Antioxidant activity of hydroxystilbene derivatives in homogeneous solution

Article abstract of DOI:10.1021/jo0497860

The antioxidant activity of the cis and trans isomers of several analogues of resveratrol and pterostilbene has been investigated, especially with regard to the effect of the stereochemistry about the olefinic double bond. The antioxidant power of these c

Full text of DOI:10.1021/jo0497860

An tioxid a n t Activity of Hyd r oxystilben e Der iva tives in
Hom ogen eou s Solu tion
Riccardo Amorati, Marco Lucarini, Veronica Mugnaini, and Gian Franco Pedulli*
Dipartimento di Chimica Organica “A. Mangini”, Via S. Donato 15, 40127 Bologna, Italy
Marinella Roberti* and Daniela Pizzirani
Dipartimento di Scienze Farmaceutiche, Via Belmeloro 6, 40126 Bologna, Italy
gianfranco.pedulli@unibo.it
Received February 6, 2004
The antioxidant activity of the cis and trans isomers of several analogues of resveratrol and
pterostilbene has been investigated, especially with regard to the effect of the stereochemistry about
the olefinic double bond. The antioxidant power of these compounds was estimated by measuring
the rate constants for their reactions with peroxyl radicals and, with two of them, the bond
dissociation enthalpy (BDE) of the phenolic O-H bond which is cleaved in the inhibition reaction.
The present data show that in homogeneous solution the various hydroxystilbenes investigated
behave as mild antioxidants with the notable exceptions of the trans isomer of 4 and 6, whose
activities are only slightly lower than that of R-tocopherol (vitamin E). The rate constants of the
inhibition reaction show that the antioxidant activity of the cis-hydroxystilbene is in all the examined
cases worse, by a factor ranging between 2 and 6, than that of the corresponding trans isomers.
This lower reactivity depends on enthalpy factors as it can be inferred by the experimental values
of the O-H bond dissociation enthalpy in the two geometric isomers of 3′,5′-di-tert-butyl-4′-hydroxy-
3,5-dimethoxystilbene showing that the strength of the O-H bond in the cis isomer is larger by
1.8 kcal/mol. DFT calculations provide a rationalization of this result, indicating that, although
the cis geometry implies a destabilization with respect to the trans species of both phenoxyl radical
and parent hydroxystilbene, the destabilization of the radical is larger because the folding of the
structure strongly reduces the delocalization of the unpaired electron on the styryl group. A
comparison of these results with previously reported data on the proapoptotic activity of these
stilbenoids suggests that these two properties are not correlated.
In tr od u ction
that other phenolic derivatives present in red wine might
be more likely responsible for its antioxidant activity.4
Resveratrol has also been suggested as a potential
chemopreventive agent based on its striking inhibitory
effect on cellular events associated with cancer initiation,
promotion and progression.5 This phenolic stilbene has
also displayed in vitro growth inhibition in a number of
human cancer cell lines, and in this context, some of us
have recently investigated a series of cis- and trans-
resveratrol and pterostilbene derivatives with the aim
of discovering new lead compounds with clinical poten-
tial. The synthesized derivatives were screened in vitro
for cell growth inhibition and ability to induce apoptosis.
The cis were generally more effective than their corre-
sponding trans isomers, and some of them were found to
have potent proapoptotic activity.6
Many investigations have been recently reported on
the antioxidant properties of trans-resveratrol (1), pteros-
tilbene (2), and related derivatives. Some of these reports
attribute to trans-resveratrol (3,4′,5-trihydroxy-trans-
stilbene) an antioxidant activity comparable to (or even
higher than) R-tocopherol,1 one of the most effective lipid-
soluble chain-breaking antioxidants,2 while others indi-
cate that trans-resveratrol is the main contributor to the
total antioxidant power of red wine despite the fact that
it may not be the most abundant phenol in wine.3
Actually, some of us have recently demonstrated that
resveratrol in homogeneous solution is not an outstand-
ing antioxidant since it behaves only as a mild retarder
of the thermally initiated autoxidation of styrene and
(1) Fauconneau, B.; Waffo-Teguo, P.; Huguet, F.; Barrier, L.; De-
cendit, A.; Merillon, J .-M. Life Sci. 1997, 61, 2103-2110. Soares, D.
G.; Andreazza, A. C.; Salvador, M. J . Agric. Food Chem. 2003, 51,
1077-1080.
(2) Burton, G. W.; Ingold, K. U. Acc. Chem. Res. 1986, 19, 194-
201.
(3) Alonso, A. M.; Dom`ınguez, C.; Guille´n, D. A.; Barroso, C. G. J .
Agric. Food. Chem. 2002, 50, 3112-3115.
(4) Amorati, A.; Ferroni, F.; Pedulli, G. F.; Valgimigli, L. J . Org.
Chem. 2003, 68, 9654-9658.
To further explore the biological activity of these
compounds, we thought it could be of interest to check
10.1021/jo0497860 CCC: $27.50 © 2004 American Chemical Society
Published on Web 09/15/2004
J . Org. Chem. 2004, 69, 7101-7107
7101

Amorati et al.
SCHEME 1
SCHEME 2 a
their antioxidant activity, especially with regard to the
effect that the geometry about the olefinic double bond
has on it and in view of the conflicting results about the
antioxidant behavior of the cis and trans isomers of hy-
droxystilbenes. It has been reported that the experimental
results show no important differences between the E and
Z structures of these molecules,7a while in the case of
resveratrol a better activity is found for the trans isomer.7
a
Reagents and conditions: (a) n-BuLi, THF, -78 °C.
