Synthesis and spin-trapping properties of a trifluoromethyl analogue of DMPO: 5-methyl-5-trifluoromethyl-1-pyrroline N-Oxide (5-TFDMPO)

Article abstract of DOI:10.1002/chem.201303774

The 5-diethoxyphosphonyl-5-methyl-1-pyrroline N-oxide superoxide spin adduct (DEPMPO-OOH) is much more persistent (about 15 times) than the 5,5-dimethyl-1-pyrroline N-oxide superoxide spin adduct (DMPO-OOH). The diethoxyphosphonyl group is bulkier than the methyl group and its electron-withdrawing effect is much stronger. These two factors could play a role in explaining the different half-lifetimes of DMPO-OOH and DEPMPO-OOH. The trifluoromethyl and the diethoxyphosphonyl groups show similar electron-withdrawing effects but have different sizes. We have thus synthesized and studied 5-methyl-5-trifluoromethyl-1-pyrroline N-oxide (5-TFDMPO), a new trifluoromethyl analogue of DMPO, to compare its spin-trapping performance with those of DMPO and DEPMPO. 5-TFDMPO was prepared in a five-step sequence by means of the Zn/AcOH reductive cyclization of 5,5,5-trifluoro-4-methyl-4- nitropentanal, and the geometry of the molecule was estimated by using DFT calculations. The spin-trapping properties were investigated both in toluene and in aqueous buffer solutions for oxygen-, sulfur-, and carbon-centered radicals. All the spin adducts exhibit slightly different fluorine hyperfine coupling constants, thereby suggesting a hindered rotation of the trifluoromethyl group, which was confirmed by variable-temperature EPR studies and DFT calculations. In phosphate buffer at pH-7.4, the half-life of 5-TFDMPO-OOH is about three times shorter than for DEPMPO-OOH and five times longer than for DMPO-OOH. Our results suggest that the stabilization of the superoxide adducts comes from a delicate balance between steric, electronic, and hydrogen-bonding effects that involve the β group, the hydroperoxyl moiety, and the nitroxide. New pieces to the puzzle! A new fluorinated 5,5-dimethyl-1-pyrroline N-oxide (DMPO)-based spin trap (5-TFDMPO; see figure) was synthesized and studied by EPR/spin-trapping techniques. The properties of the new spin trap are reported for various radicals. The role of the trifluoromethyl group on the spin-trapping properties of the superoxide radical is discussed on the basis of a comparison with other spin traps.

Full text of DOI:10.1002/chem.201303774

DOI: 10.1002/chem.201303774
Full Paper
&
Spin Trapping
Synthesis and Spin-Trapping Properties of a Trifluoromethyl
Analogue of DMPO: 5-Methyl-5-trifluoromethyl-1-pyrroline
N-Oxide (5-TFDMPO)**
Hakim Karoui,[a] Cꢀline Nsanzumuhire,[a] FranÅois Le Moigne,[a] Micael Hardy,[a] Didier Siri,[a]
Etienne Derat,[b] Antal Rockenbauer,[c] Olivier Ouari,*[a] and Paul Tordo*[a]
In memory of FranÅois Le Moigne
Abstract: The 5-diethoxyphosphonyl-5-methyl-1-pyrroline N-
oxide superoxide spin adduct (DEPMPOÀOOH) is much more
persistent (about 15 times) than the 5,5-dimethyl-1-pyrroline
N-oxide superoxide spin adduct (DMPOÀOOH). The dieth-
oxyphosphonyl group is bulkier than the methyl group and
its electron-withdrawing effect is much stronger. These two
factors could play a role in explaining the different half-life-
times of DMPOÀOOH and DEPMPOÀOOH. The trifluorometh-
yl and the diethoxyphosphonyl groups show similar elec-
tron-withdrawing effects but have different sizes. We have
thus synthesized and studied 5-methyl-5-trifluoromethyl-1-
pyrroline N-oxide (5-TFDMPO), a new trifluoromethyl ana-
logue of DMPO, to compare its spin-trapping performance
with those of DMPO and DEPMPO. 5-TFDMPO was prepared
in a five-step sequence by means of the Zn/AcOH reductive
cyclization of 5,5,5-trifluoro-4-methyl-4-nitropentanal, and
the geometry of the molecule was estimated by using DFT
calculations. The spin-trapping properties were investigated
both in toluene and in aqueous buffer solutions for oxygen-,
sulfur-, and carbon-centered radicals. All the spin adducts ex-
hibit slightly different fluorine hyperfine coupling constants,
thereby suggesting a hindered rotation of the trifluorometh-
yl group, which was confirmed by variable-temperature EPR
studies and DFT calculations. In phosphate buffer at pH 7.4,
the half-life of 5-TFDMPOÀOOH is about three times shorter
than for DEPMPOÀOOH and five times longer than for
DMPOÀOOH. Our results suggest that the stabilization of the
superoxide adducts comes from a delicate balance between
steric, electronic, and hydrogen-bonding effects that involve
the b group, the hydroperoxyl moiety, and the nitroxide.
Introduction
In the last three decades, much effort has been devoted to the
development of methods to detect and characterize the super-
oxide radical, and among these methods EPR/spin-trapping
techniques have proven to be one of the most reliable.[1] 5,5-
Dimethyl-1-pyrroline N-oxide (DMPO, 1; Figure 1) is one of the
[a] Dr. H. Karoui, Dr. C. Nsanzumuhire, Dr. F. Le Moigne, Dr. M. Hardy,
Prof. D. Siri, Dr. O. Ouari, Prof. P. Tordo
Aix-Marseille Universitꢀ, CNRS UMR7273
Institut de Chimie Radicalaire (ICR)
Case 521, Avenue Escadrille Normandie-Niemen
13397 Marseille Cedex 20 (France)
Fax: (+33)491-288-758
Figure 1. Spin-trap structures.
E-mail: olivier.ouari@univ-amu.fr
paul.tordo@univ-amu.fr
most popular spin traps, and EPR features of its spin adducts
are well known.[2] However, major limitations, such as a low su-
peroxide trapping rate constant and a poor stability (t1/2
ꢀ1 min at physiological pH) of the DMPOÀOOH spin adduct
restrict the use of DMPO in trapping the superoxide anion
[b] Dr. E. Derat
UPMC Sorbonne Universitꢀs
Institut Parisien de Chimie Molꢀculaire, UMR CNRS 7201
4 place Jussieu, 75005 Paris (France)
[c] Prof. A. Rockenbauer
Molecular Pharmacology, Research Center for Natural Sciences
Budapest (Hungary)
C
(O2 ), one of the most important free-radical species formed in
aerobic organisms.[3,4] We have shown that the replacement of
a methyl group of DMPO with a diethoxyphosphonyl group to
generate 5-diethoxyphosphonyl-5-methyl-1-pyrroline N-oxide
[**] DMPO=5,5-dimethyl-1-pyrroline N-oxide
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201303774.
