Kinetic study on the free radical-scavenging and vitamin E-regenerating actions of caffeic acid and its related compounds

Article abstract of DOI:10.1246/bcsj.79.1501

A kinetic study involving 4-hydroxycinnamic acid derivatives (HCAs) was performed in order to clarify the mechanism for free radical-scavenging and vitamin E-regenerating. The second-order rate constants of the radical-scavenging reaction (k_s) observed for caffeic acid derivatives are larger than those for ferulic acid derivatives. The result may be explained by i) the lower oxidation potential coming from the existence of the more electron-donating hydroxy group comparing with methoxy group, and ii) the stabilization effect of the corresponding radicals produced in the radical-scavenging reaction through intramolecular hydrogen bonding. The radical-scavenging activity of HCAs in ethanol mainly occurs via the hydrogen atom transfer from the phenolic OH-protons. The pH dependence of k_s for ferulic acid in the aqueous Triton X-100 micelle solutions suggests the importance of the phenolic OH proton on the radical-scavenging reaction. On the other hand, the pH dependence of k_s for caffeic acid and chlorogenic acid suggests the occurence of intramolecular hydrogen bonding and an electron-transfer process in the radical-scavenging reaction.

Full text of DOI:10.1246/bcsj.79.1501

Ó 2006 The Chemical Society of Japan
Bull. Chem. Soc. Jpn. Vol. 79, No. 10, 1501–1508 (2006) 1501
Kinetic Study on the Free Radical-Scavenging and Vitamin
E-Regenerating Actions of Caffeic Acid and Its Related Compounds
ꢀ
Shin-ichi Nagaoka, and Kazuo Mukai
Keishi Ohara, Yoko Ichimura, Kumi Tsukamoto, Mayumi Ogata,
Department of Chemistry, Faculty of Science, Ehime University, Matsuyama 790-8577
Received March 14, 2006; E-mail: ohara@chem.sci.ehime-u.ac.jp
A kinetic study involving 4-hydroxycinnamic acid derivatives (HCAs) was performed in order to clarify the mech-
anism for free radical-scavenging and vitamin E-regenerating. The second-order rate constants of the radical-scavenging
reaction (ks) observed for caffeic acid derivatives are larger than those for ferulic acid derivatives. The result may be
explained by i) the lower oxidation potential coming from the existence of the more electron-donating hydroxy group
comparing with methoxy group, and ii) the stabilization effect of the corresponding radicals produced in the radical-
scavenging reaction through intramolecular hydrogen bonding. The radical-scavenging activity of HCAs in ethanol
mainly occurs via the hydrogen atom transfer from the phenolic OH-protons. The pH dependence of ks for ferulic acid
in the aqueous Triton X-100 micelle solutions suggests the importance of the phenolic OH proton on the radical-scav-
enging reaction. On the other hand, the pH dependence of ks for caffeic acid and chlorogenic acid suggests the occurence
of intramolecular hydrogen bonding and an electron-transfer process in the radical-scavenging reaction.
A group of 4-hydroxycinnamic acid derivatives (HCAs) in-
cluding caffeic acid (CA), ferulic acid (FA), chlorogenic acid
(CGA), and rosmarinic acid (RA) (Fig. 1) are a few of the
polyphenols found in a variety of plants, such as beans, fruits,
herbs, and vegetables.1–5 For example, CGA is found in coffee
beans, burdocks, apples, pears, berries, and tomatoes. These
HCAs are produced from phenylalanine or L-tyrosine via an
organic pathway in plants. The total content of CGA and CA
in roasted coffee beans is about 2%, and the amount of RA
gested HCAs are of great interest to many researchers. Recent
pharmacological reports suggested that plant-origin HCAs
have pharmacological activity, such as the inhibition of lipid-
peroxidation6–10 and DNA-damage,11 reduction of non-heme
iron in blood,12 antiallergic activities,13 and cancer-preventive
effects.14,15 Many of these effects are considered to result from
antioxidant activity toward active-oxygen species (AOS), such
as singlet oxygen and free radicals that are produced in the
body. The antioxidant activity of natural materials is generally
considered to be one of the most important factors in pharma-
cological activities, because it protects living bodies and tis-
sues from diseases and injuries caused by AOS.6–24 Kinetic
studies on the antioxidant activities of these HCAs versus sev-
eral free radicals, such as DPPH (2,2-diphenyl-1-picrylhydra-
in the herb rosemary is 10 mg gꢁ1 2,3 Human being and animals
.
usually take a certain amount of these HCAs from daily meals
and drinks.
Because a considerable amount of HCAs is always taken
and absorbed from a diet into human, the health effects of in-
O
O
OH
OH
OH
O
HO
HO
COOH
HO
HO
COOCH3
HO
HO
(a) Caffeic acid
(b) Methyl caffeate
(c) Rosmarinic acid
O
HO
HO
C O
COOH
OH
H3CO
HO
COOH
H3CO
OH
HO
OH
(d) Chlorogenic acid
(e) Ferulic acid
COOH H3CO
HO
(g) Half-curcumin (h) 4-Hydroxycinnamic acid (i) Guaiacol
(f) Isoeugenol
O
H3CO
HO
H3CO
HO
CH3
HO
(j) Creosol
Fig. 1. Molecular structures of HCAs and the related phenols.