SCHEME 3 a
Thus, stilbene derivatives bearing a 4′-hydroxyl func-
tion in both cis and trans geometry were selected and
their antioxidant activity was evaluated quantitatively
by measuring the two parameters which are generally
believed to provide better quantitative estimates of the
antioxidant power of a given compound, i.e., the rate
constant for the reaction with peroxyl radicals and the
bond dissociation enthalpy (BDE) of the phenolic O-H
bond. The former parameter was determined in hydro-
carbon solution by measuring the rate constants for the
inhibition of the thermally initiated autoxidation reaction
of styrene (kinh)8 and the latter one by using the EPR
radical equilibration technique.9
The investigated compounds (see Scheme 1) include the
cis and trans isomers of 4′-hydroxy-3,5-dimethoxystilbene
(2c and 2t), 3′-methoxy-4′-hydroxy-3,5-dimethoxystilbene
(3c and 3t), and 3′,4′-dihydroxy-3,5-dimethoxystilbene (4c
and 4t), prepared as described previously.5 Since the
determination of the O-H BDE value by means of the
EPR method requires phenols giving reasonably long-
lived phenoxyl radicals, we also synthesized the trans and
cis isomers of 3′,5′-di-tert-butyl-4′-hydroxy-3,5-dimethox-
ystilbene (5c and 5t) and the trans isomer 3′,4′-dihy-
droxy-5′-tert-butyl-3,5-dimethoxystilbene (6t) which, due
to the presence in the positions ortho to the OH group of
bulky substituents, afford by irradiation phenoxyl radi-
cals persistent enough to be studied by EPR spectroscopy.
a
Reagents and conditions: (a) pyridinium p-toluensulfonate,
CH3OH, reflux; (b) anhydrous KF, HBr 48%, DMF, rt.
Resu lts a n d Discu ssion
Syn th etic P r oced u r es. The preparation of new de-
rivatives 5c-t and 6t was accomplished by means of the
Wittig reaction between the appropriate aromatic alde-
hydes 8 and 9 and the suitable aromatic ylide, in turn
obtained from the phosphonium salt 7 (Scheme 2). Bond
geometries of compounds were established by comparing
1
the HNMR of the isomeric pairs. Z isomers showed the
(5) J ang, M.; Cai, L.; Udeani, G. O.; Slowing, K. U.; Thomas, C. F.;
Beecher, C. W. W.; Fong, H. H. S.; Farnworth, R. N.; Kinghorn, A. D.;
Metha, R. G.; Moon, R. C.; Pezzuto, J . M. Science 1997, 275, 218-220.
Fre´mont, L. Life Sci. 2000, 66, 663-673. Matsuda, H.; Morikawa, T.;
Toguchida, I.; Park, J .-Y.; Harima, S.; Yoshikawa, M. Bioorg. Med.
Chem. 2001, 9, 41.
(6) Roberti, M.; Pizzirani, D.; Simoni, D.; Rondanin, R.; Baruchello,
R.; Bonora, C.; Buscemi, F.; Grimaudo, S.; Tolomeo, M. J . Med. Chem.
2003, 46, 3546-3554.
(7) (a) Waffo Teguo, P.; Fauconneau, B.; Deffieux, G.; Huguet, F.;
Vercauteren, J .; Merillon, J .-M. J . Nat. Prod. 1998, 61, 655-657. (b)
Stivala, L. A.; Savio, M.; Carafoli, F.; Perucca, P.; Bianchi, L.; Maga,
G.; Forti, L.; Pagnoni, U. M.; Albini, A.; Prosperi, E.; Vanini, V. J . Biol.
Chem. 2001, 276, 22586-22594.
(8) Burton, G. W.; Doba, T.; Gabe, E. J .; Hughes, L.; Lee, F. L.;
Prasad, L.; Ingold, K. U. J . Am. Chem. Soc. 1985, 107, 7053-7065
and references therein.
olefinic protons at 0.3-0.4 ppm higher field compared
with the E isomers. The coupling constants of the vinylic
protons of the E isomers were about 16 Hz, whereas the
Z isomers showed coupling constants of 12 Hz. Depro-
tection of the methoxymethoxy derivatives 10c-t, per-
formed with pyridinium p-toluensolfonate (Scheme 3),
gave the desired compounds 5t and 5c, whereas removal
of the tert-butyldimethylsilyl (TBDMS) group from stil-
bene 11t using potassium fluoride afforded compound 6t
in appreciable yield. Under the same conditions, the cis
isomer 11c isomerized to the more stable trans derivative
6t.
(9) (a) Lucarini, M.; Pedrielli, P.; Pedulli, G. F.; Cabiddu, S.;
Fattuoni, C. J . Org. Chem. 1996, 61, 9259-9263. (b) Brigati, G.;
Lucarini, M.; Mugnaini, V.; Pedulli, G. F. J . Org. Chem. 2002, 67,
4828-4832.
Kin etic Mea su r em en ts. The determination of the
rate constants for the reaction with peroxyl radicals of
the above derivatives (kinh) was made by studying the
7102 J . Org. Chem., Vol. 69, No. 21, 2004

Antioxidant Activity of Hydroxystilbene Derivatives
Rox,0
Rox
Rox
nkinh[AH]0
-
)
(7)
Rox,0
2k R
x
t
i
The term n represents the stoichiometric coefficient,
i.e., the number of peroxyl radicals trapped by each
antioxidant molecule, and can be determined from eq 8
by measuring the length of the induction period τ.
Riτ
F IGURE 1. Oxygen consumption observed during the au-
toxidation of styrene (4.30 M) in chlorobenzene initiated by
AMVN (5 × 10-3 M) at 30 °C in the absence of antioxidant
(dashed line) and in the presence of R-tocopherol (R-TOH) and
of one of the various trans (plot A) and cis (plot B) isomers of
the hydroxystilbenes 2-4, each 5.5 × 10-6 M.
n )
(8)
[AH]
In the present cases, the stoichiometric factor was
determined experimentally for several compounds, as
reported in Table 1, by studying the inhibited autoxida-
tion of cumene which, due to the smaller kp value, gives
measurable inhibition periods more easily than styrene.12
An examination of the two graphs shows that all
hydroxystilbenes behave as weak antioxidants with the
exception of the derivative 4t, where the good antioxidant
activity is due to the presence of the catechol ring.13 When
comparing the geometrical isomers of each compound, it
appears that the trans isomers are characterized by
inhibition rate constants approximately twice as large
as those of the cis isomers. These data are consistent with
a previous7b rough determination of the relative antioxi-
dant activity of cis- and trans-resveratrol 1, which was
found to be higher for the latter isomer. It was therefore
important to confirm the above result by an independent
experimental method.