Chem. Eur. J. 2014, 20, 4064 – 4071
4064
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
(DEPMPO, 2; Figure 1) dramatically improved the spin-trapping
properties.[5] Under the same experimental conditions, the half-
lifetime of the DEPMPOÀOOH adduct is at least 15 times
oxide (TMINO) and M3PO, 6) were reported to lead to superox-
ide adducts with short lifetimes (<1 min), thus highlighting
the destabilizing effect of C-2 substitution and the decrease in
longer than that of DMPOÀOOH. Furthermore, unlike DMPOÀ the trapping rate.[23]
OOH, the DEPMPOÀOOH adduct does not decay to the corre-
sponding hydroxyl adduct in phosphate buffer.[5,6] In the
course of our studies on the parameters that control the spin-
adduct stability, we investigated the spin-trapping properties
Results and Discussion
of the 5-diethoxyphosphonyl-5-phenyl-1-pyrroline N-oxide spin
trap (DEPPPO, 3; Figure 1).[7] The replacement of a methyl
group with a phenyl was shown to change the stereoselectiv-
ity of the trapping reaction and the conformations of the nitro-
xide spin adducts,[8] as illustrated by the significant changes
observed for the hyperfine coupling constants. However, no
significant change was observed for the half-lifetime of the
DEPPPOÀOOH spin adduct relative to DEPMPOÀOOH
Synthesis
According to the method described by Shepard and Sciaraf-
fa,[24] we prepared 1,1,1-trifluoroacetone oxime (9) starting
from commercially available 1,1,1-trifluoroacetone (8). Nitro
compound 10 was obtained by means of oxidizing oxime 9
with H2O2 in the presence of HNO3 and NH4NO3 as reported by
Bissel and Fields.[25] Treatment of nitro compound 10 with
acrolein in the presence of triethylamine afforded nitroalde-
hyde 11, and reductive cyclization of 11 yielded nitrone 5 (5-
TFDMPO) as a colorless oil after distillation. 5-TFDMPO samples
stored at À208C were stable for several months, and we never
observed any additional paramagnetic signal when they were
used for spin-trapping studies either in phosphate buffers or in
organic solvents (Scheme 1).
(DEPPPOÀOOH t1/2 =(13.1Æ0.1) min and DEPMPOÀOOH t1/2
=
(14.8Æ1.4) min at pH 7.4).[7] Additionally, we reported an EPR/
spin-trapping study of the 5-ethoxycarbonyl-5-methyl-1-pyrro-
line N-oxide spin trap (EMPO, 4; Figure 1) in which one b sub-
stituent is an ethoxycarbonyl group.[9] The half-lifetime of the
EMPOÀOOH spin adduct was estimated from 4.8[9] to
8.6 min.[10] Several articles reporting the spin-trapping proper-
ties of EMPO derivatives were subsequently published by
Kalyanaraman et al.,[11] Nohl et al.,[10,12] Rosen et al.,[13] and
Stolze et al.[14] From these studies, the superoxide spin-adduct
stability appeared to depend on the chain length of the alkoxy
moiety and/or on the pyrrolidine ring substitution.[11–14] Since
these works, numerous studies have been devoted to the de-
velopment of new pyrroline N-oxide-based spin traps with im-
proved superoxide trapping properties,[15–20] and DEPMPO de-
rivatives remain one of the most efficient structures for spin
traps for in vitro and ex vivo trapping of oxygen-centered radi-
cals.
Scheme 1. Synthesis of the 5-TFDMPO spin trap.
In our continuing efforts to understand the influence of the
diethoxyphosphonyl group on the stabilization of the
DEPMPOÀOOH adduct, we prepared the 5-methyl-5-trifluoro-
methyl-1-pyrroline N-oxide spin trap (5-TFDMPO, 5; Figure 1)
and performed spin-trapping experiments for oxygen-, car-
bon-, and sulfur-centered radicals. It has been previously re-
ported that the trifluoromethyl group has a similar electron-
withdrawing effect to the diethoxyphosphonyl group although
its size is significantly smaller.[21a] It can be assumed that the
general pattern of the spin-adduct EPR spectra of 5-TFDMPO
should be similar to those of DMPO adducts. However, further
splittings owing to the fluorine atoms (mI =1/2) can be expect-
ed. For instance, it has been reported that the i-amyloxy
adduct of an isoindoline-N-oxide spin trap that bears a ÀCF3
DFT molecular geometry of DEPMPO, EMPO, and 5-TFDMPO
To obtain information on the molecular geometry of 5-
TFDMPO, EMPO, and DEPMPO, the energetically favored con-
formations were calculated for each molecule. The optimized
geometries of their most stable conformer are shown in
Figure 2, and the main geometrical features are reported in
Table 1. The net atomic charges were calculated using the
CHelpG scheme.[26]
group exhibited additional fluorine hyperfine splittings.[21b]
A
DMPO derivative that bears a ÀCF3 group on the nitronyl
double bond (2-TFDMPO, 7; Figure 1) has been previously de-
scribed by Janzen et al.[22] For 2-TFDMPO (7), EPR spectra of
most spin adducts exhibited a three-line signal further split
into 1:3:3:1 quartets owing to the g coupling with fluorine
atoms. No indication of the half-lifetime of 2-TFDMPOÀOOH
was reported. However, nitrone spin traps that bear a substitu-
ent on the nitronyl double bond (1,1,3-trimethylisoindole-N-
Figure 2. DFT-calculated geometries of the most stable conformer of a) 5-
TFDMPO, b) DEPMPO, and c) EMPO.
Chem. Eur. J. 2014, 20, 4064 – 4071
www.chemeurj.org
4065
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Table 1. Main geometrical features of DEPMPO, 5-TFDMPO, and EMPO.
