck reaction’ of
linalool 1 was expected to lead mostly to compound 3, having the
hydroperoxide on the more substituted carbon.
1
Most commonly, O2 is generated by photosensitization of
molecular oxygen in its ground state, but it can also be generated
in high yields using a chemical source. A mild chemical method to
quantitatively generate 1O2, developed by Aubry and Cazin in the
1980s, consists of the disproportionation of hydrogen peroxide
induced catalytically by sodium molybdate in alkaline aqueous
solutions.19 With this method, a high flux of 1O2 may be
generated in situ at room temperature without the need for light
and photochemical equipment. However, it is known that the
’ RESULTS AND DISCUSSION
2À
H2O2/MoO4 system also acts as an efficient double bond
Synthesis of Allylic Hydroperoxides. In order to carry out
the EPR spin-trapping and reactivity studies, hydroperoxides 3
and 4 were synthesized, as the formation of hydroperoxides by
autoxidation of terpenes, not always completely reproducible,
gives low yields and the products are not easy to isolate from the
epoxidizing agent under acidic and neutral conditions.20 In order
to avoid this pH-dependent competitive reaction, we decided to
use a microemulsion consisting of water, dichloromethane
(organic phase), sodium dodecyl sulfate (surfactant), and
6189
dx.doi.org/10.1021/jo200948x |J. Org. Chem. 2011, 76, 6188–6200

The Journal of Organic Chemistry
ARTICLE
butanol (cosurfactant) as reaction medium, previously described
as particularly suitable for the peroxidation of hydrophobic
substrates. Also, this water-in-oil microemulsion was reported
to limit the direct interaction between the substrate and the
peroxomolybdate intermediates in a way in which the olefin
Scheme 3. Synthesis of 4
epoxidation is minimized.21,22 The treatment of linalool 1 with
H2O2/MoO4
reactants, generating 1O2 in the described
2À
microemulsion, afforded not only the expected allylic hydroper-
oxide 3 but also the structural isomer 5. In a 80% yield reaction, a
ratio 40:60 of 3 and 5, respectively, was measured by 1H NMR.
This result was in opposition to the previous Alberti and
Orfanopoulos observations, in which the presence of a large
alkyl substituent on the double bond should have favored the
formation of 3.18 Yet, the hypothetical steric hindrance of this
group seemed to be minimized because of the possibility for free
rotation. Much work has been dedicated to the elucidation of the
mechanism of the ene reaction for substrates with several allylic
hydrogen atoms, and even if significant regioselectivity can be
achieved by choosing appropriate structural features within the
allylic substrate,18 there are still examples where allylic hydrogen
abstraction may take place at all possible sites, resulting in
mixtures of isomeric allylic hydroperoxides.
All attempts to separate 3 from isomer 5 by column chroma-
tography (silica gel, silica gel permeated with silver nitrate,
alumina) were unsuccessful. On the basis of a study of Clark
et al. in which regioisomer allylic hydroperoxides were selectively
protected by using tert-butyldimethylsilyl chloride and imidazole
in DMF, we suggested the chemical derivatization of 3 and 5 by
protection of the hydroperoxide functionality as silyl ethers.23
We employed tert-butyldiphenylsilyl chloride (TBDPSCl) as a
protective agent because we thought the steric hindrance of the
bulky TBDPS group could favor the selective protection of one
of the hydroperoxides. The first trials ended with the protection
of both hydroperoxides, and again it was not possible to isolate
the peroxysilanes formed. However, it was possible to resolve
kinetically the selective protection of the hydroperoxides when
treating much more diluted DMF solutions of the 3 and 5
mixture (0.13 M) with TBDPSCl (1.2 equiv) and imidazole
(2 equiv), as 3 was protected much faster than 5, affording the
peroxysilane 6, which was easy to separate from nonprotected 5.
However, it was very difficult to completely remove the excess
TBDPSCl. Thus, no absolute yield is given for the formation of 6,
which was purified as much as possible and then deprotected
with a methanolic solution of potassium hydroxide to afford 3.
The optimized yield for this two-step method for protection and
deprotection of the hydroperoxide was 96%. In summary, a
simple three-step route was thus developed for the synthesis of 3
with an overall yield of 31%.
10-fold excess of hydrogen peroxide in a THF solution
(Scheme 3). But we only recovered 10% of expected hydroper-
oxide 4 together with starting product 7. Finally, 4 was surprisingly
obtained in a higher yield and in one step by using the most
classical method for preparation of tertiary allylic hydroperoxides,
that is the quenching with hydrogen peroxide of a carbocation
generated from a hydroxyl function under acidic conditions.25
Even if secondary carbocations are less favorable than tertiary
carbocations for such reactions, the treatment of (À)-carveol 7
with hydrogen peroxide in the presence of H2SO4 afforded 4 in a
48% yield (Scheme 3). To our knowledge, this hydroperoxide had
never been isolated in more than 30% yield.
Radical Species Issued from 3 and 4: EPR Spin-Trapping
Studies. Our initial investigations on the identification of radical
species issued from allergenic allylic hydroperoxides started several
years ago by combining two radical trapping techniques, the first
leading to neutral compounds and the second allowing the detec-
tion of radical intermediates by EPR.11,12,26 For conversion into
neutral molecule radical species obtained from the hydroperoxides
by the action of several radical initiators (light, heat, TPP-Fe3+),
the stable TMIO was used as a trap. On the other hand, combined
with EPR spectroscopy, the spin-trapping technique is considered
one of the best methods for the detection of short-lived radicals and
is commonly employed for the detection of free radicals in
biological media. The EPR spin-trapping technique we initially
employed was based on the use of the allylic alcohol precursor of
the hydroperoxide. The same allyloxyl radical that could derive
from the hydroperoxide was generated in situ by photolysis of the
nitrite formed by reacting the alcohol with tert-BuONO, which also
played the role of the spin trap. Major drawbacks of these methods
are that they are specific for the identification of carbon-centered
radicals and do not allow direct identification of derived oxygen-
centered radicals (e.g., alkoxyl, peroxyl, and hydroxyl). Also, the
parent alcohol, and not directly the hydroperoxide of interest, was
used for the EPR studies. A solution was initially found by making
use of the EPR spin-trapping technique and of DEPMPO as a spin
trap, as one of its key features is the high persistence of its oxygen-
centered radical adducts, especially the superoxide and the alkylper-
oxyl radicals.27 The formation of alkylperoxyl radicals from direct
photolysis of hydroperoxide 4 was in this way demonstrated.14
However, although photochemical cleavage of hydroperoxides is
quite facile, it is unlikely to happen in vivo. Reductive cleavage of
peroxides and hydroperoxides has been reported to occur not only
with reduced metals, in particular Fe(II), but also with cytochrome
b. Allylic Hydroperoxide 4. The use of the Schenck reaction for
the synthesis of hydroperoxide 4 from R-(+)-limonene 2 resulted
in the generation of multiple isomeric allylic hydroperoxides of
similar polarity, from which 4 was a minor product that was
difficult to isolate. As previously reported, secondary allylic
hydroperoxides can be synthesized by SN2 nucleophilic substitu-
tion from the corresponding alcohol, such as the formation of 4
from (À)-carveol 7.14 The alcohol functionality is converted into
a good leaving group and then substituted by hydrogen peroxide.