EP R Sp ectr a a n d BDE Va lu es. Another experimen-
tal parameter closely related to the antioxidant activity
of phenolic derivatives is the bond dissociation enthalpy
(BDE) of the O-H bond. The method we have been
extensively using during the last 10 years consists of the
determination of the O-H BDE value of a given phenol,
ArOH, by measuring, by means of EPR spectroscopy, the
equilibrium constant, Ke, for its reaction with a phenoxyl
radical, Ar′O•, whose heat of formation is known (eq 9)
under the assumption that the entropic term can be
neglected (eq 10).9a,13
inhibition of the thermally initiated autoxidation of
styrene10 (eqs 1-6).
Ri
9
initiator
8R•
(1)
(2)
R• + O2 f ROO•
kp
9
ROO• + RH
8 ROOH + R•
(3)
(4)
2kt
ROO• + ROO• 8 products
kinh
ROO• + ArOH
8 ROOH + ArO•
(5)
(6)
ROO• + ArO• f products
The reaction was followed by monitoring the oxygen
consumption during the autoxidation with an automatic
recording gas absorption apparatus, built in our labora-
tory and described previously,11 which contains as detec-
tor a commercial differential pressure transducer. The
reactions (initiated by the thermal decomposition of
AMVN (2,2′-azobis(2,4-dimethylvaleronitrile)) were car-
ried out at 30 °C under controlled conditions in an air-
saturated solution of styrene, both in the absence and in
the presence of various amounts of antioxidants. R-To-
copherol was used as reference chain breaking inhibitor.
Figure 1 shows that most of the investigated com-
pounds do not give origin to a definite induction period
(τ), but rather to a retarding of the autoxidation reaction.
The inhibition rate constants kinh (eq 5) of these com-
pounds were determined by using a kinetic treatment11
which consisted of measuring the initial rates of oxidation
of styrene both in the presence (-d[O2]/ dt ) Rox) and in
the absence ((-d[O2]/ dt)0 ) Rox,0) of antioxidant, AH, and
calculating kinh from these data by using eq 7. To use eq
7, we also need the termination constant 2kt (eq 4), which
for styrene is known from the literature10 as 4.2 × 107
M-1 s-1 at 30 °C and the initiation rate Ri determined in
preliminary experiments as described in the Experimen-
tal Section.
ArOH + Ar′O• h ArO• + Ar′OH
(9)
BDE(ArO-H) = BDE(Ar′O-H) - RT ln(Ke) (10)
However, this technique requires that the investigated
compounds give phenoxy radicals reasonably long-lived,
while the hydroxystilbenes investigated so far (1-4)
afford, by photolysis, transient species. We therefore
synthesized the cis and trans isomers of two other
derivatives, 5 and 6, containing tert-butyl substituents
ortho to the OH group, since this kind of substitution
generally increases the persistency of the corresponding
phenoxyl radicals by preventing, for steric reasons, their
decay by dimerization.
Actually, the phenoxyl radical from the trans isomer
of the di-tert-butyl-substituted derivative 5t, obtained by
photolyzing at room temperature a deoxygenated benzene
(10) Howard, A. In Free Radicals; Kochi, J . K., Ed.; Wiley: New
York, 1973; Chapter 12, Vol. 2. Denisov, E. T. Liquid-Phase Reaction
Rate Constants; Plenum: New York, 1974.
(11) Amorati, R.; Pedulli, G. F.; Valgimigli, L.; Attanasi, O. A.;
Filippone, P.; Fiorucci, C.; Saladino, R. J . Chem. Soc., Perkin Trans. 2
2001, 2142-2146. Chatgilialoglu, C.; Timokhin, V. I.; Zaborovskyi, A.
B.; Lutsyk, D. S.; Prystansy, R. E. Chem Commun. 1999, 405-406.
Foti, M.; Ruberto, G. J . Agric. Food Chem. 2001, 49, 342-348.
(12) Horswell, E. C.; Howard, J . A.; Ingold, K. U. Can. J . Chem.
1966, 44, 986-991.
(13) Lucarini, M.; Mugnaini, V.; Pedulli, G. F. J . Org. Chem. 2002,
67, 928-931. Lucarini, M.; Pedulli, G. F.; Guerra, M. Chemistry Eur.
J . 2004, 10, 933-939. Amorati, R.; Ferroni, F.; Lucarini, M.; Pedulli,
G. F.; Valgimigli, L. J . Org. Chem. 2002, 67, 9295-9303.
J . Org. Chem, Vol. 69, No. 21, 2004 7103

Amorati et al.
TABLE 1. Ra te Con sta n ts kIn h for th e Rea ction w ith
P er oxyl Ra d ica ls in P h Cl a t 30 °C a n d OH Bon d
Dissocia tion En th a lp ies (BDE) of Hyd r oxystilben es
since photolysis of a solution of benzene, di-tert-butyl
peroxide and 5c afforded a spectrum identical to that
obtained from 5t and assigned to the corresponding trans
radical. This result is not unexpected since it is known
that the radical anion and cation of stilbene undergo cis-
trans interconversion much faster than the corresponding
parent molecule due to the lower barrier to internal
rotation.14-18
On the basis of these results, it is conceivable that also
the phenoxyl radicals from hydroxyl stilbenes are char-
acterized by a rate of cis to trans conversion larger than
that of the parent compounds, thus making the detection
of the less stable cis radical difficult at room temperature.