C5ÀR NÀO C=N C2NC5R C2NC5C4 Charge[a]
[ꢃ]
[ꢃ]
[ꢃ]
[8]
[8]
C5(R)
DEPMPO (R=P(O)(OEt)2) 1.865 1.272 1.306 101.1 14.4
0.243
0.234
0.182
5-TFDMPO (R=ÀCF3)
1.538 1.263 1.310 103.7 13.1
1.541 1.264 1.307 102.6 12.4
EMPO (R=ÀCO2Et)
[a] The net atomic charges were calculated using the CHelpG scheme.[27]
All the calculations were performed with the Gaussian 98
molecular-orbital package.[27] The geometry optimizations were
carried out without constraints at the B3LYP/6-31G(d) level. For
DEPMPO and 5-TFDMPO, very close values of net charge at the
C-5 position (the carbon atom that bears the electron-with-
drawing group) have been obtained by calculation, thus indi-
cating that ÀCF3 and ÀP(O)(OEt)2 groups have similar electronic
effects on the nitronyl function. For EMPO, the net charge is
smaller at the C-5 position, thus indicating a less important
electron-withdrawing effect. Vibrational frequencies were cal-
culated at the B3LYP/6-31G(d) level to determine the nature of
the located stationary points.
Figure 3. EPR spectra of the hydroxyl and superoxide spin adducts of 5-
TFDMPO. a) UV photolysis of a solution containing 5-TFDMPO (50 mm) and
H2O2 (1%) in phosphate buffer (0.1m, pH 7.4). b) Incubation mixture contain-
ing hypoxanthine (0.4 mm), xanthine oxidase (0.4 UmlÀ1), DTPA (1 mm), and
5-TFDMPO (100 mm) in phosphate buffer (0.1m, pH 7.4). The computer sim-
ulations are below the experimental spectra. Spectrometer settings: micro-
wave power 10 mW, modulation amplitude a) 0.05 mT and b) 0.1 mT, time
constant 0.128 s, gain 5ꢂ104, scan range 10 mT, scan time a) 120 s and b)
240 s.
The three spin traps display very similar geometries: the
conformation of the ring is an envelope (E4), and the measured
C2NC5C4 dihedral angles are 14.4, 13.1, and 12.48 for DEPMPO,
5-TFDMPO, and EMPO, respectively. It is noteworthy that the
ring conformation imposes a pseudoaxial position of the dieth-
oxyphosphonyl, trifluoromethyl, and carboxyethyl moieties
(C2NC5P and C2NC5C6 dihedral angles are 101.1, 103.7, and
102.68 for DEPMPO, 5-TFDMPO, and EMPO, respectively). Thus,
in contrast to DMPO in which the gem-dimethyl groups have
Spin trapping of superoxide radical anions and tert-butyl-
peroxyl radicals
Spin-trapping of superoxide with 5-TFDMPO was performed at
a bisecting conformation, the pseudoaxial position of the ÀCF3, pH 7.4 and 5.6 using hypoxanthine (HX; 0.4 mm)/xanthine ox-
ÀCO2Et, and ÀP(O)(OEt)2 groups strongly increases the steric
crowding of one face of the molecule. Thus a stereoselective
trapping reaction is expected, with the trans spin adducts
formed primarily, most notably for DEPMPO because of the
larger size of the diethoxyphosphonyl group.
idase (XOD; 0.4 UmlÀ1) as the superoxide-generating system. A
complex EPR spectrum was observed (Figure 3b), which was
cancelled in the presence of superoxide dismutase (SOD;
1250 UmlÀ1), thus indicating that the signal can be assigned to
the 5-TFDMPOÀOOH spin adduct. The same signal was ob-
served by UV photolysis of a solution that contained 5-
TFDMPO and H2O2 (30%) in phosphate buffer. During the
decay or after the total disappearance of the 5-TFDMPOÀOOH
spin adduct, no 5-TFDMPOÀOH signal was observed. The ex-
perimental spectrum (Figure 3b) was satisfactorily calculated
(Figure 3b, bottom and Table 2) by assuming the superimposi-
tion of the signals of the cis (7.4%) and trans (92.6%) diaste-
reomeric spin adducts. The best fit was obtained by assuming
three nonequivalent fluorine atoms for each diastereoisomer.
5-TFDMPOÀOOtBu was generated by irradiation of a solution
of tBuOOH in benzene (1.5m) in the presence of 5-TFDMPO.
For both 5-TFDMPOÀOOH (Figure 3b) and 5-TFDMPOÀOOtBu
(see the Supporting Information) spin-adduct signals, a com-
plex and similar spectral pattern was obtained. The 5-
TFDMPOÀOOtBu signal was calculated by considering the pres-
ence of two diastereoisomers (cis and trans addition, 28.4 and
71.6%, respectively) that exhibited different coupling constants
(Table 2). As for 5-TFDMPOÀOOH, the major species is sup-
posed to be the trans diastereoisomer owing to the steric
crowding of the face that displays the ÀCF3 moiety.
EPR studies
Spin trapping of hydroxyl radicals
The hydroxyl radical spin adduct (5-TFDMPOÀOH) was generat-
ed by incubating 5-TFDMPO (50 mm) with H2O2 (1 mm) and
FeSO4 (0.3 mm) in 0.1m phosphate buffer at pH 7.4. A persis-
tent and intense EPR signal (Figure 3a) was observed that ex-
hibited a general 1:2:2:1 pattern (aN =1.38 mT; aH =1.34 mT)
but with each line of the quartet being further split into a quar-
tet due to coupling with three nonequivalent fluorine atoms
(aF =0.217, 0.203, and 0.159 mT; Table 3 and Figure 3a). The
same EPR signal was also observed either by Fe3+-catalyzed
nucleophilic addition of water[28] or by UV photolysis of a solu-
tion that contained 5-TFDMPO (50 mm) and H2O2 (1%) in phos-
phate buffer at pH 7.4. The 5-TFDMPOÀOH adduct signal was
inhibited by adding catalase (50 UmlÀ1) in the Fenton system
incubation mixture.
Chem. Eur. J. 2014, 20, 4064 – 4071
www.chemeurj.org
4066
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
tylperoxide (0.15m) in tert-butylbenzene at 298 K. In
the presence of 5-TFDMPO, the 5-TFDMPOÀOtBu
adduct spectrum was observed as a very persistent
signal and was best calculated as a single diastereo-
isomer adduct (only trans addition) with three none-
quivalent fluorine atoms (Figure 4a). The hyperfine
coupling constants are reported in Table 3.
Table 2. Calculated EPR features of 5-TFDMPOÀOOH and 5-TFDMPOÀOOtBu.