We tried this methodology via the known silver ion-assisted
reaction of the alcohol-corresponding halide and hydrogen
peroxide.24 (À)-Carveol 7 was treated with CBr4 and PPh3 in
dichloromethane. Unstable crude product 8 was immediately
treated with silver trifluoroacetate (1.2 equiv) and at least a
6190
dx.doi.org/10.1021/jo200948x |J. Org. Chem. 2011, 76, 6188–6200

The Journal of Organic Chemistry
ARTICLE
Figure 1. Experimental EPR spectra obtained from the reaction of 3 or 4 with FeSO4, (a and c) or hemin (b and d), in the presence of DMPO and in
CH3CN. Computer simulation of spectra led to the following parameters. (a) Signal (b) aN = 13.57 G, aHβ = 11.50 G, aHγ = 1.37 G (g = 2.0065) for
DMPO/RO• or DMPO/ROO•; signal (1) aN = 15.23 G, aHβ = 23.93 G (g = 2.0061) for DMPO/R• carbon-centered adduct; signal (O) aN = 14.01 G
(g = 2.0065) for 10. (b) Signal (b) aN = 13.57 G, aHβ = 11.50 G, aHγ = 1.37 G (g = 2.0065) for DMPO/RO• or DMPO/ROO•; signal (1) aN = 15.23 G,
aHβ = 23.93 G (g = 2.0061) for DMPO/R• carbon-centered adduct; signal (]) aN = 15.53 G, aHβ = 1.60 G (g = 2.0063) for 12. (c) Signal (b) aN = 13.28
G, aHβ = 7.52 G, aHγ = 1.86 G (g = 2.0064) for DMPO/RO• or DMPO/ROO•. (d) Signal (b) aN = 13.28 G, aHβ = 7.52 G, aHγ = 1.86 G (g = 2.0064) for
DMPO/RO• or DMPO/ROO•; signal (]) aN = 15.05 G, aHβ = 1.65 G (g = 2.0064) for 12.
enzymes and related Fe(III)-heme-containing models and de-
scribed to proceed via both homolytic and heterolytic pathways.28,29
As Fe(II)/Fe(III) redox cycles are common in biological media, we
have therefore started to study the behavior of allergenic allylic
hydroperoxides in the presence of Fe(II)/Fe(III) systems by EPR
spectroscopy and the spin-trapping technique. Preliminary studies
were conducted with a model monocyclic tertiary allylic hydroper-
oxide together with DMPO and DEPMPO, both spin traps being
known to distinguish oxygen-centered (peroxyl/alkoxyl) and
carbon-centered radical adducts.30À32 In the presence of FeSO4, a
typical reagent for the reduction of hydroperoxides in Fenton’s
reaction, and Fe(III) derivatives such as FeCl3 and because of the
ubiquitous nature of heme complexes such as Fe(III)-porphyrins
TPP-Fe3+ and hemin, described to induce cleavage of the hydro-
peroxide functionality,33,34 we confirmed the formation of carbon-
centered radicals and of oxygen-centered allyloxyl and peroxyl
radicals in organic/aqueous media.10 We have now applied this
methodology to 7-hydroperoxy-3,7-dimethylocta-1,5-dien-3-ol 3
and 2-hydroperoxylimonene 4. Experiments were carried out in
CH3CN or in a 1:1 (v/v) H2O/CH3CN mixture. Nonpolar organic
solvents can mimic the reaction conditions in a membrane for
example, but because many biological reactions proceed at aqueous/
organic medium interfaces the reactivity of allylic hydroperoxides in
aqueous solutions is also of interest. The skin could also be seen as a
semiorganic medium with hydrophilic and hydrophobic properties.
a. Influence of Fe(II)/Fe(III) Radical Initiation. Reactions were
carried out in CH3CN and in the presence of FeSO4, TPP-Fe3+,
or hemin. In some cases, the recorded EPR spectra when using
DEPMPO corroborated the results obtained with DMPO.
However, in many other cases the superimposition of all EPR
signals could not help to elucidate the adducts obtained. There-
fore, results shown here are mainly those obtained with DMPO.
Figure 1 shows the EPR spectra obtained from the reaction of 3
and 4 with FeSO4 or hemin and in the presence of DMPO.
DMPO-trapped radicals with hfc values of aN = 13.57 G, aHβ
=
11.50 G, aHγ = 1.37 G (g = 2.0065) and aN = 13.28 G, aHβ = 7.52
G, aHγ = 1.86 G (g = 2.0064) indicate the formation of oxygen-
centered radicals, regardless of the Fe species used for radical
initiation. While DMPO/•OH and DMPO/•OOH adducts are
well-defined with characteristic hfc values, the attribution of
other structures of DMPO oxygen-centered radical adducts, such
as alkoxyl and peroxyl adducts, has been described to be contro-
versial.35,36 The observed hfc values were, in principle, in good
agreement with those reported for the trapping of tertiary peroxyl
radicals such as those derived from cumene hydroperoxide.37
However, according to the literature, many DMPO peroxyl
radical adducts (DMPO/ROO•) have virtually the same hfc as
alkoxyl radical adducts (DMPO/RO•), raising the issue of
correct assignement to peroxyl radical adducts. The same phe-
nomenon has been described to happen with DEPMPO spin-
trapping.30,38 The accurate discrimination between spin-trapping
of ROO• and of RO• seems to be therefore uncertain when using
DMPO or DEPMPO. However, it is well-known that when
alkylperoxyl radicals are trapped by DEPMPO, the EPR spectra
6191
dx.doi.org/10.1021/jo200948x |J. Org. Chem. 2011, 76, 6188–6200

The Journal of Organic Chemistry
ARTICLE
Chart 2. Structures of DMPO Degradation Products
Detected in the EPR Spectra
various isomer forms for which the hfc values are different. The
same is not observed when trapping alkoxyl RO• radicals.38,40
These results corroborated the presence of alkylperoxylradicals
ROO• issued from the hydroperoxides in the reaction mixtures.
However, the simultaneous presence of alkoxyl RO• radicals for
which spin-trapping signals could be subjacent to those of the
ROO• spin adducts cannot be excluded.
Most interestingly, when the experiments were conducted with
compound 3, a signal corresponding to the trapping of carbon-
centered radicals was also observed, especially when using Fe(III)-
porphyrins such as hemin (Figure 1a and 1b). The experimental
signal was in good agreement with simulations that used the calc-
ulated parameters aN = 15.23 G, aHβ = 23.93 G (g = 2.0061), aH >
aN, being characteristic of trapped carbon-centered radicals derived
from the decomposition of organic hydroperoxides.41 Finally, degra-
dation of the spin-trap itself was also detected in the spectra with the
identification of known oxidation products 10 (aN = 14.01 G,
g = 2.0065) and 11 (aN = 6.90, aHγ = 3.45 G, g = 2.0073; spectra not
shown), and probably sec-alkyl-tert-alkyl nitroxide derivative 12
(aN = 15.53 G, aHβ = 1.60 G, g = 2.0063) (Chart 2).42,43
b. Influence of the Solvent. The same experiments described
above were conducted in 1:1 (v/v) H2O/CH3CN solutions.
Compared to the results obtained in CH3CN, the outlines of the
spectra were modified considerably (Figure 3). In all cases, the
EPR spectra showed the presence of DMPO/R• carbon-cen-
tered radical adducts characterized by hfc of aH > aN. As carbon-
centered radicals were in general not trapped when CH3CN was
used as solvent, it might be expected that in CH3CN oxygen-
centered radicals were trapped before subsequent rearrangement
processes. However, in H2O/CH3CN solutions the spin-trap-
ping process seemed to be slower, allowing enough time for
rearrangements to take place and thus accounting for the
formation of carbon-centered radicals mainly. Only the reaction
of hydroperoxide 4 with FeSO4 permitted the identification of
DMPO/RO•/ROO• adducts, formed before potential rearran-
gement of the oxygen-centered radicals (Figure 3c).
Figure 2. Experimental EPR spectra of DEPMPO spin adducts formed
from reaction of 4 with FeSO4. (a) The trans-DEPMPO/ROO• isomer
is observed with coupling constants aP = 46.45 G, aN = 12.80 G, aHβ
=
7.54 G (g = 2.0065). (b) Half EPR spectrum of the trans-DEPMPO/
ROO• spin adduct recorded at low modulation amplitude (1 G). The
lines were split by weak couplings with the pyrrolidine ring hydrogens.
exhibit a dramatic alternate line width when recorded with low
modulation amplitude.38 In order to test if alkylperoxylradicals were
present in the reaction mixtures, we also performed the spin-trapping
experiments with DEPMPO at low modulation amplitude. As an
example, Figure 2 shows the EPR spectra obtained from the reaction
of 4 with FeSO4 in CH3CN and in the presence of DEPMPO
(modulation frequence 100 kHz, modulation amplitude 1 G).
