Radiolysis of kaempferol in water/methanol mixtures. Evaluation of antioxidant activity of kaempferol and products formed

Article abstract of DOI:10.1021/jf020836g

Oxidative reaction between hydroxymethyl radical (CH₂OH) and kaempferol, in methanol and methanol/water mixtures, was studied by γ-radiolysis using a ⁶⁰Co source. Radiolysis was performed with concentrations and doses ranging from 5 × 10^-5 M to 5 × 10^-3 M and from 0.5 kGy to 14 kGy, respectively. Kaempferol degradation was followed by HPLC, Results showed that CH₂OH reacts with kaempferol at the 3-OH group and produces two depsides (K1 and K2) and other products including K3. K1, K2, and K3 were identified by NMR, LC-MS, and HRMS. The kaempferol degradation pathway leading to the K1, K2, and K3 formation is proposed. It was observed that the more water concentration in the irradiation medium increases, the more K2 concentration increases. Comprehension of food preservation is not clear because many phenomena occurring during irradiation are not established. Radiolysis of kaempferol in water/methanol mixtures helps to elucidate the phenomenon and it is possible that during the treatment of nutriments by γ-irradiation, a series of products such as depside K2 could be formed. Antioxidant properties of kaempferol radiolysis products were evaluated according to their capacity to decrease the EPR DPPH (1,1 -diphenyl2-picrylhydrazil) signal and to inhibit superoxide radicals formed by the enzyme reaction xanthine + xanthine oxidase .

Full text of DOI:10.1021/jf020836g

1270 J. Agric. Food Chem. 2003, 51, 1270−1277
Radiolysis of Kaempferol in Water/Methanol Mixtures.
Evaluation of Antioxidant Activity of Kaempferol and Products
Formed
ABDELGHAFOUR MARFAK,† PATRICK TROUILLAS,† DAOVY-PAULETTE ALLAIS,‡
YVES CHAMPAVIER,§ CLAUDE-ALAIN CALLISTE,† AND JEAN-LUC DUROUX*,†
UPRES EA 1085, “Biomole´cules et Cibles Cellulaires Tumorales”, Faculte´ de Pharmacie,
2 rue du Dr. Marcland, 87025 Limoges Cedex, France
Oxidative reaction between hydroxymethyl radical (•CH2OH) and kaempferol, in methanol and
methanol/water mixtures, was studied by γ-radiolysis using a 60Co source. Radiolysis was performed
with concentrations and doses ranging from 5 × 10-5 M to 5 × 10-3 M and from 0.5 kGy to 14 kGy,
•
respectively. Kaempferol degradation was followed by HPLC. Results showed that CH2OH reacts
with kaempferol at the 3-OH group and produces two depsides (K1 and K2) and other products
including K3. K1, K2, and K3 were identified by NMR, LC-MS, and HRMS. The kaempferol degradation
pathway leading to the K1, K2, and K3 formation is proposed. It was observed that the more water
concentration in the irradiation medium increases, the more K2 concentration increases. Comprehen-
sion of food preservation is not clear because many phenomena occurring during irradiation are not
established. Radiolysis of kaempferol in water/methanol mixtures helps to elucidate the phenomenon
and it is possible that during the treatment of nutriments by γ-irradiation, a series of products such
as depside K2 could be formed. Antioxidant properties of kaempferol radiolysis products were
evaluated according to their capacity to decrease the EPR DPPH (1,1-diphenyl-2-picrylhydrazil) signal
and to inhibit superoxide radicals formed by the enzyme reaction “xanthine + xanthine oxidase”.
KEYWORDS: Kaempferol; gamma irradiation; depsides; antioxidant activity
INTRODUCTION
phenyl}(oxo) methyl acetate (a depside) (14). To understand
flavonoid radiolysis and then preservation of foods by γ-ir-
radiation, we have studied the radiolysis of kaempferol and have
evaluated the antioxidant activity of radiolysis products.
Flavonols, belonging to the flavonoid natural compounds,
have biological activities that may be beneficial to the health
(1-3). Flavonols are present in fruits and vegetables (4, 5). They
have been a subject of increasing research concerning their
antioxidant capacity including radical scavenging, metal che-
lation, and enzyme inhibition. Structure-activity relationships
to scavenge some reactive species including superoxide, hy-
droxyl, and peroxyl radicals have been proposed repeatedly (6-
9). Quercetin and kaempferol are the most efficient radical
scavengers.
At the present time, comprehension of food preservation is a
major problem since certain phenomena occurring during ir-
radiation are not totally established. Some studies have observed
that the levels of polyphenols, including flavonoids, decline
when plant foods undergo γ irradiation treatment (10-12). It
was also observed that γ irradiation can cause a change of
coloration of some irradiated food stuffs (12, 13). Recently, we
have demonstrated that radiolysis of quercetin in methanol
solution forms {2-[(3′,4′-dihydroxybenzoyl)oxy]-4,6-dihydroxy-
MATERIALS AND METHODS
Chemicals and Reagents. HPLC Methanol grade (99.8%) was
purchased from SdS (Peypin, France), acetic acid from Merck (Darm-
stadt, Germany), and kaempferol from Sigma (St. Louis, MO).