Radical
Source
Solvent
water[a]
[%]
aN [mT]
aHb [mT]
aF [mT]
C
HOO
HX/XOD
trans
92.6[b]
cis
1.279
1.120
0.237, 0.236,
0.183
0.251, 0.168,
0.168
1.290
1.156
1.212
0.863
0.950
0.816
7.4[b]
trans
71.6[b]
cis
C
tBuOO
tBuOOH/hn
benzene
0.266, 0.242,
0.213
0.310, 0.192,
0.147
28.4[b]
Sulfur-centered radicals
[a] Phosphate buffer (0.1m), pH 7.4. [b] Percentage contribution of diastereoisomer.
Sulfur-centered radicals were generated by UV pho-
tolysis of a degassed solution that contained different
disulfides (RS–SR, 1.5m in benzene). Spin adducts ob-
tained by trapping of thiyl radicals were persistent,
and their corresponding spectra were calculated by assuming
the presence of only one diastereoisomer (only trans addition)
Kinetic decay of the superoxide adduct
5-TFDMPO-OOH was generated by incubating HX (0.4 mm),
XOD (0.04 UmlÀ1), DTPA (0.3 mm), and 5-TFDMPO (50 mm) in
0.1m phosphate buffer (pH 7.4 and 5.6). After 7 min, a steady-
state concentration of adduct was reached, and a large excess
amount of SOD (1250 UmlÀ1) was added to stop its formation.
Decay of 5-TFDMPO-OOH was monitored by measuring the de-
crease in the low-field line intensity of the EPR spectrum. The
kinetic decay was calculated as a pseudo-first-order process.
Half-lifetimes were determined to be (4.5Æ0.3) min at pH 7.4
and (14.3Æ0.3) min at pH 5.6. Under the same conditions,
DEPMPOÀOOH exhibited a longer half-lifetime, with values of
(14.8Æ1.4) min at pH 7.4 and 30.4 min at pH 5.6.[5,29] It was not
possible to perform an accurate comparison with DMPO under
the same experimental conditions because of the very short
half-lifetime of DMPOÀOOH (t1/2 ꢀ50 s at pH 7.4). Study of 5-
TFDMPOÀOOH decay was performed at two different pH
values (5.6 and 7.4). At pH 5.6, the 5-TFDMPOÀOOH adduct
was more persistent than at pH 7.4 ((14.3Æ0.3) and (4.5Æ
0.3) min, respectively), but the signal intensity was weaker. It is
interesting to mention that unlike DMPOÀOOH, 5-TFDMPOÀ
OOH adduct decays by giving diamagnetic byproducts. Thus,
5-TFDMPO exhibited a similar performance to EMPO, even if
the ÀCF3 group is more electron-withdrawing than the ÀCO2Et
Figure 4. EPR spectra from spin trapping of tert-butoxyl radical with 5-
TFDMPO at different temperatures. a) UV photolysis of a solution containing
5-TFDMPO (50 mm) and tBuOOtBu (1.5m) in tert-butylbenzene at 298 K.
b) Same as in (a) but at 373 K. The computer simulations are below the ex-
perimental spectra. Spectrometer settings: microwave power 10 mW, modu-
lation amplitude 0.015 mT, time constant 0.064 s, gain 5ꢂ104, scan range
10 mT, scan time 168 s.
group. Both spin adducts have similar half-lifetimes (t1/2
=
(4.5Æ0.3) min for 5-TFDMPOÀOOH and (4.8Æ1.1) min for
EMPOÀOOH at pH 7.4).[9]
Spin trapping of other radicals
Other (C-, O-, S-centered) radi-
Table 3. EPR parameters of different 5-TFDMPO spin adducts at 298 K.
cals were trapped with 5-
TFDMPO in water and in organic
solvents. The calculated EPR pa-
rameters of the spin adducts are
reported in Table 3.
Radical
Source
Solvent
aN [mT]
aH [mT]
aF [mT]
aHg [mT]
HO
H2O2/Fe2+
(tBuO)2/hn
H2O2/DMSO/Fe2+
CH3I/hn
water[a]
tert-butylbenzene
water[a]
benzene
benzene
water[a]
benzene
benzene
benzene
water[a]
1.382
1.227
1.490
1.328
1.282
1.419
1.260
1.269
1.224
1.291
1.341
0.813
2.233
1.981
1.921
1.739
1.070
1.127
1.339
1.522
0.217, 0.203, 0.159
0.224, 0.212, 0.184
0.250, 0.218, 0.174
0.272, 0.235, 0.187
0.284, 0.255, 0.225
0.266, 0.235, 0.208
0.230, 0.211, 0.186
0.225, 0.199, 0.176
0.235, 0.212, 0.192
0.278, 0.254, 0.212
–
C
C
tBuO
0.094
–
–
–
C
C
CH3
CH3
Ph·
PhI/hn
À
CO2
H2O2/Fe2+/HCO2Na
(EtS)2/hn
–
C
Oxygen-centered radical in
organic solvent
C
EtS
0.090, 0.120
tBuS·
(tBuS)2/hn
(PhS)2/hn
H2O2/Fe2+/NaHSO3
–
–
–
C
PhS
CÀ
C
tert-Butoxyl radical (tBuO ) was
SO3
produced by UV photolysis of
[a] Phosphate buffer (0.1m), pH 7.4.
a degassed solution of di-tert-bu-
Chem. Eur. J. 2014, 20, 4064 – 4071
www.chemeurj.org
4067
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
with three nonequivalent fluorine atoms (see the Supporting
Information). 5-TFDMPOÀSO3H spin adduct was obtained by
using the Fenton reaction in the presence of NaHSO3 (0.1m).
EPR parameters of 5-TFDMPO-thiyl spin adducts are reported
in Table 3.
obtained by assuming the presence of three nonequivalent flu-
orine hyperfine coupling constants. In his previous work on
the 2-TFDMPO spin trap (6), Janzen and co-workers reported
nonequivalent fluorine atoms in the case of carbon dioxide
anion, phenyl, and alkyl adducts.[22] Our calculations suggest
that the trapping of radicals with 5-TFDMPO is highly stereose-
lective. The high stereoselectivity observed could be explained
by the pseudoaxial position of the ÀCF3 group (Figure 2) and
its steric hindrance, which, because of the solvent sphere, is re-
ported to be in between that of isopropyl and tert-butyl
groups.[30] The observation of a small amount (7.4%) of cis
adduct for 5-TFDMPOÀOOH could be accounted for by the
presence of a network of hydrogen bonds from water mole-
cules that link the ÀCF3 group, the nitroxide, and the peroxyl
moiety (Figure 6).