The EPR spectrum shown in Figure 2a was attributed to spin
adducts arising from the addition of formed alkylperoxyl radicals
ROO• on both faces of DEPMPO, as described in the literature.39,40
This addition can lead to trans and cis diastereoisomers. The main
signal observed, with twelve EPR lines (•) and exhibiting an
alternating line width effect, corresponds to the trans diastereoi-
somer resulting from addition of the alkylperoxyl radical on the less
hindered face of DEPMPO. Indeed, the hfc values found aP = 46.45
G, aN = 12.80 G, and aHβ = 7.54 G (g = 2.0065) matched well with
patterns described for example for the trans spin adducts of tert-
BuOO• or superoxide with DEPMPO and analogues.39,40 In the
recorded spectrum, the small shoulder seen on the left side of some
of the twelve lines probably corresponds to the minor cis diaster-
eoisomer, hidden by the signals of the major trans-DEPMPO/
ROO• isomer.14 When the spectra were recorded at low modula-
tion amplitude (Figure 2b), the trans-DEPMPO/ROO• adduct
clearly displayed the alternating line width mentioned above and
showed the characteristic resolution of the long-range γ-hydrogen
splittings from the pyrrolidine ring (i.e., aHγ = 0.45 and 1.01 G).
This phenomenon has been described to be generated by a slow
down rotation around the OÀO peroxyl bond only possible in the
case of the trans-DEPMPO/ROO• spin adduct. In this way, the
spin adduct acquires different conformational changes resulting in
InH2O/CH3CN, while it was not observed at all in CH3CN, the
monohydroxyl DMPO adduct 9 was clearly identified (Figure 3a,
3c, and 3d), together with the dihydroxyl adduct 10 in the case of
hydroperoxide 4 reactions (Figure 3c and 3d) (Chart 2).
c. Mechanistic Interpretations. The formation of oxygen-
centered radicals ROO•/RO• derived from the reactions of hydro-
peroxides 3 and 4 with Fe(II)/Fe(III) species was demonstrated
by EPR spin-trapping using DMPO and DEPMPO as spin-traps.
However, it was not possible to determine the preferential forma-
tion of alkylperoxyl ROO• or alkoxyl RO• radicals but rather to
6192
dx.doi.org/10.1021/jo200948x |J. Org. Chem. 2011, 76, 6188–6200

The Journal of Organic Chemistry
ARTICLE
Figure 3. Experimental EPR spectra obtained from the reaction of 3 or 4 with FeSO4, (a and c) or hemin (b and d), in the presence of DMPO and in 1:1
(v/v) H2O/CH3CN. Computer simulation of spectra led to the following parameters: (a) Signal (1) aN = 15.92 G, aHβ = 24.35 G (g = 2.0059) for
DMPO/R• carbon-centered adduct; signal (4) aN = 14.20 G, aHβ = 14.20 G (g = 2.0062) for 9. (b) Signal (1) aN = 15.92 G, aHβ = 24.35 G (g = 2.0059)
for DMPO/R• carbon-centered adduct. (c) Signal (1) aN = 15.47 G, aHβ = 23.36 G (g = 2.0062) for DMPO/R• carbon-centered adduct; signal (b) aN =
13.28 G, aHβ = 7.52 G, aHγ = 1.86 G (g = 2.0064) for DMPO/RO• or DMPO/ROO•; signal (4) aN = 14.20 G, aHβ = 14.20 G (g = 2.0062) for 9; signal
(O) aN = 13.74 G (g = 2.0064) for 10; (d) signal (1) aN = 15.47 G, aHβ = 23.36 G (g = 2.0062) for DMPO/R• carbon-centered adduct; signal (4) aN =
14.20 G, aHβ = 14.20 G (g = 2.0062) for 9; signal (O) aN = 13.74 G (g = 2.0064) for 10.
Scheme 4. Competitive Formation of Oxygen-Centered
Radicals in ROOH/Fe(II) Reactions
Scheme 5. Homolytic/Heterolytic OÀO Bond Cleavage of
Hydroperoxides by Fe(III) Porphyrin Complexes
accept the competitive existence of both in the reaction mixtures.
Indeed, it is commonly accepted that hydroperoxides decompose in
the presence of ferrous salts, simultaneously affording alkoxyl RO•
radicals and hydroxide ions in a process similar to the Fenton reaction
(Scheme 4, entry a). However, these alkoxyl radicals can easily
further abstract the hydrogen of the hydroperoxide ROOH function-
ality to lead to alkylperoxyl ROO• radicals (Scheme 4, entry b).
Furthermore, ROO• radicals are known to undergo bimolecular
self-reactions to form tetroxide intermediates (R-OOOO-R)
which decompose irreversibly, leading again to the formation of
alkoxyl RO• radicals and dioxygen (Scheme 4, entry c).10,41 ROO•
and RO• can thus coexist together in the reaction mixtures.
In the case of Fe(III) porphyrin complexes, there is no clear
mechanistic consensus on both the possible homolytic and hetero-
lytic OÀO bond cleavage of the hydroperoxide ROOH function-
ality. Almarsson and Bruice proposed a straight homolytical
cleavage of the OÀO bond of an initially formed Fe(III) porphyr-
in-ROOH complex, leading directly to RO• radicals and to a ferryl-
oxo complex (Scheme 5, pathway A).44 However, Traylor et al.
proposed that the OÀO bond would be cleaved in an heterolytic
way, leading to the generation of ROH and the formation of a high-
valent Fe(IV) oxo porphyrin cation radical intermediate (Scheme 5,
pathway B).45 They also suggested a different mechanism for the
product distribution of OÀO bond homolysis.46 The heterolytic
OÀO bond cleavage being the starting process, they proposed it
can be followed by a fast reaction between the Fe(IV) oxo
porphyrin cation radical intermediate and the hydroperoxide,
leading to the product distribution of OÀO bond homolysis and
to ROO• radicals (Scheme 5, pathway B followed by C). Also, as
shown in Scheme 4 (entry c), ROO• radicals can evolve and result
in the formation of RO• radicals. ROO• and RO• could thus coexist
together in the reaction mixtures.
6193
dx.doi.org/10.1021/jo200948x |J. Org. Chem. 2011, 76, 6188–6200

The Journal of Organic Chemistry
ARTICLE
Scheme 6. Reactivity of Hydroperoxides 3 and 4 with Fe(II)/Fe(III): Potential Oxygen/Carbon-Centered Radicals Revealed by
Spin-Trapping EPR Studies in H2O and H2O/CH3CN Media
For many other authors, one pathway would not necessarily rule
out the other pathway. Partitioning between heterolysis and homo-
lysis would basically be affected by the electronic nature of the iron
porphyrin complexes (i.e., electronic nature of axial ligands) and also
by the electron-withdrawing or -donating character of the ROOH
substituent.47,48 Electron-deficient iron porphyrin complexes show a
propensity to cleave the OÀO bond heterolytically, while electron-
rich iron porphyrin complexes such as TPP-Fe3+ do it homolytically.
It has also been shown that hydroperoxides containing electron-
donating tert-alkyl groups such as t-BuOOH tend to be cleaved
homolytically, whereas electron-withdrawing substituents such the
acyl group in m-CPBA seem to facilitate OÀO bond heterolysis.
This approach leads us initially to favor the hypothesis of a homolytic
cleavage of the OÀO bond of compounds 3 and 4 by TPP-Fe3+ for
example, nevertheless without excluding the heterolysis possibility.
Indeed, unfortunately the present data do not allow us to make a
definite statement concerning a homolytic or heterolytic mechanism
for OÀO bond cleavage when using hemin, for instance, and
therefore on the precise nature of oxygen-centered radicals formed.
The coexistence of both the peroxyl ROO• and the alkoxyl RO•
radicals could thus also be possible in the presence of Fe(III) species.
The rearrangement of oxygen-centered radicals to form R•
carbon-centered radicals is thoroughly discussed in the literature.
Concerning allyloxyl radicals, there is a competition between the
3-exo-trig cyclization leading to the formation of oxiranylcarbinyl
radicals and the β-scission leading to the formation of carbonyl
compounds and alkyl radicals.49 These two processes are ex-
tremely fast, and the relative rate of obtention of each one has
been described to depend on the number and nature of alkyl
substituents present on the carbon atom in a position R to the
oxygen atom. An allyloxyl radical having a dimethylated R-C, such
as the one derived from 3, would have a strong preference for the
3-exo-trig cyclization, while a secondary allyloxyl radical, such as
the one derived from 4, would have a preference for the β-scission
favored by the alkyl chain length in R. Concerning allylperoxyl
radicals, they can follow a 4-exo-trig cyclization by addition to the
double bond and lead to the formation of carbon-centered radicals
with an unstable dioxetane structure known to cleave to yield
carbonyl compounds.10,12 Finally, the abstraction of hydrogens
from different allylic positions is also an important pathway for the
formation of carbon-centered radicals from RO• and ROO•
radicals issued from hydroperoxides.12 In this work, the spin-
trapping experiments carried out with 3 and 4 seemed to show,
except if other signals were overlaid, the formation of a single species
of carbon-centered radicals. Unfortunately, we could not determine
precisely the chemical structure corresponding to that species.