Published on the web October 10, 2006; doi:10.1246/bcsj.79.1501

1502 Bull. Chem. Soc. Jpn. Vol. 79, No. 10 (2006)
Antioxidant Activity of Caffeic Acid
OH
OH
HO
H3CO
H3CO
H3CO
O•
C16H33
O
H
10
(a) Aryloxyl radical
(b)
α
-Tocopherol
(c) Ubiquinol-10
OH
OH
OH
OH
H
H
H
H
HO
O
HO
O
OH
OH
OH
OH
OH
OC
O
OH
OH
(d) Epicatechin
(e) Epigallocatechin gallate
Fig. 2. Molecular structures of aryloxyl radical (ArOꢂ) and some antioxidants.
ks
kr
ArO · + HCA
ArOH + HCA ·
α
-Toc · + HCA
α -TocH + HCA ·
Scheme 1.
Scheme 2.
zyl), hydroxyl (ꢂOH), superoxide (ꢂO2ꢁ), and azide radical
(ꢂN3), in solutions, have been reported.16–24 It has been sug-
gested that HCAs have a certain antioxidant activity as well
as other naturally existing polyphenols.
ArOꢂ scavenging reaction have been measured in aqueous
Triton X-100 (TX-100) micelle solutions at various pH values.
TX-100 micelle systems are frequently used as a model for
biomembranes in the kinetic studies for the reaction between
water-soluble antioxidants and lipophilic free radicals, such
as Tocꢂ and ArOꢂ. The mechanism and the structure–activity
relationship for HCAs are discussed on the basis of results
obtained from these rate constants.
The plant-origin HCAs are classified into caffeic acid type
which has an o-hydroxyphenol (catechol) moiety and ferulic
acid type which has an o-methoxyphenol (guaiacol) moiety.
The former contains CA, CGA, and RA, and the latter contains
FA and curcumin. The o-hydroxy- and o-methoxy-HCAs are
superior to 4-hydroxycinnamic acid (p-coumaric acid), and
CA derivatives usually have higher activity than FA deriva-
tives.7,19,23 Among the natural HCAs, the dietary contents,
the adsorption in living bodies, the distribution in tissues, the
solubility, and the antioxidant activities are different for each
HCA. Further systematic studies involving HCAs on the rela-
tionship between their chemical structure and activity are
needed for understanding their health and pharmacological
effects and their antioxidant mechanisms.
In the present work, a kinetic investigation on the free radi-
cal-scavenging and vitamin E (VE)-regenerating reactions in-
volving a group of HCAs and related phenols was carried
out in order to clarify their structure–activity relationship. To
investigate the difference in antioxidant activities between
o-methoxyphenol (guaiacol) and o-hydroxyphenol (catechol)
moieties, and the effect of the conjugated double bond in the
side chain of HCAs, we used CA, methyl caffeate, CGA, RA,
FA, isoeugenol, 4-hydroxycinnamic acid, guaiacol (2-meth-
oxyphenol), and creosol (2-methoxy-4-methylphenol) (Fig. 1).
The second-order rate constants of the HCAs and related phe-
nols for the reactions with the model free-radical, aryloxyl
(2,6-di-tert-butyl-4-(4-methoxyphenyl)phenoxyl, ArOꢂ, Fig. 2a)
(Scheme 1), and the regeneration reaction of VE (ꢀ-tocoph-
erol, ꢀ-TocH) from ꢀ-tocopheroxyl (ꢀ-Tocꢂ) (Scheme 2)
have been measured using a stopped-flow spectrophotometer
in ethanol. Furthermore, to investigate the contribution of the
acid–base dissociation equilibrium to the antioxidant activity,
the second-order rate constants of FA, CA, and CGA for the
Experimental
Caffeic acid, chlorogenic acid, rosmarinic acid, ferulic acid,
isoeugenol, 4-hydroxycinnamic acid, guaiacol, and creosol were
purchased from Nacalai Tesque and were used as received. Meth-
yl caffeate was synthesized from CA and methanol. ꢀ-Tocopherol
was obtained from Wako and was used as received. ArOꢂ was
prepared according to a published procedure.25–29 Ethanol (Wako)
and acetonitrile (Aldrich) were dried and purified by distillation.
Triton X-100 is an extra-pure grade reagent commercially availa-
ble from Nacalai Tesque and was used as received. All buffer
solutions were prepared using deionized water purified with a
Millipore-Q system. The pH of the solutions was adjusted using
the following 0.1 M buffer solutions; pH 4.0–5.0, CH3COONa–
CH3COOH; pH 6.0–8.0, Na2HPO4–KH2PO4; pH 9.0–11.0,
Na2CO3–NaHCO3.25,26 The concentration of TX-100 in the mi-
celle solutions was kept at 5.0 wt %. The sample solutions were
deoxygenated by bubbling nitrogen gas before the experiments.
The kinetic data were obtained using a Unisoku RS-450 or
RSP-1000 stopped-flow spectrophotometer by mixing equal vol-
umes of a HCA solution and ArOꢂ or ꢀ-Tocꢂ solution at 25 ꢃC.27
The ꢀ-Tocꢂ radical was prepared by mixing equal volumes of ꢀ-
tocopherol and ArOꢂ solutions 2 seconds prior to the mixing of
HCA and ꢀ-Tocꢂ solutions using the double-mixing unit of the
RSP-1000.27 The reactions were studied under pseudo-first-order
conditions for HCA, and the absorption decay of ArOꢂ (ꢀ-Tocꢂ)
was well-characterized as a single-exponential decay. Detailed
experimental procedures have been reported elsewhere.25–29
Pseudo-first-order rate constants (kobsd) for the scavenging reac-
tion of ArOꢂ or ꢀ-Tocꢂ by HCAs were estimated by using a non-

K. Ohara et al.
Bull. Chem. Soc. Jpn. Vol. 79, No. 10 (2006) 1503
The ks values were widely spread (0.1 to 5:21 ꢅ 103 Mꢁ1 sꢁ1).