Carbon-centered radicals
Very intense signals (Figure 5) were detected for the trapping
C
C
of Ph and CH3 generated by UV photolysis of a benzene solu-
À
C
C
C
tion of the corresponding iodides, and for CO2 , CH3, and CH-
(CH3)OH generated by Fenton reaction in water in the pres-
ence of sodium formate, dimethylsulfoxide, and ethanol, re-
spectively. The EPR signals consisted of six lines (aN <aH) fur-
ther split into four lines because of the presence of three
slightly nonequivalent fluorine atoms (Table 3).
Figure 6. Calculated DFT structures for cis and trans adducts of 5-TFMPOÀ
OOMe at the UB3LYP/SVP level with PCM modeling water. The isosurface de-
picting the spin density of the radical is given at 0.005 eꢃÀ3. Hydrogen-
bond lengths are given in ꢃ. The cis adduct is 2.15 kcalmolÀ1 more stable
than the trans adduct.
Nonequivalence of the fluorine atoms
Figure 5. EPR spectra from spin trapping of CO2À radical with 5-TFDMPO.
C
For all the spin adducts, the best EPR-calculated parameters
were obtained by assuming three nonequivalent fluorine cou-
pling constants. To verify the possible ÀCF3 hindered rotation
at the origin of the nonequivalency of the three fluorine
atoms, quantum mechanical (QM) simulations and EPR varia-
ble-temperature experiments were performed. To assess the
hypothesis of a conformational restraint induced by water,
quantum chemical calculations were carried out on a model
system including, specifically, a water molecule (Figure 6). The
level of calculation chosen (UB3LYP/SVP, with implicit water
solvent modeled by the polarizable continuum model (PCM))
is sufficient to describe this type of radical.[31]
C
CO2À was generated by means of a Fenton standard system (2 mm H2O2,
1 mm FeSO4) in the presence of sodium formate (0.2m) and 5-TFDMPO
(50 mm). The black line below represents a computer simulation of the ex-
perimental spectrum. Spectrometer settings: microwave power 10 mW,
modulation amplitude 0.05 mT, time constant 0.064 s, gain 5ꢂ104, scan
range 10 mT, scan time 240 s.
Stereoselectivity of the trapping reaction
Two diastereoisomers (cis and trans) are expected when trap-
ping radicals with 5-TFDMPO because of the chirality of the C-
5 atom. For DEPMPO in which the chiral C-5 atom is linked to
an EPR-active phosphorus nucleus (mI =1/2), when they are
observed the two diastereoisomers exhibit different EPR spec-
tra owing to the large value and the conformational depend-
ence of the b-phosphorus coupling constant. For the 5-
TFDMPO spin adducts, the two diastereoisomers are expected
to exhibit close EPR signals because of the small values of the
g-fluorine hyperfine coupling constants (aF). For all the spin ad-
ducts, except for 5-TFDMPOÀOOH and 5-TFDMPOÀOOtBu
(Table 2), we obtained a good fit between the experimental
and calculated EPR spectra, assuming the presence of only one
diastereoisomer. For all the calculated spectra, a better fit was
The simulation for both diastereoisomers (cis and trans) indi-
cates that a water molecule can simultaneously establish hy-
drogen bonding between the nitroxide and the trifluoromethyl
part of the molecule. In both configurations, the distance be-
tween water and the nitroxide was found to be 1.90 ꢃ, where-
as the hydrogen-bond length between water and ÀCF3 is
around 2.4 ꢃ. In the cis adduct, a second water molecule can
establish a bridge between the trifluoromethyl group and the
peroxide (hydrogen-bond lengths: 2.33 and 2.02 ꢃ). Thus, the
hydrogen-bonding network can slow the rotation of the ÀCF3
group and affect the EPR spectra. Variable-temperature EPR ex-
Chem. Eur. J. 2014, 20, 4064 – 4071
www.chemeurj.org
4068
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
over Na2SO4, then the organic phase was distilled to yield the
Table 4. 5-TFDMPOÀOtBu spin adduct in tert-butylbenzene. Evolution of
fluorine coupling constants (aF) with temperature.
1
oxime (19 g, 75%). B.p. 101–1048C; H NMR (100 MHz, CDCl3, TMS):
d=1.76 (s; CH3), 7.36 ppm (s; NOH); 19F NMR (94 MHz, CDCl3): d=
À71.4 ppm.
T [K] 298
323
353
373
403
aF 0.224, 0.212, 0.223, 0.208, 0.224, 0.205, 0.224, 0206, 0.214, 0.214,
1,1,1-Trifluoro-2-nitropropane (10)
[mT] 0.184 0.186 0.188 0.187 0.200
Compound 9 (10 g, 78.7 mmol) was added over 30 min at 108C to
a stirred mixture of HNO3 (13.20 g) and NH4NO3 (6.75 g). The mix-
ture was stirred at 108C for 2 h. H2O2 (8 mL, 30%) was added over
15 min, and the mixture was stirred at 108C for 30 min, then at
room temperature for 2.5 h. We added another 4 mL of H2O2
(30%). The mixture was poured into crushed ice and water. The or-
ganic phase was removed, and the aqueous layer was extracted
with four portions of CH2Cl2 (10 mL). The organic layer was dried
over Na2SO4. The solvent was then removed to yield a mixture
(2.00 g, 2:1) of 1,1,1-trifluoro-2-nitropropane and 1,1,1-trifluoro-2,2-
periments were also performed to monitor the evolution of
the EPR spectra of 5-TFDMPOÀOtBu (Figure 4) from 298 to
403 K, and the fluoride hyperfine coupling constant (aF) was
examined (Table 4).
At 403 K, the three hyperfine coupling constants of the fluo-
rine atoms become nearly identical (DaF =0.014 mT), thus indi-
cating that at 400 K the rotation of the ÀCF3 group is fast in
regard to EPR.