As described above, different potential carbon-centered radicals could
thus be formed and are shown in Scheme 6. At this stage, it was not
possible to distinguish between oxiranylcarbinyl radicals, dioxetane
radicals, and radicals derived from allylic hydrogen abstraction.
6194
dx.doi.org/10.1021/jo200948x |J. Org. Chem. 2011, 76, 6188–6200

The Journal of Organic Chemistry
ARTICLE
Figure 4. Reaction of 3 with N-Ac-His-OMe studied by LC-ESI-MS/MS, and potential chemical structures proposed for adducts formed. (a) LC
chromatogram for the reaction catalyzed by FeSO4. Peaks at retention times (tR) of 0.97, 4.12, and 4.52 min had associated m/z values of 226, 396, and
380, respectively. (b) ESI-MS/MS spectrum obtained from the analysis of m/z 380. (c) ESI-MS/MS spectrum obtained from the analysis of m/z 396.
(d) ESI-MS/MS spectrum obtained from the analysis of m/z 226.
Aside from detected known DMPO oxidation products 10 and 11,
the hydroxylated DMPO derivative 9 was also observed but only in
H2O/CH3CN mixtures. This nitroxide radical would result from the
nucleophilic addition of water to the double bond of DMPO
complexed to ferric ions present in solution, as previously described.50
Reactivity with N-Acetylhistidine Methyl Ester Catalyzed
by Fe(II)/Fe(III). Analyses of amino acid residues from enzymes
and of their related antioxidant activity have demonstrated that
those having an aromatic moiety in the side chain are prone to
radical reactions. The aromatic imidazole ring of histidine, a
common coordinating ligand in metalloproteins and radical
enzymes, has two unsaturated double bonds and therefore could
react with metabolic free radicals (R•) by a free radical addition
reaction. It has been proposed that in electron-transfer processes
and biological redox reactions, including oxidative stress, histidine
radicals are formed as transient intermediates. EPR spectroscopic
evidence of the oxidation of histidine in aqueous solution by
addition of OH• radical to position C5 of the imidazole ring has
been reported in an H2O2ÀFenton system over a large range of
pH values, attesting to its antioxidant potential.51 The formation of
OH• radicals and other reactive oxygen species in the reaction of
nickel(II)-histidine complexes with hydroperoxides has also been
reported.52 Very recently, the reactivity of a large excess of
hydroperoxide 4 with protected histidine in the presence of a
catalytic amount of TPP-Fe3+ has been reported in the literature.53
The authors showed for the first time the formation of an adduct
with this amino acid and, up to previous studies, proposed the
probable addition of a carvone moiety (major decomposition
product of 4) to histidine. Studies with angiotensin I, having two
histidine residues in the peptide sequence, corroborated this result
through the detection of an adduct bound at one of the two
residues. Thus, we also tried to trap radical intermediates issued
from allylic hydroperoxides 3 and 4 using a histidine derivative as
trapping agent. The reactions were conducted in the presence of a
catalytic amountofeitherFe(II) or Fe(III). An excess ofthe amino
acid was used, as these are conditions supposed to happen in vivo.
Both the amine and carboxyl groups of histidine were protected to
mimic its structure as part of a protein sequence but also to favor
the reactivity of the side chain exclusively and to avoid the eventual
reactivity of the amine terminus with decomposition products
of the test compounds that contained carbonyl chemical groups
(i.e., carvone for hydroperoxide 4 via Schiff base formation).
Compounds 3 or 4 and N-acetylhistidine methyl ester (N-Ac-
His-OMe, 2 equiv) in a deaerated 1:1 (v/v) H2O/CH3CN
mixture were treated with a catalytic amount of FeSO4 or FeCl3
(0.1 equiv), at room temperature. TLC indicated that the hydro-
peroxides were not completely consumed even after 7 days. The
reactions were then stopped, and the solvent was removed under
reduced pressure. The crude products were analyzed by LC-ESI-
MS/MS.
6195
dx.doi.org/10.1021/jo200948x |J. Org. Chem. 2011, 76, 6188–6200

The Journal of Organic Chemistry
ARTICLE
Figure 5. Reaction of 4 with N-Ac-His-OMe as studied by LC-ESI-MS/MS, and potential chemical structures proposed for adducts formed. (a) LC
chromatrogram for the reaction catalyzed by FeCl3. Peaks at retention times (tR) of 21.30 and 41.78 min had associated m/z values of 380 and 362,
respectively. (b) ESI-MS/MS spectrum obtained from the analysis of m/z 362. (c) ESI-MS/MS spectrum obtained from the analysis of m/z 380.
a. Allylic Hydroperoxide 3. The reaction of 3 with N-Ac-His-
OMe in the presence of Fe(II)/Fe(III) was rather slow (reaction
stopped after 7 days without complete disappearance of 3). The
LC-MS spectra profiles of the crude products were practically
identical under both experimental conditions and showed the
formation of a large number of compounds. Among all recorded
LC-peaks, three of them had a related MS spectrum that could fit
with the generation of N-Ac-His-OMe adducts, with [M + H]+
molecular ions at m/z 226, m/z 396, and m/z 380 values,
respectively (Figure 4a). The peak corresponding to m/z 226
was quite important, whereas peaks at m/z 396 and 380 were of
lower magnitude. To get further information, complementary
ESI-MS/MS analyses of these signals were carried out. First, all
ESI-MS/MS spectra showed a fragment ion at m/z 110 due to
the immonium ion related to histidine (Figure 4bÀd).54 Im-
monium ions, of general structure [H2NdCH-R]+ where R is
the amino acid side chain, are characteristic fingerprints of amino
acid fragmentations in tandem mass spectrometry. This result
was a first indication of the formation of adducts involving N-Ac-
His-OMe and hydroperoxide 3. Also, [M + H]+ molecular ions at
m/z 380 and at m/z 396 produced a fragment ion with an
associated m/z value of 212, characteristic of a loss of N-Ac-His-
OMe (Figure 4b and 4c). From a mechanistic point of view, and
on the basis of the EPR results described above, the OÀO bond
cleavage of the hydroperoxide functionality by Fe(II) and Fe(III)
can form allyloxyl radicals RO•. These can further evolve to
carbon-centered radicals either by allylic hydrogen abstraction or
by 3-exo-trig cyclization leading to the formation of oxiranylcar-
binyl radicals. In both cases, a radical-based reaction with a side
chain of N-Ac-His-OMe would give adducts with an associated
value of m/z 380. Therefore, 13 and 14 could be principally
suggested for m/z 380 (Figure 4). However, further studies
allowed neither the elucidation of the exact chemical structure of
the adduct nor the precise position of addition on the side chain
of N-Ac-His-OMe. The molecular ion at m/z 396 could corre-
spond to compound 15, the mass increment of 16 with regard to
m/z 380 indicating the presence of an oxygen atom in the
addition. Indeed, the hydroperoxide OÀH bond cleavage can
form allylperoxyl radicals ROO• that can also evolve through
hydrogen abstraction to form carbon radicals in allylic positions.