Irradiation Treatment. Different solutions of kaempferol (5 × 10-5
M, 10-4 M, 5 × 10-4 M, 5 × 10-3 M) dissolved in methanol were
prepared and were irradiated in aliquots of 1 mL with different doses
(0.5-20 kGy) at a dose rate of 800 Gy/h in the 60Co source carrier
type Oris experimental irradiator.
HPLC Analyses. For each analysis of pure kaempferol (5 × 10-5
M, 10-4 M, 5 × 10-4 M, 5 × 10-3 M), as control, and irradiated
kaempferol, 20 µL were injected into the analytical HPLC system, a
Waters model equipped with a 600 model pump, a variable wavelength
photodiode array detector (PDA 996) and a 600 model controller. The
column used was a 250 × 4.6 mm i.d.,10 µm, µBondapak C18 cartridge
(Waters). The mobile phase consisted of 100% methanol (A) and 1%
aqueous acetic acid (B). Analyses were performed using a linear
gradient from 20% A to 80% A during 40 min at 1 mL/min.
Isolation Procedure Using Preparative HPLC. Purification was
performed by semipreparative HPLC using a 100 × 25 mm i.d.,10
µm, µBondapak C18 cartridge (Waters) and the same solvent system
as above. The gradient was 20-80% A at 5 mL/min during 60 min.
Solutions were evaporated at 45 °C under vacuum.
* Corresponding author. Tel.: 33 (0)5 55 43 58 45. Fax.: 33 (0)5 55 43
58 44. E-mail: duroux@pharma.unilim.fr.
† Laboratoire de Biophysique.
‡ Laboratoire de Pharmacognosie.
§ Service commun de RMN.
10.1021/jf020836g CCC: $25.00 © 2003 American Chemical Society
Published on Web 01/21/2003

Radiolysis of the Antioxidant Kaempferol
J. Agric. Food Chem., Vol. 51, No. 5, 2003 1271
Figure 1. HPLC chromatograms, recorded at 280 nm, of A. kaempferol and irradiated kaempferol (5 × 10-3 M) at different doses: B, 2 kGy; C, 8 kGy;
and D, 14 kGy.
Identification of Products K1, K2, and K3. The identity of these
radiolysis products was achieved by NMR, ES-MS, and FAB-LSIMS.
1H NMR and 13C NMR spectra were measured in CD3OD at 400 MHz
where ref is the reference signal (DPPH + methanol), test is the test
signal, and bg is the background signal. The data were the mean of
four measurements.
1
for H NMR and 100 MHz for 13C NMR on a Bruker DPX Avance
Superoxide Radical Scavenging Test. Oxidation of lucigenin by
superoxide radicals was assayed by the chemiluminescence (CL)
response of lucigenin (7). Superoxide radicals (O2•-) were generated
by the action of 20 µL xanthine oxidase (0.5 U) on 60 µL xanthine (2
mM) in a reaction mixture containing 400 µL phosphate buffer (pH
7.4), 50 µL lucigenin (1 mM), and 5 µL test reagent (methanol as ref-
erence signal). The mixture was quickly vortexed, and CL was measured
in a luminometer analyzer (Lumat LB 9507sBerthold) at room
temperature for 60 s using 75 × 12 mm tubes. Percent inhibition was
calculated using formula 1. Data were the mean of four measurements.
Spectrometer using tetramethylsilane as internal standard. Complete
proton and carbon assignments were based on 1D (1H standard, 13C
Jmod) and 2D (1H-1H correlation spectroscopy (COSY), 1H-13C
heteronuclear multiple quantum coherence (HMQC), and 1H-13C
heteronuclear multiple bond correlation (HMBC, Bruker pulse program
for acquisition: inv4gslplrnd)) NMR experiments. ES-MS analysis was
performed on a Waters Alliance system equipped with a Waters
electrospray interface. The source was operated in the negative and
positive ion modes (ES-) and (ES+) with a 40 V cone voltage. The
FAB-LSIMS mass spectroscopy was performed at “Centre Re´gional
de Mesures Physiques de l’Ouest” (Rennes, France), using metani-
trobenzylic alcohol (mNBA) matrix. The determination of molecular
formula was based on the observation of ion peaks corresponding to
[M - H]-, [M + H]+, and [M + Na]+.
DPPH Scavenging Test. Antioxidant activity was assessed on the
basis of the scavenging activity of stable 1,1-diphenyl-2-picrylhydrazil
(DPPH, from Sigma) free radical using EPR spectroscopy (15). Reaction
mixtures contained 100 µL test samples and 100 µL DPPH ethanolic
solution (5 × 10-4 M). EPR spectra were obtained with a Bruker
ESP300E spectrometer using microsampling pipets at a room temper-
ature under the following conditions: frequency modulation, 100 kHz;
amplitude modulation, 0.197 mT; scanning field, 349.7 mT; receiver
gain, 1.25 105; sweep time, 11 s; microwave power, 4 mW; microwave
frequency, 9.78 GHz. All spectra were recorded in 3 min after
homogenization by agitation. Percent inhibition was calculated using
the double integral of the signal and eq 1
RESULTS AND DISCUSSION
γ Irradiation of Kaempferol in Methanol. Radiolysis of
kaempferol in methanol solution was performed with different
kaempferol concentrations irradiated at doses ranging from 0.5
to 20 kGy. Chromatograms obtained for kaempferol (5 × 10-3
M) irradiated at 2, 8, and 14 kGy are given in Figure 1. These
HPLC data show a decrease in concentration of kaempferol
(retention time of 34 min) between 2 and 14 kGy. A dose of 6
kGy decomposes 50% of kaempferol (5 × 10-3 M) and 14 kGy
is required for a complete degradation. For low concentrations
(<10-4 M), 3 kGy is sufficient to cause total degradation of
kaempferol. Similar results were obtained for quercetin radi-
olysis (14). Figure 1B, 1C, and 1D revealed that degradation
of kaempferol led to the formation of several compounds and
their concentrations increase when those of kaempferol decrease.