The ks value of CA (5:54 ꢅ 10 Mꢁ1 sꢁ1) was much smaller
than those of ꢀ-tocopherol (5:21 ꢅ 103 Mꢁ1 sꢁ1) and EGCG
(ks ¼ 3:36 ꢅ 102 Mꢁ1 sꢁ1), and on the same order as that for
BHT (butylated hydroxytoluene, 2,6-di-t-butyl-4-methylphe-
nol, ks ¼ 3:50 ꢅ 10 Mꢁ1 sꢁ1).27 The ks values for methyl caf-
feate and CGA were on the same order as that of CA, and
RA (ks ¼ 9:92 ꢅ 10 Mꢁ1 sꢁ1) had the highest activity among
the HCAs. The ks values for FA (1:44 ꢅ 10 Mꢁ1 sꢁ1) was com-
parable to the value reported for half-curcumin (4-(4-hydroxy-
3-methoxyphenyl)-3-buten-2-one, ks ¼ 1:53 ꢅ 10 Mꢁ1 sꢁ1),26
and 1/4 of that for CA. The activity of isoeugenol which has
no carboxyl group in its side-chain was higher than FA and
half-curcumin. 4-Hydroxycinnamic acid which has no sub-
stituent at the position ortho to its phenol OH had almost no
activity (ks ¼ 1:0 ꢅ 10ꢁ1 Mꢁ1 sꢁ1). The ks value of creosol
(2-methoxy-4-methylphenol) was 9 times larger than that of
guaiacol (2-methoxyphenol). The results suggest that the free
radical-scavenging activities of these phenol compounds
strongly depend on both ortho- and para-substituents on the
phenol moiety. Existence of strong electron-donating substitu-
ents, such as hydroxy, methoxy, and methyl groups, causes an
increase in the ks values.
linear least-squares fitting method to a single-exponential curve
from the decrease in the absorbance at 385 or 580 nm of ArOꢂ
or the absorbance at 420 nm of ꢀ-Tocꢂ (Fig. 3). The rate constant
(kobsd) is given by25–27
kobsd ¼ k0 þ ks½HCAꢄ;
ð1Þ
where k0 is the first-order rate constant for the natural decay of
ArOꢂ or ꢀ-Tocꢂ in the medium. The second-order rate constant
(ks or kr) was obtained as a slope of plots of kobsd versus the con-
centration of HCA ([HCA]).
Cyclic voltammetry measurements were performed in acetoni-
trile containing 0.10 M tetrabutylammonium perchlorate under a
nitrogen atmosphere at 15 ꢃC using platinum electrodes and an
Ag/Agþ reference electrode with a BAS CV-50W cyclic voltam-
metric analyzer. Under these conditions, ferrocene, as a standard
sample, had a half-peak potential (E1=2) of þ40 mV vs Ag/Agþ.
Results and Discussion
Aryloxyl Radical-Scavenging Rate Constant (ks) in
Ethanol. The second-order rate constants (ks) for the ArOꢂ
scavenging reaction involving HCAs and related phenols
(Fig. 1) are listed in Table 1, together with those reported for
epicatechin and epigallocatechin gallate (EGCG) (Fig. 2).27
The free radical-scavenging activities of these phenolic
compounds should be related to their oxidation potentials.
Figure 4 shows the semi-logarithm plots for ks versus the an-
odic peak oxidation potential (Ep) for HCAs and related phe-
nols. A pseudo-linear relationship between log ks and Ep was
observed for the HCAs except for RA. This behavior is similar
to one of the features of hydrogen atom transfer (HAT) reac-
tions. Therefore, similar to half-curcumin,26 the free radical-
scavenging reactions involving HCAs and related phenols in
ethanol progress as a typical HAT. As described above, higher
radical-scavenging activity was observed for CA derivatives
comparing with FA derivatives. It can be explained by the low-
er oxidation potential of CA derivatives due to the existence of
the o-hydroxy group which is more electron-donating than the
o-methoxy group. The large activity of isoeugenol, which is
one of the o-methoxyphenols, also comes from its lower oxi-
dation potential compared to those of FA and half-curcumin.
Isoeugenol has an electron-donating methyl group in place
0.5
376 nm
Interval : 120 ms
0.4
0.3
0.2
580 nm
427nm
0.1
0
350 400 450 500 550 600 650
Wavelength / nm
Fig. 3. Time evolution of the absorption spectrum of ArOꢂ
in the reaction with ꢀ-tocopherol in ethanol at 25 ꢃC.
Table 1. Second-Order Rate Constants for the Scavenging Reactions of ArOꢂ (ks) and ꢀ-Tocopher-
oxyl (kr) in Ethanol and Peak Oxidation Potentials (Ep)
ks/Mꢁ1 sꢁ1
kr/Mꢁ1 sꢁ1
in Ethanol
Ep/mV vs Ag/Agþ
in Acetonitrile
Caffeic acid (CA)
Methyl caffeate
Chlorogenic acid (CGA)
Rosmarinic acid (RA)
Isoeugenol
Ferulic acid (FA)
4-Hydroxycinnamic acid
Half-curcumin
Guaiacol
Creosol
5:54 ꢅ 10
4:56 ꢅ 10
6:83 ꢅ 10
9:92 ꢅ 10
9:76 ꢅ 10
1:44 ꢅ 10
1:0 ꢅ 10ꢁ1
1:53 ꢅ 10
4.8
2:31 ꢅ 102
900
937
881
939
844
966
1200
953
1050
986
865
798
—
2:63 ꢅ 102
9:05 ꢅ 102
9:28 ꢅ 102
3.68
—
—
ꢆ10ꢁ1
4:23 ꢅ 10
1:32 ꢅ 102
3:36 ꢅ 102
3:5 ꢅ 10ꢁ1
7:58 ꢅ 102 aÞ
2:39 ꢅ 104 aÞ
Epicatechin (EC)
Epigallocatechin gallate (EGCG)
a) The values in 5:1 (v/v) ethanol–water mixture.27

1504 Bull. Chem. Soc. Jpn. Vol. 79, No. 10 (2006)
Antioxidant Activity of Caffeic Acid
of the electron-withdrawing carboxyl group in the side chain.
