Antioxidant properties of natural and synthetic chromanol derivatives: Study by fast kinetics and electron spin resonance spectroscopy

Article abstract of DOI:10.1021/jo047927s

(Chemical Equation Presented) Chromanol-type compounds act as antioxidants in biological systems by reduction of oxygen-centered radicals. Their efficiency is determined by the reaction rate constants for the primary antioxidative reaction as well as for disproportionation and recycling reactions of the antioxidant-derived radicals. We studied the reaction kinetics of three novel chromanols: cis- and trans-oxachromanol and the dimeric twin-chromanol, as well as ubichromanol and ubichromenol, in comparison to α-tocopherol and pentamethylchromanol. The antioxidant-derived radicals were identified by optical and electron spin resonance spectroscopy (ESR). The kinetics of the primary antioxidative reaction and the disproportionation of the chromanoxyl radicals were assessed by stopped-flow photometry in different organic solvents to simulate the different polarities associated with biomembranes. Furthermore, the reduction of the chromanoxyl radicals by ubiquinol and ascorbate was measured after laser-induced one-electron chromanol oxidation in ethanol and in a micellar system, respectively. The rate constants showed that twin-chromanol had better radical scavenging properties than α-tocopherol and a significantly slower disproportionation rate of its corresponding chromanoxyl radical. In addition, the radical derived from twin-chromanol is reduced by ubiquinol and ascorbate at a faster rate than the tocopheroxyl radical. Finally, twin-chromanol can deliver twice as many reducing equivalents, which makes this compound a promising new candidate as artificial antioxidant in biological systems.

Full text of DOI:10.1021/jo047927s

Antioxidant Properties of Natural and Synthetic Chromanol
Derivatives: Study by Fast Kinetics and Electron Spin Resonance
Spectroscopy
Wolfgang Gregor,† Gottfried Grabner,‡ Christian Adelwo¨hrer,§ Thomas Rosenau,§ and
Lars Gille*,†
Research Institute for Pharmacology and Toxicology of Oxygen Radicals, University of Veterinary
Medicine Vienna, Vienna, Austria, Max F. Perutz Laboratories, Department of Chemistry,
University of Vienna, Vienna, Austria, and Department of Chemistry,
University of Natural Resources and Applied Life Sciences, Vienna, Austria
Lars.Gille@vu-wien.ac.at
Received November 23, 2004
Chromanol-type compounds act as antioxidants in biological systems by reduction of oxygen-centered
radicals. Their efficiency is determined by the reaction rate constants for the primary antioxidative
reaction as well as for disproportionation and recycling reactions of the antioxidant-derived radicals.
We studied the reaction kinetics of three novel chromanols: cis- and trans-oxachromanol and the
dimeric twin-chromanol, as well as ubichromanol and ubichromenol, in comparison to R-tocopherol
and pentamethylchromanol. The antioxidant-derived radicals were identified by optical and electron
spin resonance spectroscopy (ESR). The kinetics of the primary antioxidative reaction and the
disproportionation of the chromanoxyl radicals were assessed by stopped-flow photometry in
different organic solvents to simulate the different polarities associated with biomembranes.
Furthermore, the reduction of the chromanoxyl radicals by ubiquinol and ascorbate was measured
after laser-induced one-electron chromanol oxidation in ethanol and in a micellar system,
respectively. The rate constants showed that twin-chromanol had better radical scavenging
properties than R-tocopherol and a significantly slower disproportionation rate of its corresponding
chromanoxyl radical. In addition, the radical derived from twin-chromanol is reduced by ubiquinol
and ascorbate at a faster rate than the tocopheroxyl radical. Finally, twin-chromanol can deliver
twice as many reducing equivalents, which makes this compound a promising new candidate as
artificial antioxidant in biological systems.
Introduction
physiologically evolving radicals by chemical antioxidants
on several levels of complexity, ranging from tests in
mono- and biphasic chemical systems and subcellular or
cellular model systems to tests on the organ and in whole
organisms.4 Vitamin E (R-tocopherol, 1), ubiquitous in
biomembranes, is believed to be the most important
lipophilic antioxidant, inhibiting lipid peroxidation by a
chain-breaking radical scavenging mechanism.5-7 Even
though its full repertoire of activities (which also includes
nonantioxidant functions)8 is far from being fully under-
stood, the most complete pharmacological and toxicologi-
The role of reactive oxygen species (ROS; e.g., perox-
ides, free oxygen-centered radicals) in the pathogenesis
of many diseases has been recognized over the past
decades.1-3 This stimulated the exploration of counter-
acting strategies such as inhibition of ROS-forming
enzymes (e.g., xanthine oxidase) or the control of patho-
* Corresponding author: Research Institute for Pharmacology and
Toxicology of Oxygen Radicals, University of Veterinary Medicine
Vienna, Veterina¨rplatz 1, A-1210 Vienna, Austria; tel +43 1 25077
4407; fax +43 1 25077 4490; e-mail Lars.Gille@vu-wien.ac.at.
† University of Veterinary Medicine Vienna.
‡ University of Vienna.
(4) Halliwell, B. Free Radical Res. Commun. 1990, 9, 1-32.
(5) Burton, G. W.; Ingold, K. U. Ann. N.Y. Acad. Sci. 1989, 570,
7-22.
§ University of Natural Resources and Applied Life Sciences.
(1) Sies, H. Am. J. Med. 1991, 91S, 31-38.
(2) Beal, M. F. Ann. Neurol. 1995, 38, 357-366.
(3) McCall, M. R.; Frei, B. Free Radical Biol. Med. 1999, 26, 1034-
1053.
(6) Brigelius-Flohe, R.; Traber, M. G. FASEB J. 1999, 13, 1145-
1155.
(7) Pryor, W. A. Free Radical Biol. Med. 2000, 28, 141-164.
10.1021/jo047927s CCC: $30.25 © 2005 American Chemical Society
Published on Web 03/30/2005
3472
J. Org. Chem. 2005, 70, 3472-3483

Antioxidant Properties of Chromanol Derivatives
cal profile is available for this chromanol, among all
natural and synthetic antioxidants.5-7 Strategies to
discover and design new antioxidants by derivatization
of the evolutionarily favored chromanol structure used
R-tocopherol as a starting structure, leading to new
molecules such as Trolox.9,10 Also the reported regulation
of gene expression and enzyme activity by tocopherol8
stimulated the development and study of new compounds
in this class. While the medical application could profit
from more active antioxidants overcoming transport
limitations of natural antioxidant vitamins, the food and
packaging industry is limited by the complicated tech-
nological properties of natural R-tocopherol and its
uneconomic synthesis.
Unfortunately, in the past the quality of an antioxidant
was often reduced to its radical scavenging activity.
However, detailed studies on the oxidation products of
vitamin E and ubiquinol have demonstrated that both
antioxidant-derived radical intermediates (chromanoxyl
and especially semiquinone radicals) and stable two-
electron oxidation products (quinoid compounds) can
compromise the overall benefit of an antioxidant.11-14
Therefore, not only the bimolecular disproportionation
of these radicals but also their recycling by hydrophilic
or lipophilic cellular reductants must be considered. Since
its discovery in 1979,15 the reduction of the tocopheroxyl
radical by ascorbate (vitamin C) has been widely covered
in biochemical studies,16,17 and evidence for its occurrence
in vivo has been presented.18,19 Later the reduction of this
radical by ubiquinol (reduced coenzyme Q) has been
documented as well.20,21 Yet kinetic information on the
reaction of these and related radicals is scarce. Further-
more, no single study so far has presented kinetic data
for both the primary antioxidative reaction and the
disproportionation and recycling reactions of the chro-
manoxyl radical, measured under comparable conditions,
not even for vitamin E. Here we report a complete set of
rate constants for these reactions to characterize novel
synthetic antioxidants derived from the chromanol core
of vitamin E (twin-chromanol and oxachromanols) and
similar compounds synthesized from naturally occurring
ubiquinone (ubichromanol and ubichromenol).
been synthesized by Rosenau and co-workers to explore
the redox properties of new chromanol structures.
Ubichromenol was isolated from various animal tissues
as well as synthesized from ubiquinone25 and received
some attention as a hypothetical intermediate in the
overall process of oxidative phosphorylation. Whereas
this idea was rejected, the reported antioxidant proper-
ties of this compound26 seem more conceivable, although
it is still unknown if ubichromenol is biosynthesized as
a functional compound or if it is just a minor degradation
(“waste”) product. On the other hand, the related ubichro-
manol was never isolated from nature, but has also been
synthesized from ubiquinone.25
The aim of this study was a comprehensive description
of the antioxidant-related kinetic properties of the novel
as well as ubiquinol-derived chromanols in comparison
to vitamin E; to this end we investigated the influence
of chromanol derivatization on the reactivities of both the
parent chromanol and the chromanoxyl radical. Since in
biological systems such antioxidants may function in
environments with different physicochemical properties
(e.g., the core of the lipid membranes or the polar
headgroup region of phospholipids), we extended this
kinetic investigation to various solvents, which were
employed to mimic the environment in different mem-
brane regions. On the basis of these results, the most
promising antioxidant will be selected for future tests in
biological systems.
Results
Compounds Studied. We studied the antioxidant
properties of phenolic compounds (PhOH) that have
recently been synthesized de novo s twin-chromanol
(3)22 and oxachromanols (4, 5)23,24 s or derived from natu-
ral ubiquinone (UQ-10) by chemical modification s
ubichromanol (6) and ubichromenol (7). Dimeric chro-
manol structures and benzodioxines have not been
(9) Giulivi, C.; Romero, F. J.; Cadenas, E. Arch. Biochem. Biophys.
1992, 299, 302-312.
(10) Iuliano, L.; Pedersen, J. Z.; Camastra, C.; Bello, V.; Ceccarelli,
S.; Violi, F. Free Radical Biol. Med. 1999, 26, 858-868.
(11) Siegel, D.; Bolton, E. M.; Burr, J. A.; Liebler, D. C.; Ross, D.
Mol. Pharmacol. 1997, 52, 300-305.
(12) Gille, L.; Staniek, K.; Nohl, H. Free Radical Biol. Med. 2001,
30, 865-876.
(13) Moore, A. N.; Ingold, K. U. Free Radical Biol. Med. 1997, 22,
931-934.
(14) Nohl, H.; Gille, L.; Kozlov, A. V. Free Radical Biol. Med. 1998,
25, 666-675.
(15) Packer, J. E.; Slater, T. F.; Willson, R. L. Nature 1979, 278,
737-738.
(16) Bisby, R. H.; Parker, A. W. FEBS Lett 1991, 1, 205-208.
(17) Mukai, K.; Nishimura, M.; Kikuchi, S. J. Biol. Chem. 1991, 266,
274-278.
(18) Huang, J.; May, J. M. Mol. Cell. Biochem. 2003, 247, 171-176.
(19) Kagan, V. E.; Kuzmenko, A. I.; Shvedova, A. A.; Kisin, E. R.;
Li, R.; Martin, I.; Quinn, P. J.; Tyurin, V. A.; Tyurina, Y. Y.; Yalowich,
J. C. Biochim. Biophys. Acta 2003, 1620, 72-84.
(20) Kagan, V.; Serbinova, E.; Packer, L. Biochem. Biophys. Res.
Commun. 1990, 169, 851-857.
(21) Mukai, K.; Kikuchi, S.; Urano, S. Biochim. Biophys. Acta 1990,
1035, 77-83.
(22) Rosenau, T.; Potthast, A.; Hofinger, A.; Kosma, P. Angew.
Chem., Int. Ed. 2002, 41, 1171-1173.
(23) Rosenau, T.; Potthast, A.; Elder, T.; Lange, T.; Sixta, H.; Kosma,
P. J. Org. Chem. 2002, 67, 3607-3614.
Twin-chromanol22 and several oxachromanol deriva-
tives including two dimethyl isomers23,24 have recently
(24) Adelwo¨hrer, C.; Rosenau, T.; Gille, L.; Kosma, P. Tetrahedron
2003, 59, 2687-2691.
(25) Hatefi, Y. Adv. Enzymol. 1963, 25, 275-328.
(26) Burlakova, E. B.; Storozhok, M. M.; Khrapova, N. G. Vopr.
Pitan. 1990, 4, 53-58.
(8) Azzi, A.; Stocker, A. Prog. Lipid Res. 2000, 39, 231-255.
J. Org. Chem, Vol. 70, No. 9, 2005 3473