K1 (eluted at 30 min) was detected as the major product formed
during radiolysis. Other products were formed in low yields. A
ref - test
ref - bg
percent inhibition )
× 100
(1)

1272 J. Agric. Food Chem., Vol. 51, No. 5, 2003
Marfak et al.
yield increases with the water percentage. This indicates that
the radiolysis of water led to the formation of some radical
species which favor the K2 formation.
Table 1. Radiation Chemical Yields of Kaempferol, K1, K2, and K3
versus the Initial Concentration of Kaempferol and versus the Different
Percentage of Water/ Methanol Mixtures Utilized as System Solvent in
the Radiolysis of Kaempferol
UV/Visible Spectra of K1, K2, and K3. The UV spectra of
these compounds exhibited two major bands in the 222-227
nm range, and in the 270-280 nm range (Figure 3). Compared
with the kaempferol spectrum, we concluded that both the
characteristic absorption bands at 265 nm and at 365 nm
(attributed to A-C and B-C ring conjugation, respectively)
disappeared. This indicated that the C-ring of kaempferol was
destroyed. A similar result was established for quercetin (14).
Identification of K1, K2, and K3. These three radiolysis
products, formed during the radiolysis of kaempferol, were
kaempferol
concentration (M)
G(µmol/J) water/methanol mixtures 5 × 10-5 1 × 10-4 5 × 10-4 5 × 10-3
G(-kaempferol)
0/100%
10/90%
15/85%
20/80%
25/75%
0/100%
10/90%
15/85%
20/80%
25/75%
0/100%
10/90%
15/85%
20/80%
25/75%
0/100%
10/90%
15/85%
20/80%
25/75%
0.0416 0.0679 0.1660 0.4070
0.0369 0.0611 0.1490 0.3540
0.0312 0.0490 0.1310 0.3100
0.0270 0.0428 0.1080 0.2530
0.0229 0.0370 0.0923 0.2120
0.0314 0.0420 0.1550 0.3060
0.0220 0.0304 0.1120 0.2170
0.0144 0.0193 0.0690 0.1490
0.0094 0.0130 0.0465 0.0920
0.0063 0.0084 0.0323 0.0640
0.0017 0.0028 0.0105 0.0291
0.0020 0.0032 0.0131 0.0340
0.0022 0.0036 0.0146 0.0372
0.0024 0.0038 0.0159 0.0401
0.0026 0.0041 0.0197 0.0447
0.0054 0.0081 0.0303 0.0469
0.0046 0.0068 0.0288 0.0397
0.0038 0.0056 0.0220 0.0317
0.0029 0.0045 0.0171 0.0260
0.0021 0.0032 0.0119 0.0188
G(K1)
1
identified by H, 13C, COSY, HMQC, and HMBC NMR and
by negative and positive ES mass data, and their molecular
formula were determined by HRMS.
G(K2)
The observed signals from the K2 1H NMR spectrum (Table
2) corresponded to six aromatic protons: two meta coupled (J
) 2.2 Hz) at 6.27 ppm and at 6.21 ppm, and four ortho coupled
(J ) 8.8 Hz) at 7.96 ppm (2H) and at 6.92 ppm (2H). These
data indicated the presence of an asymmetrical A ring and a
symmetrical B ring, tetra- and disubstituted, respectively. The
positions of all protons were deduced from 1H-1H COSY
spectrum of the proton-proton coupling connectivities.
The 13C Jmod spectrum of K2 gave fifteen carbons signals
(Table 2): six tertiary carbons at δ 133.9 ppm (C-2′/C-6′), 116.6
ppm (C-3′/C-5′), 105.0 ppm (C-5), and 101.7 ppm (C-3) and
nine quaternary carbons. The latter were identified as three
carbonyl carbons at δ 188.6 ppm (C-7), 165.7 ppm (C-8), and
165.9 ppm (C-7′) representing ketone, acid, and ester groups,
respectively; four O-linked aromatic carbons at δ 168.2 ppm
(C-4), 167.7 ppm (C-2), 164.8 ppm (C-4′), and 155.3 ppm (C-
6) and two C-linked aromatic carbons at δ 120.5 ppm (C-1′)
and 105.7 ppm (C-1). The positions of the six tertiary carbons
were assigned from the 1H-13C HMQC experiments (Table 2):
the carbons at 133.9, 116.6, 105.0, and 101.7 ppm correlated
with protons at 7.96, 6.92, 6.21, and 6.27 ppm, respectively.