Another factor for larger activity of CA derivatives is the sta-
bilization effect of the corresponding radicals produced in the
radical-scavenging reaction through intramolecular hydrogen
bonding.31,32 As shown in Fig. 5, intramolecular hydrogen
bonding between two hydroxy groups should exist in the cat-
echol moiety of CA derivatives. Through the intramolecular
hydrogen migration, the corresponding phenoxyl radical gen-
erated by the radical-scavenging reaction is stabilized by the
delocalization of the unpaired electron. This was also suggest-
ed in the theoretical studies for CA derivatives.31,32 This type
stabilization of the radical is not expected in the case of FA
derivatives. This might be the reason why the activities of
CA derivatives are larger than those of FA derivatives.
The ks value of RA was about twice of those of CA and
methyl caffeate, while their Ep values were similar. This might
suggest that RA almost acts as a dimer of CA molecules in the
free radical-scavenging reaction. In the literature, the high ac-
tivity of EGCG in the VE-regeneration reaction was explained
by intramolecular ꢁ–ꢁ stacking interaction between epigallo-
catechin and gallate moieties in EGCG,25 and the condensed
structures of polyphenols sometimes contribute to the free rad-
ical-scavenging.26
second-order rate constants (kr) for HCAs on the natural VE
(ꢀ-tocopherol)-regeneration reaction are listed in Table 1.
The kr value for CA (2:31 ꢅ 102 Mꢁ1 sꢁ1) in ethanol is 3–4 or-
ders smaller than those of ubiquinol-10 (2:15 ꢅ 105 Mꢁ1 sꢁ1
)
and sodium ascorbate (2:68 ꢅ 106 Mꢁ1 sꢁ1).33 RA has the larg-
est kr (9:05 ꢅ 102 Mꢁ1 sꢁ1) in the HCAs, while the kr value of
FA is very small (3.68 Mꢁ1 sꢁ1). o-Methoxyphenols (guaiacol
and creosol) have almost no activity for the VE-regeneration
reaction (kr / 10ꢁ1 Mꢁ1 sꢁ1). This behavior for the VE-regen-
erating activity of HCAs was slightly different from that for
the previous free radical scavenging. For CA derivatives, the
kr values were 5–10 times larger than the ks values. On the
other hand, a notable decrease in kr was observed for FA
and o-methoxyphenols. In other words, for efficient VE-regen-
eration, the antioxidants are required to have lower oxidation
potentials than the threshold value. From the kr value, the ox-
idation potential of FA is nearly equal to the threshold.
The free radical-scavenging and VE-regenerating activities
of these HCAs are not high compared to those of other natural-
ly existing antioxidants. However, taking into account the
large daily intakes by animals and human, HCAs may also
act as good antioxidants for both the free radical and the VE
radical which is produced by the antioxidant reaction involv-
ing VE towards AOS, such as lipid-peroxyl radicals in living
bodies.
VE-Regeneration Rate Constant (kr) in Ethanol. The
pH Dependence of the Aryloxyl Radical-Scavenging
Rate (ks) in TX-100 Micelle Solutions. The second-order
rate constants (ks) of the ArOꢂ scavenging reaction by FA,
CA, and CGA in the TX-100 micelle solutions were measured
at various pH values. Previous kinetic studies showed that the
second-order rate constants for the radical-scavenging reac-
tion of vitamin C, flavonoids, and curcumin have notable pH
dependence according to the variation of the mole fractions
of some species produced by the acid–base dissociation equi-
librium.25–30 Figure 6a shows the plots of ks observed for FA
versus pH in the TX-100 micelle solutions. The ks values in
the micelle solution at pH 6–8 are one order of magnitude
larger than that obtained in ethanol (see Table 2). With an
increase in pH, the ks values in the micelle solution decrease
gradually between pH 4–6, are constant at pH 6–8, and
decreased rapidly at pH 8–10. The measurements at pH < 4
were unsuccessful, because ArOꢂ is unstable in these condi-
tions. This pH-dependent behavior of ks is similar to that
observed for half-curcumin reported previously.26
103
EGCG
EC
102
101
100
RA
Methyl caffeate
Creosol
Isoeugenol
CGA
CA
Half-curcumin
Guaiacol
FA
800
900
1000
1100
Ep / mV vs Ag/Ag+
Fig. 4. Semi-logarithm plots of ks versus Ep for CA deriv-
atives ( ), FA derivatives ( ), and the related phenols
(
).
3
4
O
O
R
H
H
HO
O
R
HO
O
R
R
HO
O
R
HO
O
R
R
HO
O
O
O
R
O
R
O
R
HO
HO
HO
HO
Fig. 5. Possible resonance structure of the radical produced from caffeic acid.