Gregor et al.
studied as antioxidants so far, and for 6 and 7 detailed
kinetic information is still missing.27 For reference
purposes, R-tocopherol (1) and pentamethylchromanol (2)
were included in the study. The structures of 6 and 7
have been confirmed by 1H and 13C NMR. The peak
assignments are in agreement with chemical shifts
reported for ubichromanol-0,28 ubichromenol-1,27 ubi-
chromenol-6,29 and the polyprenyl chain of UQ-10.30
For transport processes in biological systems, the
distribution of drugs between the aqueous phase and lipid
membranes is of importance. Therefore, the lipophilicity
of all compounds was assessed both by a theoretical
approach and by experimental measurement of the
octanol/water partition coefficients (P). The theoretical
approach using the structure-activity relationship mod-
ule of HyperChem31 based on an incremental method of
Ghose and co-workers32,33 gave log P values of 9.6, 15.7,
and 15.6 for the extremely lipophilic compounds 1, 6, and
7 (including the side chains), respectively. In contrast,
experimental values for these compounds (1, ca. 6.2; 6
and 7, ca. 5.2) were much smaller but still indicated a
high lipophilicity. For short-chain compounds a better
agreement of experimental and theoretical log P values
was observed. Compounds 2, 3, 4, and 5 gave theoretical
values of 3.8, 5.7, 3.3, and 3.3, respectively. The corre-
sponding experimental values were 3.7, 3.6, 3.2, and 3.2,
respectively.
Identification of Radical Intermediates. It is
known that lipophilic antioxidants exert their protective
function in biological membranes by interception of
reactive oxygen and nitrogen species. Since most of these
species are radicals (R•), the antioxidant reaction can be
characterized as a formal one-electron oxidation (eq 1)
of the radical scavenging compounds including H-atom
transfer and electron-transfer mechanisms.34 Therefore,
this reaction leads to the formation of an antioxidant
radical (PhO•), which in turn can have a significant
influence on the total benefit of an antioxidant.
FIGURE 1. ESR spectra of phenoxyl radicals generated by
UV irradiation of 50 mM solutions of 1-5 in benzene or by
oxidation of 3 mM solutions of 6 and 7 in benzene with PbO2,
all under anaerobic conditions at room temperature. For each
compound, the center multiplets of the experimental (top) and
simulated spectra (bottom) are shown in the inset at an
expanded field scale.
constants was difficult. Therefore, quantum chemical cal-
culations were performed to predict the distribution of
the unpaired electron and the coupling constants of the
attached hydrogen atoms in the double-ring structure37,38
(Table 1). Although these calculations do not take into
account solvent effects, this starting approximation was
useful to fit the experimental ESR spectra iteratively
(Figure 1) and the quality of the simulations can be
assessed from the comparison of an expanded section of
the experimental spectra (insets of Figure 1, top traces)
with the simulated spectra (bottom traces). The resulting
coupling constants are also given in Table 1. The cou-
plings are assigned according to our calculations (Table
1); for 1 and 2 they are in agreement with the litera-
ture.28,39 The novel oxachromanol isomers exhibited coup-
ling constants for the methyl protons similar to those in
1 and 2. Interestingly, the coupling constants for the pro-
tons in the oxa ring (positions 2 and 4) were different for
the cis (4) and trans derivatives (5). This assignment was
confirmed by the simulation of ESR spectra from radi-
cals of 4 and 5 deuterated in positions 2 and 4 (CD
and CD3), giving coupling constants for the respective
PhOH + R• f PhO• + RH
(1)
To identify the phenoxyl radical intermediates, these
were generated either by UV irradiation or by oxida-
tion with the stable radical diphenyl picryl hydrazyl
(DPPH•).35,36 The ESR (electron spin resonance) spectra
of the radicals are shown in Figure 1, clearly demon-
strating the formation of paramagnetic one-electron oxi-
dation products. Due to the highly resolved complex
hyperfine structure, an intuitive guess of the coupling
(27) Mukai, K.; Okabe, K.; Hosose, H. J. Org. Chem. 1989, 54, 557-
560.
(28) Mukai, K.; Ikeuchi, T.; Morimoto, C.; Ishizu, K. Tetrahedron
Lett. 1984, 25, 1929-1932.
(37) Xie, C. P.; Lahti, P. M.; George, C. Org. Lett. 2000, 2, 3417-
3420.
(29) Imada, M.; Morimoto, H. Chem. Pharm. Bull. 1964, 12, 1047-
1051.
(38) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A.;
Stratman, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; 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.; Nananyakkara, A.; Gonzalez, C.;
Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.;
Andres, J. L.; Head-Gordon, M.; Replogle, E. S.; Pople, J. A. Gaussian
98 [revision A.6]; Gaussian Inc.: Pittsburgh, PA, 1998.
(39) Tsuchiya, J.; Niki, E.; Kamiya, Y. Bull. Chem. Soc. Jpn. 1983,
56, 229-232.
(30) Boers, R. B.; Gast, P.; Hoff, A. J.; de Groot, H. J. M.; Lugtenburg,
J. Eur. J. Org. Chem. 2002, 1, 189-202.
(31) Hyperchem [6.0]; HyperCube, Inc.: Gainesville, FL, 1994.
(32) Ghose, A. K.; Pritchett, A.; Crippen, G. M. J. Comput. Chem.
1988, 9, 80-90.
(33) Viswanadhan, V. N.; Ghose, A. K.; Revankar, G. R.; Robins, R.
K. J. Chem. Inf. Comput. Sci. 1989, 29, 163-172.
(34) Foti, M.; Ruberto, G. J. Agric. Food Chem. 2001, 49, 342-348.
(35) Boguth, W.; Niemann, H. Int. Z. Vitaminforsch. 1969, 39, 429-
437.
(36) Boguth, W.; Niemann, H. Biochim. Biophys. Acta 1971, 248,
121-130.
3474 J. Org. Chem., Vol. 70, No. 9, 2005