Experiments were therefore carried out at -65 °C in
degassed toluene containing either 5t or 5c and 10% di-
tert-butyl peroxide. In the case of 5t, an EPR spectrum
similar to that recorded at room temperature and inter-
preted in terms of the following hyperfine splitting
constants, a1H ) 6.51 G, a1H ) 2.67 G, a3H ) 1.65 G, a3H
) 1.45 G, was observed. When the solution containing
the cis isomer 5c was photolyzed at -65 °C, a spectrum
clearly consisting of the superimposition of two species
was detected. One of them (ca. 50%) was easily identified
as the trans-phenoxyl radical on the basis of its spectral
parameters, while the other one was attributed to the
cis radical having the following hyperfine splitting con-
cis
trans
a
a
kinh
compd nb (M-1 s-1
BDE
kinh
nb (M-1 s-1
BDE
)
(kcal mol-1
)
)
(kcal mol-1
)
1
2
3
4
5
6
7.4 × 104
1.9 1.4 × 105 83.7c
1.9 1.3 × 105 83.7c
2.0 1.0 × 105 83.5c
1.9 1.1 × 106 79.3b
2.2 6.3 × 104
1.9 5.7 × 104
1.4 7.6 × 105
2.3 × 104
80.4e
-
6.8 × 104 78.6d (78.9c)
1.6 2.4 × 106 77.7d (77.6c)
a
Inhibition rate constants are believed to be accurate within
10%. Measured using cumene as the oxidizable substrate. c BDE
value calculated using the additive rule.9b The contribution for the
(3,5-dimethoxyphenyl)ethenyl substituent was taken as -3.9 kcal/
b
d
mol. Value measured at room temperature by means of the EPR
radical equilibration technique. e The value for 5c was measured
at -65 °C.
stants a1H ) 8.36 G, a1H ) 3.22 G, a2H ) 1.62 G, a2H
)
F IGURE 2. Room-temperature EPR spectrum of the phenoxyl
radical obtained by photolyzing a deoxygenated solution of 5t
and di-tert-butyl peroxide in benzene.
1.86 G, a1H ) 0.91 G. The EPR spectra changed gradually
upon irradiation, with the amount of the trans species
increasing and that of the cis isomer decreasing with
time.
solution of the precursor containing some di-tert-butyl
peroxide, was reasonably persistent. The observed EPR
spectrum (see Figure 2), indicating that the radical is
formed by abstraction of the hydrogen atom from the OH
group in position 4′, was interpreted in terms of the
following spectral parameters: a1H ) 6.44 G, a1H ) 2.92
G, a2H ) 1.60 G, a3H ) 1.51 G, g ) 2.0040, where the
first two splitting constants are assigned to the vinyl
protons and the remaining ones to the couple of protons
in positions 2′, 6′ of the B ring and to the three protons
in positions 2, 4, and 6 of the A ring, respectively. Since
the shape of the EPR spectrum remains unchanged with
time, it seems reasonable to admit that this radical
retains on formation the same trans geometry of the
diamagnetic precursor and that it is characterized by a
higher geometrical stability than the cis radical. This
conclusion was confirmed by the study carried out on the
cis isomer 5c described below.
The measure of the BDE value of the O-H bond in 5t
was performed by studying the equilibrium constant for
the hydrogen atom exchange reaction with the phenoxyl
radical from 2,6-di-tert-butyl-4-methoxyphenol (BHA),
whose BDE in benzene solution is 78.3 kcal/mol.9a This
phenol was chosen because its BDE value is expected to
be similar to that of 5t (vide infra) and because only a
small portion of the EPR spectrum of the phenoxyl radical
from 5t overlaps with that of the radical from BHA. Due
to this separation, the relative concentrations of the two
radical species could be easily measured by double
integration of the EPR spectra in order to obtain Ke. The
BDE value for the O-H bond of 5t was 78.6 ( 0.5 kcal/
mol.
There are two reasons for the detection of such a large
amount of the trans isomeric radical: one is that the
synthesis of 5c afforded a product containing 5t as an
impurity (although its percent was too small to be
detected by NMR), the second one is that, since 5t is
characterized by an O-H BDE value lower than 5c (vide
infra), the hydrogen atom exchange reaction involving
5c, 5t, and the corresponding phenoxyl radicals is shifted
toward the formation of the trans isomeric radical. Thus,
even in the presence of a very low concentration of 5t, a
remarkable amount of the corresponding radical was
observed.
To obtain the BDE value of the cis isomer we photo-
lyzed at -65 °C a toluene and ButOOBut solution
containing 95% 5c and 5% 5t. Under these conditions,
the ratio between the cis- and trans-phenoxyl radicals
dropped from ca. 1 to 0.2. From these data, we could
obtain the percent of 5t (1.3%) contained as an impurity
in 5c, as well as the equilibrium constant for the
hydrogen atom exchange reaction (eq 9) involving 5c and
5t and the corresponding radicals (Ke ) 0.014). Thus,
from the difference between the BDEs of the cis- and
trans-hydroxyl stilbenes (1.8 kcal/mol) the O-H BDE
value of 5c was obtained as 80.4 kcal/mol, consistent with
the lower reactivity toward peroxyl radicals of the cis
isomers of pterostilbenes (see Table 1).
(14) Taylor, T. W. J .; Murray, A. R. J . Chem. Soc. 1938, 2078.
(15) Kistiakowsky, G. B.; Smith, W. R. J . Am. Chem. Soc. 1934, 56,
638.
(16) Gerson, F.; Ohya-Nishiguchi, H.; Szwarc, M.; Levin, G. Chem.
Phys. Lett. 1977, 52, 587-589.
(17) Pedulli, G. F. Res. Chem. Intermed. 1993, 7, 617-634.
(18) Courtneidge, J . L.; Davies, A. G.; Gregory, P. S. J . Chem. Soc.,
Perkin Trans. 2 1987, 1527-1532.
Attempts to record at room temperature the EPR
spectrum of the corresponding cis-phenoxyl radical failed
7104 J . Org. Chem., Vol. 69, No. 21, 2004

Antioxidant Activity of Hydroxystilbene Derivatives
The higher BDE value found in the cis isomers can be
rationalized as follows. Being the O-H BDE given by the
difference between the energies of the phenoxyl radical
(plus that of the hydrogen atom) and of the starting
phenol, the cis geometry for resveratrol and related
compounds implies that the folding of the structure19
causes a greater destabilization of the phenoxyl radical
than of the parent phenol. It will be shown in the next
section that a density functional theory (DFT) study of
the geometrical isomers of resveratrol correctly predicts
these experimental results.