K. Ohara et al.
Bull. Chem. Soc. Jpn. Vol. 79, No. 10 (2006) 1505
Table 2. pH Dependence of the Second-Order Rate Con-
stants (ks) for the ArOꢂ Scavenging Reactions of Ferulic
Acid, Caffeic Acid, and Chlorogenic Acid in TX-100
Micelle Solution (5.0 wt %) at 25.0 ꢃC
(a)
FerH2
FerH-
Fer2-
1
0.8
0.6
0.4
0.2
0
4
3
2
1
0
ks/Mꢁ1 sꢁ1
pH
Ferulic acid
Caffeic acid
6:15 ꢅ 102
Chlorogenic acid
1:40 ꢅ 102
4.0
4.5
5.0
5.5
6.0
6.5
7.0
7.5
8.0
8.5
9.0
9.5
10.0
10.5
11.0
3:13 ꢅ 102
3:13 ꢅ 102
2:84 ꢅ 102
2:65 ꢅ 102
2:62 ꢅ 102
2:64 ꢅ 102
2:65 ꢅ 102
2:64 ꢅ 102
2:63 ꢅ 102
2:60 ꢅ 102
2:13 ꢅ 102
8:32 ꢅ 10
4:62 ꢅ 10
3:34 ꢅ 10
1:63 ꢅ 10
4:92 ꢅ 102
3:94 ꢅ 102
3:68 ꢅ 102
3:75 ꢅ 102
4:43 ꢅ 102
1:38 ꢅ 103
2:31 ꢅ 103
1:48 ꢅ 102
1:57 ꢅ 102
1:67 ꢅ 102
2:37 ꢅ 102
4
4
4
6
6
6
8
pH
10
10
10
12
(b)
5:39 ꢅ 102
7:96 ꢅ 102
1:44 ꢅ 103
1:73 ꢅ 103
3:09 ꢅ 103
1
0.8
0.6
0.4
0.2
0
30
25
20
15
10
5
of the total reaction rate (kstotal) may be represented as the sum
of contributions for these species (Eq. 2),26
kstotal ¼ ksa fðFerH2Þ þ ksb f ðFerHꢁÞ þ ksc fðFer2ꢁÞ;
ð2Þ
a
b
c
where ks , ks , and ks are second-order rate constants (inde-
pendent of pH) for FerH2, FerHꢁ, and Fer2ꢁ, respectively,
and f(FerH2), f(FerHꢁ), and f(Fer2ꢁ) are pH dependent
mole-fractions of FerH2, FerHꢁ, and Fer2ꢁ, respectively. Us-
ing a similar procedure as reported for curcuminoids,26 the de-
0
8
pH
12
(c)
a
b
c
termination of these rate constants (ks , ks , and ks ) and the
dissociation constants (pKa1 and pKa2) for FA was performed.
1
0.8
0.6
0.4
0.2
0
35
30
25
20
15
10
5
The ks value was determined to be 2:64 ꢅ 102 Mꢁ1 sꢁ1 by
b
a
averaging the observed ks values at pH 5.5–8. The ks value
was assumed to be slightly larger than the ks value at pH 4.0
(3:13 ꢅ 102 Mꢁ1 sꢁ1), because the mole fraction of the neutral
c
form (FerH2) will be more than 0.8 at pH 4. The ks value is
considered to be slightly smaller than the ks value at pH 11
(1:63 ꢅ 10 Mꢁ1 sꢁ1) because the di-anion form (Fer2ꢁ) shows
the smallest activity amongst the three species and the largest
contribution at pH 11. Simulations of the pH dependence of
a
c
ks were performed by varying ks , ks , pKa1, and pKa2 values.
From the best fit, each value for FA was estimated to be
0
8
pH
12
pKa1 ¼ 4:70, pKa2 ¼ 9:50, ks ¼ 3:25 ꢅ 102 Mꢁ1 sꢁ1
, and
a
ks ¼ 1:0 ꢅ 10 Mꢁ1 sꢁ1. These pKa values estimated in the
present study agree well with those reported (pKa1 ¼ 4:58,
pKa2 ¼ 9:39).34 Solid lines in Fig. 6a show mole fractions of
three FA species (FerH2, FerHꢁ, and Fer2ꢁ) versus pH, calcu-
lated using pKa1 ¼ 4:70 and pKa2 ¼ 9:50. As shown in Fig. 6a,
good accordance between the measured rate constants and sim-
ulated curve (the broken line) was obtained. In the pH > 5:5
region, the pH-dependent behavior of FA was similar to that
of half-curcumin reported.26 As reported for half-curcumin,
the phenolic OH proton of the FA derivatives is necessary to
scavenge free radicals, indicating that the free radical scaveng-
ing is HAT reaction.26 The free radical scavenging of Fer2ꢁ
should progress as an electron-transfer (ET) process because
Fer2ꢁ has no phenol-proton for the HAT activity. The radi-
c
Fig. 6. Plots of ks versus pH in TX-100 micelle solutions
for (a) ferulic acid, (b) caffeic acid, and (c) chlorogenic
acid. Solid lines show calculated mole-fraction of neutral
and dissociated anion forms versus pH. Broken lines show
total
the simulated ks
versus pH plots (see text).