Antioxidant Properties of Chromanol Derivatives
TABLE 1. Characterization of Phenoxyl Radicals by
Quantum Chemical Calculations and by ESR and UV/Vis
Spectroscopy
ꢀb/L
a0a/G (ab initio
calculations)
a0a/G
(exptl)
λmax/nm mol-1 cm-1
PhO•
(ethanol)
(ethanol)
1
3H(5aCH3)
3H(7aCH3)
3H(8aCH3)
1H(4CH2)
1H(4CH2)
3H(5aCH3)
3H(7aCH3)
3H(8aCH3)
1H(4CH2)
1H(4CH2)
3H(5aCH3)
3H(7aCH3)
3H(8aCH3)
1H(4CH)
6.49c 3H 6.02
4.52c 3H 4.62
1.68c 3H 0.98
2.79c 3H 1.58
1.98c 1H 1.49
6.49 3H 6.01
4.52 3H 4.63
1.68 3H 0.98
2.79 1H 1.56
1.98 1H 1.55
5.99 3H 5.60
5.15 3H 5.45
1.87 3H 0.55
0.71 1H 1.23
0.38 1H 1.18
0.22 1H 0.99
0.02 1H 0.98
6.13 3H 6.15
5.40 3H 4.63
1.89 3H 1.03
1.67 1H 2.38
1.44 1H 1.48
6.03 3H 6.32
5.34 3H 4.83
1.86 3H 0.92
1.84 1H 1.33
0.01 1H 1.17
4.75c 3H 5.72
426
3800 ( 800
2
3
426
415
4800 ( 700
2200 ( 200
1H(3CH2)
1H(2CH)
1H(3CH2)
3H(5aCH3)
3H(7aCH3)
3H(8aCH3)
1H(4CH)
4
5
417
421
4000 ( 800
2400 ( 100
FIGURE 2. UV/Vis spectra of phenoxyl radicals generated
by UV irradiation of 50 mM solutions of 1-5 or by reaction of
1 mM solutions of 6 and 7 with 20 µM DPPH•, in benzene (6)
or acetonitrile (others) under aerobic conditions.
1H(2CH)
3H(5aCH3)
3H(7aCH3)
3H(8aCH3)
1H(4CH)
of our method. These data were used for further kinetic
analysis.
1H(2CH)
Efficiency of the Phenolic Antioxidants. The re-
activity of the phenolic compounds as antioxidants was
tested by characterizing their reaction with the model
radical DPPH• according to eq 2. In contrast to highly
oxidizing radicals, fully substituted phenoxyl radicals
(PhO•) only poorly react with DPPH• and do not interfere
with the decay of this model radical.41 Although DPPH•
is a nitrogen-centered radical, it is known that its
reaction rates with antioxidants give good approxima-
tions for the relative radical scavenging activities with
lipid peroxyl radicals.42,43 The decay of the DPPH• radical
during the rapid reaction with the antioxidants was
followed photometrically by the stopped-flow technique,
as shown for 3 in Figure 3 (solid trace).
6
7
3H(5aCH3)
425
475
3400 ( 200
1000 ( 300
3H(7aOCH3) 2.04c 3H 0.45
1H(4CH2)
1H(4CH2)
3H(5aCH3)
1.01c 3H 1.00
0.69c 1H 1.00
3.89c 3H 4.62
3H(7aOCH3) 2.26c 3H 0.60
3H(8aOCH3) 0.06c 3H 0.17
1H(4CH)
0.34c 1H 0.60
a Isotropic hyperfine coupling constant. b Molar extinction coef-
ficient; statistics are derived from sets of 3-10 measurements.c 2,
ubichromanol-1, and ubichromenol-1 were calculated instead of
1, 6, and 7, respectively.
atoms about 6 times smaller than for the H-substituted
compounds (not shown). Thus the coupling constants
scale with the gyromagnetic ratios (H/D ) 6.15) as
expected.
k1
9
DPPH• + PhOH
8 DPPH-H + PhO•
(2)
To use optical detection methods for kinetic measure-
ments, characterization of the radical absorption in the
visible range was required. The chromanoxyl radical of
1 (tocopheroxyl) absorbs in the visible range, peaking at
425 nm.35,40 The resulting spectra (Figure 2) of all radical
intermediates obtained by oxidation of the phenolic
compounds by UV irradiation, DPPH•, or PbO2 are very
similar, except that the ubichromenoxyl radical shows a
significant red shift and broadening of its peak due to
its extended conjugated π-system. For the other com-
pounds the shifts are small (see also Table 1). For each
compound the extinction coefficient in ethanol was
estimated (Table 1). In acetonitrile, similar values ((15%)
were obtained (not shown). The extinction coefficients
of the radicals of 1 and 2 are in reasonable agreement
with the values reported for similar chromanoxyl radi-
cals in ethanol by Mukai et al.40 (which are the only
published values to date, except ca. 7000-8000 L mol-1
cm-1 in an apolar solvent),35 demonstrating the reliability
The second-order rate constants (k1) obtained for this
reaction in three organic solvents are listed in Table 2.
The large excess (10-35-fold) of antioxidant over DPPH•
results in a stoichiometric factor of 1.0 since formally only
one electron of each antioxidant molecule is contributing
to the observed reaction. Control experiments under
anaerobic conditions gave the same results as experi-
ments with air-saturated solvents.
Fate of the Antioxidant-Derived Radical. The
phenoxyl radical still possesses one reducing equivalent.
However, as known for R-tocopherol, the chromanoxyl
radicals are less effective as radical scavengers than the
(40) Mukai, K.; Watanabe, Y.; Ishizu, K. Bull. Chem. Soc. Jpn. 1986,
59, 2899-2900.
(41) Gille, L.; Pro¨sch, U.; Stoesser, R. Radiat. Phys. Chem. 1992,
40, 461-468.
(42) Brand-Williams, W.; Cuvelier, M. E.; Berset, C. Lebensm.-Wiss.
Technol. 1995, 28, 25-30.
(43) Foti, M. C.; Johnson, E. R.; Vinqvist, M. R.; Wright, J. S.;
Barclay, L. R.; Ingold, K. U. J. Org. Chem. 2002, 67, 5190-5196.
J. Org. Chem, Vol. 70, No. 9, 2005 3475

Gregor et al.
TABLE 3. Second-Order Rate Constants Obtained from
Optical Stopped-Flow and ESR Measurements at 20 °C
for Disproportionation of Phenoxyl Radicals in
Homogeneous Solutions of Different Solvents, Generated
a
with DPPH• or PbO2
2k2/L mol-1 s-1
DPPH•,
optical,
DPPH•,
optical,
DPPH•,
optical,
DPPH•,
ESR,
PbO2,
optical,
PhO• acetonitrileb ethanolb
hexanec acetonitrilec acetonitrilec
1
2
3
4
5
6
7
1890 ( 50 1060 ( 80 3120 ( 540 1170 ( 70 1730 ( 140
1610 ( 30 1070 ( 80 3310 ( 230 1020 ( 60 2340 ( 80
243 ( 21 267 ( 11 275 ( 38 820 ( 40
ndd
3560 ( 80 5240 ( 220 2070 ( 100 2100 ( 120 3770 ( 100
452 ( 41 614 ( 6 207 ( 25 630 ( 85 430 ( 30
1830 ( 360 1570 ( 220 1500 ( 90 650 ( 90 1260 ( 40
16.4 ( 0.7 57 ( 6 31 ( 9 19 ( 3 67 ( 1
FIGURE 3. Kinetic traces obtained by stopped-flow photo-
metry for the reaction of 8 mM 3 with 200 µM DPPH• in aerobic
ethanol at 20 °C.
a Column headings list method of generation, method of detec-
tion, and solvent. b Statistics are derived from sets of 3-5 mea-
surements. c Statistics are derived from sets of 2-4 measurements.
d Phenoxyl peak not detected.
TABLE 2. Second-Order Rate Constants Obtained from
Stopped-Flow Measurements at 20 °C for the Reaction of
DPPH• with Phenolic Antioxidants in Homogenous
Solutions of Different Solvents
and is a prerequisite for their successful function as
biological antioxidants. Decay traces after generation of
all phenoxyl radicals with an independent method (PbO2
oxidation) revealed rate constants (Table 3) within the
same range compared to those measured by stopped-flow
photometry.
Recycling by Cellular Reductants. For antioxidant
recycling studies we chose a hydrophilic (ascorbate) and
a lipophilic reductant (ubiquinol), which have both been
shown to reduce the tocopheroxyl radical.15,17,20,21,47
1. Recycling by Ascorbate. The active ascorbate
species in biological systems has been reported to be its
deprotonated anionic form (Asc-),17 which predominates
at the pH (6.8) chosen in our experiments (pK1 ) 4.2,
pK2 ) 11.6).17 Since Asc- is water-soluble, we have chosen
a biphasic model system consisting of an aqueous buffer
containing ascorbate and a lipophilic phase formed by
TTAB micelles harboring the lipophilic antioxidants:
k1/L mol-1 s-1
PhOH
AcNa,b
ethanolc
hexanec
1
2
3
4
5
6
7
350 ( 7
325 ( 25
280 ( 19
52 ( 2
245 ( 15
253 ( 9
376 ( 7
107 ( 5
73 ( 13
122 ( 15
136 ( 53
6850 ( 150
6750 ( 250
50 100 ( 2100d
2250 ( 250
1770 ( 50
39 ( 2
92 ( 31
54 ( 12
450 ( 90
370 ( 40
a Acetonitrile. b Statistics are derived from sets of 3-7 measure-
ments. c Statistics are derived from sets of 2-4 measurements.
d Statistics are derived from a set of 20 measurements.
corresponding chromanol.44 Therefore, its reduction to the
antioxidatively more active form (the phenol) is of major
importance. Two principal pathways have been identi-
fied for this reduction: (1) the disproportionation (“self-
decay”) of two phenoxyl molecules to one molecule of
phenol and one molecule of a quinoid oxidation product
(tocopheryl quinone in the case of 1) in a multistep
reaction45 and (2) the reduction of the phenoxyl radical
by cellular reductants, such as ascorbate and ubiquinol
(antioxidant recycling). Therefore, model experiments
were performed to measure the rate constants for both
pathways.
Disproportionation. The formation of phenoxyl radi-
cals by oxidation of the phenols with DPPH• or PbO2 and
the decay of these radicals due to their disproportionation
(eq 3) was followed at their respective absorption maxima
(Figure 2 and Table 1), as shown for 3 in Figure 3 (dotted
trace).
k3
8
PhO•micelle + Asc-aqueous phase + H+ 9
PhOHmicelle + Asc•aqueous phase (4)
Preliminary experiments revealed that the reaction
with DPPH• was too slow to generate measurable amounts
of phenoxyl radicals in micelles. Since UV laser-induced
monophotonic H-abstraction to generate phenoxyl radi-
cals has been used by other groups,16 we employed a laser
emitting at 266 nm to generate the chromanoxyl radicals
by flash photolysis. Following the laser pulse, the radical
decay was detected photometrically. At this excitation
wavelength the extinction coefficients of the studied
chromanols are ca. 10-20% of their maximal absorptivity
at 290 nm. For each phenolic compound the absorption
maximum of the phenoxyl radical in micelles was deter-
mined in the absence of ascorbate by probing the visible
spectrum in 5 nm steps. The absorption maxima observed
(not shown) agree well ((5 nm) with the absorption
maxima detected with the diode-array photometer (vide
supra), except for 7, which produced a nonreducible
species with a considerably broader spectrum than
k2
2 PhO• 98 PhOH + PhQ
(3)
Rate constants (2k2) obtained from stopped-flow mea-
surements in three organic solvents are shown in Table
3. The decay of all radicals studied was independent of
the presence or absence of O2 in the solutions. This poor
reactivity with O2 is well-known for radicals derived from
highly substituted (“hindered”) phenolic antioxidants46
(46) Doba, T.; Burton, G. W.; Ingold, K. U.; Matsuo, M. J. Chem.
Soc., Chem. Commun. 1984, 461-462.
(47) Bisby, R. H.; Parker, A. W. Arch. Biochem. Biophys. 1995, 317,
170-178.
(44) Liebler, D. C.; Burr, J. A. Biochemistry 1992, 31, 8278-8284.
(45) Bowry, V. W.; Ingold, K. U. J. Org. Chem. 1995, 60, 5456-
5467.
3476 J. Org. Chem., Vol. 70, No. 9, 2005