These data confirmed the asymmetrical and the symmetrical
structures of A and B rings. The contributions of chemical shifts
to the different O-linked and C-linked aromatic carbons were
deduced from the carbon-proton long-range couplings obtained
from the HMBC experiments (Table 2).
G(K3)
similar result was also observed for quercetin radiolysis (14).
K1 and two minor radiolysis products named K2 (retention time
of 16 min) and K3 (retention time of 25 min) were separated,
from a 5 × 10-3 M kaempferol methanol solution (30 mg)
irradiated at 14 kGy, by repeated reversed-phase semipreparative
HPLC and gave 14.5 mg (K1), 6.2 mg (K2), and 4.9 mg (K3).
During the evaporation step, we observed that K1 was trans-
formed into compound K2. Concentrations of K1, K2, and K3
were calculated by the use of calibration curves performed with
purified materials. Radiation chemical yields of radiolysis
products were calculated from the slope of linear concentration-
dose equations and with the formula (2) (16)
106 ∆M
0.79 D
G(µmol/J) )
(2)
where D is the dose (Gy), ∆M the concentration (M) of the
molecules formed, and it is well known that G is the radiation
chemical yield, which corresponds to the number of molecules
liberated per 100 eV of absorbed energy.
G(-K), G(K1), G(K2), and G(K3) are summarized in Table
1. The radiation chemical yield of each compound increases
with initial concentration of kaempferol and seemed to saturate
for high concentrations (>10-3 M) (data not shown). This
agreement with a classic radiolytic process where the solute
reacts with the radiolytic species because of solvent, until all
species are consumed.
γ Irradiation of Kaempferol in Water/Methanol Mixtures.
Kaempferol in different mixtures of water/methanol (10/90, 15/
85, 20/80, 25/75) was irradiated. Figure 2 reports HPLC
chromatograms of kaempferol (5 × 10-3 M) irradiated at 14
kGy versus the different percentage of water/methanol. Com-
pared with the result obtained for radiolysis of kaempferol in
methanol (Figure 1D), the chromatograms are quite identical,
indicating that in the presence of water, K1, K2, and K3 are
also formed. Radiation chemical yields of K1, K2, K3, and K
were calculated and summarized in Table 1. We observed that
kaempferol consumption and K1 and K3 formation decreases
when methanol percentage decreases. K2 radiation chemical
In the ROESY and the NOESY spectra of K2, no correlation
was observed between H3 and H6′ or H2′. This indicates that
the stable conformation is obtained for a torsion angle between
the A and B rings which implicates a high distance (g4-5 Å)
between the cited protons. However, the connection between
the two rings was established from the HMBC spectrum. First,
3
the J correlation between C-7′ (165.9 ppm) and the protons
H2′/H6′ (7.96 ppm), together with the chemical shift of C-1′
(120.5 ppm) which corresponds to a C-linked carbon, demon-
strates the C1′-C7′ linkage. Second, C-7′ (165.9 ppm) was
identified as an ester group and C-2 (167.7 ppm) as a non-
hydroxylated O-linked carbon because of the A ring structure.
This led us to establish the connection between A and B rings
through the C7′-O-C2 linkage.
The absence of three bond couplings between any protons
and the C-7 (188.6 ppm) and C-8 (165.7 ppm) carbons, together
with the presence of the quaternary C-linked carbon (105.7
ppm, C-1), led us to propose the C1-C7-C8 linkage. Fin-
ally, K2 was identified as {2-[(4′-hydroxybenzoyl)oxy]-4,6-
dihydroxyphenyl}(oxo) acetic acid (Figure 4).
ES-MS analysis of K2 suggested 318 as a molecular weight.
The high-resolution mode of the FAB-MS revealed that the

Radiolysis of the Antioxidant Kaempferol
J. Agric. Food Chem., Vol. 51, No. 5, 2003 1273
Figure 2. HPLC chromatograms, recorded at 280 nm, of irradiated kaempferol methanol/water solutions (5 × 10-3 M) at dose of 14 kGy. A, 90/10; B,
85/15; C, 80/ 20; and D, 75/25.
3
because of a J correlation observed between the methoxyl
protons (3.46 ppm) and the C-8 carbon (165.4 ppm). Conse-
quently, K1 was identified as {2-[(4′-hydroxybenzoyl)oxy]-4,6-
dihydroxyphenyl}(oxo) methyl acetate (Figure 4).
The ES-MS and the high-resolution mode of the FAB-MS
demonstrated that the molecular formula of K1 was C16H12O8
(calculated for C16H13O8 and C16H12O8Na, 333.0610 and
355.0429; found, 333.0607 and 355.0428, respectively) confirm-
ing the structure obtained from the NMR analysis.
1
The H spectrum of K3 consisted of three signals. Two,
corresponding to four aromatic protons, detected at 7.84 ppm
(H2/H6) and 6.88 ppm (H3/H5), were ortho coupled (J ) 8.8
Hz), and one signal at 3.93 ppm revealed the presence of a
methoxyl group. The 13C spectrum consisted of four tertiary
carbons at δ 133.8 ppm (C-2/C-6) and 116.9 ppm (C-3/C-5),
three quaternary signals at δ 125.2 ppm (C-1), 165.8 ppm (C-
Figure 3. UV spectra of products K1, K2, K3, and kaempferol.