1
dinitropropane after distillation. B.p. 102–1038C; H NMR (100 MHz,
CDCl3): d=1.78 (d, J=6.2 Hz; CH3), 5.01 ppm (m; CH); 19F NMR
(94 MHz, CDCl3): d=À74.7 ppm (d, J(F,H)=7 Hz)
Conclusion
5,5,5-Trifluoro-4-methyl-2-nitropentanal (11)
5-TFDMPO is a new trifluoromethyl analogue of DMPO. It is
very stable for several months when stored at low temperature
and does not give any paramagnetic impurity. 5-TFDMPO has
shown efficient trapping of oxygen-, carbon-, and sulfur-cen-
tered radicals and a higher persistency than DMPO concerning
the superoxide spin adduct (t1/2 =4.5 min and 50 s at pH 7.4
for 5-TFDMPOÀOOH and DMPOÀOOH, respectively). The super-
oxide spin adduct does not decompose spontaneously to give
the hydroxyl adduct or any paramagnetic byproduct. Despite
extra couplings owing to the presence of three fluorine atoms,
5-TFDMPO spin adducts could be easily characterized. Relative
to DEPMPO and EMPO superoxide adducts, the persistency of
the 5-TFDMPOÀOOH adduct is lower or close, respectively. This
result indicates that the electron-withdrawing effect plays
a role in the stabilization of the superoxide spin adducts, but it
does not constitute a key parameter like the conformational
effect and the hydrogen-bond network do.
Triethylamine (0.25 g, 2.45 mmol) was added at 08C to a cooled
mixture of acrolein (3.30 g, 58.8 mmol) and the previous mixture of
1,1,1-trifluoro-2-nitropropane and 1,1,1-trifluoro-2,2-dinitropropane
(3.55 g, 24.8 mmol) in acetonitrile (15 mL). The mixture was stirred
for 3 h at room temperature, and the solvent was removed under
reduced pressure. A colorless oil was obtained (1.88 g, 43%), which
was further purified by distillation (428C at 0.1 mm Hg). 19F NMR
(94 MHz, CDCl3): d=À76.2 ppm.
5-Trifluoromethyl-5-methyl-1-pyrroline N-oxide (5)
A solution of acetic acid (2.00 g, 33.3 mmol) in ethanol (17 mL) was
added to
a mixture of 5,5,5-trifluoro-4-methyl-2-nitropentanal
(3.55 g, 9.6 mmol) and zinc (1.00 g, 15.3 mmol) in ethanol (50 mL).
The medium was allowed to warm to room temperature under
stirring and was stirred for 4 h. After removal of the solvent and ex-
traction by CH2Cl2 (2ꢂ50 mL), crude oil (1.12 g) was obtained and
purified by distillation (608C at 10À1 mm Hg) to yield 5-TFDMPO
1
(0.56 g, 3.36 mmol, 35%). H NMR (400 MHz, CDCl3): d=1.72 (s, 3H;
CH3), 2.19–2.77 (m, 4H; 2CH2), 7.05 ppm (t, J=2.55 Hz, 1H; CH);
13C NMR (50.25 MHz, C6D6): d=18.51 (s; CH3), 24.53 (s; CH2), 28.68
(s; CH2), 76.62 (q, J=27.98 Hz; CÀCF3), 125.99 (q, J=284.90 Hz;
CF3), 133.48 ppm (s; CH); 19F NMR (94 MHz, CDCl3): d=À78.2 ppm;
elemental analysis calcd (%) for C6H8F3NO·0.25H2O (171.63): C
41.99, H 4.99, N 8.16; found: C 42.02, H 5.07, N 8.07.
Experimental Section
Synthesis and characterization
1H NMR spectra were recorded at 100 and 400 MHz in CDCl3 using
TMS as an internal reference. 13C NMR spectral measurements
were performed at 100.6 MHz using CDCl3 and C6D6 as internal
standards. 19F NMR spectral measurements were performed at
94.21 MHz in CDCl3 and C6D6 using CClF3 as an external standard.
Chemical-shift (d) values are given in ppm, and J values in Hz. Ele-
mental analyses were carried out at the Spectropole of the Aix-
Marseille University.
Spin-trapping studies
Xanthine oxidase (XOD), bovine erythrocyte superoxide dismutase
(SOD), and catalase were purchased from Boehringer Mannheim
Biochemical Co. Glutathione (GSH), glutathione peroxidase (Gpx),
diethylenetriaminepentaacetic acid (DTPA), di-tert-butyl peroxide,
dibenzoyl peroxide, tert-butyl hydroperoxide, riboflavin, and other
chemicals were from Sigma Chemical Co.
1,1,1-Trifluoroacetone oxime (9)
Compound 8 (22.4 g, 0.20 mol) was added to a stirred solution of
hydroxylamine hydrochloride (27.80 g, 0.40 mol) and sodium ace-
tate (65 g, 0.79 mol) in water (300 mL) at 08C. The mixture was
then stirred at room temperature for 2 d and finally heated to
reflux for 18 h. The mixture was extracted with diethyl ether (2ꢂ
100 mL), and the organic phase was washed with water (50 mL) sa-
turated with sodium bicarbonate, with water (50 mL), and dried
EPR measurements
EPR spectra were recorded at room temperature using computer-
controlled Varian E-3 and Bruker ESP 300 EPR spectrometers at
9.5 GHz (X-band) and employing 100 kHz field modulation. Reac-
tion mixtures were prepared in a Chelex-treated phosphate buffer
Chem. Eur. J. 2014, 20, 4064 – 4071
www.chemeurj.org
4069
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
(0.1m, pH 7.4). EPR spectra were simulated using EPR software de-
veloped by D. Duling from the Laboratory of Molecular Biophysics,
NIEHS, USA (this software is available from the World Wide Web at
http://epr.niehs.nih.gov/PEST.html) and by A. Rockenbauer from
the Central Research Institute of Chemistry, Hungary.[32] UV photoly-
sis was performed by using a 1000 W Xe/Hg Oriel lamp.
ed from the signal amplitude at time tnÀ1 by using the chosen rate
equation. The standard least-squares method was then applied to
fit the calculated curves with the experimental ones. In these calcu-
lations, the monitored EPR peak intensity is related to the actual
radical concentration [SA] by a scale factor. The first-order constant
ka and the product kb[SA]0 are independent of this scale factor.
Half-lifetimes are given as the meanÆSD (nꢁ3).
Superoxide trapping hypoxanthine/xanthine oxidase system:
Xanthine oxidase (0.4 UmlÀ1) was added to a solution of 5-
TFDMPO (0.1m), diethylenetriaminepentaacetic acid (DTPA; 1 mm),
and hypoxanthine (0.4 mm) in phosphate buffer (0.1m, pH 7.4).
The EPR spectrum was recorded 60 s after the addition of XOD.