The proton-dissociation equilibrium for FA is shown in
Fig. 7a. Proton dissociation from the carboxyl group of FA
due to acid–base equilibrium was reported to occur around
pH 4.6 (pKa1 ¼ 4:58), and the proton dissociation from its
phenolic OH was to do around pH 9.4 (pKa2 ¼ 9:39).34 The
neutral form (FerH2), dissociated monoanion form (FerHꢁ),
and dissociated dianion form (Fer2ꢁ) of FA should have differ-
ent rate constants for radical scavenging. The pH dependence

1506 Bull. Chem. Soc. Jpn. Vol. 79, No. 10 (2006)
Antioxidant Activity of Caffeic Acid
(a)
H3CO
COOH
H3CO
HO
COO
H3CO
O
COO
HO
pKa1
pKa2
FerH2
FerH-
ksb = 2.64 × 102 M-1s-1
Fer2-
ksc = 1.0 × 10 M-1s-1
ksa = 3.25 × 102 M-1s-1
(b)
HO
HO
COOH
HO
HO
COO
HO
COO
O
O
COO
O
pKa1
pKa2
pKa3
ksa = 6.40 × 102 M-1s-1 ksb = 3.60 × 102 M-1s-1
(c)
O
O
O
O
HO
HO
C O
HO
HO
C O
HO
O
C O
O
O
C O
COOH
OH
COO
OH
COO
OH
COO
OH
OH
OH
OH
OH
OH
OH
OH
OH
pKa1
pKa2
pKa3
ksa = 1.2 × 102 M-1s-1
ksb = 1.60 × 102 M-1s-1
Fig. 7. Acid–base dissociation equilibriums for (a) ferulic acid, (b) caffeic acid, and (c) chlorogenic acid.
cal-scavenging activity of Fer2ꢁ by ET was quite lower than
that of FerHꢁ by HAT.
ation constants were reported as pKa1 ¼ 4:38, pKa2 ¼ 8:67,
and pKa3 ¼ 12:6 for CA, and pKa1 ¼ 3:36, pKa2 ¼ 8:25, and
pKa3 ¼ 12:3 for CGA.34 The value of pKa2 ¼ 8:67 for CA
(8.25 for CGA) for the dissociation of the first phenolic OH
proton is smaller than those (9.2–9.7) for FA, half-curcumin,
dihydrocaffeic acid, and catechol.26,34 This might be due to
both the stabilization effect of the catechol anion structure
through intramolecular hydrogen bonding with the neighbor-
ing OH proton (Fig. 5) and a resonance effect induced by
the conjugated double bond in the side chain. The decrease
in ks at low pH (pH < 5) for CA is explained the same way
as that in the case of FA, i.e., it is due to the difference in
the molecular activity and the solubility into the micelle
between the neutral and monoanion forms generated by the
acid–base dissociation equilibrium of the carboxyl-proton of
CA.
On the other hand, a small decrease in ks at low pH
(pH < 5) is caused by the ArOꢂ radical which exists inside
the micelle because of its lipophilic property. It is difficult
for ionic species, such as FerHꢁ and Fer2ꢁ, to access the inside
of a micelle because of poor solubility in the hydrophobic
phase.35 The rate constant for the ArOꢂ scavenging reaction
in the micelle solution is controlled by both the amount of
the antioxidant close to ArOꢂ inside the micelle and the reac-
tivity of the antioxidant in a molecular unit. Although the
reactivity of the neutral form (FerH2) is almost the same as
or less than that of the monoanion form (FerHꢁ), the ks value
a
for the neutral form at the low pH range are apparently larger
b
than the ks value for the monoanion form because the neutral
form has a higher solubility to the micelle.27,35 Therefore, in
the micelle system, the neutral form has the highest activity
among three forms of FA arising from its acid–base dissocia-
tion equilibrium. The dissociation of the carboxyl-proton in
FA has little effect on the antioxidant activity.
On the other hand, the large increases in ks at high pH
(pH > 9) for CA and CGA cannot be explained via the HAT
mechanism. The solid lines in Figs. 6b and 6c show the pH
dependence of the mole fractions calculated for four kinds of
species generated from CA and CGA by the acid–base equilib-
riums, using the pKa values reported, respectively. The pH
dependence curves of ks for CA and CGA in the high pH
region (pH > 9) are not consistent with the corresponding
mole-fraction curves. Tentatively, assuming that the pH de-
pendent behavior of ks for CA and CGA occurs the same
Figures 6b and 6c show the plots of ks for CA and CGA ver-
sus pH in TX-100 micelle solutions, respectively. The pH-
dependent behaviors of ks observed for these o-dihydroxycin-
namic acid derivatives are different from those of FA and
half-curcumin.26 The ks values of CA (Fig. 6b) decrease grad-
ually at pH 4–6, are constant at pH 6–8, and increase rapidly
at pH 9–11 with an increase in pH. In the case of CGA
(Fig. 6c), with an increase in pH, the ks values increase slightly
from 1:40 ꢅ 102 at pH 4.0 to 1:67 ꢅ 102 at pH 7.0, and then
from 2:37 ꢅ 102 Mꢁ1 sꢁ1 at pH 8.0 to 3:09 ꢅ 103 Mꢁ1 sꢁ1 at
pH 11.0.
a
way as for FA, the determination of the rate constants (ks ,
b
c
ks , and ks ) and the dissociation constants (pKa1 and pKa2)
total
was done by fitting ks calculated from Eq. 2 to ks observed.