We also generated inside the EPR cavity the phenoxyl
radical from 6t, containing the catechol group, by pho-
tolyzing this compound in a 9:1 mixture of benzene/
di-tert-butyl peroxide. The room-temperature EPR spec-
trum was interpreted in terms of the following spectro-
F IGURE 3. DFT-computed energy differences (kcal/mol) for
the cis and trans isomers of 1 and of the corresponding
phenoxyl radicals at the B3LYP/6-31G* level, in parentheses
with ZPVE corrections.
scopic parameters: a1H ) 5.58 G, a1H ) 2.63 G, a1H
)
1.81 G, a1H ) 0.46 G, a1H(OH) ) 0.98 G, a2H ) 1.32 G,
a1H ) 1.20 G, g ) 2.0042. The first two constants were
assigned to the vinyl protons, the last two to the three
protons of the A ring, and the 1.81 and 0.46 G splittings
to the two protons of the B ring. The determination of
the OH bond dissociation enthalpy of this catechol
derivative was made by studying its equilibration with
2,6-di-tert-butyl-4-methoxyphenol (BHA) under continu-
ous photolysis. A BDE value of 77.7 kcal/mol was
obtained in benzene solution for 6t.
errors. Absolute BDEs are obtained by adding to this
term the experimental BDE value for phenol (eq 13).
∆BDEcalc(4-XC6H4OH) ) [E(4-XC6H4O•) -
E(4-XC6H4OH)] - [E(C6H5O•) - E(C6H5OH)] (12)
BDEcalc(4-XC6H4OH) ) BDEexp(C6H5OH) +
∆BDEcalc(4-XC6H4OH) (13)
With those hydroxystilbenes giving by photolysis
scarcely persistent phenoxyl radicals, the O-H BDE
values could not be determined experimentally but the
O-H bond strength could be estimated using the addi-
tivity rule.9b Actually, it is known that the change of the
O-H bond strength due to a given substituent is ap-
proximately constant in the variously substituted phenols
and that this contribution may be used to estimate the
BDE of polysubstituted phenols. The contributions of
o-tert-butyl, o-methoxyl, and o-hydroxyl substituents have
been reported in previous papers,9b,13 while that due to
the (3,5-dimethoxyphenyl)ethenyl substituent could be
estimated from the experimental BDEs of 5t (-4.2 kcal/
mol) and 6t (-3.8 kcal/mol). Since these values are quite
close to that reported for the phenylethenyl substituent
(-3.9 kcal/mol),9b the two methoxy groups in the meta
positions of the A phenyl ring have only a negligible effect
on the BDE of hydroxystilbenes. It is therefore possible
to estimate the O-H BDE value for the hydroxyl group
in the 4′ position of the B ring in all the investigated
compounds, as shown in Table 1.
The DFT results indicate that substitution of the para
hydrogen atom of phenol with either a cis- or a trans-
3,5-dimethoxystyryl substituent causes a decrease of the
O-H BDE value, since this reduction is somewhat higher
for the trans (-5.4 kcal/mol)21 than for the cis isomer
(-4.2 kcal/mol). These data should be compared to the
corresponding experimental ∆BDE values, i.e., -4.2 and
-2.4 kcal/mol for the trans and cis isomer, respectively.
Although the computed ∆BDE are larger than the
experimental ones, their differences are similar, being
1.2 kcal/mol for the calculated and 1.8 kcal/mol for the
measured values.
To rule out any uncertainty about the effect of remote
substituents in positions 3 and 5, similar calculations
were carried out also on the cis and the trans isomers of
resveratrol (1) and 4-hydroxystilbene. The results, re-
ported in Table S1 (Supporting Information), clearly show
that the influence of remote substituents is small on the
calculated BDE values (<0.3 kcal/mol) as well as on the
energy difference between the cis and the trans isomers
(<0.3 kcal/mol).
The reason the cis isomer possesses a slightly lower
∆BDE is shown in Figure 3, which reports the differences
between the energies of the isomeric phenols and those
of the isomeric phenoxyl radicals. Actually, the folded and
therefore strained structure of the cis isomer phenol (2c)
is characterized by an energy value higher by 5.3 kcal/
mol than that of the trans isomer (2t). On the other hand,
in the corresponding phenoxyl radicals the difference be-
tween the energies of the two isomers increases to 6.4
kcal/mol, indicating that the destabilization associated
Com p u ta tion a l Stu d ies. The gas-phase bond dis-
sociation enthalpies of the phenolic OH bond of both trans
and cis isomers of 4′-hydroxy-3,5-dimethoxystilbene were
also calculated by means of the DFT method at the
B3LYP/6-31G* level using the isodesmic reaction showed
in eq 11,20 where X represents either the cis- or trans-
3,5-dimethoxystyryl substituent.
4-XC6H4OH + C6H5O• h C6H5OH + 4-XC6H4O•
(11)
This method yields reliable ∆BDE values (eq 12) for
the various substituted phenols,20 due to cancellation of
(21) The ∆BDE values obtained in this work for the trans-3,5-
dimethoxystyryl substituent (-5.4 kcal/mol) is significantly lower than
what obtained by Wright et al.22 for the trans-styryl (-8.5 kcal/mol)
and the trans-3,5-dihydroxystyryl (-8.2 kcal/mol) substituent by using
the locally dense basis sets (LDBS) method.
(19) Traetteberg, M.; Frantsen, E. B. J . Mol. Struct. 1975, 26, 69.
(20) Lucarini, M.; Mugnaini, V.; Pedulli, G. F.; Guerra, M. J . Am.
Chem Soc. 2003, 125, 8318-8329.
J . Org. Chem, Vol. 69, No. 21, 2004 7105

Amorati et al.
layer chromatography (TLC) on precoated silica gel plates
(Kieselgel 60 F254). Flash column chromatographies were
performed on silica gel (particle size 40-63 µM). Melting points
were determined in open capillary tube and are uncorrected.
1H NMR and 13C NMR spectra were recorded at 300 and 75
MHz, respectively, in CDCl3 solution unless otherwise noted.
Chemical shifts (δH) are reported relative to TMS as internal
standard. IR-FT spectra were recorded in CCl4 solution; νmax
with the adoption of the cis geometry is larger in the rad-
ical than in the phenol. Calculations also show that the
spin densities on the A ring of 2c are lower than those of
2t while for the B ring the reverse is true (Figure 1S,
Supporting Information). The above considerations are
not modified by ZPVE corrections (see Figure 3 and Table
S1, Supporting Information). In conclusion, the BDE
difference between cis and trans isomers can be explained
in terms of reduced conjugative interactions between the
two rings of 2c, this having a greater effect on the radical,
where conjugation is more effective than in the phenol.
is expressed in cm-1
.