4), and 166.5 ppm (C-8), a carbonyl group at δ 186.5 ppm (C-
7) and one signal at 53.0 ppm, confirming the presence of a
methoxyl group. The positions of the tertiary and the quaternary
carbons were identified by the HMQC and HMBC spectra,
respectively, and the position of OCH3 was deduced according
to the 3J correlation between the methoxyl protons (3.93 ppm)
and the C-8 carbon. Finally, K3 was identified as [4-hydrox-
yphenyl](oxo) methyl acetate (Figure 4) or methyl 4-hydrox-
yphenylglyoxylate (17). Mass spectroscopy analyses demon-
strated that the molecular weight of K3 is 180 corresponding
to the molecular formula C9H8O4 and confirming the structure
obtained by NMR analyses.
molecular formula of K2 was C15H10O8 (calculated for C15H10O8-
Na, 341.02734; found, 341.02730) corresponding to the structure
obtained from the NMR analysis. In addition, the positive ES-
MS spectrum displayed other ion peaks at m/z 121 and 199
corresponding to the fragment C7′O linked to the B ring and to
the residue of K2 (+2H), respectively (Figure 4). This confirms
the connection between the A and B rings through C2-O-
C7′.
1
The H NMR and 13C spectra (Table 2) of K1 were very
similar to those of K2, and it appeared that they belonged to
the same analogous series, differing only by the presence of a
methoxyl group represented by a carbon at 53.3 ppm and by a
three-proton singlet (3.46 ppm). The chemical structure of K1
was determined by procedures similar to those used for
identification of K2. The methoxyl position was assigned
Formation of K1 and K3. To understand the radiolysis of
solutes, two effects are proposed, direct and indirect. The first
type concerns direct interactions between molecules in solution

1274 J. Agric. Food Chem., Vol. 51, No. 5, 2003
Marfak et al.
Table 2. 1H (400 MHz); 13C (100 MHz) NMR Data and Correlations
Observed in COSY, HMQC, and HMBC Spectra of Products K1 and
K2 in CD OD/TMS (δ ppm; J Hz)
3
position
1H
COSY
13C Jmod/ HMQC
HMBC
Product K1
1
2
3
4
5
6
7
8
105.5
167.5
101.6
168.0
104.8
155.4
188.6
165.4a
53.3
6.26 d (2.2)
6.19 d (2.2)
H-5
H-3
C-5; C-1
C-1; C-3
8-OCH3
3.46 s
C-8
1′
2′
3′
4′
5′
6′
7′
120.9
133.9
116.5
164.8
116.5
133.9
165.9a
7.99 d (8.7)
6.88 d (8.7)
H-3′
H-2′
C-4′; C6′/2′; C-7′
C-1′; C
5′/3′
6.88 d (8.7)
7.99 d (8.7)
H-6′
H-5′
C-1′; C
3′/5′
C-4′; C2′/6′; C-7′
Product K2
1
2
3
4
5
6
7
8
1′
2′
3′
4′
5′
6′
7′
105.7
167.7
101.7
168.2
105.0
155.3
188.6
165.7a
120.5
133.9
116.6
164.8
116.6
133.9
165.9a
6.27 d (2.2)
6.21 d (2.2)
H-5
H-3
C-5; C-1
C-1; C-3
7.96 d (8.8)
6.92 d (8.8)
H-3′
H-2′
C-4′; C6′/2′; C-7′
C-1′; C
5′/3′
6.92 d (8.8)
7.96 d (8.8)
H-6′
H-5′
C-1′; C
3′/5′
C-4′; C2′/6′; C-7′
a Assignments may be interchangeable.
Figure 4. Selected 1H−13C long-range correlations (3J) (T) observed in
HMBC spectra of products K1, K2, and K3.
and γ-rays. This leads to excitation and ionization of molecules.
In fact, direct effects could occur when using solutions with
high concentrations or with large molecules such as polymers
or DNA. In our study, we worked with concentrations of
kaempferol lower than 5 × 10-3 M, so direct effects could only
be considered as a minor contribution. The second effect is
responsible for chemical degradation of molecules. Indeed, the
reactive species formed during the radiolysis of solvent react
with the solute and lead to the formation of other radiolysis
products. In our investigation, methanol and water were utilized
as solvents. Radiolysis of these two solvents has been well
established for many years (6, 18). It has been shown that 50
ns after the beginning of irradiation, several reactive species
including solvated electron (eS-), H2, H•, CH3O•, •CH2OH, (CH2-
OH)2, H2CO, H+ such as CH3OH2+ are formed from methanol,
and (eS-), H2, H•, HO• from water. For low concentrations (<5
× 10-3 M), these radical species react with the solute and lead
to its degradation. The strong potential of electron transfer of
the cited species implicates that the solute degradation is
attributed to redox reactions, including stereospecific oxidation,
hydroxylation, and scission of molecules.