Antioxidant Properties of Chromanol Derivatives
experiment in which a dye (DPPH•) was mixed with pure
solvent (acetonitrile) and monitored at 520 nm (not
shown).
2. Recycling by Ubiquinol. Using the same tech-
nique, we studied the reduction of photochemically
generated phenoxyl radicals by decylubiquinol, an ana-
logue of natural ubiquinol, in ethanolic solution (eq 8).
Due to the similar absorption spectra of chromanols and
ubiquinol around 266 nm and thus interference of the
cogenerated ubisemiquinone radical, a direct photolytic
generation of phenoxyl radicals turned out to be inap-
propriate. Therefore the phenoxyl radicals were gen-
erated indirectly by photolytic cleavage of the azo
initiator 2,2′-azobis(2,4-dimethylvaleronitrile) (AMVN,
R-NdN-R) at 355 nm according to eqs 5-7:
FIGURE 4. (A) Kinetic traces of the transient absorption of
the radicals generated by laser flash photolysis of 1 mM 3,
incorporated into 40 mM aqueous TTAB micelles, in the
absence and presence of ascorbate (Asc) at room temperature.
(B) Plot of the measured pseudo-first-order rate constant kpseudo
(from which k3 was derived) versus ascorbate concentration.
R-NdN-R
9
hν8 2R• + N2
(5)
(6)
(7)
TABLE 4. Second-Order Rate Constants Obtained
from Laser Flash Photolysis at 20 °C for the Reduction
of Phenoxyl Radicals by Ascorbate (k3) and
Decylubiquinol (k4)
R• + O2 f ROO•
PhOH + ROO• f PhO• + ROOH
k3a/107 L mol-1 s-1
k4a/104 L mol-1 s-1
PhO•
(micelles)
(ethanol)
Reaction of PhOH with rapidly formed ROO• is the
major source of PhO•. At this wavelength the extinction
coefficient of AMVN is ca. 90% of its peak absorption at
348 nm. A chromanol:ubiquinol concentration ratio of at
least 10:1 favored the reaction of the initiator radicals
with the chromanol (assuming a similar reactivity based
on literature data of R-tocopherol and ubiquinol).49
Control experiments with only azo initiator and ubiquinol
(concentrations as used for kinetic experiments) did not
yield significant absorptions in the 400-500 nm range.
Although neutral ubisemiquinones have also been re-
ported to show absorptions in this spectral range,50 their
disproportionation (eq 9) is so rapid (k ≈ 4 × 107 L mol-1
1
2
3
4
5
6
7
1.1 ( 0.2
2.6 ( 0.1
0.64 ( 0.14
1.96 ( 0.21
3.4 ( 2.0
12.6 ( 1.4
17.4 ( 1.5
<1500 ( 500b
>3 ( 1c
3.7 ( 0.1
4.2 ( 0.2
10.4 ( 0.4
0.90 ( 0.07
>0.003 ( 0.001c
a Statistics are derived from sets of 5-9 measurements. b The
second-order decay of the radical in the absence of quinol was too
fast to resolve the quinol reaction, but control experiments showed
the reducibility of the radical (see Results). c The expected radical
species could not be generated photochemically; lower limits of
the rate constants were estimated from experiments with PbO2
-generated radicals.
s-1 51
) that they do not interfere with the phenoxyl decay
kinetics. The rate constants (k4) obtained for the reduc-
tion of PhO• by decylubiquinol (eq 8) are shown in Table
4. For 6, the second-order decay of the radical was too
fast to resolve the quinol reaction; from this only an upper
limit can be given. However, with a diode-array photo-
meter the spectra from radicals of 6, generated with PbO2
in ethanol, disappeared within 30 s after admixing of 100
µM decylubiquinol (but not in the absence of reductant;
not shown). For 7, only a lower limit can be estimated
as for the ascorbate reaction (vide supra).
expected. Similar effects were observed upon continuous
UV irradiation of 7. We assume that the chromenol ring
structure is opened to a polyunsaturated isomer. Such a
photochromic effect has been described for chromenes.48
Due to the short pulse length of 10 ns, the phenoxyl
radicals were generated much faster than their consump-
tion by the reaction with ascorbate. Ascorbate concentra-
tions were limited to a range that does not significantly
influence the absorption of chromanols at 266 nm.
Control experiments with ascorbate solutions, omitting
the antioxidants, revealed no transient absorptions in the
range of 410-480 nm, which was used for the detection
of phenoxyl radicals. The resulting kinetic traces showed
increasing decay rates with increasing concentrations of
ascorbate, as shown for 3 in Figure 4A. The linear
relationship between the observed pseudo-first-order rate
constant (kpseudo) and the ascorbate concentration (Figure
4B) confirms the bimolecular reduction of the radicals
by ascorbate. The rate constants (k3) obtained for the
phenoxyl reduction are shown in Table 4. For 7, only a
lower limit could be estimated from conventional mixing
of PbO2-generated chromenoxyl with ascorbate (see Ex-
perimental Section). This limit corresponds to the kinetics
of the mixing process itself, as evident from a control
k4
9
PhO• + UQH2
8 PhOH + UQH•
(8)
(9)
UQH• + UQH• f UQ + UQH2
Number of Reducing Equivalents. Besides kinetic
aspects of antioxidant efficiency, the total amount of
reduction equivalents per molecule is of interest. We
approached this question experimentally by completely
oxidizing the phenolic compounds with an excess of
DPPH• in acetonitrile. This integrates the number of
(49) Shi, H.; Noguchi, N.; Niki, E. Free Radical Biol. Med. 1999,
27, 334-346.
(50) Land, E. J.; Swallow, A. J. J. Biol. Chem. 1970, 245, 1890-
1849.
(48) Becker, R. S.; Michl, J. J. Am. Chem. Soc. 1966, 88, 5931-
5933.
(51) Roginsky, V. A.; Pisarenko, L. M.; Bors, W.; Michel, C. J. Chem.
Soc., Perkin Trans. 2 1999, 4, 871-876.
J. Org. Chem, Vol. 70, No. 9, 2005 3477