Acknowledgements
C
HO trapping Fenton system: Hydroxyl radical was generated by
We thank the ANR (SpinBioRad project, ANR-09-blan-0193),
Rꢀgion PACA, and the Hungarian Scientific Research Fund for
financial support.
adding FeSO4 (0.3 mm) to a solution of 5-TFDMPO (50 mm) and
H2O2 (1 mm) in phosphate buffer (0.1m, pH 7.4). The EPR spectrum
of the hydroxyl adduct was recorded 60 s after the addition of fer-
rous sulfate. No EPR signal was observed in the presence of cata-
lase (50 UmlÀ1) in the incubation mixture
Keywords: EPR spectroscopy · fluorinated ligands · radicals ·
spin trapping · substituent effects
UV-photolysis system:
A solution that contained 5-TFDMPO
(50 mm) and H2O2 (1%) in phosphate buffer (0.1m, pH 7.4) was irra-
diated with UV light, and the EPR spectrum was recorded after
15 s of irradiation.
[1] a) F. A. Villamena, J. L. Zweier, Antioxid. Redox Signaling 2004, 6, 619–
629; b) O. Ouari, M. Hardy, H. Karoui, P. Tordo EPR in Electron Spin Reso-
nance (Eds.: B. C. Gilbert, D. M. Murphy, V. Chechik), The Royal Society
of Chemistry, Cambridge, 2010, 1–40; c) M. P. Murphy, A. Holmgren,
N. G. Larsson, B. Halliwell, C. J Chang, B. Kalyanaraman, S. G. Rhee, P. J.
Thornally, L. Partridge, D. Gems, T. Nystrçm, V. Belousov, P. T. Schumack-
er, C. C. Winterbourn, Cell Metab. 2011, 13, 361–364; d) R. P. Mason,
P. M. Hanna, M. J. Burkitt, M. B. Kadiiska, Environ. Health Perspect. 1994,
102, 33–36.
[2] a) G. R. Buettner, Free Radical Biol. Med. 1987, 3, 259–303; b) K. Rangue-
lova, R. P. Mason, Magn. Reson. Chem. 2011, 49, 152–158.
[3] a) G. R. Buettner, L. W. Oberley, Biochem. Biophys. Res. Commun. 1978,
83, 69–74; b) M. T. Lin, M. F. Beal, Nature 2006, 443, 787–795; c) S. G.
Rhee, Science 2006, 312, 1882–1883.
Nucleophilic addition of water: The hydroxyl spin adduct was ob-
tained by adding FeCl3 (0.5 mm) to a solution of 5-TFDMPO
(50 mm) in double-distilled water.
C
C
C
C
CO2 ( CO2H), CH3, and CH(CH3)OH trapping: A Fenton system in
the presence of HCO2Na, DMSO, and ethanol was used to generate
CO2 ( CO2H), CH3, and CH(CH3)OH, respectively. FeSO4 (0.5 mm)
was added to a solution of 5-TFDMPO (50 mm) and H2O2 (1 mm),
and HCO2Na (0.1m) and ethanol (10%), respectively, in phosphate
buffer (0.1m, pH 7.4). The EPR spectrum of the corresponding spin
adduct was recorded 60 s after the addition of ferrous sulfate.
À
C
C
C
C
C
C
C
C
CH3 and Ph trapping in organic solvent: CH3 and Ph were gen-
erated by UV photolysis of a solution of methyl and phenyl iodide
(1m), respectively, in the presence of 5-TFDMPO (50 mm) in ben-
zene.
[4] a) E. Finkelstein, G. M. Rosen, E. J. Rauckman, J. Paxton, Mol. Pharmacol.
1979, 16, 676–685; b) E. Finkelstein, G. M. Rosen, E. J. Rauckman, Mol.
Pharmacol. 1982, 21, 262–265; c) E. Finkelstein, G. M. Rosen, E. J. Rauck-
man, J. Am. Chem. Soc. 1980, 102, 4994–4999; d) E. Finkelstein, G. M.
Rosen, E. J. Rauckman, Arch. Biochem. Biophys. 1980, 200, 1–16; e) G. R.
Buettner, Free Radical Res. Commun. 1993, 19, s79–87.
[5] C. Frꢀjaville, H. Karoui, F. Le Moigne, M. Culcasi, S. Piꢀtri, R. Lauricella, B.
Tuccio, P. Tordo, J. Chem. Soc. Chem. Commun. 1994, 1793–1794.
[6] C. Frꢀjaville, H. Karoui, F. Le Moigne, M. Culcasi, S. Piꢀtri, R. Lauricella, B.
Tuccio, P. Tordo, J. Med. Chem. 1995, 38, 258–265.
[7] H. Karoui, C. Nsanzumuhire, F. Le Moigne, P. Tordo, J. Org. Chem. 1999,
64, 1471–1477.
[8] A. Rockenbauer, L. Korecz, K. Hideg, J. Chem. Soc. Perkin Trans. 2 1993,
2149–2156.
[9] G. Olive, A. Mercier, F. Le Moigne, A. Rockenbauer, P. Tordo, Free Radical
Biol. Med. 2000, 28, 403–408.
[10] K. Stolze, N. Udilova, T. Rosenau, A. Hofinger, H. Nohl, Biol. Chem. 2003,
384, 493–500.
C
C
tBuOO trapping: tBuOO was produced by UV photolysis of a solu-
tion of tert-butyl hydroperoxide (1.5m) and 5-TFDMPO (50 mm) in
benzene.
C
C
C
Thiyl radical trapping: CH3CH2S , PhS , and tBuS were produced
by UV photolysis of the respective dialkyl disulfide (1.5m) and 5-
TFDMPO (50 mm) in benzene.
C
(SO3) trapping: A Fenton system in the presence of NaHSO3 was
C
C
used to generate SO3 ((SO3H) ). FeSO4 (0.5 mm) was added to a so-
lution of 5-TFDMPO (50 mm), H2O2 (1 mm), and NaHSO3 (0.1m) in
phosphate buffer (0.1m, pH 7.4). The EPR spectrum of the corre-
sponding spin adduct was recorded 60 s after the addition of fer-
rous sulfate.
C
tert-Butoxyl radical trapping: tBuO was generated by UV photoly-
[11] H. Zhao, J. Joseph, H. Zhang, H. Karoui, B. Kalyanaraman, Free Radical
Biol. Med. 2001, 31, 599–606.
sis of a solution of di-tert-butyl peroxide (0.15m) and 5-TFDMPO
(50 mm) in toluene.
[12] a) K. Stolze, N. Udilova, H. Nohl, Biol. Chem. 2002, 383, 813–820; b) K.