The broken lines in Figs. 6b and 6c show plots of the simulat-
total
ed ks
versus pH using the estimated values (for CA: pKa1
a
¼
b
The proton-dissociation equilibriums for CA and CGA are
shown in Figs. 7b and 7c, respectively. The acid–base dissoci-
4:70, pKa2 ¼ 10:15, ks ¼ 6:40 ꢅ 102 Mꢁ1 sꢁ1, ks ¼ 3:60 ꢅ
c
102 Mꢁ1 sꢁ1, and ks ¼ 2:6 ꢅ 103 Mꢁ1 sꢁ1; for CGA: pKa1
¼

K. Ohara et al.
Bull. Chem. Soc. Jpn. Vol. 79, No. 10 (2006) 1507
•O
R
-
+ H+
H3CO
H3CO
O
+
pH > 10.1
O
ET
O
O
R
H
H3CO
+
O•
HAT
•O
R
pH ≤10.1
+
OH
O
Fig. 8. Possible mechanism of pH-dependent antioxidant action of CA and CGA.
a
b
3:4, pKa2 ¼ 10:15, ks ¼ 1:2 ꢅ 102 Mꢁ1 sꢁ1, ks ¼ 1:60 ꢅ 102
should be ArOꢁ (the anion form of ArOH) (Fig. 8). Accord-
ingly, it is reasonable to consider that the ET process to ArOꢂ
from CA (CGA) can occur at pH ꢇ 10. The reaction rate of
this ET process may be faster than that of the HAT process
producing ArOH because the product is more stable.
On the other hand, for FA and half-curcumin, the contribu-
tion of ET in ArOꢂ scavenging was very small. The results
suggest that the threshold oxidation potential exists in the
ArOꢂ scavenging action by ET. In fact, the oxidation potential
of CA and CGA are 50–90 mV smaller than those for FA and
half-curcumin.
According to the present kinetic study, we concluded that
the radical-scavenging activity of HCAs in ethanol mainly
comes from the HAT reaction of the phenolic OH-protons.
The higher radical-scavenging activity observed for CA deriv-
atives than for FA derivatives is explained by the lower oxida-
tion potential arising from the existence of more electron-
donating hydroxy groups and the stabilization of the corre-
sponding radicals induced by intramolecular hydrogen bond-
ing. The pH dependence of ks for FA in the aqueous micelle
solutions suggests that phenolic OH proton is important in
the radical-scavenging reaction. On the other hand, the pH
dependence of ks for CA and CGA suggests that other factors
are involved in the radical-scavenging action; for example, the
intramolecular hydrogen bonding and ET. The free radical-
scavenging and VE-regenerating activities of these HCAs are
not that remarkable compared to those of other naturally exist-
ing antioxidants, such as tocopherols and catechins. However,
taking into account the large daily intake by animals and
human and high solubility in both aqueous and organic media,
HCAs may act as good antioxidants in living bodies both in
water-rich regions, such as body fluids, and in hydrophobic
regions, such as membranes inside of living bodies.
c
M
ꢁ1 sꢁ1, and ks ¼ 3:5 ꢅ 103 Mꢁ1 sꢁ1). There was good ac-
cordance between the measured rate constants and simulated
curves, however, the estimated pKa2 values are different from
the reported values. Furthermore, the UV–visible absorption
spectrum measured for CGA in the TX-100 (5.0 wt %) micelle
solution red-shifted from 325 nm at pH 7.3 to 370 nm at
pH 9.0, probably due to the phenolic proton dissociation. This
result is consistent with pKa2 ¼ 8:25 reported for the dissocia-
tion of the first phenolic OH proton of CGA.34 In other words,
it cannot be determined that the rapid increases of ks in high
pH region observed for CA and CGA are due to the dissocia-
tion of the phenolic OH-proton.
The pH dependent behavior is probably due to strong intra-
molecular hydrogen bonding in the catechol monoanion moie-
ty of CA (CGA) and the contribution of the ET process in the
antioxidant action. As previously described, several antioxi-
dants that have multiple OH groups in one molecule, such as
ascorbic acid, catechins, flavonoids, and curcumin, show an in-
crease in the free radical-scavenging activity with an increase
of pH.25–30 This behavior was explained by the decrease in the
oxidation potential of the antioxidants in the anion forms by
the electron-donating effect of the minus charge. In the case
of CA, while the first OH-proton of its catechol moiety disso-
ciates around pH 8.7, an increase of ks was not observed there.
This might be explained as follows. When the first OH proton
dissociation occurs, the second OH proton, which contributes
to the HAT reaction, may be trapped via the intramolecular
hydrogen bonding between 3- and 4-phenolate-oxygens. The
hydrogen bonding is known to stabilize protons strongly. As
a result, the pKa value for the dissociation of the second OH
proton is much larger (pKa3 ¼ 12:6) than that for the first
one (pKa2 ¼ 8:67). Therefore, HAT activity for the second
OH does not increase with the first OH dissociation.
In this situation, the increase in ks starting around pH 9 can
be explained by the ET process from CA (CGA) to ArOꢂ. As
reported before, the ArOꢂ scavenging through the ET process
occurs in the cases where the antioxidants have comparatively
low oxidation potential.26,27 For example, the anion forms of
curcumin and edaravone, which have no reactive hydrogen
for the HAT reaction, can scavenge ArOꢂ by ET. Why does
the increase in ks start around pH 9, while the OH-proton of
CA (CGA) starts to dissociate around pH 7? The mechanism
of ArOꢂ scavenging reaction changes. The dissociation con-
stant of the phenol OH-proton of ArOH, which is the reduced
form of ArOꢂ, is estimated to be 10.0–10.5. Therefore, at
pH > 10, the product from ArOꢂ in the scavenging reaction
This work was supported by Grant-in-Aids for the Scientific
Research C (No. 16550016) and Scientific Research on Prior-
ity Areas ‘‘Application of Molecular Spins: Nanomagnets to
Biological Spin Systems’’ (Area No. 769, 15087104) from
the Japanese Ministry of Education, Culture, Sports, Science
and Technology (MEXT).
References
1
2
K. Herrmann, Crit. Rev. Food Sci. Nutr. 1989, 28, 315.
H. Wang, G. J. Provan, K. Helliwell, Food Chem. 2004,
87, 307.