Gregor et al.
reducing equivalents over the primary reaction and sub-
sequent reactions, such as delivery of additional re-
ducing equivalents via disproportionation of phenoxyl
radicals. The presence of multiple OH groups in polyphe-
nols such as flavonoids was often considered as evidence
for higher antioxidative activity, which is not necessarily
true.52
taking into account that the quantum chemical calcula-
tions performed do not consider solvent effects. Further-
more, ESR data and quantum chemical calculations
clearly demonstrate that the unpaired electron is not only
located in the conjugated π-ring system but also extends
to the oxa ring. In detail, the major coupling constants
of the radical of 3 were significantly different from those
of 1 and 2, showing a more uniform distribution of the
unpaired electron over positions 5 and 7 as well as more
contribution from coupling hydrogens of the heterocyclic
bridge, suggesting a higher degree of electron delocal-
ization and hence a more stable radical. The smaller
methyl coupling in 7 (4.6 G versus 5.7 G in 6) indicates
a higher degree of electron delocalization into the unsat-
urated heterocycle, again predicting a more stable radi-
cal. These findings suggest that chemical modifications
of the oxa ring should modulate both the antioxidant
activity and the radical stability, and possibly also the
reactivity toward cellular reductants.
The DPPH• decay we obtained (not shown), until
completion of the reaction, took about the same time as
the decay of the phenoxyl radicals (as measured in our
disproportionation experiment), as expected. Surprisingly
we observed that for 3 and for the oxachromanol isomers
the total number of reducing equivalents per molecule
was related to the water content in the solvent. In dry
solvent, 1, 2, 4, 5, 6, and 7 delivered 2.0 ( 0.1, 1.9 ( 0.1,
2.0 ( 0.1, 2.1 ( 0.1, 2.0 ( 0.1, and 2.0 ( 0.1 equiv,
respectively, as expected for a chromanol. In solvent
containing 10% water the number was raised to 2.4 (
0.1 for both oxachromanol isomers but not for 1, 2, 6,
and 7. This suggested that an oxidation product was
formed which can act as an antioxidant itself. Indeed
Rosenau et al.23 showed that oxachromanol undergoes
water-dependent oxidative heterocycle opening to yield
a hydroxyethylquinone, which isomerizes to a quinol
form. Compound 3 delivered 2.8 ( 0.1 equiv in dry
solvent and 4.1 ( 0.1 equiv in the presence of 10% water.
Above 10% water, no further increase was observed for
all compounds. The numbers obtained in aqueous solvent
are certainly more relevant to biological systems.
Primary Antioxidant Reaction. The use of DPPH•
as a model radical represents a convenient method to
evaluate the radical scavenging properties of antioxida-
tive molecules and antioxidant plant extracts.42 Since
DPPH• and peroxyl radicals have similar electronic
structures (the unpaired electron is delocalized over both
N atoms of hydrazyl and both O atoms of peroxyl),54
relative reactivities of antioxidants with DPPH• generally
show the same order as with the more relevant lipid
peroxyl radicals.42,43 In our study we considered the
variation of k1 in dependence on both the antioxidant
structure and the solvent in order to mimic different
polarities in biomembranes (Table 2).
Discussion
Distribution of Chromanol Antioxidants. All com-
pounds studied are highly lipophilic (log P ca. 3-6), and
therefore, they are expected to partition strongly into
biomembranes. Thus they should be able to protect
biological membranes from lipid peroxidation, although
short- (3, 4, and 5) and long-chain compounds (6 and 7)
may function in different regions of the membrane. Short-
and long-chain tocopherol derivatives have been shown
to exert similar antioxidative activities. However, their
faster diffusional exchange rates between micelles and
biomembranes could be a significant advantage of short-
chain antioxidants with respect to pharmacological tar-
geting.
Structure of Phenoxyl Radicals. To characterize
the reactivities of the different phenoxyl radicals, we first
identified them by ESR (Figure 1) and UV/Vis spectros-
copy (Figure 2). ESR revealed the proton hyperfine
coupling constants (Table 1), which reflect the electronic
structure of the radicals, that is, the degree of electron
delocalization, which influences radical stability and also
antioxidant reactivity, because radical stability supports
H-abstraction.53 The proton coupling constants obtained
for 1 and 2 are very similar to the ones reported.39 The
values obtained for the ubiquinone derivatives are similar
to those reported for short-chain analogues.28 Also, the
reasonable agreement between experimental and theo-
retical coupling constants confirms the classification of
all radicals as chromanoxyl or chromenoxyl radicals,
For the different structures we found a high-activity
group (1-3) and a low-activity group (4-7) in ethanol
and acetonitrile, while the two oxachromanol isomers (4
and 5) showed more intermediate activities in hexane.
The k1 values for 1 and 2 in hexane are almost identical
with the ones measured in hydrocarbons,55,56 whereas our
value for 1 in acetonitrile is somewhat smaller than the
one (490 L mol-1 s-1) reported for this solvent.55 Com-
pound 3 reacts with the model radical approximately as
fast as 1 and 2 in acetonitrile and ethanol or even 7 times
faster in hexane. This cannot be accounted for by a
statistical factor of 2, which would be expected for a
dimeric chromanol in which the two OH groups act
independently of each other as confirmed for a substi-
tuted bisphenol.57 Therefore the higher reactivity of 3
must originate from its special electronic properties, in
agreement with the larger stability of its radical (vide
infra). In contrast, 4 and 5 react ca. 3-16 times more
slowly than methyl-substituted compounds with only one
O atom in the attached ring (1-3). While Mukai et al.27
reported a 10 times lower reactivity of short-chain
ubichromanol and ubichromenol analogues with a stable
substituted phenoxyl radical in ethanol, in our experi-
ments the rates of 6 and 7 were as high as 50% that of
1 in the same solvent. Therefore a biological antioxidant
(54) Lucarini, M.; Pedulli, G. F.; Valgimigli, L. J. Org. Chem. 1998,
63, 4497-4499.
(55) Valgimigli, L.; Banks, J. T.; Ingold, K. U.; Lusztyk, J. J. Am.
Chem. Soc. 1995, 117, 9966-9971.
(52) Metodiewa, D.; Jaiswal, A. K.; Cenas, N.; Dickancaite, E.;
Segura-Aguilar, J. Free Radical Biol. Med. 1999, 26, 107-116.
(53) 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.
(56) Barclay, L. R. C.; Edwards, C. E.; Vinqvist, M. R. J. Am. Chem.
Soc. 1999, 121, 6226-6231.
(57) Amorati, R.; Lucarini, M.; Mugnaini, V.; Pedulli, G. F. J. Org.
Chem. 2003, 68, 5198-5204.
3478 J. Org. Chem., Vol. 70, No. 9, 2005

Antioxidant Properties of Chromanol Derivatives
function of the naturally occurring 726 seems possible.
The lower activity of the ubiquinone derivatives can be
explained by intramolecular H-bonding between the
phenolic OH group and the neighboring methoxy group,58
similar to the intermolecular solvent effects outlined
below. In contrast to the larger stability of the radical of
7 with respect to 6 (vide infra) which should enhance the
antioxidant activity of 7, we find comparable or even
slightly smaller rates for 7. Pentamethylchromenol also
showed only 66% activity toward peroxyl radicals in an
apolar solvent compared to its saturated analogue 2.53
While the antioxidant reaction between phenols (PhOH)
and radicals (R•) can proceed via H-atom transfer in
apolar media (eq 10), in polar solvents the electron-
transfer mechanism (eq 11) is more likely, and the rates
are modulated by interference of the solvent, by inhibit-
ing the formation of the intermediate complex or by
accelerating the electron transfer within the caged
complex between PhOH and R• (DPPH• in our experi-
ments).34
FIGURE 5. Correlation between the logarithm of the rate
constants (ln k1) of the antioxidant reaction with DPPH• in
acetonitrile (Table 2) and the energy levels of the highest
occupied molecular orbital (HOMO; panel A) or the O-H bond
dissociation enthalpies (BDE) of the antioxidant (panel B).
especially if they transfer an electron in the antioxidative
mechanism. HOMO energies as well as O-H bond
dissociation enthalpies (BDE) of the chromanols were
calculated by the semiempirical quantum chemical method
PM3. By plotting the logarithm of the rate constants k1
(in acetonitrile) versus the HOMO energies and BDE, the
correlations shown in Figure 5 were obtained. The figure
shows a fairly linear increase of reaction rates with
increasing HOMO energies (panel A) and decreasing
O-H bond dissociation enthalpies (panel B), confirming
the thermodynamic control of this reaction in polar
solvents. For the disproportionation (k2) and recycling
reactions (k3, k4) no correlation with the frontier orbital
energies or BDE was found, suggesting that steric factors
control these rates.
Fate of Antioxidant-Derived Radicals. Although
several studies in the past dealt with the radical scav-
enging properties of chromanols,27,53,55,56 none of these
studies considered primary radical reaction, dispropor-
tionation, and radical recycling together. This is a major
problem when different investigators use different sol-
vents, since rate constants may differ by several orders
of magnitude, as demonstrated by Table 2, thereby
preventing adequate comparison of antioxidants with
reference compounds, such as 1 and 2. We therefore
measured the rate constants of the disproportionation
and radical recycling reactions in addition to those of the
DPPH• reaction. For the disproportionation the same
solvents were used as for the DPPH• reaction, including
ethanol, which was also used for the ubiquinol recycling,
whereas a heterogeneous micellar system was chosen for
the ascorbate recycling.
R• + PhOH f (R•‚‚‚H‚‚‚OPh)cage f R-H + PhO•
(10)
R• + PhOH f (R-‚‚‚+•HOPh)cage f R-H + PhO•
(11)
Since the rate constants k1 (Table 2) for all compounds
are considerably smaller in polar solvents (acetonitrile
and ethanol) than in hexane, obviously hindrance of
complex formation by H-bonding between solvent and the
phenolic hydroxyl group seems to prevail, as has been
found for tocopherol.55,58 That the relative solvent effects
can be quite different for different antioxidants, as seen
in Table 2, has already been found for other compounds
and has been explained by differential H-donor/acceptor
interactions with the solvent.56 Interestingly, 6 and 7
show the smallest solvent effect, suggesting that the
intramolecular H-bond between the methoxy and OH
groups is either retained in the more polar solvents or
replaced by an H-bond of comparable strength. Several
studies placed the prenyl chain of 1 parallel to the
membrane lipid chains and the phenolic OH group
H-bonding to the lipid carbonyl or phosphate groups.59
The short-chain chromanols of the present study should
insert into membranes even closer to the membrane-
water interface. Therefore the kinetic properties obtained
in H-bond donor (ethanol) or acceptor solvents (acetoni-
trile) are more relevant to the antioxidant performance
in biomembranes than those obtained in hexane.
The primary reaction of the phenolic antioxidant with
lipid peroxyl radicals (LOO•) will determine in part its
biological efficiency, since it has to compete with the
H-abstraction reaction of these radicals with polyunsatu-
rated fatty acids, which was estimated to be rather slow
(k ≈ 10 L mol-1 s-1).62 In the case of successful detoxifi-
cation of such radicals, the phenolic antioxidant is
oxidized to a phenoxyl radical. Now essentially two
pathways determine its fate. To recover the first anti-
oxidant equivalent, the phenoxyl radical can undergo a
redox reaction either with another phenoxyl molecule
(disproportionation) or with a cellular reductant, such as
Studies of Hasford and Rizzo60 and Caffery et al.61 have
demonstrated that the reduction potential of flavin model
compounds is correlated with the energy of their lowest
unoccupied molecular orbitals (LUMOs). Vice versa, the
oxidation of chromanols should depend on the energy of
their highest occupied molecular orbitals (HOMOs),
(58) Barclay, L. R. C.; Vinqvist, M. R. In The chemistry of phenols;
Rappoport, Z., Ed.; John Wiley and Sons: Chichester, U.K., 2003; pp
839-908.
(59) Wang, X. Y.; Quinn, P. J. Mol. Membr. Biol. 2000, 17, 143-
156.
(60) Hasford, J. J.; Rizzo, C. J. J. Am. Chem. Soc. 1998, 120, 2251-
2255.
(61) Caffery, M. L.; Dobosh, P. A.; Richardson, D. M. Laboratory
Exercises using Hyperchem; Hypercube, Inc.: Gainesville, FL, 1998.
(62) Buettner, G. R. Arch. Biochem. Biophys. 1993, 300, 535-543.
J. Org. Chem, Vol. 70, No. 9, 2005 3479