The m/z value of 226 fitted well with the addition of a methyl
group (Δmass 15) into the side chain of N-Ac-His-OMe to form
adduct 16. Mechanistically, the RO• radicals formed could in
parallel follow a β-scission process, releasing, in this way, reactive
methyl radicals that could be trapped by the amino acid. But this
result remains just a hypothesis, as we were not able to exactly
elucidate the precise position of methyl addition on the side
chain of N-Ac-His-OMe, and ESI-MS/MS analysis of the mo-
lecular ion at m/z 226 only corroborated the loss of the acetyl
group of methylated N-Ac-His-OMe that formed the fragment
ion at m/z 184 (Figure 4d). Even so, reaction of 3 with N-Ac-His-
OMe in Fe(II)/Fe(III) systems could be representative of the
three main possible mechanisms explaining the evolution of the
initially formed RO• radicals to carbon-centered radicals, i.e.,
allylic hydrogen abstraction, 3-exo-trig cyclization, and β-scission
processes. Finally, no oxidation products of N-Ac-His-OMe were
detected in any of the experimental conditions tested.
b. Allylic Hydroperoxide 4. As indicated above, the reaction of
4 with N-Ac-His-OMe, catalyzed by Fe(II)/Fe(III) ions, was also
rather slow (reaction stopped after 7 days without complete
disappearance of 4). The LC-MS spectra profiles of the crude
products were practically identical in both experimental condi-
tions. Among the recorded peaks, two of them had an associated
MS spectrum that could fit with the formation of a N-Ac-His-
OMe adduct. The observed value of m/z 362 suggested the
formation of an adduct in the presence of either FeSO4 or FeCl3,
detected at relatively high amounts. Another potential adduct at
m/z 380 was only seen in the presence of FeCl3, detected at
smaller amounts (Figure 5a). Complementary ESI-MS/MS
analyses of these signals were carried out. As described for N-
Ac-His-OMe adducts of hydroperoxide 3, all ESI-MS/MS spec-
tra showed the fragment ion at m/z 110 due to the immonium
ion related to histidine (Figure 5b and 5c). Morever, both
[M + H]+ molecular ions at m/z 362 and at m/z 380 produced
a fragment ion with an associated m/z value of 212, characteristic
of a loss of N-Ac-His-OMe (Figure 5b and 5c). Following the
same reasoning as for hydroperoxide 3, chemical structures 17
and 18 could be suggested for m/z 362. The OÀO bond cleavage
6196
dx.doi.org/10.1021/jo200948x |J. Org. Chem. 2011, 76, 6188–6200

The Journal of Organic Chemistry
ARTICLE
of the hydroperoxide functionality by Fe(II) and Fe(III) can
form allyloxyl radicals RO•. These can further evolve to carbon-
centered radicals either by allylic hydrogen abstraction or by
3-exo-trig cyclization, leading to the formation of oxiranylcarbi-
nyl radicals. In both cases, a radical-based reaction with the side
chain of N-Ac-His-OMe would give adducts with an associated
value of m/z 362 (Δmass 150). Therefore, 17 and 18 could be
principally suggested for m/z 362 (Figure 5). However, further
studies allowed neither the elucidation of the exact chemical
structure of the adduct nor the precise position of addition on the
side chain of N-Ac-His-OMe. In the case of m/z 380, the mass
increment of 18 with regard to m/z 362 would correspond to the
addition of a molecule of H2O. It may then be suggested that
nucleophilic addition of H2O, present in the reaction solvent, on
the electrophilic epoxide functionality of 18 afforded 19 at m/z
380 (Figure 5). Adduct 19 could thus be indirect evidence of the
formation of 18. Being only observed when FeCl3 catalyzed the
reaction, it could be proposed that the formation of oxiranylcar-
binyl radicals was favored when using these experimental condi-
tions. However, the simultaneous formation of allyloxyl radicals
which would afford 17-like adducts cannot be excluded. Finally,
as for hydroperoxide 3, no oxidation products of N-Ac-His-OMe
were detected under any of the experimental conditions tested.
protein modifications may thus arise that could lead to
sensitization.
’ EXPERIMENTAL SECTION
Caution: Skin contact with allylic hydroperoxides must be avoided
because these compounds are skin-sensitizing substances. Hydrogen
peroxide and allylic hydroperoxides must be handled with care, as
hazardous side reactions are possible.
General Experimental Chemical Procedures. All reagents
were obtained from commercial sources and used as received. Com-
mercial linalool 1 and (À)-carveol 7 were available as mixtures of
isomers. DMF was dried over phosphorus pentoxide prior to vacuum
distillation. Air- or moisture-sensitive reactions were conducted in flame-
dried glassware under an atmosphere of dry argon. All reactions leading
to the formation of hydroperoxides were carried out protected from
daylight. The reactions were followed by TLC performed on 0.25 mm
silica gel plates (60F254). After migration, the TLC plates were inspected
under UV light (254 nm) and then sprayed with a solution containing
phosphomolybdic acid (5 g), cerium(IV) sulfate (2 g), and sulfuric acid
(12 mL) in water (188 mL), or a solution containing p-anisaldehyde
(0.5 mL), o-anisaldehyde (0.5 mL), sulfuric acid (5 mL), and ethanol
(8 mL) in acetic acid (100 mL), followed by heating. Column chromato-
graphy purifications were performed using silica 60 (Geduran, 40À63
μm) or previously neutralized silica gel. Neutralized silica was prepared
by adding to an homogeneous water solution of silica 60 a saturated
solution of NaHCO3 until a pH of about 8. After decantation, the silica
precipitate was washed with water to reach a pH of 7, filtered, and then
dried in a drying oven for at least 24 h. 1H and 13C NMR spectra were
recorded at 300 and 75 MHz, respectively. The chemical shifts (δ) are
reported in ppm and are indirectly referenced to TMS via the solvent
’ CONCLUSIONS
To complement our previous studies on allergenic allylic
hydroperoxides and their supposed ability to form immunogenic
structures by reacting with skin proteins through radical mecha-
nisms, we now confirm, by using the spin-trapping technique
combined with EPR spectroscopy, the formation of potentially
reactive oxygen- and carbon-centered radicals derived from
7-hydroperoxy-3,7-dimethylocta-1,5-dien-3-ol 3 and 2-hydro-
peroxylimonene 4. Both hydroperoxides have been identified
as being responsible for the sensitizing potential of autoxidized
linalool and R-(+)-limonene, respectively. Radical initiation by
using Fe(II)/Fe(III) redox cycles, common in biological media,
confirmed that the generation of these radical intermediates
could be possible in vivo, in the epidermis, once the intact
hydroperoxide molecules have penetrated into the skin. Even if
with the spin-trapping EPR technique the exact chemical struc-
ture of carbon-centered radicals formed could not be completely
elucidated, when histidine was used as a model amino acid for
radical reactions, adducts were formed by reaction of different
radical intermediates issued from 3 and 4 with the aromatic
imidazole ring. The LC-ESI-MS/MS studies conducted on the
radical reactivity of hydroperoxides 3 and 4 with histidine thus
completed the spin-trapping EPR data. Reactions with histidine
indicated that not just a single species of carbon-centered radicals
was formed, as suggested when analyzing the EPR data, but
rather many of them. The acyclic tertiary allylic hydroperoxide 3
illustrated ideally the case in which the first formed allyloxyl
radical evolves subsequently through the three main mechanisms
(hydrogen abstraction, cyclization, and β-scission) to the forma-
tion of carbon-centered radicals (allylic, oxiranylcarbinyl,
methyl), all reactive toward histidine, whereas for the cyclic
and secondary allylic hydroperoxide 4, the possibilities of re-
arrangement were more limited. Significantly, this indicates that
if a certain number of reactive radical intermediates can issue
from these compounds depending on their chemical structure,
different potentially immunogenic protein chemical modifica-
tions could be induced by these hydroperoxides. Different
1
signal (CDCl3: δ 1H = 7.26, δ 13C = 77.16). In the H NMR spectra
multiplicities are denoted as s (singlet), d (doublet), dd (doublet of
doublets), m (multiplet), and br (broad). The different types of carbon
in the structures were identified by the DEPT-135 technique. The
analytical data for known compounds matched the literature data.
Schenck Reaction: (5E)-7-Hydroperoxy-3,7-dimethylocta-1,5-
dien-3-ol (3) and 6-Hydroperoxy-3,7-dimethylocta-1,7-dien-3-ol
(5). A microemulsion was prepared by adding, drop by drop, an aqueous
solution of Na2MoO4 (3.38 g in 18 mL of distilled water) into a suspension
of sodium dodecyl sulfate (27.9 g) in butanol (34 mL) and dichloro-
methane (166 mL). The microemulsion became clear after 5 min stirring.
A solution of linalool 1 (2.83 g, 17.8 mmol, 1 equiv) in the microemulsion
(297 g) was treated at rt with a first portion of hydrogen peroxide (35 wt %
in water, 1 mL). The brick red mixture was stirred for about 20 min, until it
became light yellow. Another 14 portions of hydrogen peroxide (35 wt % in
water, approximately 1 mL each) were successively added every 10 min. In
the end, a total of 15.3 mL of aqueous hydrogen peroxide (35%) was used
(179 mmol, 10 equiv). The light yellow mixture was stirred overnight at rt
and became completely clear. The solvent was removed in vacuo and the
crude product obtained dissolved in dichloromethane (900 mL). The
suspension was stirred vigorously during 2 days and filtered. The filtrate,
partially evaporated in vacuo, was washed with water (3 Â 100 mL), and
the aqueous layers were extracted with dichloromethane (1 Â 100 mL). The
combined organic layers were dried over MgSO4, filtered, and evaporated in
vacuo. Purification by flash chromatography (petroleum ether/EtOAc 8:2,
then 6:4) furnished a mixture 4:6 of compounds 3 and 5 as a yellowish
oil (2.64 g, 14.2 mmol, 80%). Rf = 0.24 (petroleum ether/EtOAc 7:3).