It has been reported (6, 8, 19) that antioxidant potential of
kaempferol is attributed to the presence of hydroxyl substituents
in positions 3 and 4′ which permit redox transfers and to the π
electron delocalization between the B ring, the 2,3-double bond,
and the 4-keto group which is responsible for stabilization of
flavonoxy radicals produced after oxidative reactions. Therefore,
kaempferol has potential for redox reactions occurring during
radiolysis. This kaempferol chemical degradation led to the
formation of depsides K1 and K2. Flavonols transformation to
depsides was observed, more than 30 years ago, by Gottlieb
and co-workers (20). They demonstrated that aerial oxidation
of quercetin leads to saturation of the 2,3-double bond, scission
of C-ring, and consequently formation of depside. To understand
the radiolysis process as well as the radical reactions occurring
in the irradiated medium, we propose hypothetical degradation
mechanism for kaempferol (Figure 5). We have observed that
the consumption of kaempferol decreased with the presence of
water in the irradiated medium. Radiolysis of both solvents
methanol and water leads to the formation of •H. We conclude
that the degradation of kaempferol is probably not attributed to
•H radicals and propose that the first step of the degradation
pathway comprises the abstraction of the hydrogen atom of the
•
3-OH group, by the CH2OH (•OCH3) radical, leading to the
flavonoxy radical (K•) and to the methanol formation. The
second step is the formation of a keto group in position 3 by
mesomeric effect. The formed radical (K11) reacts with
methanol and leads to formation of an intermediate product K12.
Another •CH2OH (•OCH3) radical could react with K12 yielding
the K13 radical. The formation of keto group in position 2 of
radical K13 requires scission of the C-ring. Two possible
pathways are proposed: (1) Scission of the C2-C3 bond and
consequently formation of an intermediate radical K14 which
could react with the methanol to give the depside K1 and (2)
scission of the C2-O1 bond (Figure 4) to give the radical K15
followed by the scission of the C3-C4 bond and then formation
of the radical K16. This latter reacts with methanol yielding
the compound K3. K1 was detected as the major product formed

Radiolysis of the Antioxidant Kaempferol
J. Agric. Food Chem., Vol. 51, No. 5, 2003 1275
Figure 5. Proposed kaempferol-degradation mechanism yielding to the K1 and K3 formation.
during radiolysis. This result demonstrated that cleavage of K13
at the C2-C3 bond is more probable than that at the C2-O1
bond.
The same result has been observed for the quercetin radiolysis
products (14).
Concerning the K2 formation in the irradiated medium,
attacking of K1 by the radiolytic species could be suggested,
but this proposition is not confirmed because we have not
observed any product after irradiation of K1. Therefore, we
postulate that the K2 formation in the irradiated medium is
attributed to the presence of low concentrations of •OH radicals
which probably react with the K14 radical (Figure 5) to form
K2. This hypothesis is confirmed because during kaempferol
radiolysis in water/methanol mixtures, the more water concen-
tration increases, the more K2 concentration increases.
In conclusion, we observed that the presence of the OH group
in position 3 was the key for the γ-degradation of kaempferol
in methanol solution. This was confirmed because no transfor-
mation was observed on the flavones, apigenin and luteolin,
analogous to the flavonols kaempferol and quercetin, respec-
tively, when we studied radiolysis in the same experimental
conditions used for kaempferol.
Flavonoid metabolism and absorption has been the subject
of investigation and certain studies reported the role of intestinal
microflora (8, 21). Several phenolic compounds and aromatic
acids have been observed which were formed by detachment
of the A ring from the flavonoid molecule and by opening of
the heterocyclic C-ring. For example, Winter and co-workers
(22) reported that four Clostridium strains, recovered from
Kaempferol and quercetin are flavonols distinguished only
by hydroxylation of position 3′ for quercetin. The mechanism
leading to the formation of K1 and the major radiolysis product
of quercetin (14) do not implicate the B ring. Hence, we
conclude that 3′ and 4′ hydroxylation do not influence formation
of the major radiolysis products obtained by irradiation of both
quercetin and kaempferol. However, it is possible that these
groups react with radicals of the irradiated medium to form
minor radiolytic compounds. It is well known that oxidation of
the B ring leads to flavonoxy radicals and quinone forms.
•
Presence of OCH3 radicals in the irradiated medium can also
lead to methoxylation of the B ring.
Formation of K2. K2 was detected as a minor metabolite in
the irradiated kempferol solution, and it was also formed during
evaporation of K1, out of the irradiated medium. Indeed, the
eluate containing K1 was collected with approximately 70%
methanol and 30% water. During the evaporation phase, water
seemed to hydrolyze the methoxyl group of K1 and forms the
product K2 through an reverse esterification reaction. In
presence of methanol, we observed esterification of K2 and the
following chemical equilibrium is established:
K2 + CH3OH h K1 + H2O

1276 J. Agric. Food Chem., Vol. 51, No. 5, 2003
Marfak et al.
superoxide scavenging activity of K2 is 6 times as high as K1.
This demonstrates the importance of the 8-OH group in K2.
Table 3. IC50 Values of Kaempferol as Control, K1, K2, and K3
According to the Inhibition of Superoxide Anion and the Reduction of
EPR DPPH Signal
Fruits, vegetables, and beverages are rich sources of flavonols.
The irradiation process has been reported as a potential treatment
for extending the shelf life of foods (28-30). γ-irradiation of
Citrus clementina increases the synthesis of total phenolic
compounds such as phenolic acids and flavonoids (31) but little
is known concerning the degradation behavior of these products
when foods undergo irradiation treatment. Low solubility of
most flavonoids in neutral aqueous media prevented us from
performing the radiolysis of kaempferol in water. Therefore,
we have chosen methanol as solvent. Although this is not the
exact medium of flavonoids in fruits and there are not the same
radicals constituting the irradiated medium during the preserva-
tion treatment, we have also studied kaempferol radiolysis in
different water/methanol mixtures and have demonstrated that
formation of depside K2 increased in the presence of water. It
is possible that during the preservation process of foods by
γ-irradiation, a series of products such as depside K2 could be
formed.
superoxide anion
IC50 (µΜ) SD
DPPH radical
IC50 (mΜ)
SD
kaempferol
1.18
18.12
5.37
1.22%
2.45%
1.60%
2.14%
0.19
3.12
0.79%
1.50%
1.31%
1.42%
K1
K2
K3
2.76
83.91
13.10
human feces, synthesize enzymes capable of cleaving the C-ring
of some flavonoids such as quercetin, kaempferol, and narin-
genin at the C3-C-4 bond. This type of C-ring opening was
observed for the anaerobic degradation of flavonoids by
Eubacterium ramulus (23). Opening the pyrone ring was also
found during the degradation of quercetin by methanogenic
Consortia (24).
Our study has shown that γ-irradiation allows the C-ring
cleavage at C2-C3 double bond to form the depsides K1 and
K2 and at the C2-O1 bond giving K3 via oxidative reactions
similar to those occurring in microorganisms. For example,
oxygenation of flavonols obtained by copper (II) complexes
reproduced the in vivo quercetin and kaempferol transformation
to depsides. This reaction is catalyzed by dioxygenase quer-
cetinase (25). Recently, it has been demonstrated that in the
presence of free radicals such as 2,2,6,6-tetramethyl-1-piperidi-
nyloxyl (TEMPO) and 2,6-di-tert-butyl-a-(3,5-di-tert-butyl-4-
oxo-2,5-cyclohexadien-1-ylidene)-p-totylox yl (galvinoxyl), fla-
vonols undergo catalytic oxygenation to the corresponding
depsides with concomitant CO release (26). Compared with our
study, the loss of CO was not observed in the K1 and K2
ACKNOWLEDGMENT
The authors are grateful to H. Lotfi (Faculte´ de Pharmacie de
Limoges, France) for his help in LC-MS measurements through-
out this work.
LITERATURE CITED
(1) Brasseur, T. Anti-inflammatory properties in flavonoids. J.
Pharm. Belg. 1989, 44, 235-241.
(2) Huang, M.; Ferraro, T. Phenolic compounds in foods and cancer
prevention. In Phenolic compounds in food and their effect on
health II, Huang, M. T., Ho, C. T., Lee, C. Y., Eds.; ACS
Symposium Series; American Chemical Society: Washington,
DC, 1992; p 507, Chapter 2.
•
formation mechanism, but the CH2OH radicals play the same
role as that of TEMPO and galvinoxyl because all these radicals
begin the flavonol degradation process at the 3-OH group.
Antioxidant Capacity of Kaempferol Radiolysis Products.
We observed that γ-irradiation transformed kaempferol to a
major depside K1 and to other minor radiolysis products, among
which we identified K2 and K3. We have discussed the depside
formation in microorganisms but no published work has dealt
with the antioxidant activity of these products. We isolated K1,
K2, and K3 from kaempferol methanol solution (5 × 10-3 M)
irradiated at 14 kGy, and we evaluated their antioxidant ability
to scavenge DPPH and superoxide radicals using the EPR and
CL, respectively. IC50 values were calculated by linear regres-
sion and are listed in Table 3. The highest superoxide and DPPH
scavenging activities were observed for kaempferol, followed
by depsides K2 and K1. The metabolite K3 possessed the lowest
scavenging capacity.
Flavonoid structure radical-scavenging abilities relationships
have been reported repeatedly (6-9, 27). More specifically,
three criteria for radical-scavenging capacity are suggested (19).
First, presence of the catechol group in B ring, which has better
electron-donating properties. Second, the 2,3-double bond
conjugated with the 4-oxo group, which is responsible for
electron delocalization. Third, presence of a 3-hydroxyl group
in the heterocyclic ring. Therefore, the antioxidant activity of
kaempferol was attributed to the 4′, 3, 5, and 7-OH groups and
to the C2-C3 double bond. 4′, 5, and 7-OH remain during the
kaempferol transformation to K1 and K2. The H-donor capacity
of these three groups (4′-OH, 4-OH, and 6-OH in Figure 4)
can explain the antioxidant activity of K1 and K2. π Electron
delocalization can exist from the A ring to the 8-keto group
and from the B ring to the 7′-keto group. Additionally, the
(3) Hertog, M. G. L.; Feskens, E. J. M.; Hollman, P. C. H.; Katan,
M. B.; Kromhout, D. Dietary antioxidant flavonoids and risk of
coronary heart disease: The Zutphen elderly study. Lancet 1993,
342, 1007-1011.
(4) Herrmann, K. Flavonols and flavones in food plants: A review.