Compound 8 was obtained from the commercially available
3,5-di-tert-butyl-4-hydroxybenzaldheyde hemihydrate by pro-
tection of phenol with a methoxymethyl group (MOM) follow-
ing standard procedures. Compound 9 was obtained from the
known 3-tert-butyl-4,5-dihydroxybenzaldehyde23 by protection
of the hydroxyl groups with tert-butyldimethylsilyl groups
(TBDMS) following standard procedures.
Con clu sion s
The kinetic investigation of the reaction between
peroxyl radicals and substituted hydroxystilbenes, here
reported, has shown that the antioxidant activity of the
cis geometrical isomers of these compounds is in all
examined cases worse, by a factor of about 2, than that
of the corresponding trans isomers. This lower reactivity
is fully justified on a thermodynamic basis by the
experimental determination of the O-H bond dissocia-
tion enthalpy (BDE) in the two geometric isomers of 3′,5′-
di-tert-butyl-4′-hydroxy-3,5-dimethoxystilbene (5c and
5t). The BDE value for the cis isomer has been actually
found to be larger by 1.8 kcal/mol than that for the trans
isomer, this making the transfer of the hydroxyl hydrogen
from the latter one to peroxyl radicals easier. DFT
calculations, besides predicting a difference (1.2 kcal/mol)
between the strengths of the O-H bond in the two
isomers of 4′-hydroxy-3,5-dimethoxystilbene (2) similar
to the experimental value (1.8 kcal/mol) determined for
the two isomers of 5, provide an explanation of this result.
In fact, the DFT data indicate that, although the cis
geometry implies a destabilization of both phenoxyl
radical and parent hydroxystilbene, destabilization of the
phenoxyl radical is larger as a result of a reduced
delocalization of the unpaired electron on the styryl group
(see the Supporting Information).
The present data also corroborate our previous finding
that in homogeneous solution resveratrol, despite its
excellent reputation, is only a mild antioxidant.4 Between
the various hydroxystilbenes presently investigated, a
remarkable antioxidant efficacy is only exhibited by the
trans isomers of 4 and 6, which, however, should be
attributed to the presence in these compounds of the
catecholic function rather than of the arylvinyl group.
A comparison of the present results with previously
reported data on the proapoptotic activity of these
stilbenoids, for which the cis stereochemistry has been
shown to be associated with a stronger activity than the
trans geometry,6 suggests that the reverse is true for the
antioxidant ability of these compounds. Thus, it seems
that from the structural point of view there is no
correlation between these two properties and that stil-
benoids characterized by a better antioxidant behavior
possess lower proapoptotic activities. To further explore
this point, studies of structure-activity relationships
(SAR) are under way.
Gen er a l P r oced u r e for P r ep a r in g Stilben es 10c-t a n d
11c-t. To a suspension of the phosphonium bromide salt 7
(1.0 equiv) in anhydrous tetrahydrofuran at -78 °C was added
n-butyllithium (2.5 M in hexanes, 1.0 equiv), and the resulting
red solution was stirred under nitrogen for 1 h. A solution of
aldehydes 8 and 9 (1.0 equiv) in tetrahydrofuran was added
dropwise over 30 min, and the reaction mixture was stirred
for 2 h at room temperature. The resulting suspension was
poured into water and extracted with CH2Cl2. The organic
phase was washed with brine (2 × 15 mL) and dried over
Na2SO4, and the solvent was removed in vacuo to afford a
mixture of the cis/trans-stilbenes 10c-t and 11c-t that were
separated by flash chromatography. The cis-stilbenes were
eluted first, followed by the trans isomers.
3′,5′-Di-ter t-b u t yl-4′-m et h oxym et h oxy-3,5-d im et h ox-
ystilben es (10c-t). Reaction of phosphonium bromide salt 7
and 3,5-di-tert-butyl-4-methoxymethoxybenzaldehyde 8 (1.2 g,
4.31 mmol) gave a mixture of cis-stilbene 10c and trans-isomer
10t. Flash chromatography: petroleum ether/ethyl acetate
9.98:0.02.
10c: 0.85 g (yield 50%); red oil; 1H NMR δ 1.43 (s, 18H),
3.71 (s, 3H), 3.75 (s, 6H), 4.98 (s, 2H), 6.43-6.37 (m, 3H), 6.50
(m, 2H), 6.60 (m, 2H); 13C NMR δ 31.9, 35.6, 55.1, 57.2, 99.4,
100.4, 106.6, 127.5, 129.1, 130.9, 131.5, 139.6, 143.8, 153.5,
160.5.
1
10t: 0.65 g (yield 38%); white solid; mp 62 °C; H NMR δ
1.55 (s, 18H), 3.71 (s, 3H), 3.88 (s, 6H), 4.98 (s, 2H), 6.43-
6.45 (t, 1H, J ) 2.2 Hz), 6.74-6.73 (m, 2H), 6.85 (d, 1H, J )
16 Hz), 7.13 (d, 1H, J ) 16.2 Hz), 7.48 (m, 2H); 13C NMR δ
32.0, 35.8, 55.3, 57.3, 99.7, 100.5, 104.3, 124.8, 127.3, 129.4,
131.7, 139.5, 144.5, 154.3, 160.8.
3′,4′-Di-(ter t-b u t yld im et h ylsilyloxy)-5′-ter t-b u t yl-3,5-
d im eth oxystilben es (11c, 11t). Reaction of phosphonium
bromide salt 7 and 3-tert-butyl-4,5-di-(tert-butyldimethylsily-
loxy) benzaldehyde 9 (0.28 g, 0.66 mmol) gave a mixture of
the cis-11c and trans-11t isomers. Flash chromatography:
petroleum ether/ethyl acetate 9.5:0.5.