1
Compound 3 (mixture of enantiomers): H NMR (300 MHz, CDCl3)
δ 1.29 (s, 3 H), 1.32 (s, 6 H), 2.29 (m, 2 H), 5.06 (dd, J = 10.8, 1.1 Hz, 1 H),
5.19 (dd, J = 17.4, 1.1 Hz, 1 H), 5.58À5.73 (m, 2 H), 5.92 (dd, J = 17.3, 10.7
Hz, 1 H); 13C NMR (75 MHz, CDCl3) δ 24.3, 24.5, 27.8, 45.3, 72.9, 82.1,
112.3, 126.6, 138.1, 144.8; CAS registry number [51276-32-5]. Compound
1
5 (mixture of 4 diastereomers): H NMR (300 MHz, CDCl3) δ 1.28
6197
dx.doi.org/10.1021/jo200948x |J. Org. Chem. 2011, 76, 6188–6200

The Journal of Organic Chemistry
ARTICLE
(s, 6 H), 1.51À1.66 (m, 8 H), 1.72 (bs, 6 H), 4.29 (m, 2 H), 4.99 (m, 4 H),
5.07 (dd, J = 10.8, 1.1 Hz, 2 H), 5.20 (dd, J = 17.4, 1.1 Hz, 1 H), 5.21 (dd, J =
17.4, 1.6 Hz, 1 H), 5.87 (dd, J = 17.4, 10.8 Hz, 1 H), 5.88 (dd, J = 17.4, 10.8
Hz, 1 H); 13CNMR(75MHz, CDCl3) δ17.6 (2 C), 25.2 (2 C), 28. 1, 28.2,
37.7, 37.9, 73.3 (2 C), 89.5, 89.6, 112.3 (2 C), 114.2, 114.3, 143.7 (2 C),
144.7, 144.8; CAS registry number [51276-31-4]. These spectroscopic data
are in agreement with those reported.7,17
[1045725-54-9]. These spectroscopic data are in agreement with those
reported.14
EPR Spin-Trapping Studies. DEPMPO was synthesized as
reported in the literature.55 All other reagents were obtained from
commercial sources and used as received. Spectrophotometric grade
CH3CN (99.8%) was used, and aqueous solutions were prepared with
deionized water. EPR spectra were recorded on a spectrometer equipped
with an X-band microwave bridge (10 GHz). The standard spectrometric
settings were modulation frequency 100 kHz, microwave power 7 mW,
modulation amplitude 1.0 G, 2.2, or 7.1 G, receiver gain 2 Â 106, scan
width 100 or 120 G, and conversion time 25 ms. The g calibration
standard was a strong pitch of known g factor (2.0028).56 For these EPR
spin-trapping experiments in iron systems, special glassware was used
where the hydroperoxide and the spin-trap in solution were initially in a
separate compartment (round-bottom container) from that of the iron
reagent (capillary tube). Compound 3 (5 mg, 27 μmol) or compound 4
(5 mg, 30 μmol) and the spin-trap, DMPO(5 mg, 43μmol) or DEPMPO
(5 mg, 21 μmol), were dissolved in 500 μL of deaerated CH3CN or in
500 μL of a deaerated 1:1 (v/v) H2O/CH3CN mixture. The solution was
placed in the round-bottom compartment of the EPR glassware. In
the capillary tube the Fe(II)/Fe(III) reagent was introduced, FeSO4
(heptahydrate, 3 mg, 11 μmol), TPP-Fe3+ (3 mg, 4 μmol), or hemin
(3 mg, 4 μmol). The compartments were interlinked to allow the
subsequent mix of the reagents. To avoid mixing the hydroperoxides
too rapidly when being poured over the iron derivative, a layer of polyvinyl
alcohol was used as a filter. In order to carry out the experiments in the
absence of oxygen, to avoid the formation of secondary oxidation pro-
ducts, a vacuum line using the standard freezeÀthaw technique degassed
the whole system. Then, the solution containing 3 or 4 was poured into
the capillary tube containing the layer of polyvinyl alcohol and iron
underneath. An EPR spectrum of the reaction mixture was recorded
directly by placing the capillary tube in the spectrometer cavity. Hyperfine
splitting assignments were obtained by means of computer simulation
using the Bruker WINEPR SimFonia software (version 1.25; Rheinstet-
ten, Germany) and were in agreement with those reported in the
literature.14,57À60
Reaction with N-Acetylhistidine Methyl Ester Catalyzed by
Fe(II)/Fe(III). Allylic hydroperoxide 3 (35 mg, 0.19 mmol) or 4 (300 mg,
1.78 mmol) were dissolved in a deaerated 1:1 (v/v) H2O/CH3CN
mixture (10 mL for 3, 100 mL for 4). N-Ac-His-OMe was added to the
solutions (2 equiv) together with a catalytic amount of FeCl3 or FeSO4
heptahydrate (0.1 equiv). The reactionmixtures were continuouslystirred
at room temperature. The reactions were monitored by TLC (0.25 mm
silica gel plates, 60F254). After migration, the TLC plates were inspected
under UV light (254 nm) and then sprayed with a solution containing
phosphomolybdic acid (5 g), cerium(IV) sulfate (2 g) and sulfuric acid
(12 mL) in water (188 mL), or a solution containing p-anisaldehyde
(0.5 mL), o-anisaldehyde (0.5 mL), sulfuric acid (5 mL), and ethanol
(8 mL) in acetic acid (100 mL), followed by heating. The reactions were
stopped after 7 days, as the starting hydroperoxides were still present in the
reaction mixtures and the reactions did not evolve after that time. In order
to remove iron salts, the mixtures were filtered on Celite (Celite 545).
The solvent was removed under reduced pressure, and the crude products
obtained were analyzed by LC-ESI-MS associated with ESI-MS/MS. LC-
ESI-MS analyses were performed using a HPLC system equipped with a
binary pump, an automatic sample injector, and a diode array absorbance
detector scanning 190 to 700 nm. The samples were subjected to reverse
phase chromatography on a C18 column (1 Â 100 mm; Thermo Hypersil
Gold; 1.9 μm particle size) at a flow rate of 0.2 mL/min. Samples were
eluted from the column using a mobile phase B (0.01% formic acid
in H2O) and a mobile phase A (CH3CN) with a gradient. The gradient
started at 98% B and was decreased to 10% B after 20 min, was at 10% B
during 10 min, and was increased again to 98% B in the last 10 min. The
injection volume was 0.1 mL. After passing through the diode array
SelectiveProtection:(5E)-7-(tert-Butyldiphenylsilylperoxy)-
3,7-dimethylocta-1,5-dien-3-ol (6). In a flame-dried, two-necked,
round-bottomed flask, under argon, were introduced a mixture of 3 and
5 (200 mg, 1.07 mmol, 1 equiv), dry DMF (5 mL), and imidazole
(99.5%, 147 mg, 2.15 mmol, 2 equiv). tert-Butyldiphenylsilyl chloride
(98%, 0.34 mL, 1.28 mmol, 1.2 equiv) was added, and the reaction
mixture was stirred at rt. After 10 days, the yellow reaction mixture was
quenched by the addition of water (30 mL), followed by extraction with
pentane (3 Â 50 mL). The organic layers were combined, dried over
MgSO4, filtered, and evaporated in vacuo. The crude product was
purified by flash chromatography (petroleum ether/EtOAc 9:1) to obtain
a yellow oil (377 mg) corresponding to compound 6 together with a tiny
amount of tert-butyldiphenylsilyl chloride. This mixture was directly used
in the next step. Rf = 0.26 (petroleum ether/EtOAc 9:1); mixture of
enantiomers; 1H NMR (300 MHz, CDCl3) δ 1.11 (s, 9 H), 1.24 (s, 3 H),
1.25 (s, 3 H), 1.26 (s, 3 H), 2.20À2.28 (m, 2 H), 5.02 (dd, J = 10.8, 1.3 Hz,
1 H), 5.17 (dd, J = 17.2, 1.3 Hz, 1 H), 5.51À5.68 (m, 2 H), 5.91 (dd, J =
17.4, 10.8Hz, 1 H), 7.39 (m, 6 H), 7.71 (d, J = 1.5 Hz, 2 H), 7.74(d, J = 1.3
Hz, 2 H);13CNMR(75MHz, CDCl3) δ 19.7, 24.6, 24.9, 26.7, 27.5(3 C),
45.5, 72.5, 82.6, 112.1, 124.8, 127.6 (4 C), 129.8 (2 C), 133.4 (2 C), 136.0
(4 C), 139.7, 144.8; HRMS (ESI) m/z calcd for C26H36O3Si [M + Na+]
447.2330, found 447.2300.