Stolze, N. Udilova, T. Roseneau, A. Hofinger, H. Nohl, Biochem.Pharmacol.
2005, 69, 297–305; c) K. Stolze, N. Rohr-Udilova, T. Roseneau, R. Stadt-
muller, H. Nohl, Biochem. Pharmacol. 2005, 69, 1351–1361.
[13] P. Tsai, K. Ichikawa, C. Mailer, S. Pou, H. J. Halpern, B. H. Robinson, R.
Nielsen, G. M. Rosen, J. Org. Chem. 2003, 68, 7811–7817.
[14] K. Stolze, N. Rohr-Udilova, A. Patel, T. Roseneau, Bioorg. Med. Chem.
2011, 19, 985–993.
[15] S.-U. Kim, Y. Liu, K. M. Nash, J. L. Zweier, A. Rockenbauer, F. A. Villamena,
J. Am. Chem. Soc. 2010, 132, 17157–17173.
[16] Y. Han, B. Tuccio, R. Lauricella, A. Rockenbauer, J. L. Zweier, F. A. Villame-
na, J. Org. Chem. 2008, 73, 2533–2541.
Kinetics of decay of superoxide spin adducts
We used the hypoxanthine/xanthine oxidase system described pre-
viously to generate superoxide in phosphate buffer (0.1m, pH 7.4).
The spin-trap concentration was 0.05m for 5-TFDMPO. The super-
oxide generation was initiated by incubating xanthine oxidase in
the reaction mixture for 7 min and suppressed by adding SOD
(1250 UmlÀ1). The spin-adduct decay was followed by monitoring
the decrease in an appropriate line of the spin adduct. Computer
simulations were performed using DAPHNIS Labs software devel-
oped by R. Lauricella. The signal amplitude at time tn was calculat-
[17] a) M. Hardy, F. Chalier, O. Ouari, J.-P. Finet, A. Rockenbauer, B. Kalyanara-
man, P. Tordo, Chem. Commun. 2007, 1083–1085; b) F. Chalier, M.
Chem. Eur. J. 2014, 20, 4064 – 4071
www.chemeurj.org
4070
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Hardy, O. Ouari, A. Rockenbauer, P. Tordo, J. Org. Chem. 2007, 72, 7886–
7892.
Montgomery, R. E. Stratmann, J. C. Burant, S. Dapprich, J. M. Millam,
A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J. Tomasi, V. Barone, M.
Cossi, R. Cammi, B. Mennucci, C. Pomelli, C. Adamo, S. Clifford, J. Och-
terski, G. A. Petersson, P. Y. Ayala, Q. Cui, K. Morokuma, D. K. Malick,
A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz,
A. G. Baboul, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi,
R. Gomperts, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng,
A. Nanayakkara, C. Gonzalez, M. Challacombe, P. M. W. Gill, B. Johnson,
W. Chen, M. W. Wong, J. L. Andres, C. Gonzalez, M. Head-Gordon, E. S.
Replogle, J. A. Pople, Gaussian, Inc., Pittsburgh PA, 1998.
[18] M. Hardy, D. Bardelang, H. Karoui, A. Rockenbauer, J.-P. Finet, L. Jicsinsz-
ky, R. Rosas, O. Ouari, P. Tordo, Chem. Eur. J. 2009, 15, 11114–11118.
[19] a) H. Karoui, J-L. Clꢀment, A. Rockenbauer, D. Siri, P. Tordo, Tetrahedron
Lett. 2004, 45, 149–152; b) M. Hardy, O. Ouari, L. Charles, J. P. Finet, G.
Iacazio, V. Monnier, A. Rockenbauer, P. Tordo, J. Org. Chem. 2005, 70,
10426–10433.
[20] C. Nsanzumuhire, J-L. Clꢀment, O. Ouari, H. Karoui, J.-P. Finet, P. Tordo,
Tetrahedron Lett. 2004, 45, 6385–6389.
[21] a) J. Katzhendler, I. Ringel, R. Karaman, H. Zaher, E. Bruer, J. Chem. Soc.
Perkin Trans. 2 1997, 341–349; b) B. Hatano, K. Miyoshi, H. Sato, T. Ito,
T. Ogata, T. Kijima, Tetrahedron Lett. 2010, 51, 5399–5401.
[22] E. G. Janzen, Y. K. Zhang, M. Arimura, J. Org. Chem. 1995, 60, 5434–
5440.
[28] a) K. Makino, T. Hagiwara, A. Hagi, M. Nishi, A. Murakami, Biochem. Bio-
phys. Res. Commun. 1990, 172, 1073–1080; b) P. M. Hanna, W. Chamuli-
trat, R. P. Mason, Arch. Biochem. Biophys. 1992, 296, 640–644.
[29] B. Tuccio, R. Lauricella, C. Frꢀjaville, J. C. Bouteiller, P. Tordo, J. Chem.
Soc., Perkin Trans. 2 1995, 2, 295–298.
[23] a) D. Barasch, M. C. Krishna, A. Russo, J. Katzhendler, A. Samuni, J. Am.
Chem. Soc. 1994, 116, 7319–7324; b) S. E. Bottle, A. S. Micallef, Org.
Biomol. Chem. 2003, 1, 2581–2584; c) S. E. Bottle, G. R. Hanson, A. S. Mi-
callef, Org. Biomol. Chem. 2003, 1, 2585–2589.
[30] T. Fujita, C. Takayama, M. Nakajima, J. Org. Chem. 1973, 38, 1623–1630.
[31] C. Houriez, N. Ferrꢀ, D. Siri, P. Tordo, M. Masella, J. Phys. Chem. B 2010,
114, 11793–11803.
[32] A. Rockenbauer, L. Korecz, Appl. Magn. Reson. 1996, 10, 29–43.
[24] R. A. Shepard, P. L. Sciaraffa, J. Org. Chem. 1996, 31, 964–965.
[25] E. R. Bissell, D. B. Fields, Tetrahedron 1970, 26, 5737–5743.
[26] C. M. Breneman, K. B. Wiberg, J. Comput. Chem. 1990, 11, 361–373.
[27] Gaussian 98, Revision A11.4, M. J. Frisch, G. W. Trucks, H. B. Schlegel,
G. E. Scuseria, M. A. Robb, J. R. Cheeseman, V. G. Zakrzewski, Jr. J. A.
Received: September 26, 2013
Published online on March 3, 2014
Chem. Eur. J. 2014, 20, 4064 – 4071
www.chemeurj.org
4071
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim