Kinetic and thermochemical study of the antioxidant activity of sulfur-containing analogues of vitamin E

Article abstract of DOI:10.1002/chem.200700309

Sulfur-containing analogues of vitamin E (thiachromanols), either linked or not to a catechol moiety, were synthesized and their hydrogenatom donating ability evaluated. The determination of the O-H bond dissociation enthalpy (BDE) of the a-tocopherol analogue 4 by the electron paramagnetic resonance (EPR) equilibration technique provided a value of 78.9 kcal mol^-1, that is, approximately 1.8 kcal mol^-1 higher than that of α-to-copherol. The kinetic rate constants for the reaction with peroxyl radicals (k_inh), measured by inhibited autoxidation studies, showed that thiachromanols react 2.5 times slower than the corresponding tocopherols, in agreement with the higher BDE value. This behavior was explained, on the basis of crystallographic analyses and DFT calculations, in terms of a change in the molecular geometry caused by insertion of a sulfur atom into the frame-work of vitamin E. This behavior implies a greater deviation of the condensed ring from coplanarity with the aromatic ring, thus giving rise to a decrease in the conjugative stabilization of the phenoxyl radical and consequently to an increase in the O-H bond strength. Although less reactive than tocopherols, thiachromanols may, however, act as bimodal antioxidants as a result of the hydroperoxide decomposing ability of the sulfur atom.

Full text of DOI:10.1002/chem.200700309

FULL PAPER
DOI: 10.1002/chem.200700309
Kinetic and Thermochemical Study of the Antioxidant Activity of Sulfur-
Containing Analogues of Vitamin E
Riccardo Amorati,[a] Andrea Cavalli,[c] Maria Grazia Fumo,[a] Matteo Masetti,[c]
Stefano Menichetti,*[b] Chiara Pagliuca,[b] Gian Franco Pedulli,*[a] and
Caterina Viglianisi[b]
Abstract: Sulfur-containing analogues
of vitamin E (thiachromanols), either
linked or not to a catechol moiety,
were synthesized and their hydrogen-
atom donating ability evaluated. The
tion studies, showed that thiachroma-
nols react 2.5 times slower than the cor-
responding tocopherols, in agreement
with the higher BDE value. This be-
havior was explained, on the basis of
crystallographic analyses and DFT cal-
culations, in terms of a change in the
molecular geometry caused by inser-
tion of a sulfur atom into the frame-
work of vitamin E. This behavior im-
plies a greater deviation of the con-
densed ring from coplanarity with the
aromatic ring, thus giving rise to a de-
crease in the conjugative stabilization
of the phenoxyl radical and conse-
À
determination of the O H bond disso-
À
ciation enthalpy (BDE) of the a-toco-
pherol analogue 4 by the electron para-
magnetic resonance (EPR) equilibra-
tion technique provided a value of
78.9 kcalmolÀ1, that is, approximately
1.8 kcalmolÀ1 higher than that of a-to-
copherol. The kinetic rate constants for
the reaction with peroxyl radicals
(kinh), measured by inhibited autoxida-
quently to an increase in the O H
bond strength. Although less reactive
than tocopherols, thiachromanols may,
however, act as bimodal antioxidants
as a result of the hydroperoxide de-
composing ability of the sulfur atom.
Keywords: autoxidation
·
bond
EPR spectroscopy
transfer radical
energy
·
·
hydrogen
·
reactions · sulfur heterocycles
Introduction
reasons and for its relevance in biological environments,
since this process accounts for the damage caused by free
radicals to organic and bioorganic systems. This reaction can
be inhibited, or at least retarded, by chain-breaking antioxi-
dants capable of trapping free radicals without transforming
themselves in reactive intermediates. Phenols, which repre-
sent the main antioxidant family, can donate the phenolic
hydrogen atom to the chain-carrying peroxyl radicals to
form a phenoxyl radical stabilized by resonance, hence they
are relatively unreactive toward oxygen and organic materi-
als.[1] The extent to which the autoxidation is retarded by
phenols depends upon the rate constant of the inhibition re-
action between antioxidant and peroxyl radicals kinh
[Eq. (1)].
The autoxidation of unsaturated hydrocarbons and lipids is
one of the most studied reactions, both for technological
[a]Dr. R. Amorati, Dr. M. G. Fumo, Prof. G. F. Pedulli
Dipartimento di Chimica Organica “A. Mangini” Università di Bolo-
gna, Via San Giacomo 11
40126 Bologna (Italy)
Fax : (+39)051-209-5688
E-mail: gianfranco.pedulli@unibo.it
[b]Prof. S. Menichetti, Dr. C. Pagliuca, Dr. C. Viglianisi
Dipartimento di Chimica Organica e Laboratorio di Progettazione
Sintesi e Studio di Eterocicli Bioattivi (HeteroBioLab)
Polo Scientifico e Tecnologico Università di Firenze
Via della Lastruccia 13, 50019 Sesto Fiorentino (Italy)
Fax : (+39)055-457-3531
ArOH þ ROOC ! ArOC þ ROOH
ð1Þ
E-mail: stefano.menichetti@unifi.it
[c]Dr. A. Cavalli, Dr. M. Masetti
Dipartimento di Scienze Farmaceutiche
Università di Bologna
Over recent decades, much work has been done to clarify
the basic principles that determine the rates of reaction (1)
and to synthesize compounds that show kinh values higher
than a-tocopherol,[2] the most effective lipid-soluble antioxi-
dant present in nature. Actually, several of the synthetic
Via Belmeloro 6, 40126 Bologna (Italy)
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
Chem. Eur. J. 2007, 13, 8223 – 8230
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
8223

phenols developed show excellent features, such as very
high reactivity towards peroxyl radicals[3,4] and improved air
stability.[4]
Another active area of research is the study of preventive
antioxidants,[5] compounds often containing divalent S, Se,
or Te atoms in the molecular skeleton,[6] which inhibit the
formation of initiating free radicals by decomposing hydro-
peroxides by a nonradical process. Chalcogen-substituted
phenols have been found to possess both chain-breaking
and preventive antioxidant activities.[7]
Recently, an original hetero-Diels–Alder approach, con-
sisting of the reaction of electron-poor ortho-thioquinones
with suitable alkenes (Scheme 1),[8] was used to prepare hy-
droxylated sulfur-containing heterocycles that resemble nat-
ural polyphenols present in plants. The main appeal of this
new family of phenolic compounds is that both quinone and
alkene moieties can be readily modified to confer the de-
sired attributes to the final products. In previous studies,[9]
several sulfur-containing analogues of flavonoids were syn-
thesized and tested for their antiradical activity both in
model and biological systems. Herein, the chain-breaking
activity of such molecules has been optimized (compounds
1–7) both by substituting the A ring protons with methyl
groups or by introducing a catechol moiety into the A or B
rings. In the latter case, the cooperative antioxidant effect of
the two moieties was investigated.
Scheme 1. Access to 4-thiaflavan structure and natural phenolic antioxi-
dants containing the chromane skeleton by using the hetero-Diels–Alder
(HDA) reaction.
Et3N in CHCl3 at 60–708C.[12] The inverse electron-demand
hetero-Diels–Alder reaction of thioquinones with 4-methoxy
or 3,4-bis-(tert-butyldimethylsilyloxy)styrene and 2-methyl-
1,3-pentadiene[13] led to the formation of the expected ben-
zoxathiin cycloadducts, which were transformed into the
final 4-thiaflavanes by deprotection of the hydroxy groups.
The use of 2-methyl-1,3-pentadiene as a dienophile required
the cycloaddition to be performed at 708C to drive the reac-
tion exclusively towards the thermodynamic oxathiin cyclo-
adducts 3 and 6.[13] The pyrrolidine-mediated deprotection,
carried out as the final step in the preparation of derivative
7,[11] required careful deoxygenation of all reagents and sol-
vents to prevent the oxidation of 7 to the corresponding
ortho-quinone. Preparation of the analogue of 7, bearing
two hydroxy groups in positions 6 and 7 (see Scheme 1), was
also tried without success, as the final product was too reac-
tive toward molecular oxygen to be stored and analyzed. As
a further endorsement of the validity of the procedure for
the preparation of these valuable heterocycles, hydroxy-4-
thiaflavanes 1–7 were isolated in overall yields of 25–40%
and were fully characterized, as reported in the experimen-
tal section.
Results and Discussion
Synthesis: Derivatives 1–7 were prepared by using readily
available phenols (2,3-dimethylhydroquinone, trimethylhy-
droquinone, and pyrogallol) as starting materials. These phe-
nols were protected as tert-butyldimethylsilyl (TBDMS)
ethers[10] or 2-ethylidene acetals[11] by using previously re-
ported procedures (Scheme 2). Monohydroxy arenes were
sulfenylated with PhtNSCl to give the ortho-hydroxythioph-
thalimides regiospecifically as suitable precursors for the
corresponding ortho-thioquinones obtained by reaction with
Antioxidant activity: The rate
constant for the reaction with
peroxyl radicals kinh of the
above derivatives was deter-
mined by studying the inhibi-
tion of the thermally initiated
autoxidation of a hydrocarbon
[Eqs (2)–(7)]under controlled
conditions, as described else-
where.[14] As styrene was
chosen as an oxidizable sub-
strate, the propagation step
[Eq. (4)]actually consists of the
addition of peroxyl radicals to
the olefinic double bond of sty-
rene.[15]
8224
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 8223 – 8230

Antioxidant Properties of Thiachromanols
FULL PAPER
related
tocopherols
(see
Table 1) and the inspection of
Figure 1a, which reports the
oxygen-uptake traces recorded
in the presence of the various
antioxidants, allows us to make
some considerations about the
contributions of groups or sub-
stituents to the activity of the
various antioxidants. The com-
parison of 1 and 3 with g-toco-
pherol and 4 and 6 with a-toco-
pherol (g- and a-TOH) shows
that the introduction of a sulfur
atom in the C ring causes a de-
crease in the kinh values by a
factor of approximately 2.5.
The reasons for the decreased
antioxidant efficacy of thiafla-
vanes will be discussed later.
Scheme 2. Reagents and conditions: a) Et3N, CHCl3, 60–708C; b) TBAF·3H2O, THF, 08C; c) pyrrolidine,
CH3CN, RT. PG=protecting group, TBAF=tetra-n-butylammonium fluoride.
Ri
[a]
Initiator
RC
ð2Þ
ð3Þ
ð4Þ
ð5Þ
ð6Þ
ð7Þ
ƒ!
Table 1. Kinetic rate constant for the reaction with peroxyl radicals kinh
and number of radicals trapped by each antioxidant molecule n.[b]
RC þ O2 ! ROOC
ROOC þ RH
ROOC þ ROOC
ROOC þ ArOH kinh ROOH þ ArOC
Phenol
kinh(PhCl) ”106
[mÀ1 sÀ1
kinh(MeCN) ”106
[mÀ1 sÀ1
k
inh(MeCN)/kinh(PhCl)
n
G
]
G
]
kp
ƒ!
ROOH þ RC
1
2
0.51
0.43[c]
0.42[c]
0.48
1.3
0.26
–
0.5
1.7
3.6
2kt
ƒ!
Products
3
4
5
–
1.7
1.8
3.6
ƒ!
0.95
0.73[c]
0.18[c]
–
0.7
1.0[c]
0.41[c]
1.2
0.7[c]
0.4[c]
ROOC þ ArOC ! Products
6
1.8
1.9
2[d]
2[d]
1.9
7
0.42
1.4[d]
3.2[d]
0.52
0.26
–
–
0.6
The number of radicals trapped by each antioxidant mole-
cule n is related to the length of the inhibited period t
through Equation (8), in which Ri is the initiation rate that
can be determined by using a reference antioxidant. The
values of kinh were calculated by using Equation (9): a plot
g-TOH
a-TOH
Catechol
0.16
0.3
[a]Error within ꢀ 10%. [b]Measured in a mixture of styrene/chloroben-
zene in the presence of 0.48m acetonitrile (MeCN). Mean value of three
determinations. [c]From numerical simulations. [d]From reference [3].
of D[O2]t versus ln
(1Àt/t) gives a straight line of slope kp-
[styrene]/kinh from which kinh is obtained by using the known
kp value at 308C of styrene, that is, 41 mÀ1 sÀ1.[15]
Ri ¼ n½ArOHŠ=t
ð8Þ
ð9Þ
ÀD½O2Š ¼ ðkp=kinhÞ½styreneŠ lnð1Àt=tÞ
In the case of the bifunctional antioxidants 2 and 5, the
kinh values were obtained by numerical simulation of the
oxygen-consumption traces, which also consider the equili-
bration between phenols and phenoxyl radicals [Eq. (10)]
and the radical-radical reactions involving the semiquinone
unit from the catechol group[16] (the complete list of equa-
tions is reported in the Supporting Information).
Figure 1. Oxygen-consumption traces observed during the autoxidation
of styrene (4.3m) in chlorobenzene initiated by AMVN (5”10À3 m) at
308C in the presence of the investigated antioxidants (5”10À6 m).
ArOC þ Ar0OH Ð ArOH þ Ar0OC
ð10Þ
The comparison of 1 with 4 and of 3 with 6 shows instead
that the antioxidant activity of phenols increases with in-
creasing the number of methyl substituents on the aromatic
The examination of the measured values of the inhibition
rate constants kinh for the investigated thiaflavanes and the
Chem. Eur. J. 2007, 13, 8223 – 8230
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8225

G. F. Pedulli, S. Menichetti et al.
A ring. This behavior, common to other antioxidants, has
been explained in terms of a favorable balance between the
electron-donating effect of methyl groups that decrease the
À
Additional evidence that supports this interpretation was
obtained by studying solvent effects on kinh caused by com-
plexation of the investigated thiaflavanes by a hydrogen-
bond-accepting solvent, such as acetonitrile (MeCN).[19] Ac-
tually, antioxidants that behave as good inhibitors in polar
solvents may give poor performances in polar environments,
as the formation of hydrogen bonds between the solvent
and phenolic OH group decreases the reactivity toward free
radicals.[20] The data in Table 1 show that the decrease in
kinh values observed in the presence of a small amount of
MeCN depends on the phenol structure since the ratio
kinh(MeCN)/kinh(PhCl) decreases in the order 4>7>1>catechol.
The smaller solvent effect observed in 4 can be explained in
terms of the higher steric hindrance around the phenolic
OH group in 4, which decreases its interaction with the sol-
vent with respect to the less-crowded compound 1.[19] Cate-
chol shows the larger solvent effect, as the OH group, not
involved in intramolecular hydrogen bonding, is more acidic
than the hydroxy groups of simple phenols.[20] Derivative 7,
on the other hand, shows a decrease in kinh values much
smaller than catechol upon addition of MeCN, thus indicat-
ing that neither hydroxy group is readily solvated as a result
of the presence of two intramolecular hydrogen bonds.[18]
Notably, for the bifunctional thiaflavane 5 the variations
of the kinh values with the solvent measured for the two anti-
oxidant moieties are similar to those observed in 4 and cate-
chol.
BDE(O H) and the steric crowding around the phenolic
N
Other noteworthy features are that compounds 2 and 5,
containing both thiachromanol and catechol groups, show
oxygen-uptake traces characterized by longer inhibition pe-
riods and lower slopes (see Figure 1b for 5). This behavior
is indicative of enhanced antioxidant activity with respect to
the other thiaflavanes and is in line with the simultaneous
presence of two groups capable of inhibiting the autoxida-
tion of styrene. To obtain the experimental values of kinh
,
the oxygen-decay curves were simulated by numerical solu-
tion of the simultaneous differential equations for a model
system containing two reactive sites with different inhibition
rate constants. The two kinh values determined for
2
(Table 1) are quite similar to those determined in separate
experiments for 1 and catechol, whereas those for 5 resem-
ble those of 4 and catechol, thus suggesting that the thiach-
romanol and catechol moieties behave independently. The
absence of synergistic effects was also confirmed by studying
equimolar mixtures of catechol with 1 or 4 (Figure 1b),
which gave inhibition traces practically superimposable to
those obtained in the presence of 2 or 5.
Another approach adopted to improve the antioxidant
properties of thiaflavanes was to introduce two adjacent hy-
droxy groups into the A ring. As already mentioned, the
isomer with the OH substituents at positions 6 and 7 was
not stable in air, whereas the 7- and 8-disubstituted com-
pound 7 (kinh =4.2”105 mÀ1 sÀ1) was characterized by an anti-
À
O H bond dissociation enthalpy (BDE): To obtain a ther-
mochemical estimate of the antioxidant activity of thiachro-
manols and of the effect of the heterocyclic sulfur atom in
À
the C ring, we determined the BDE of the O H bond in 4,
which is structurally related to a-tocopherol. The BDE was
determined by using the electron paramagnetic resonance
(EPR) radical equilibration technique,[17] which measures
the equilibrium constant Ke of two equilibrating phenols
and the corresponding phenoxyl radicals [Eq. (10)]. This
measurement was made by determining the relative concen-
trations of the phenoxyl radicals generated by photolyzing a
mixture of 4 (ArOH) and 2,6-di-tert-butyl-4-methoxyphenol
(BHA). The latter is used as reference phenol (Ar’OH) and
oxidant activity slightly lower than that of catechol (kinh
=
5.2”105 mÀ1 sÀ1), despite the presence of the electron-donat-
ing atoms oxygen and sulfur in the condensed C ring. An ex-
planation of the relatively low reactivity of 7 toward peroxyl
radicals can be given in terms of the adopted structure as in-
ferred by its IR spectrum, which shows a single absorption
at 3563 cmÀ1 in the O H stretching region. This frequency
À
value is characteristic of intramolecularly hydrogen-bonded
OH groups similar to that of ortho-methoxyphenol
(3557 cmÀ1). The absence of an absorption as a result of a
free OH group (for comparison, the two OH stretching vi-
brations of catechol are observed at 3618 and 3556 cmÀ1 for
the free and hydrogen-bonded OH groups, respectively) in-
dicates that both hydroxy protons form intramolecular hy-
drogen bonds with the adjacent oxygen atoms (Scheme 3)
and that their vibrational stretching frequencies are too
close to be resolved. As intra-
À1
À
is characterized by a BDE
A
(recently revised).[17,21]
BDEðArOÀHÞ ꢁ BDEðAr0OÀHÞÀRT lnðKeÞ
ð11Þ
The phenoxyl radical from thiachromanol 4, produced pho-
tolytically in deoxygenated benzene, was highly persistent
and showed an EPR spectrum interpretable, by similarity
with the related radical from a-tocopherol,[3] in terms of the
proton hyperfine splitting constants reported in Table 2.
Figure 2 shows the full EPR spectrum observed on photoly-
sis of a mixture of 4 and BHA together with its computer
simulation; the insert also shows part of the high-resolution
spectrum of the radical from 4.
molecularly hydrogen-bonded
hydroxy groups are known to
be characterized by low reactiv-
ity toward attacking radicals,[18]
the behavior of 7 is entirely ex-
pected.
Scheme 3. Intramolecular hy-
drogen bonds in thiachroma-
The measurements were repeated under different light in-
tensities to check the constancy of Ke and, therefore, the po-
nol 7.
8226
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 8223 – 8230

Antioxidant Properties of Thiachromanols
FULL PAPER
Table 2. EPR spectral parameters[a,b] of the phenoxyl radicals from 4 and
À
BHA and the BDE
A
Phenol
aH
A
BDE
[kcalmolÀ1
]
4
5.94 (5-Me)
5.17 (7-Me)
1.29 (8-Me)
–
2.61 (1H)
0.38 (1H)
0.19 (1H)
0.93 (2H)
1.53 (OMe)
2.0047
78.9ꢀ0.3
BHA[c]
2.00475
77.2
[a]In benzene/di- tert-butyl peroxide (9:1, v/v) at room temperature.
[b]Hyperfine splitting constants are given in Gauss; peak-to-peak line
width=0.12 G. [c]Data for the radical from BHA is taken from refer-
ence [17].
Figure 3. Logarithm of the rate constant for the reaction with peroxyl
À
*
radicals against the BDE
(O H) of 4 ( ) and other substituted 2,6-dime-
[3,17,21]
*
(
thylphenols
).
1) a-TOH, 2) 2,6-dimethyl-4-methoxyphenol,
3) 2,3,5,6-tetramethyl-4-methoxyphenol, 4) 2,4,6-trimethylphenol, 5) 2,6-
dimethylphenol.
DBDEðArOÀHÞ ¼ EðArOCÞÀEðArOHÞÀ½EðPhOCÞÀEðPhOHފ
ð12Þ
The dihedral angles # formed
by
atoms
10-9-1-2
(see
Scheme 4) in both the parent
phenols and phenoxyl radicals
were determined for 4 and 8, as
indicators of the extent of the
conjugation between the hetero-
cyclic oxygen atom and the phe-
nolic aromatic ring.
Scheme 4. Geometry of 4 (X=
S; R=H; R1 =Ar) and 8 (X=
CH2; R=R1 =CH3).
Figure 2. a) Experimental EPR spectrum obtained by photolysis of a mix-
ture of 4 (7.5”10À3 m) and BHA (2.1”10À4 m) in degassed benzene con-
taining 10% di-tert-butyl peroxide at room temperature and c) the com-
puter simulated spectrum. b) The expansion of the first part of the of the
high-resolution EPR spectrum of the phenoxyl radical from 4.
À
sition of the equilibrium. The BDE
lated by making the reasonable assumption that the entropic
G
The computed results indicate that thiachromanol 4 pos-
sesses a slightly higher bond-dissociation enthalpy value
term can be neglected[17] and by using Equation (11). The
À
À
BDE
(O H) of 4 was found to be 78.9 kcalmol , that is,
than 8 (Table 3). The calculated BDE
C
1.8 kcalmolÀ1 larger than that of a-tocopherol. This differ-
À
ence, which indicates a stronger O H bond in 4, is consis-
tent with its weaker antioxidant activity and is in line with
the kinetic results. Actually, the inhibition rate constant and
the BDE value obtained for thiaflavane 4 lie on the correla-
À
tion line between the log
A
À
Table 3. DBDE
C
of other ortho-dimethylphenols (Figure 3).[17]
C
#(ArO )
[8]
Phenol
DBDE
#
G
G
[b]
[kcalmolÀ1
]
[8]
Computational studies: The energies and relevant geome-
tries of 4, 8, and the corresponding phenoxyl radicals were
4
8
9
À9.7
À11.6
À0.70
À29 (À30[c]
)
À17
À12
–
À17[d]
calculated at the B3LYP/6-31+G(2d,p)//B3LYP/6–31G(d)
G
–
level to explain the effect of introducing a heterocyclic
[a]Calculated at the B3LYP/6-31 +G(2d,p)//B3LYP/6–31G(d) level by
G
sulfur atom in the condensed ring of the chromanol struc-
ture.[22,13b] The gas-phase differences DBDE
A
using the DFT method [b]Values include thermal correction to enthalpy
(298 K). [c]Experimental value measured by X-ray crystallographic anal-
yses. [d]Experimental values of À15.9 and À198 were determined for
each of the two molecules of 8 contained in the same unit crystallograph-
ic cell.[3]
À
the BDE values of the investigated antioxidants ArOH and
that of the unsubstituted phenol (PhOH) were calculated by
using the isodesmic approach shown in [Eq. (12)].[23]
Chem. Eur. J. 2007, 13, 8223 – 8230
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8227

G. F. Pedulli, S. Menichetti et al.
phenolic OH group was also estimated by repeating these
calculations on meta-thiometylphenol (9).
The larger distortion of the condensed ring of thiaflavanes
with respect to chromanols is responsible for the larger
BDE value observed in the former compounds.[3] The
reason depends on the different effects that departure from
coplanarity has in the phenoxyl radical and in the parent
phenol. Actually, the decrease in conjugative stabilization as
a result of the decreased overlap between the p orbitals of
the aromatic ring and the 2pz orbital of the heterocyclic
oxygen atom is higher in the phenoxyl radical than in the
parent phenol. Thus, little destabilization of the phenol will
occur upon deviation from planarity, whereas a much larger
effect will be experienced by the corresponding phenoxyl
radical in which delocalization of the unpaired electron in
the para-alkoxy substituent is a very important stabilizing
factor. Experimental evidence of this different behavior is
provided by the barrier to internal rotation of the methoxy
group, which in the 4-methoxyphenoxyl radical ranges from
8.3 to 10.4 kcalmolÀ1 depending on the solvent,[23] whereas
in anisole it does not exceed 2–3 kcalmolÀ1.[27] Thus, in-
creased distortion of the C ring, even if small, can substan-
tially decrease the stability of the phenoxyl radical with a
À
Both kinetic and thermochemical studies agree that
thiachromanols are less reactive toward peroxyl radicals
than the corresponding chromanols. The reason might
depend on electronic effects, as a result of the substitution
of a sulfur atom for a CH2 group, or a change in the pre-
ferred conformation of the C ring. Disappointingly, to the
best of our knowledge there are no reported data on the ex-
perimental effects of meta-alkylthio-substituents on the
BDE of phenols. This effect can be approximately estimated
by considering the effect of alkoxy substituents, which are
expected to behave similarly to SR groups.[24] There are two
reports on the effect of a meta-OMe group: Bordwell and
Cheng[25] used an electrochemical method in dimethyl sulf-
oxide (DMSO) and Lucarini et al.[17] used the EPR equili-
bration technique in benzene. The values are positive
(+0.4 kcalmolÀ1) and negative (À0.5 to À0.8 kcalmolÀ1), re-
spectively, the latter value being presumably more reliable
as DMSO gives rise to strong interactions with the investi-
gated compound. However, the two reports agree on the
conclusion that the effect of a meta-OMe group is small and
comparable to the uncertainties of these measurements.
Therefore, as a SR substituent is less perturbing than a OR
group, because of the lower overlap between the sulfur 3p
orbital and the p system of the aromatic ring,[25] the contri-
À
consequent increase in the BDE(O H) value.
G
Reactivity toward hydroperoxides: The lower efficacy of
thiaflavanes as chain-breaking antioxidants with respect to
tocopherols might be compensated by the presence of a bi-
valent sulfur atom in the molecular skeleton. This sulfur
atom can confer the properties of preventive antioxidants,
capable of decomposing hydroperoxides to give nonradical
products, to these derivatives. In the case of sulfides, the
products of this reaction are known to be sulfoxides and the
alcohol from the peroxide. Thus, we treated 4 (0.32 mmol)
in methanol with tButOOH (3.2 mmol) at room temperature
bution to the BDE(O H) of a meta-SMe group is expected
G
to be smaller than that of a meta-OCH3 group. This conclu-
sion is substantiated by DFT calculations performed on the
model phenol 9, and these calculations also suggest that the
introduction of a SCH3 substituent at the meta position to
the hydroxy group has a small negative effect on the BDE-
À
1
A
for 68 h. Analysis of the reaction mixture by H NMR spec-
A more reasonable explanation for the increase in BDE
troscopy allowed the identification of the residual 4 and
trans-sulfoxide 10 in a 10:1 ratio. The rate constant for the
bimolecular reaction between 4 and tert-ButOOH (3”
10À3 mÀ1 sÀ1), estimated from these data, is not far from that
reported for the reaction of other sulfides with hydroperox-
ides. For example, the rate constants for the reactions of di-
benzylsulfide and diphenylsulfide with cumylhydroperoxide
are 1.3”10À3 and 1.0”10À2 mÀ1 sÀ1 (at 353 and 423 K), re-
spectively.[5]
It must be therefore concluded that thiaflavanes are able
to decompose the peroxides formed during the oxidation of
hydrocarbons, that is, they may also behave as preventive
antioxidants.
observed on passing from chromanols to thiachromanols can
be given instead by considering the change in the # angle in-
duced by replacing the sulfur with the carbon atom in the
C ring. Actually, it is well known that the presence of the
condensed, saturated ring in chromanols is essential for
their good antioxidant activity, as it forces the heterocyclic
oxygen atom to conjugate better with the aromatic ring.[3]
The introduction of the sulfur atom into chromanol affects
the geometries of both phenol and the phenoxyl radical, as
the computed # values are larger by 12 and 58, respectively.
This difference implies that the preferred geometry of
thiachromanol 4 deviates appreciably from planarity, so that
the conjugation between the aromatic ring and lone pair on
the oxygen atom should be considerably decreased with re-
spect to a-TOH in both phenol and the phenoxyl radical. To
confirm experimentally the significant differences between
the condensed-ring geometries predicted by calculations on
tocopherol and 4, a determination of the structure of the
latter compound was made by X-ray crystallographic studies
(see the Experimental Section).[26] The more relevant result
concerns the value of the # angle, which was found to be
À308 and is, therefore, very close to that predicted computa-
tionally.
Conclusion
It has been shown herein that the introduction of methyl
groups in the thiachromanol A ring affords tocopherol-like
phenols with a chain-breaking activity almost as good as
that of the related tocopherols. Furthermore, the presence
of a sulfur atom in these molecules represents an important
property that is typical of preventive antioxidants. The
reason why thiachromanols are characterized by a lower re-
8228
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 8223 – 8230

Antioxidant Properties of Thiachromanols
FULL PAPER
of an EPR spectrometer, and photolyzed with the unfiltered light from a
500-W high-pressure mercury lamp at room temperature. The EPR spec-
tra were recorded on a spectrometer equipped with a microwave fre-
quency counter for the determination of the g factors, which were cor-
rected with respect to the perylene radical cation in concentrated H2SO4
(g=2.00258). The BDE value of 4 was determined by studying mixtures
of BHA and 4. The molar ratio of the two equilibrating radicals, obtained
from the EPR spectra, was used to calculate the equilibrium constant Ke.
Different concentration ratios of the starting phenols were used to check
if the equilibrium had been reached. The spectra were recorded a few
seconds after beginning the irradiation to avoid significant consumption
of the phenols during the course of the experiment. Relative radical con-
centrations were determined by comparison of the digitized experimental
spectra with computer-simulated spectra, as previously described.[17]
activity toward free radicals than their natural counterparts
was explained in terms of greater nonplanarity of the molec-
ular skeleton, which causes a decrease in the conjugative
stabilization of the phenoxyl radicals. The ready insertion of
other functional groups into these compounds is an impor-
tant feature that has been exploited to introduce supplemen-
tary antioxidant functions, such as the catechol group on the
B ring. Although in homogeneous solutions no advantages
can be obtained with respect to equimolecular mixtures of
the parent derivatives, it can not be excluded that in more
complex systems, such as low-density lipoproteins and lipo-
somes, in which diffusion of inhibitors among particles be-
comes the rate-determining step,[28] the covalent link be-
tween the two antioxidant moieties may be relevant to the
overall inhibiting activity.
IR measurements: FT-IR spectra of 7 were measured in dilute solutions
(0.01–0.05m) of tetrachloromethane in a sealed KBr cell with a 0.5-mm
optical path.
Computational details: The 2R configuration of 4 was arbitrarily chosen
for all calculations. A Monte Carlo conformational analysis was per-
formed by means of MacroModel9[31] to investigate the relative minima
adopted by the C ring of the chromane skeleton in both molecules. The
analysis was done under vacuum employing 5000 MC steps per molecule
and using an energy window of 50 kJmolÀ1. The MMFF94* force-field[32]
was used for the geometry optimization, which was performed by means
of a conjugate gradient algorithm[33] with a derivative convergence criteri-
on of 0.05 kJŠmolÀ1. The local minima found by the conformational
analysis were used as starting structures for the following quantum me-
chanical calculations. Geometry optimization was calculated with the
B3LYP hybrid functional,[34] with a 6–31G(d) basis set, and single point
Experimental Section
Materials: The protection of 2,3,5-trimethylhydroquinone[10] and pyrogal-
lol[11] was carried out following previously reported procedures. The sul-
fenylation of protected phenols with phthalimidesulfenyl chloride[8,9] and
the cycloaddition with styrenes[8,9] and 1,3-dienes[13] were performed as
reported elsewhere. The experimental data are available in the Support-
ing Information. All other compounds used are commercially available.
Solvents of the highest-purity grade were used as received. Styrene was
percolated on alumina before each experiment to remove traces of inhib-
itor.
energies at the level 6-31+G(2d,p) using the Gaussian03 program
G
suite.[35] Unrestricted wave functions were used for radical species. Ge-
ometry optimizations were carried out in the gas phase by using the
Berny algorithm along with the default convergence criteria,[36] whereas
frequency calculations at the reference temperature of 298.15 K were
performed at the same level of theory to both characterize stationary
points and calculate their thermodynamic properties. The zero-point vi-
brational energies (ZPVE) were corrected using a scale factor of 0.9806.
Oxidation reactions of sulfoxides with tButOOH: A solution of tBu-
tOOH in hexane (5m, 640 mL, 3.2 mmol) was added to a solution of cy-
cloadduct 4 (100 mg, 0.32 mmol) in MeOH (5 mL). The reaction mixture
was stirred at room temperature for 68 h, diluted with CH2Cl2 (80 mL)
and washed with a solution of Na2S2O3 (10%, 2”20 mL) and brine
(30 mL). The organic layer was dried over anhydrous Na2SO4 and con-
centrated in vacuo. 1H NMR spectroscopic analysis of the crude mixture
allowed the identification of the residual cycloadduct 4 and the trans-sulf-
oxide 10 in a 10:1 ratio.
X-ray crystallographic analyses: RX analysis was carried out with a Goni-
ometer Oxford Diffraction KM4 Xcalibur2 at room temperature. Graph-
ite-monochromated MoKa radiation (40 mA/À40 KV) and a KM4 CCD/
SAPPHIRE detector were used for cell-parameter determination and
data collection. The integrated intensities were measured using the w
scan mode and corrected for Lorentz and polarization effects.[37] The sub-
stantial redundancy in the data allows empirical absorption corrections to
be applied using multiple measurements of symmetry-equivalent reflec-
tions. The structure was solved by direct methods of SIR97[38] and refined
using the full-matrix least squares on F2 provided by SHELXL97.[39] The
non-hydrogen atoms were refined anisotropically, the methyl hydrogen
atoms were assigned in calculated positions, and the other hydrogen
atoms were found in the Fourier synthesis (all of them were refined as
isotropic).
Kinetic measurements: The rate constants for the reaction of the title
compounds with peroxyl radicals were measured by following the autoxi-
dation of styrene at 308C with 2,2’-azobis(2,4-dimethylvaleronitrile)
(AMVN; 5”10À3 m) as the initiator. The reaction was performed in an
oxygen-uptake apparatus built in our laboratory and based on a differen-
tial pressure transducer. The entire apparatus was immersed in a thermo-
statted bath, which ensured a constant temperature within 0.18C. In a
typical experiment, an air-saturated solution of styrene in chlorobenzene
(4.3m) containing the antioxidant was equilibrated with the reference so-
lution containing an excess of a-tocopherol (1”10À3 m) in the same sol-
vent. After equilibration, a concentrated solution of AMVN in chloro-
benzene was injected in both the reference and sample flasks, and the
oxygen consumption in the sample was measured, after calibration of the
apparatus, from the differential pressure recorded over time between the
two channels. This instrumental setting allowed us to have the N2 produc-
tion and oxygen consumption derived from the azo-initiator decomposi-
tion already subtracted from the measured reaction rates. Initiation rates
Ri were determined for each condition in preliminary experiments by
using the inhibitor method[15] with a-tocopherol as the reference antioxi-
Acknowledgements
Financial support from MIUR (Research projects “Radical Processes in
Chemistry and Biology: Synthesis, Mechanism, Application”, contract
2004038243 and “Stereoselezione in Sintesi Organica. Metodologie ed
Applicazioni” contract 2005035330) are gratefully acknowledged by
G.F.P and S.M. We also wish to thank Dr. Elisabetta Mileo for technical
assistance, Dr. Cristina Faggi for X-ray crystallographic analysis, and
Prof. Marco Lucarini for helpful suggestions.
dant: Ri =2[a-TOH]/t. The length of the induction period t was deter-
G
mined using an integration procedure suggested by Roginsky et al.[29] Nu-
merical simulation of the oxygen-consumption traces were performed by
using the software Gepasi, version 3.30,[30] freely available on the Internet
at http://www.gepasi.org, as described in the Supporting Information.
EPR and thermochemical measurements: Deoxygenated solutions of 4
and di-tert-butyl peroxide (10% v/v) in benzene were sealed under nitro-
gen in a suprasil quartz EPR tube. The sample was inserted in the cavity
[1]P. Mulder, H.-G. Korth, K. U. Ingold, Helv. Chim. Acta 2005, 88,
370–374.
Chem. Eur. J. 2007, 13, 8223 – 8230
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8229

G. F. Pedulli, S. Menichetti et al.
[2]G. W. Burton, K. U. Ingold, J. Am. Chem. Soc. 1981, 103, 6472–
6477.
[26]Compound 4 crystallizes with a molecule of 1,4-dioxane in the asym-
metric unit; X-ray crystal structure data: formula C22H28O5S
(C18H20O3S+C4H8O2), triclinic, crystal size: 0.20”0.10”0.05 mm,
[3]G. W. Burton, T. Doba, E. J. Gabe, L. Hughes, F. L. Lee, L. Prasad,
K. U. Ingold, J. Am. Chem. Soc. 1985, 107, 7053–7065.
[4]M. Wijtmans, D. A. Pratt, L. Valgimigli, G. A. DiLabio, G. F. Pedulli,
N. A. Porter, Angew. Chem. 2003, 115, 4506–4509; Angew. Chem.
Int. Ed. 2003, 42, 4370–4373.
[5]E. T. Denisov, I. B. Afanas ’ev, Oxidation and Antioxidants in Organ-
ic Chemistry and Biology, CRC Press, Boca Raton, 2005.
[6]C. Jacob, G. I. Giles, N. M. Giles, H. Sies, Angew. Chem. 2003, 115,
4890–4907; Angew. Chem. Int. Ed. 2003, 42, 4742–4758.
[7]a) D. Shanks, R. Amorati, M. G. Fumo, G. F. Pedulli, L. Valgimigli,
L. Engman, J. Org. Chem. 2006, 71, 1033–1038; b) J. Malmstrom,
M. Jonsson, I. A. Cotgreave, L. Hammarstrom, M. Sjodin, L.
Engman, J. Am. Chem. Soc. 2001, 123, 3434–3440.
¯
space group: P1, a=9.454(1), b=10.652(4), c=12.433(3) Š, a=
93.78(2), b=108,71(2), g=114.30(2)8, V=1052.0(5) Š3, Z=2 1calcd
1.277, m=0.183 mmÀ1, F
(000)=432; 11090 reflections were collected
=
AHCTREUNG
with a 3.88<q<26.00 range; 4068 were independent; the parame-
ters were 317 and the final R index was 0.0462 for reflections with
I>2sI and 0.0888 for all data. CCDC-629272 contains the supple-
mentary crystallographic data for this compound. These data can be
obtained free of charge from the Cambridge Crystallographic Data
Centre via www.ccdc.cam.ac.uk/data_request/cif.
[27]S. Tsuzuki, H. Houjou, Y. Nagawa, K. Hiratani, J. Chem. Soc.
Perkin Trans. 2 2002, 1271–1273.
[28]M. Alessi, T. Paul, J. C. Scaiano, K. U. Ingold, J. Am. Chem. Soc.
2002, 124, 6957–6965.
[29]D. Loshadkin, V. Roginsky, E. Pliss, Int. J. Chem. Kinet. 2002, 34,
162–171.
[30]a) P. Mendes, Comput. Appl. Biosci. 1993, 9, 563–571; b) P. Mendes,
Trends Biochem. Sci. 1997, 22, 361–363; c) P. Mendes, D. B. Kell,
Bioinformatics 1998, 14, 869–883.