3 M. R. Olthof, P. C. H. Hollman, M. B. Katan, J. Nutr.
2001, 131, 66.

1508 Bull. Chem. Soc. Jpn. Vol. 79, No. 10 (2006)
M. J. del Bano, J. Lorente, J. Castillo, O. Benavente-
Garcia, J. A. del Rio, A. Ortuno, K.-W. Quirin, D. Gerard, J.
Agric. Food Chem. 2003, 51, 4247.
Antioxidant Activity of Caffeic Acid
4
20 H. Kikuzaki, M. Hisamoto, K. Hirose, K. Akiyama, H.
Taniguchi, J. Agric. Food Chem. 2002, 50, 2161.
21 W. Bors, C. Michel, K. Stettmaier, Y. Lu, L. Y. Foo, Biol.
Res. 2004, 37, 301.
22 W. Bors, C. Michel, K. Stettmaier, Y. Lu, L. Y. Foo,
Biochim. Biophys. Acta 2003, 1620, 97.
5
121.
6
M. Petersen, M. S. J. Simmonds, Phytochemistry 2003, 62,
M. Nardini, M. D’Aquino, G. Tomassi, V. Gentili, M. Di
Felice, C. Scaccini, Free Radical Biol. Med. 1995, 19, 541.
J. Laranjinha, O. Vieira, V. Madeira, L. Almeida, Arch.
Biochem. Biophys. 1995, 323, 373.
23 M. C. Foti, C. Daquino, C. Geraci, J. Org. Chem. 2004, 69,
2309.
24 F. A. M. Silva, F. Borges, C. Guimaraes, J. L. F. C. Lima,
C. Matos, S. Reis, J. Agric. Food Chem. 2000, 48, 2122.
25 K. Mukai, S. Mitani, K. Ohara, S. Nagaoka, Free Radical
Biol. Med. 2005, 38, 1243.
26 K. Ohara, W. Mizukami, A. Tokunaga, S. Nagaoka, H.
Uno, K. Mukai, Bull. Chem. Soc. Jpn. 2005, 78, 615.
27 K. Ohara, A. Fujii, Y. Ichimura, K. Sato, K. Mukai, Bull.
Chem. Soc. Jpn. 2006, 79, 421.
7
8
9
J. Laranjinha, E. Cadenas, Life 1999, 48, 57.
V. Roginsky, Arch. Biochem. Biophys. 2003, 414, 261.
10 V. Raneva, H. Shimasaki, Y. Ishida, N. Ueta, E. Niki,
Lipids 2001, 36, 1111.
11 W. Wenfeng, L. Jian, Y. Side, L. Zhirui, Z. Zhihua, Z.
Jiashan, L. Nianyun, Radiat. Phys. Chem. 1995, 46, 41.
12 S. Samman, B. Sandstrom, M. B. Toft, K. Bukhave, M.
¨
Jensen, S. S. Sørensen, M. Hansen, Am. J. Clin. Nutr. 2001, 73,
607.
28 K. Mukai, W. Oka, K. Watanabe, Y. Egawa, S. Nagaoka,
J. Terao, J. Phys. Chem. A 1997, 101, 3746.
13 H. Ito, T. Miyazaki, M. Ono, H. Sakurai, Bioorg. Med.
Chem. 1998, 6, 1051.
29 K. Mukai, M. Nishimura, S. Kikuchi, J. Biol. Chem. 1991,
266, 274.
14 C. A. Gomes, T. G. Gira˜o Da Cruz, J. L. Andrade, N.
Milhazes, F. Borges, M. P. M. Marques, J. Med. Chem. 2003,
46, 5395.
15 M. Kampa, V. I. Alexaki, G. Notas, A. P. Nifli, A.
Nistikaki, A. Hatzoglou, E. Bakogeorgou, E. Kouimtzoglou, G.
Blekas, D. Boskou, A. Gravanis, E. Castanas, Breast Cancer
Res. 2004, 6, R63.
30 K. Mukai, Y. Kanesaki, Y. Egawa, S. Nagaoka, Phyto-
chemicals and Phytopharmaceuticals, ed. by F. Shahidi, C.-T.
Ho, AOCS Press, Champaign, Illinois, 2000, Chap. 20.
31 L. Kong, Z.-L. Sun, L.-F. Wang, H.-Y. Zhang, S.-D. Yao,
Helv. Chim. Acta 2004, 87, 511.
32 H. Cao, W.-X. Cheng, C. Li, X.-L. Pan, X.-G. Xie, T.-H.
Li, THEOCHEM 2005, 719, 177.
16 S. Foley, S. Navaratnam, D. J. McGarvey, E. J. Land, T. G.
Truscott, C. A. Rice-Evans, Free Radical Biol. Med. 1999, 26,
1202.
17 J. H. Chen, C.-T. Ho, J. Agric. Food Chem. 1997, 45, 2374.
18 R. H. Bisby, A. W. Parker, Free Radical Res. 2001, 35, 85.
19 H. Hotta, S. Nagano, M. Ueda, Y. Tsujino, J. Koyama, T.
Osakai, Biochim. Biophys. Acta 2002, 1572, 123.
33 K. Mukai, S. Itoh, H. Morimoto, J. Biol. Chem. 1992, 267,
22277.
34 M. L. Adams, B. O’Sullivan, A. J. Downard, K. J. Powell,
J. Chem. Eng. Data 2002, 47, 289, and reference there in.
35 K. Ohara, R. Watanabe, Y. Mizuta, S. Nagaoka, K. Mukai,
J. Phys. Chem. B 2003, 107, 11527.