(5E)-7-Hydroperoxy-3,7-dimethylocta-1,5-dien-3-ol (3) from
(5E)-7-(tert-Butyldiphenylsilylperoxy)-3,7-dimethylocta-1,5-dien-
3-ol (6). Compound 6 obtained in the previous step (377 mg), together
with a tiny amount of tert-butyldiphenylsilyl chloride, was dissolved in
methanol (25 mL) and the solution cooled to 0 °C. Potassium hydroxide
(241 mg, 4.30 mmol, 4 equiv related to a 1.07 mmol mixture of 3 and 5)
was added all at once. The reaction mixture was allowed to warm to rt
and stirred overnight. Water was added (125 mL), and the solution
became milky white. The mixture was then neutralized with an aqueous
solution of chlorhydric acid (2%, 70 mL) and extracted with diethyl
ether (3 Â 250 mL). The organic layers were combined, dried over
MgSO4, filtered, and evaporated in vacuo. The crude product was
purified by flash chromatography (petroleum ether/EtOAc 65:35) to
obtain 3 as a yellow oil (77 mg, 0.41 mmol, 96% over two steps, calculated
from 3in a 4:6 mixture of 3and 5). Rf = 0.24 (petroleum ether/EtOAc 7:3);
mixture of enantiomers; 1H NMR (300 MHz, CDCl3) δ 1.29 (s, 3 H),
1.32 (s, 6 H), 2.29 (m, 2 H), 5.06 (dd, J = 10.8, 1.1 Hz, 1 H), 5.19 (dd, J =
17.4, 1.1 Hz, 1 H), 5.58À5.73 (m, 2 H), 5.92 (dd, J = 17.3, 10.7 Hz, 1 H);
13C NMR (75 MHz, CDCl3) δ 24.3, 24.5, 27.8, 45.3, 72.9, 82.1, 112.3,
126.6, 138.1, 144.8; CAS registry number [51276-32-5]. These spectro-
scopic data are in agreement with those reported.17
2-Hydroperoxylimonene (4). To an aqueous solution of hydrogen
peroxide (50%, 85 mL) were added, at 0 °C, some drops of concentrated
sulfuric acid and (À)-carveol 7 (97%, 6.0 g, 38.2 mmol). The reaction
mixture was vigorously stirred at 0 °C during 2 days, followed by extraction
with pentane (5 Â 200 mL). The organic layers were combined, dried over
MgSO4, filtered, and evaporated in vacuo. The crude product was purified by
flash chromatography on neutralized silica gel (pentane/diethyl ether 8:2) to
obtain 4 as a colorless oil (3.08 g, 18.3 mmol, 48%). Rf = 0.21 (petroleum
ether/EtOAc 95:5); mixture of diastereomers; 1H NMR (300 MHz, CDCl3)
δ 1.40À1.51 (m, 2 H), 1.75 (s, 6 H), 1.80 (s, 6 H), 2.13À2.38 (m, 8 H),
4.36 (s, 1 H), 4.53 (m, 1 H), 4.74 (s, 2 H), 4.75 (s, 2 H), 5.64 (m, 1 H), 5.75
(d, J = 5.0 Hz, 1 H), 7.67 (s, 1 H), 8.02 (s, 1 H); 13C NMR (75 MHz,
CDCl3) δ 19.3, 21.4, 20.6, 21.0, 31.0, 31.3 (2 C), 32.6, 35.3, 40.7, 82.7, 84.4,
109.2, 109.4, 127.2, 129.4, 129.8, 132.9, 148.8, 149.3; CAS registry number
6198
dx.doi.org/10.1021/jo200948x |J. Org. Chem. 2011, 76, 6188–6200

The Journal of Organic Chemistry
ARTICLE
absorbance detector, the eluent was directed to a connected ion trap mass
spectrometer with a standard electrospray source. The ionization method
used was ESI in the positive ion mode (heated capillary temperature
325 °C, gas flow 10 L/min, nebulizer gas 40 psi). Full scan mass spectra
were acquired in the profile mode scanning m/z 50 to 1000. For the ESI-
MS/MS studies, the mass spectrometer was equipped with the ESI source
used for the ESI-MS studies described above and with the same source
parameter settings. The ionization method used was ESI in the positive
ion mode. The isolation width was of 4 m/z. All data were processed using
Agilent MassHunter Qualitative Analysis software.
(15) Johansson, S. G. H.; Emilsson, K.; Grøtli, M.; B€orje, A. Chem.
Res. Toxicol. 2010, 23, 677–688.
(16) Szpilman, A. M.; Korshin, E. E.; Hoos, R.; Posner, G. H.; Bachi,
M. D. J. Org. Chem. 2001, 66, 6531–6540.
(17) Sk€old, M.; B€orje, A.; Matura, M.; Karlberg, A.-T. Contact
Dermatitis 2002, 46, 267–272.
(18) Alberti, M. N.; Orfanopoulos, M. Tetrahedron 2006, 62,
10660–10675.
(19) Aubry, J. M.; Cazin, B. Inorg. Chem. 1988, 27, 2013–2014.
(20) Nardello, V.; Bouttemy, S.; Aubry, J. M. J. Mol. Catal. 1997,
117, 439–447.
(21) Nardello, V.; Caron, L.; Aubry, J. M.; Bouttemy, S.; Wirth, T.;
Chantu, R. S. M.; Adam, W. J. Am. Chem. Soc. 2004, 126, 10692–10700.
(22) Aubry, J. M.; Adam, W.; Alsters, P. L.; Borde, C.; Queste, S.;
Marko, J.; Nardello, V. Tetrahedron 2006, 62, 10753–10761.
(23) Clark, G. R.; Nikaido, M. M.; Fair, C. K.; Lin, J. J. Org. Chem.
1985, 50, 1994–1996.
(24) Frimer, A. A. J. Org. Chem. 1977, 42, 3194–3196.
(25) Anderson, G. H. J. Can. Chem. 1968, 46, 1561–1570.
(26) Mutterer, V.; Gimꢀenez-Arnau, E.; Karlberg, A.-T.; Lepoittevin,
J.-P. Chem. Res. Toxicol. 2000, 13, 1028–1036.
’ ASSOCIATED CONTENT
Supporting Information. 1H and 13C NMR spectra of all
S
b
compounds. HRMS (ESI) for compound 6. This material is
available free of charge through the Internet at http://pubs.acs.
org. This material is available free of charge via the Internet at
http://pubs.acs.org.
’ AUTHOR INFORMATION
(27) Clement, J.-L.; Gilbert, B. C.; Ho, W. F.; Jackson, N. D.;
Newton, M. S.; Silvester, S.; Timmins, G. S.; Tordo, P.; Whitwood,
A. C. J. Chem. Soc., Perkin Trans. 2 1998, 1715–1717.
(28) Dussault, P. In Active Oxygen in Chemistry; Foote, C. S.,
Valentine, J. S., Greenberg, A., Liebman, J. F., Eds.; Blackie Academic
& Professional: London, 1995; pp 141À203.
Corresponding Author
*E-mail: egimenez@unistra.fr.
’ ACKNOWLEDGMENT
(29) Labeque, R.; Marnett, L. J. J. Am. Chem. Soc. 1987, 109,
2828–2829.
We gratefully acknowledge Firmenich SA (Geneva, Switzerland)
for financial support to D.K. We also thank Patrick Wehrung
(Pharmaceutical Department, University of Strasbourg, France)
for his valuable technical support on the LC-ESI-MS/MS studies.
(30) Stolze, K.; Udilova, N.; Nohl, H. Free Radical Biol. Med. 2000,
29, 1005–1014.