Gregor et al.
ascorbate and ubiquinol (reduced coenzyme Q). Dispro-
portionation reactions have the disadvantage that one
of two antioxidant molecules is lost in the form of a
quinoid oxidation product. In contrast, the second path-
way can recover both reducing equivalents and both
antioxidant molecules. Therefore, the ratio of these two
reactions determines the long-term efficiency of an anti-
oxidant.58 Furthermore, the formation of quinoid me-
tabolites during disproportionation increases the risk of
toxic side effects.63 Although tocopheryl quinone (the
quinoid oxidation product of 1) in its reduced (quinol)
state was considered a cellular antioxidant itself,49 it is
also known that many quinones can cause harm to the
cell by ROS formation during redox cycling.63
The disproportionation rate constants (2k2, Table 3)
which reflect the radical stabilities, vary by more than 2
orders of magnitude in the group of compounds studied.
The small 2k2 value, that is, high stability of the radical
of 7, is in line with its ESR spectrum (vide supra) and
contrasts the relatively unstable radicals formed from 1,
2, 4, and 6, with 3 and 5 showing intermediate stability.
That 3 forms a more stable radical than 1 was also
evidenced from its ESR spectrum (vide supra). The rate
constants for 1 and 2 in hexane are virtually identical
with the values reported for the same compounds in
benzene,46 and the values in ethanol are similar to 1400
L mol-1 s-1 reported for 1 in the same solvent.64 The large
value for 4 may, at least in part, be attributed to the
larger distortion of the aromatic plane compared to the
other compounds, as seen from the calculated structure
(see Experimental Section). Nevertheless this compound
showed a slightly larger k1 value than the trans isomer
(5), suggesting that stereochemical factors play a role.
Contrary to the primary antioxidant reaction (k1), the
disproportionation rates showed only small solvent effects
and therefore should not depend on the exact location
within biomembranes. For most compounds the rate
constants are in good agreement with control values
measured by ESR.
To maintain the antioxidative capacity of lipophilic
phenolic antioxidants by avoiding their consumption as
quinoid oxidation products, antioxidant recycling is im-
portant. A major pathway is considered to be the reduc-
tion of the phenoxyl radicals by hydrophilic coantioxi-
dants such as ascorbate, which in turn can be recycled
by glutathione.58 This reaction seems to be essential for
membranes or other lipophilic compartments such as
LDL particles that have no enzymatic recycling systems
for ubiquinol (vide infra). The rate constants (k3, Table
4) determined for the phenoxyl reduction by ascorbate
are in the range of 107-108 L mol-1 s-1 for the various
antioxidants incorporated in cationic micelles (TTAB).
This correlates well with the reported value of 7 × 107 L
mol-1 s-1 for 1 incorporated in similar hexadecyltri-
methylammonium chloride micelles.16 This larger value
for cationic micelles, compared to 4 × 104 L mol-1 s-1 for
anionic SDS micelles16 and 2 × 105 L mol-1 s-1 for
zwitterionic phosphatidylcholine micelles65 was explained
FIGURE 6. Overall antioxidant efficiency coefficient (nk1k3k4/
2k2) of the novel antioxidants (3, 4, and 5) compared to 1 and
2, where n is the stoichiometric factor, k1 (Table 2) and 2k2
(Table 3) are expressed in liters per mole per second, k3 (Table
4) in 107 liters per mole per second, and k4 (Table 4) in 104
liters per mole per second. Rate constants obtained in ethanol
were chosen for the calculation, except k3, which was measured
in an aqueous micellar system.
on the basis of the ascorbate anion-micelle attraction.
Our results show that ascorbate can recycle the novel
chromanol-type antioxidants across the water-micelle
interface, which is relevant to a possible pharmacological
application.
Whereas ubiquinol, an essential constituent of the
mitochondrial inner membrane where it is continuously
recycled, reacts 10 times more slowly with lipid peroxyl
radicals compared to 1,58 it was suggested to be a major
reductant for tocopheroxyl radicals.20 The rate constants
(k4, Table 4) determined for the phenoxyl reduction by
ubiquinol in ethanol are in the range 104-105 L mol-1
s-1. The value for 1 is about 10 times smaller than the
one measured by stopped-flow spectrometry in the same
solvent.21 Still, the results showed that in the presence
of ascorbate or ubiquinol all antioxidants studied were
recycled much faster than their disproportionation to
ineffective (at least in terms of their antioxidant potency)
quinoid products. In the case of 1 this explains the limited
accumulation of R-tocopheryl quinone relative to R-toco-
pherol (1:20 to 1:100) observed in mitochondrial mem-
branes.66
Conclusion
To evaluate the potential benefit of the new chromanols
in the polar headgroup region of biomembranes, we
define a formal overall antioxidant efficiency coefficient
as (nk1k3k4/2k2) from results in ethanol/micellar systems,
where n is the stoichiometric factor (number of reducing
equivalents per molecule) and k1-k4 are the rate con-
stants discussed above. This coefficient takes into account
the various kinetic aspects associated with the antioxi-
dant reaction of such compounds. Figure 6 shows these
coefficients for those compounds for which all four rate
constants could be measured (3, 4, and 5, and as a
reference, 1 and 2). It was evident that 3 showed by far
the best overall antioxidant performance in these chemi-
cal model systems, even if the stoichiometric factor was
set to unity for all compounds. Preliminary results on
the subcellular level indicated that this may hold true
(63) Thornton, D. E.; Jones, K. H.; Jiang, Z.; Zhang, H.; Liu, G.;
Cornwell, D. G. Free Radical Biol. Med. 1995, 18, 963-976.
(64) Rousseau-Richard, C.; Richard, C.; Martin, R. FEBS Lett. 1988,
233, 307-310.
(65) Scarpa, M.; Rigo, A.; Maiorino, M.; Ursini, F. Biochim. Biophys.
Acta 1984, 801, 215-219.
(66) Gille, L.; Gregor, W.; Staniek, K.; Nohl, H. Biochem. Pharmacol.
2004, 68, 373-381.
3480 J. Org. Chem., Vol. 70, No. 9, 2005