[31]F. Mohamadi, N. G. J. Richards, W. C. Guida, R. M. J. Liskamp,
M. A. Lipton, C. E. Caulfield, G. Chang, T. F. Hendrickson, W. C.
Still, J. Comput. Chem. 1990, 11, 440–467.
[32]T. A. Halgren, J. Comput. Chem. 1996, 17, 490–519.
[33]E. Polak, G. Ribiere, Revue Francaise Inf. Rech. Oper. 1969, 16-R1,
35.
[8]G. Capozzi, P. Lo Nostro, S. Menichetti, C. Nativi, P. Sarri, Chem.
Commun. 2001, 551–552.
[9]a) S. Menichetti, M. C. Aversa, F. Cimino, A. Contini, C. Viglianisi,
A. Tomaino, Org. Biomol. Chem. 2005, 3, 3066–3072; b) M. Lodovi-
ci, S. Menichetti, C. Viglianisi, S. Caldini, E. Giuliani, Bioorg. Med.
Chem. Lett. 2006, 16, 1957–1960; c) R. Amorati, M. G. Fumo, G. F.
Pedulli, S. Menichetti, C. Pagliuca, C. Viglianisi, Helv. Chim. Acta
2006, 89, 2462–2472.
[10]T. Yoshioka, T. Fujita, T. Kanai, Y. Aizawa, T. Kurumada, K. Hase-
gawa, H. Horikoshi, J. Med. Chem. 1989, 32, 421–428.
[11]X. Ariza, O. Pineda, J. Vilarrasa, G. W. Shipps, Y. Ma, X. Dai, Org.
Lett. 2001, 3, 1399–1402.
[12]a) G. Capozzi, C. Falciani, S. Menichetti, C. Nativi, J. Org. Chem.
1997, 62, 2611–2615; b) G. Capozzi, C. Falciani, S. Menichetti, C.
Nativi, B. Raffaelli, Chem. Eur. J. 1999, 5, 1748–1754.
[13]a) S. Menichetti, C. Viglianisi, Tetrahedron 2003, 59, 5523–5530;
b) A. Contini, S. Leone, S. Menichetti, C. Viglianisi, P. Trimarco, J.
Org. Chem. 2006, 71, 5507–5514.
[14]R. Amorati, G. F. Pedulli, L. Valgimigli, O. A. Attanasi, P. Filippone,
C. Fiorucci, R. Saladino, J. Chem. Soc. Perkin Trans. 2 2001, 2142–
2146.
[15]J. A. Howard in Free Radicals, Vol. 2 (Ed.: J. K. Kochi), Wiley-Inter-
science, New York, 1975, Chapter 12.
[16]R. Amorati, F. Ferroni, M. Lucarini, G. F. Pedulli, L. Valgimigli, J.
Org. Chem. 2002, 67, 9295–9303.
[17]M. Lucarini, P. Pedrielli, G. F. Pedulli, S. Cabiddu, C. Fattuoni, J.
Org. Chem. 1996, 61, 9259–9263.
[18]M. I. de Heer, P. Mulder, H. G. Korth, K. U. Ingold, J. Lusztyk, J.
Am. Chem. Soc. 2000, 122, 2355–2360.
[19]L. Valgimigli, K. U. Ingold, J. Lusztyk, J. Am. Chem. Soc. 1996, 118,
3545–3549.
[34]A. D. Becke, J. Chem. Phys. 1993, 98, 1372–1377.
[35]Gaussian 03, Revision B.05, M. J. Frisch, G. W. Trucks, H. B. Schle-
gel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. A. Montgomer-
y, Jr., T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyen-
gar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N.
Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K.
Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda,
O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian,
J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E.
Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W.
Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth, P. Salvador, J. J.
Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C.
Strain, O. Farkas, D. K. Malick, A. D. Rabuck, K. Raghavachari,
J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cio-
slowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaro-
mi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng,
A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W.
Chen, M. W. Wong, C. Gonzalez, J. A. Pople, Gaussian, Inc., Wall-
ingford, CT, 2004.
[20]M. C. Foti, L. R. C. Barclay, K. U. Ingold, J. Am. Chem. Soc. 2002,
124, 12881–12888.
[36]C. Peng, P. Y. Ayala, H. B. Schlegel, M. J. Frisch, J. Comput. Chem.
1996, 17, 49–56.
[21]P. Mulder, H.-G. Korth, D. A. Pratt, G. A. DiLabio, L. Valgimigli,
G. F. Pedulli, K. U. Ingold, J. Phys. Chem. A 2005, 109, 2647–2655.
[22]J. S. Wright, E. R. Johnson, G. A. DiLabio, J. Am. Chem. Soc. 2001,
123, 1173–1183.
[37]N. Walker, D. Stuart, Acta Crystallogr. Sect. A 1983, 39, 158–166.
[38]A. Altomare, M. C. Burla, M. Camalli, G. L. Cascarano, C. Giaco-
vazzo, A. Guagliardi, A. G. G. Moliterni, G. Polidori, R. Spagna, J.
Appl. Crystallogr. 1999, 32, 115–119.
[23]M. Lucarini, V. Mugnaini, G. F. Pedulli, M. Guerra, J. Am. Chem.
Soc. 2003, 125, 8318–8329.
[24]R. Amorati, M. G. Fumo, S. Menichetti, V. Mugnaini, G. F. Pedulli,
J. Org. Chem. 2006, 71, 6325–6332.
[39]G. M. Sheldrick, SHELXL97, Program for Crystal Structure Refine-
ment, Institut für Anorganische Chemie de Universitat Gçttingen,
Gçttingen, Germany.
Received: February 23, 2007
[25]F. G. Bordwell, J.-P. Cheng, J. Am. Chem. Soc. 1991, 113, 1736–
1743.
Revised: May 24, 2007
Published online: July 18, 2007
8230
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 8223 – 8230