(31) Stolze, K.; Udilova, N.; Nohl, H. Acta Biochim. Pol. 2000, 47,
923–930.
(32) Davies, M. J.; Hawkins, C. L. Free Radical Biol. Med. 2004, 36,
1072–1086.
’ REFERENCES
(33) Van der Zee, J.; Barr, D. P.; Mason, R. P. Free Radical Biol. Med.
1996, 20, 199–206.
(34) Wilcox, A. L.; Marnett, L. J. Chem. Res. Toxicol. 1993, 6, 413–416.
(35) Dikalov, S. I.; Mason, R. P. Free Radical Biol. Med. 1999, 27,
864–872.
(1) Matura, M.; Sk€old, M.; B€orje, A.; Andersen, K. E.; Bruze, M.;
Frosch, P.; Goossens, A.; Johansen, J. D.; Svedman, C.; White, I. R.;
Karlberg, A.-T. Contact Dermatitis 2005, 52, 320–328.
(2) Rastogi, S. C.; Johansen, J. D.; Frosch, P.; Mennꢀe, T.; Bruze, M.;
Lepoittevin, J.-P.; Dreier, B.; Andersen, K. E.; White, I. R. Contact
Dermatitis 1998, 38, 29–35.
(36) Guo, Q.; Qian, S. Y.; Mason, R. P. J. Am. Soc. Mass Spectrom.
2003, 14, 862–871.
(3) Rastogi, S. C.; Heydorn, S.; Johansen, J. D.; Basketter, D. A.
(37) Davies, M. J.; Slater, T. F. Biochem. J. 1986, 240, 789–795.
(38) Frejaville, C.; Karoui, H.; Tuccio, B.; Le Moigne, F.; Culcasi,
M.; Pietri, S.; Lauricella, R.; Tordo, P. J. Med. Chem. 1995, 38, 258–265.
(39) Clement, J.-L.; Barbati, S.; Frejaville, C.; Rockenbauer, A.;
Tordo, P. J. Chem. Soc., Perkin Trans. 2 2001, 1471–1475.
(40) Clement, J.-L.; Finet, J.-P.; Frejaville, C.; Tordo, P. Org. Biomol.
Chem. 2003, 1, 1591–1597.
(41) Chamulitrat, W.; Takahashi, N.; Mason, R. P. J. Biol. Chem.
1989, 264, 7889–7899.
(42) Makino, K.; Hagi, A.; Ide, H.; Murakami, A.; Nishi, M. Can. J.
Chem. 1992, 70, 2818–2827.
Contact Dermatitis 2001, 45, 221–225.
(4) Groot, A. C.; Frosch, P. J. Contact Dermatitis 1997, 36, 57–86.
(5) Karlberg, A.-T.; Shao, L. P.; Nilsson, U.; G€afvert, E.; Nilsson,
J. L. G. Arch. Dermatol. Res. 1994, 286, 97–103.
(6) Sk€old, M.; B€orje, A.; Harambasic, E.; Karlberg, A.-T. Chem. Res.
Toxicol. 2004, 17, 1697–1705.
(7) B€acktorp, C.; Johnson Wass, J. R. T.; Panas, I.; Sk€old, M.; B€orje,
A.; Nyman, G. J. Phys. Chem. A 2006, 110, 12204–12212.
(8) Lepoittevin, J.-P. Molecular aspects of allergic contact dermatitis.
In Contact Dermatitis; Frosch, P. J., Mennꢀe, T., Lepoittevin, J.-P., Eds.;
Springer-Verlag: Berlin, 2006; pp 45À68.
(43) Reszka, K.; Chignell, C. F. Free Radical Res. Commun. 1991,
14, 97–106.
(9) Rustemeyer, T.; van Hoogstraten, I. M. W.; von Blomberg, B .M .
E.; Scheper, R. J. Mechanisms in allergic contact dermatitis. In
Contact Dermatitis; Frosch, P. J., Mennꢀe, T., Lepoittevin, J.-P., Eds.;
Springer-Verlag: Berlin, 2006; pp 11À43.
€
(44) Almarsson,O.;Bruice,T.C.J. Am. Chem. Soc. 1995,117,4533–4544.
(45) Traylor, T. G.; Tsuchiya, S.; Byun, Y.-S.; Kim, C. J. Am. Chem.
Soc. 1993, 115, 2775–2781.
(46) Traylor, T. G.; Kim, C.; Fann, W.-P.; Perrin, C. L. Tetrahedron
(10) Gimꢀenez-Arnau, E.; Haberkorn, L.; Grossi, L.; Lepoittevin, J.-P.
Tetrahedron 2008, 64, 5680–5691.
1998, 54, 7977–7986.
(11) Bezard, M.; Gimꢀenez-Arnau, E.; Meurer, B.; Grossi, L.; Lepoittevin,
(47) Nam, W.; Han, H. J.; Oh, S.-Y.; Lee, Y. J.; Choi, M.-H.; Han, S.-Y.;
Kim, C.; Woo, S. K.; Shin, W. J. Am. Chem. Soc. 2000, 122, 8677–8684.
(48) Labeque, R.; Marnett, L. J. J. Am. Chem. Soc. 1989, 111,
6621–6627.
(49) Grossi, L.; Strazzari, S.; Gilbert, B. C.; Whitwood, A. C. J. Org.
Chem. 1998, 63, 8366–8372.
J.-P. Bioorg. Med. Chem. 2005, 13, 3977–3986.
(12) Gimꢀenez-Arnau, E.; Haberkorn, L.; Grossi, L.; Lepoittevin, J.-P.
Tetrahedron 2002, 58, 5535–5545.
(13) Luo, Y.-R. BDEs of O-X bonds. In Comprehensive Handbook of
Chemical Bond Energies; Taylor & Francis Group: Boca Raton, 2007;
pp 255À368.
(50) Makino, K.; Hagiwara, T.; Hagi, A.; Nishi, M.; Murakami, A.
Biochem. Biophys. Res. Commun. 1990, 172, 1073–1080.
(14) Johansson, S.; Gimꢀenez-Arnau, E.; Grøtli, M.; Karlberg, A.-T.;
B€orje, A. Chem. Res. Toxicol. 2008, 21, 1536–1547.
6199
dx.doi.org/10.1021/jo200948x |J. Org. Chem. 2011, 76, 6188–6200

The Journal of Organic Chemistry
ARTICLE
(51) Lassmann, G.; Eriksson, L. A.; Lendzian, F.; Lubitz, W. J. Phys.
Chem. A 2000, 104, 9144–9152.
(52) Joshi, S.; Husain, M. M.; Chandra, R.; Hasan, S. K.; Srivastava,
R. C. Hum. Exp. Toxicol. 2005, 24, 13–17.
(53) Redeby, T.; Nilsson, U.; Altamore, T. M.; Ilag, L.; Ambrosi, A.;
Broo, K.; B€orje, A.; Karlberg, A.-T. Chem. Res. Toxicol. 2010, 23,
203–210.
(54) Falick, A. M.; Hines, W. M.; Medzihradszky, K. F.; Baldwin,
M. A.; Gibson, B. W. J. Am. Soc. Mass Spectrom. 1993, 4, 882–893.
(55) Barbati, S.; Clꢀement, J.-L.; Olive, G.; Roubaud, V.; Tuccio, B.;
Tordo, P. In Free Radicals in Biology and Environment; Minisci, F., Ed.;
Kluwer Academic Publishers: Dordrecht, The Netherlands, 1997;
pp 39À47.
(56) Poole, C. P., Jr. In Electron Spin Resonance: A Comprehensive
Treatise on Experimental Techniques; John Wiley & Sons Inc.: New York,
1983; pp 381À458.
(57) National Institute of Environmental Health Sciences, U. S.
Department of Health and Human Services. Spin Trap Database.
http://tools.niehs.nih.gov/stdb/index.cfm.
(58) Usuki, T.; Inoue, M.; Akiyama, K.; Hirama, M. Bioorg. Med.
Chem. 2005, 13, 5218–5224.
(59) Britigan, B. E.; Hamill, D. R. Arch. Biochem. Biophys. 1989,
275, 72–81.
(60) Rota, C.; Barr, D. P.; Martin, M. V.; Guenguerich, F. P.; Tomasi,
A.; Mason, R. P. Biochem. J. 1997, 328, 565–571.
6200
dx.doi.org/10.1021/jo200948x |J. Org. Chem. 2011, 76, 6188–6200