Antioxidant Properties of Chromanol Derivatives
0.005%) were used. Decylubiquinol (dUQH2) was prepared by
reduction of dUQ with sodium dithionite in water/ethanol (1:
1) followed by hexane extraction.12
also for the protection of respiratory functions of iso-
lated mitochondria against oxidative damage (unpub-
lished results). Taking together, the new compound twin-
chromanol revealed the best set of kinetic para-
meters, its overall antioxidant efficiency exceeding that
of R-tocopherol. This makes it a promising test candidate
for further biological studies.
Electron Spin Resonance Spectra of Phenoxyl Radi-
cals. Compounds 1-5 (50 mM each) were dissolved in benzene
or benzene/acetonitrile (v/v ) 10:1). Oxygen was removed by
flushing with nitrogen. The samples were transferred to a 0.9
× 5.5 cm quartz flat cell and irradiated with a high-pressure
Hg UV lamp inside the TM cavity of a Bruker ESP 300 X-band
spectrometer. Instrumental settings: microwave frequency,
9.76 GHz; microwave power, 20 mW; receiver gain, 106;
modulation frequency, 100 kHz; modulation amplitude, 0.25
G (1, 4, 5) or 0.1 G (2, 3); scan rate, 17.9 G/min; time constant,
167 ms; temperature, 298 K; 6 scans per spectrum were
averaged.
ESR spectra of the radicals of 6 and 7 were measured with
a Bruker EMX X-band spectrometer equipped with a TE
cavity. The radicals were generated by briefly stirring a 3 mM
solution of the antioxidant in argon-deaerated benzene in a 3
mm (i.d.) ESR quartz tube containing a few grains of PbO2.
Instrumental settings: microwave frequency, 9.77 GHz; mi-
crowave power, 20 mW; receiver gain, 106; modulation fre-
quency, 100 kHz; modulation amplitude, 0.05 G; scan rate, 22
G/min; time constant, 82 ms; temperature, 298 K; 20 scans
were averaged.
Using the results of the quantum chemical calculation (vide
infra) as starting parameters, analysis of the hyperfine split-
ting structure was done by two different iterative ESR simula-
tion programs (ESRSIMU of our own design and WINSIM69)
until the sum of error squares reached a minimum. Both
programs yielded essentially the same coupling constants.
Quantum Chemical Prediction of Hyperfine Coupling
Constants for Phenoxyl Radicals. The structures of the
phenoxyl radicals were assembled in the molecular modeling
program HyperChem31 and the geometries were preoptimized
on the semiempirical level (PM3/UHF). These structures were
imported into the program Gaussian9838 and the structures
were reoptimized on the DFT level [B3LYP/6-31G(d)].37 From
the output file, the Fermi contact coupling constants were
obtained. For freely rotating methyl and methoxy groups, the
three coupling constants of the H-atoms were averaged to give
one isotropic coupling constant. To keep computation times
reasonably short, the short-chain analogues 2, ubichromanol-
1, and ubichromenol-1 were calculated for the radicals derived
from 1, 6, and 7, respectively.
Optical Spectra of Phenoxyl Radicals. UV/Vis spectra
of the radicals were measured on a Spectronic 3000 diode-array
spectrophotometer (Milton Roy; slit width 1.75 nm). The
radicals were generated either photolytically by a 2-min (1, 2,
and 4) or 5-min (3 and 5) irradiation (generating saturating
steady-state concentrations) of 50 mM aerobic solutions of the
chromanols in acetonitrile, by use of a focused high-pressure
Hg UV lamp and a cuvette holder kept at 20 °C, or by reaction
of 1 mM chromanol with 20 µM (final concentration) of the
stable radical diphenyl picryl hydrazyl (DPPH•), both in
acetonitrile. Alternatively, the radicals were generated with
PbO2 as described below. Since the spectra of irradiated 6 and
7 were dominated by oxidation/decay products other than the
expected radicals, these were generated exclusively by the
DPPH• reaction. Baseline scans of the antioxidants prior to
radical formation were subtracted.
Experimental Section
Phenolic Antioxidants. all-rac-R-Tocopherol (1) and pen-
tamethylchromanol (2,2,5,7,8-pentamethyl-1-benzopyran-6-ol;
2) were obtained from commercial suppliers. The twin-chro-
manol (2,10-dihydroxy-6,12-methano-1,3,4,8,9,11-hexamethyl-
12H-dibenzo[2,1-d:1′,2′-g]dioxocin; 3) and oxachromanols
(2,4,5,7,8-pentamethyl-4H-1,3-benzodioxin-6-ol; 4 and 5) were
synthesized according to Rosenau and co-workers.23,24 Deu-
terated analogues of 4 and 5 (2,4CD and 2a,4bCD3) were pre-
pared from perdeuterated acetaldehyde.
Ubichromenol-9 (7; the index indicates the number of
isoprene units) was synthesized by cyclization of 250 mg of
ubiquinone-10 in 1 mL of deoxygenated triethylamine for 2 h
at 95 °C according to Imada and Morimoto.29 After removal of
the solvent, the product was purified by column chromato-
graphy (Florisil), with a mixture of hexane/CHCl3 (v/v ) 8:2)
as eluent. The UV spectrum (peaks at 275, 283, and 332 nm)
of the slightly yellow fraction (ca. 20-45% yield) is identical
to the published one.25 1H NMR (300.13 MHz, CDCl3): δ 1.38
(s, 3H, 2aCH3), 1.60 (s, 9 × 3H, 4′a,8′a,12′a,16′a,20′a,24′a,28′a,32′a,37′CH3),
1.68 (s, 3H, cis-36′aCH3), 1.95-2.11 (m, 18 × prenyl CH2), 2.16
(s, 3H, 5aCH3), 3.88 (s, 3H, OCH3), 3.94 (s, 3H, OCH3), 5.06-
3
5.16 (m, 9H, prenyl CH), 5.42 (s, 1H, OH), 5.58 (d, 1H, J )
10.1 Hz, 3CH), 6.50 (d, 1H, 3J ) 10.1 Hz, 4CH). 13C NMR (75.47
MHz, CDCl3): δ 10.3 (5aC), 16.0 (4′a,8′a,12′a,16′a,20′a,24′a,28′a,32′aCH3),
17.7 (cis-36′aCH3), 22.8 (2′CH2), 25.5 (2aCH3), 25.7 (trans-37′CH3),
26.6-26.8 (6′,10′,14′,18′,22′,26′,30′,34′CH2), 39.7 (5′,9′,13′,17′,21′,25′,29′,33′CH2),
40.8 (1′CH2), 61.0 (OCH3), 61.4 (OCH3), 77.3 (2C), 113.9 (5C),
116.4 (4aC), 119.8 (4C), 123.9-124.7 (3′,7′,11′,15′,19′,23′,27′,31′,35′CH),
129.0 (3C), 131.2 (36′C), 134.9-135.0 (8′,12′,16′,20′,24′,28′,32′C), 135.4
(4′C), 138.9 (6C), 139.5 (7, 8C), 140.4 (8aC). The assignments were
supported by H-H correlation spectroscopy (COSY), hetero-
nuclear multiple quantum correlation (HMQC), and hetero-
nuclear multiple bond correlation (HMBC) spectra. The ab-
sence of a carbonyl 13C resonance (δ 185)67 is indicative of the
absence of UQ-10. The purity of 7 was determined by HPLC
analysis to 95%.
Ubichromanol-9 (6), a cyclization product of ubiquinol, was
obtained by selective reduction of 7 with sodium, analogous
to a published procedure,68 followed by column chromatogra-
phy as for 7 (ca. 45-65% yield). Residual 7 was estimated by
1H NMR to 3%. 1H NMR (300.13 MHz, CDCl3): δ 1.29 (s, 3H,
2aCH3), 1.60 (s, 9 × 3H, 4′a,8′a,12′a,16′a,20′a,24′a,28′a,32′a,37′CH3), 1.68 (s,
3H, cis-36′aCH3), 1.79 (t, 2H, 3J ) 6.8 Hz, 3CH2), 1.95-2.11 (m,
18 × prenyl CH2), 2.09 (s, 3H, 5aCH3), 2.57 (t, 2H, 3J ) 6.8 Hz,
4CH2), 3.85 (s, 3H, OCH3), 3.93 (s, 3H, OCH3), 5.06-5.16 (m,
9H, prenyl CH), 5.39 (s, 1H, OH). 13C NMR (75.47 MHz,
CDCl3): δ 10.8 (5aC), 16.0 (4′a,8′a,12′a,16′a,20′a,24′a,28′a,32′aCH3), 17.7
(cis-36′aCH3), 20.2 (4C), 22.3 (1′CH2), 23.4 (2aCH3), 25.7 (trans-
37′CH3), 26.6-26.8 (2′,6′,10′,14′,18′,22′,26′,30′,34′CH2), 31.1 (3C), 39.7
(
5′,9′,13′,17′,21′,25′,29′,33′CH2), 60.8 (OCH3), 61.3 (OCH3), 75.1 (2C),
116.0, 116.1 (4a,5C), 124.1-124.4 (3′,7′,11′,15′,19′,23′,27′,31′,35′CH), 131.2
Stopped-Flow Photometry of the DPPH• Reaction. The
kinetics of the antioxidant reaction with DPPH• as well as the
decay of the resulting phenoxyl radical were measured on a
DW2000 dual-wavelength spectrophotometer (SLM Aminco)
equipped with a dual-channel MilliFlow stopped-flow reactor
(kept at 20 °C) by the same manufacturer. Acetonitrile,
ethanol, and n-hexane were used as solvents. For comparison,
argon-deaerated or freshly distilled solvents were used. De-
fined concentrations of 6 and 7 (2-7 mM, determined from
(
36′C), 134.9-135.0 (8′,12′,16′,20′,24′,28′,32′C), 135.3 (4′C), 138.2 (6C),
139.1, 139.6, 140.5 (7, 8,8aC). The assignments were supported
by H-H COSY, HMQC, and HMBC spectra.
Other Chemicals. Chromatography-grade acetonitrile
(LiChrosolv; H2O e 0.05%) and ethanol (LiChrosolv; H2O e
0.1%) as well as spectroscopy-grade n-hexane (Uvasol; H2O e
(67) Boullais, C.; Rannou, C.; Reveillere, E.; Mioskowski, C. Eur. J.
Org. Chem. 2000, 5, 723-727.
(68) Schudel, P.; Mayer, H.; Metzger, J.; Ru¨egg, R.; Isler, O. Helv.
Chim. Acta 1963, 46, 2517-2526.
(69) Duling, D. R. J. Magn. Reson. B 1994, 104, 105-110.
J. Org. Chem, Vol. 70, No. 9, 2005 3481