1
11c: 0.10 g (yield 27%); pale yellow oil; H NMR δ 0.16 (s,
6H), 0.27 (s, 6H), 0.83 (s, 9H), 0.96 (s, 9H), 1.27 (s, 9H), 3.69
(s, 6H), 6.30-6.32 (t, 1H, J ) 2.2 Hz), 6.44-6.46 (m, 4H), 6.70
(d, 1H, J ) 2.2 Hz), 6.86 (d, 1H, J ) 2.2 Hz); 13C NMR δ -2.7,
-1.2, 19.2, 19.4, 26.6, 27.2, 30.5, 35.1, 55.2, 99.7, 106.5, 118.9,
121.4, 128.3, 130.8, 139.8, 140.6, 145.0, 146.6, 160.6.
1
11t: 0.11 g (yield 30%); pale yellow oil; H NMR δ 0.32 (s,
6H), 0.35 (s, 6H), 0.95 (s, 9H), 1.00 (s, 9H), 1.44 (s, 9H), 3.86
(s, 6H), 6.30-6.33 (t, 1H, J ) 2.2 Hz), 6.67 (m, 2H), 6.82 (d,
1H, J ) 16.2 Hz), 6.98 (d, 1H, J ) 16.2 Hz), 6.99 (m, 1H), 7.07
(m, 1H); 13C NMR δ -2.4, -1.2, 19.36, 19.38, 26.7, 27.2, 30.6,
35.3, 55.4, 99.5, 104.3, 116.0, 119.3, 126.2, 128.6, 129.7, 139.9,
141.3, 145.9, 147.2, 160.9.
Gen er a l P r oced u r e for th e Clea va ge of th e Meth -
oxym eth yl Gr ou p of Com p ou n d s 10c-t. A solution of the
Exp er im en ta l Section
Ma ter ia ls a n d Meth od s. All solvents were redistilled prior
to use. All reactions were carried out under an inert atmo-
sphere. Solvent extracts of aqueous solutions were dried over
anhydrous sodium sulfate. Reactions were monitored by thin-
(22) Wright, J . S.; J ohnson E. R.; DiLabio, G. A. J . Am. Chem. Soc.
2001, 123, 1173-1183.
(23) Shultz, D. A.; Bodnar, S. H.; Vostrikova, K. E.; Kampf, J . W.
Inorg. Chem. 2000, 39, 6091-6093.
7106 J . Org. Chem., Vol. 69, No. 21, 2004

Antioxidant Activity of Hydroxystilbene Derivatives
methoxymethoxy-protected stilbenes 10c-t (1 equiv) and
pyridinium p-toluensulfonate (10 equiv) in methanol (10 mL)
was refluxed for 15 h. The reaction mixture was cooled to room
temperature and the solvent removed under reduced pressure.
The residue was triturated with diethyl ether, and the solids
were filtered off and washed with ether. The organic phase
was dried over Na2SO4 and concentrated under reduced
pressure to give the crude product which was purified by
chromatography (petroleum ether/ethyl acetate 97:3).
cis-3′,5′-Di-t er t -b u t yl-4′-h yd r oxy-3,5-d im e t h oxyst il-
ben e (5c). Compound 5c was obtained from 10c (0.425 g, 1.07
mmol): 0.13 g (yield 33%); yellow oil; 1H NMR δ 1.36 (s, 18H),
3.71 (s, 6H), 5.22 (s, 1H), 6.36-6.33 (t, 1H, J ) 2.2 Hz), 6.46
(d, 1H, J ) 11.6 Hz), 6.47 (m, 2H), 6.55 (d, 1H, J ) 12 Hz),
7.14 (m 2H); 13C NMR δ 30.2, 34.2, 55.2, 99.4, 106.5, 126.1,
127.86, 127.93, 131.3, 135.4, 140.1, 153.1, 160.6; IR νmax (CCl4)
cm-1 1015, 1069, 1098, 1155, 1206, 1260, 1437, 1458, 1595,
2958, 3642.
tr a n s-3′,5′-Di-ter t-bu tyl-4′-h yd r oxy-3,5-d im eth oxystil-
ben e (5t). Compound 5t was obtained from 10t (0.43 g, 1.08
mmol): 0.26 g (yield 65%); yellow powder; mp 106 °C; 1H NMR
δ 1.51 (s, 18H) 3 0.86 (s, 6H), 5.31 (s, 1H), 6.41-6.38 (t, J )
2.2 Hz, 1H), 6.69 (m, 2H), 6.89 (d, 1H, J ) 16 Hz), 7.08 (d, 1H,
J ) 16.2 Hz), 7.37 (m 2H); 13C NMR δ 30.3, 34.4, 55.3, 99.5,
104.2, 123.4, 125.7, 127.6, 128.3, 130.0, 136.1, 139.9, 153.8,
160.8; IR νmax (CCl4) cm-1 958, 1063, 118, 1155, 1205, 1437,
1458, 1592, 1693, 2958, 3641. Anal. Calcd for C24H32O3: C,
78.22; H, 8.75. Found: C, 78.25; H, 8.71.
tr a n s-3′,4′-Dih yd r oxy-5′-ter t-b u t yl-3,5-d im et h oxyst il-
ben e (6t). To the trans-silyloxy-protected stilbene derivative
11t (0.11 g, 0.19 mmol) in dry DMF (3-6 mL for mmol of
TBDMS group) were added anhydrous KF (0.04 g, 0.79 mmol)
and a catalytic amount of 48% aqueous HBr (7 µL, 0.06 mmol).
The reaction mixture was stirred under nitrogen at room
temperature for 4 h. The mixture was poured into 2 N HCl (2
× 15 mL), and the aqueous layer was extracted with diethyl
ether (3 × 15 mL). The combined organic layers were washed
with brine (2 × 15 mL) and dried over Na2SO4, and the solvent
was evaporated. The residue was purified by flash chroma-
tography (petroleum ether/ethyl acetate 9:1) to give 6t: (0.026
g, yield 43%): reddish-orange solid; mp 48-52 °C; 1H NMR δ
1.46 (s, 9H) 3.85 (s, 6H), 5.06 (br, 1H), 5.76 (s, 1H), 6.38-6.40
(t, 1H, J ) 2.2 Hz), 6.65-6.66 (m, 2H), 6.83 (d, 1H, J ) 16.6
Hz), 6.99 (d, J ) 15.8 Hz, 1H), 7.03-7.01 (m, 2H); 13C NMR δ
29.4, 34.7, 55.4, 99.5, 104.3, 110.1, 118.8, 126.0, 128.3, 129.5,
136.4, 139.8, 143.1, 143.7, 160.9; IR νmax (CCl4) cm-1 1068,
1155, 1205, 1260, 1294, 1429, 1457, 1507, 1596, 2958, 3546,
3615. Anal. Calcd for C20H24O4: C, 73.15; H, 7.37. Found: C,
73.18; H, 7.35.