Gregor et al.
OD293 and OD332 by use of ꢀ ) 3480 and 3230 L mol-1 cm-1 in
ethanol, respectively25) or 8 mM for all other chromanols were
mixed with 200 µM DPPH• except for 3 in hexane: because of
its low solubility, 120 µM (determined from OD290 by use of ꢀ
) 5900 L mol-1 cm-1 in ethanol) was mixed with 6 µM DPPH•.
The decay of DPPH• (eq 2) was followed at 520 nm (acetoni-
trile), 517 nm (ethanol), or 511 nm (hexane). The transient
formation and subsequent decay of the phenoxyl radicals were
followed at the peak of their absorption around 420 nm (except
475 nm for 7; see Table 1). A reference wavelength of 850 nm
was chosen at which the absorptions of both DPPH• and its
reduced product were negligible.
Laser Flash Photolysis and Antioxidant Recycling.
The rate constants (k3) of the chromanoxyl reduction by
ascorbate (eq 4) were measured by laser photolysis of the
antioxidants incorporated in aqueous cationic micelles. Com-
pound 6 or 7 (100 µL) in tert-butyl alcohol was evaporated to
dryness in a glass vial and resuspended in a 40 °C warm
solution of 200 mM TTAB and 25 mM potassium phosphate
(pH 6.8) by 5 min of vortexing, to give a final antioxidant
concentration of 1 mM. Similarly, 200 mM solutions of all other
antioxidants (except a 50 mM suspension of 3) in methanol
were added to 40 °C warm 40 mM TTAB micelles in the same
buffer under vortexing, and vortexing was continued for 5 min.
Radical concentrations of ca. 1-10 µM were generated by
flashing these micellar solutions, preloaded with ascorbate,
with a series (10-30) of 10 ns, 266 nm pulses from a Nd:YAG
laser (Quanta-Ray DCR-1) at a repetition rate of ca. 1 Hz at
room temperature. Details of the experimental setup have been
published previously.70 For each compound, three different
ascorbate concentrations (0.2, 0.4, and 0.6 mM) and three
different pulse energies (ca. 0.4-1 mJ) were applied. The
radical decay was monitored photometrically at the peak of
the respective radical spectrum around 420 nm.
The rate constants (k4) of the chromanoxyl reduction by
ubiquinol (eq 8) were measured similarly in ethanolic solution.
Compounds 1, 2, 4, and 5 (100 mM), 3 (10 mM; saturated
solution), and 6 and 7 (4 mM) were used. The laser was tuned
to 355 nm, and 10 mM azo initiator AMVN was added to the
reaction mixture. For each compound, three different decyl-
ubiquinol concentrations (1, 2, and 4 mM except 4, 8, and 12
mM for 1) were chosen; the pulse energy was in the 0.5-2 mJ
range.
The traces of the DPPH• decay were set to zero at infinite
time. Since antioxidants were in large excess over DPPH•,
these plots were fitted (by the nonlinear least-squares method
of Levenberg-Marquardt) to pseudo-first-order decay:
A ) A0 exp(-ak1t)
(12)
where A0 is the DPPH• absorption at time zero, k1 is the
second-order rate constant for the reaction of the antioxidant
with DPPH•, and a is the excess antioxidant concentration.
A0 and k1 were kept variable. Since a similar zeroing of the
slowly decaying tail of the phenoxyl decay (eq 3) would be
error-prone, the absorption plots were fitted to
A0 - Aoffset
1 + (A0 - Aoffset) 2k2ꢀ-1t
A ) Aoffset
+
(13)
where A0 is the absorption at time zero, Aoffset is the absorption
offset, 2k2 is the second-order rate constant for the dispropor-
tionation, and ꢀ is the molar extinction coefficient of the
phenoxyl radical as determined (vide infra). Only part of each
plot representing the pure decay, starting at the time of
complete DPPH• consumption (as determined from the respec-
tive DPPH• plots), was fitted. Good fits were obtained by
choosing A0, Aoffset, and k2 as variables.
ESR Measurements of the Phenoxyl Radical Decay.
Antioxidant (50 mM, except 4 mM 6 and 1.5 mM 7) was mixed
(within ca. 5 s) with 20 mM (except 2 mM for 6 and 0.6 mM
for 7) DPPH•, both in acetonitrile, with reduced pressure into
a flat cell equipped with a rubber plunger carrying two needles.
At these concentrations the DPPH• decay was complete after
5 s (20 s for 6, 80 s for 7), in agreement with the rate con-
stants determined from stopped-flow analysis (vide supra). At
15-20 s (30 s for 6, 90 s for 7) after mixing, repetitive ESR
spectra were measured. Instrumental settings: microwave
frequency, 9.67 GHz; modulation amplitude, 0.5 G; scan rate,
21 G/min; repetition time, 25 s; others as above. The peak-to-
peak intensities of the largest peak (the sum of the three
largest peaks for the noisier spectra from 4) were converted
to radical concentrations by calibration with 200 µM DPPH•
in acetonitrile, which was measured after each sample, by use
of the same instrumental settings without retuning. The decay
plots were fitted to
The traces were analyzed by a pseudo-first-order fit, taking
into account a simultaneous second-order decay of the radical
with a rate constant (2k2) measured in the absence of reduc-
tant, by use of eq 17. It should be noted that this second-order
reaction was faster than the disproportionation reaction as
measured by stopped-flow, presumably because of the presence
of additional photogenerated radicals. However, under our
conditions the half-life of this reaction was considerably longer
than that of the reaction with the reductant.
kred
PhO• + reductant + H+
8
PhOH + oxidized reductant (15)
k2
2 PhO• 98 products
(16)
k
pseudo A0
A ) Aoffset
+
(kpseudo + A02k2)exp(kpseudo t) - A02k2
(17)
where Aoffset is the absorption offset (accounting for possible
minor side reactions), A0 is the phenoxyl absorption at time
zero, 2k2 is the rate constant of the second-order decay, and
kpseudo ) credkred with cred the excess reductant concentration
and kred the second-order rate constant for the reaction of
phenoxyl with ascorbate (k3) or ubiquinol (k4). A0, Aoffset, and
kpseudo were kept variable. Comparable results were obtained
with the numerical simulation routine used for the stopped-
flow experiments (vide supra).
In the case of 7 the expected chromenoxyl radical could not
be detected. Therefore a lower limit of the rate constants was
estimated by oxidizing 1 mM 7 in micellar solution or 4 mM 7
in ethanol with ca. 1 mg of PbO2 (by vortexing for 1 min or a
few seconds, respectively), transferring the supernatant to a
cuvette after a brief centrifugation, admixing ascorbate or
decylubiquinol, respectively, from a premounted syringe under
magnetic stirring, and monitoring the radical decay at 475
c0
c )
(14)
1 + (c02k2t)
where c0 is the phenoxyl concentration at time zero and 2k2 is
the second-order rate constant for the disproportionation
(eq 3).
Other Measurements of the Phenoxyl Radical Decay.
The phenoxyl radicals were also generated by briefly vortexing
a 1 mM antioxidant solution (except 10 mM 6 and 7) in
acetonitrile with ca. 1 mg of PbO2. After a brief centrifugation
at 15000g, the supernatant was quickly transferred to a
cuvette, and time scans were measured at the peak of their
absorption around 420 nm (see Table 1) and at 850 nm (as
reference) on a diode-array photometer (vide supra). The 420
minus 850 difference traces were analyzed by use of eq 13.
(70) Baranyai, P.; Gangl, S.; Grabner, G.; Knapp, M.; Ko¨hler, G.;
Vido´czy, T. Langmuir 1999, 15, 7577-7584.
3482 J. Org. Chem., Vol. 70, No. 9, 2005

Antioxidant Properties of Chromanol Derivatives
minus 850 nm with a DW2000 photometer (SLM Aminco) on
the fast filter setting.
Semiempirical Quantum Chemistry Methods. Struc-
tures of the phenolic antioxidants were assembled in the
program HyperChem31 and subjected to a geometry optimiza-
tion by the semiempirical method PM3/UHF, yielding HOMO
and LUMO energies. Again the short-chain analogues 2,
ubichromanol-1, and ubichromenol-1 were calculated for 1, 6,
and 7. The O-H bond dissociation enthalpies (BDE) were
obtained from the calculated enthalpies of formation of PhOH,
PhO•, and H• by use of PM3/UHF optimized structures.
Stoichiometry of Antioxidant Reaction. Antioxidant (10
µM) in freshly distilled acetonitrile was oxidized with 60 µM
DPPH• (final concentration) after addition of 0-50% H2O, and
the reaction was monitored in a DW2000 photometer (SLM
Aminco) at 520 nm minus 850 nm until the DPPH• concentra-
tion was constant. From the extent of DPPH• reduction, the
number of reducing equivalents of the antioxidants was
calculated, by use of an extinction coefficient of 11 600 L mol-1
cm-1 (determined relative to 11 200 L mol-1 cm-1 in ethanol).71
Calculation and Measurement of Octanol/Water Dis-
tribution Coefficient. Experimental log P values were
calculated from the antioxidant absorptions (around 290 nm)
in the organic and water phases after equilibration of 1 mL of
500 mM 1 and 2, 100 mM 4 and 5, 30 mM 3, 50 mM 6, or 10
mM 7 in n-octanol with 1 mL water by vortexing under argon
for 25 min (1), 45 min (6 and 7), or 15 min (others) at 20°C.
Longer vortexing had no effect on log P. Prediction of log P
values was performed by an incremental method32,33 that was
used by the QSAR module of HyperChem.
Measurement of Extinction Coefficients of Phenoxyl
Radicals. Antioxidant (100 mM, except 80 mM 3 and 5 mM
6 and 7) in acetonitrile or ethanol was loaded into a cuvette
kept at 5 °C (to minimize evaporation loss, except 6, which
was kept at 20 °C) and a baseline scan was recorded on a diode-
array photometer (vide supra). After admixing of 10 µM DPPH•
(except 20 µM for 3; final concentrations) in the same solvent
within 5 s, 4-5 spectra of the resulting radicals were measured
with a repetition time of 5 s (3 nm slit width). After a linear
baseline was subtracted from each scan (to account for baseline
drifts), peak intensities were plotted against time and extrapo-
lated to time zero via a nonlinear fit for the bimolecular radical
decay (eq 14). At the chosen concentrations, the radicals
reached their maximal concentration after ca. 1-5 s and
subsequently decayed almost linearly, as predicted from the
measured rate constants k1 and k2 (vide supra), minimizing
extrapolation errors. The molar extinction coefficients (ꢀ) were
calculated from the extrapolated absorptions.
Abbreviations
AMVN, 2,2′-azobis(2,4-dimethylvaleronitrile); asc, ascor-
bate; BDE, bond dissociation enthalpy; DPPH•, diphenyl
picryl hydrazyl; dUQH2, decylubiquinol; ESR, electron
spin resonance; HOMO, highest occupied molecular orbi-
tal; LUMO, lowest unoccupied molecular orbital; PhOH,
phenolic antioxidant; PhO•, phenoxyl (chromanoxyl) radi-
cal; PhQ, quinoid oxidation product; ROS, reactive oxygen
species; TTAB, tetradecyl trimethylammoniumbromide;
UQ, ubiquinone.
Acknowledgment. This work was supported by the
Austrian Science Fund (FWF), Grant P16244-B08.
Technical assistance by Werner Stamberg and helpful
advice for the synthesis of 6 by Dr. Thomas Netscher
(DSM Nutritional Products, Basel, Switzerland) is
gratefully acknowledged.
Supporting Information Available: Cartesian coordi-
nates and total energies of the calculated molecular structures
of chromanoxyl radicals (PDF). This material is available free
of charge via the Internet at http://pubs.acs.org.
JO047927S
(71) Tampo, Y.; Yonaha, M. Arch. Biochem. Biophys 1996, 334, 163-
174.
J. Org. Chem, Vol. 70, No. 9, 2005 3483