Th er m och em istr y. To determine the BDE value of 5t,
solutions were prepared by using as solvent benzene in the
presence of a 10% v/v of di-tert-butylperoxide and the two
phenols in concentration ratios: [5t]/[BHA] ) 1.16 and [5t]/
[BHA] ) 1.76. The solutions were sealed under nitrogen in a
suprasil quartz EPR tube. The sample was inserted in the
thermostated cavity of an EPR spectrometer and photolyzed
with the unfiltered light from a 500 W high-pressure mercury
lamp. The temperature was controlled with a standard vari-
able-temperature accessory and was monitored before and
after each run with a copper-constantan thermocouple. ∆G
values of -0.30 kcal/mol and -0.37 kcal/mol were found from
the EPR spectra, corresponding to an average ∆BDE value of
- 0.34 kcal/mol and therefore to a BDE value for 5t of 78.6
kcal/mol.
A difference reference compound has also been used: 3,5-
di-tert-butyl-4-hydroxystilbene.9b Since the two phenols give
phenoxyl radicals with very similar EPR spectra, their mixture
provided only a qualitative estimate of the O-H BDE value
of 5t but could not be used for accurate measurements. To
determine the BDE of 5c, we used conditions similar to the
ones described above, apart from the temperature that was
kept at 209 K and checked before and after every measurement
with a copper-constantan thermocouple. These measurements
were made in toluene, and the trans derivative 5t was used
as reference compound at a 5% concentration.
The OH bond dissociation enthalpy of the catechol derivative
6t was made by studying its equilibration with 2,6-di-tert-
butyl-4-methoxyphenol (BHA) under continuous photolysis at
room temperature.
Com p u ta tion a l Deta ils. BDE values and energy differ-
ences between cis and trans isomers of the various stilbene
derivatives were computed following the isodesmic approach,
shown in eqs 11-13, from the total energies obtained from
DFT calculations with the B3LYP functional, using the Gauss-
ian 98 system of programs.27 The unrestricted wave function
was used for radical species. Geometries and single point
calculations were obtained at the B3LYP/6-31G* level of
theory. Stationary points were confirmed by frequency calcula-
tions at the same level of theory. Zero-point vibrational
energies corrections (ZPVE), obtained using a scale factor of
0.9806,28 had negligible effects.
Ack n ow led gm en t. Financial support from MURST
(Research project “Free Radical Processes in Chemistry
and Biology: Fundamental Aspects and Applications in
Environment and Material Sciences”) is gratefully
acknowledged.
Su p p or t in g In for m a t ion Ava ila b le: 1H and 13C NMR
spectral data for compounds 5c, 5t, and 6t. Table S1, Figure
S1: computed geometries and total energies of the cis and
trans isomers of 1, 2, and 4-hydroxystilbene (12) and their
corresponding phenoxyl radicals. This material is available free
of charge via the Internet at http://pubs.acs.org.
Kin etic Mea su r em en ts. The rate constants for the reac-
tion of hydroxystilbenes with peroxyl radicals have been
measured by following the autoxidation of styrene (4.30 M) at
303 K using as initiator AMVN (5 × 10-3 M). The antioxidant
concentration was kept constant for all measurements (5.5 ×
10-6 M) in order to compare more easily their behavior. Since
the analyzed compounds 4-6 showed a scarce solubility in
chlorobenzene, they were initially dissolved in a small amount
of methanol and then this solution was diluted in chloroben-
zene.24 Initiation rates, Ri, were determined for each condition
in preliminary experiments by the inhibitor method using
R-tocopherol as reference antioxidant: Ri ) 2[R-TOH]/τ.8
Stoichiometric coefficients, n, were determined by using
cumene (6.2 M) as the oxidizable substrate.
J O0497860
(25) Snelgrove, D. W.; Lusztyk, J .; Banks, J . T.; Mulder, P.; Ingold,
K. U. J . Am. Chem. Soc. 2001, 123, 469-477.
(26) Abraham, M. H. Chem. Soc. Rev. 1993, 73-83.
(27) Frisch, M. J .; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J . R.; Zakrzewski, V. G.; Montgomery, J . A.,
J r.; Stratmann, R. E.; Burant, J . C.; Dapprich, S.; Millam, J . M.;
Daniels, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J .;
Barone, V.; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo,
C.; Clifford, S.; Ochterski, J .; Petersson, G. A.; Ayala, P. Y.; Cui, Q.;
Morokuma, K.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.;
Foresman, J . B.; Cioslowski, J .; Ortiz, J . V.; Stefanov, B. B.; Liu, G.;
Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R. L.;
Fox, D. J .; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.;
Gonzalez, C.; Challacombe, M.; Gill, P. M. W.; J ohnson, B. G.; Chen,
W.; Wong, M. W.; Andres, J . L.; Head-Gordon, M.; Replogle, E. S.;
Pople, J . A. Gaussian 98, revision A.7; Gaussian, Inc.: Pittsburgh, PA,
1998.
(24) The small amount of added methanol is expected to decrease
the antioxidant activity of the investigated phenols except for those
containing two o-tert-butyl substituents.25 On the basis of the R and â
values reported by Abraham,26 the reduction of kinh induced by the
methanol present (0.12 M) can be estimated as 47% for the resveratrol
derivatives 1 and 2, 72% for the catechols 4 and 6, and 7% in the case
of the o-methoxy derivatives 3. Since, however, it seems reasonable
that both cis and trans isomers are equally solvated, their relative
reactivity should not be affected by methanol.
(28) Scott, P. A.; Radom, L. J . Phys. Chem. 1996, 100, 16502-16513.
J . Org. Chem, Vol. 69, No. 21, 2004 7107