J. Food Technol. 1976, 11, 433-448.
(5) Justesen, U.; Knuthsen, P.; Teth, T. Quantitative analysis of
flavonols, flavones, and flavanones in fruits, vegetables and
beverages by high-performance liquid chromatography with
photo-diode array and mass spectrometric detection. J. Chro-
matogr., A 1998, 799, 101-110.
(6) Jovanovic, S. V.; Steenken, S.; Tosic, M.; Marjanovic, B.; Simic,
M. G. Flavonoids as antioxidants. J. Am. Chem. Soc. 1994, 116,
4846-4851.
(7) Cotelle, N.; Bernier, J. L.; Catteau, J. P.; Pommery, J.; Wallet,
J. C.; Gaydou, E. M. Antioxidant properties of hydroxy-flavones.
Free Radical Biol. Med. 1996, 20, 35-43.
(8) Rice-Evans, C. A.; Miller, N.; Paganga, G. Structure-antioxidant
activity relationships of flavonoids and phenolic acids. Free
Radical Biol. Med. 1996, 20, 933-956.
(9) Guohua, C.; Emin, S.; Ronald, L. Antioxidant and prooxidant
behaviour of flavonoids: Structure-activity relationships. Free
Radical Biol. Med. 1997, 22, 749-760.
(10) Oufedjikh, H.; Mahrouz, M.; Lacroix, M.; Amiot, J.; Taccini,
M. The influence of gamma irradiation on flavonoids content
during storage of irradiated clementina. Radiat. Phys. Chem.
1998, 52, 107-112.
(11) Benoˆıt, M.; D’Aprano, G.; Lacroix, M. Effect of gamma-
irradiation on phenylalanine ammonia-lyase activity, total phe-
nolic content, and respiration of mushrooms (Agaricus bisporus).
J. Agric. Food Chem. 2000, 48, 6312-6316.

Radiolysis of the Antioxidant Kaempferol
J. Agric. Food Chem., Vol. 51, No. 5, 2003 1277
(12) Beaulieu, M.; D’Aprano, M. B.; Lacroix, M. Dose rate effect of
gamma irradiation on phenolic compounds, polyphenol oxidase,
and browning of mushrooms (Agaricus bisporus). J. Agric. Food
Chem. 1999, 47, 2537-2543.
(13) Zareena, A. V.; Variyar, P. S.; Gholap, A. S.; Bongirwar, D. R.
Chemical investigation of gamma-irradiated saffron (Crocus
satiVus L.). J. Agric. Food Chem. 2001, 49, 687-691.
(14) Marfak, A.; Trouillas, P.; Allais, D. P.; Champavier, Y.; Calliste,
C. A.; Duroux, J. L. Radiolysis of quercetin in methanol
solution: Observation of depside formation. J. Agric. Food
Chem. 2002, 49, 4827-4833.
(15) Calliste, C. A.; Trouillas, P.; Allais, D. P.; Simon, A.; Duroux,
J. L. Free radical scavenging activities measured by electron spin
resonance spectroscopy and B16 cell antiproliferative behaviors
of seven plants. J. Agric. Food Chem. 2001, 49, 3321-3327.
(16) Fricke, H.; Hart, E. J. In Chemical dosimetry, Radiation
dosimetry: Instrumentation; Attix, F. H., Roesch, W. C., Tochilin,
E., Eds.; Academic Press: 1966; Vol. II., pp 167-237.
(17) Chen, Y. T.; Seto, C. T. Divalent and trivalent R-ketocarboxylic
acids as inhibitors of protein tyrosine phosphatases. J. Med.
Chem. 2002, 45, 3946-3952.
(18) Freeman, G. R. In The radiolysis of alcohols, kinetics of
nonhomogeneous processes: A practical introduction for chem-
ists, biologists, physicists, and materials scientists; Freeman, G.
R., Ed.; Wiley-Interscience: New York, 1987; pp 73-101.
(19) Pietta, P. G. Flavonoids as antioxidants. J. Nat. Prod. 2000, 63,
1035-1042.
(24) Herrmann, A. P.; Willems M.; Janke, H. D. Degradation of
natural polyphenols by methanogenic consortia enriched from
digested municipal sludge. Water Res. 2001, 35, 2575-2582.
(25) Balogh-Hergovich, E.; Kaiser, J.; Speier, G. Synthesis and
characterization of copper (I) and copper (II) flavonolate
complexes with phthalazine, and their oxygenation and relevance
to quercetinase. Inorg. Chim. Acta 1997, 256, 9-14.
(26) Kaizer, J.; Speier, G. Radical-initiated oxygenation of flavonols
by dioxygen. J. Mol. Catal. A 2001, 171, 33-36.
(27) Cos, P.; Ying, L.; Calomme, M.; Hu, J. P.; Cimanga, K.; Van
Poel, B.; Pieters, L.; Vlietinck, A. J.; Vanden Berghe, D.
Structure-activity relationship and classification of flavonoids
as inhibitors of xanthine oxidase and superoxide scavengers. J.
Nat. Prod. 1998, 61, 71-76.
(28) Kader, A. A. Potential applications of ionizing radiation in
postharvest handling of fresh fruits and vegetables. Food Technol.
1986, 40, 117-121.
(29) Urbain, W. M. Food irradiation. Food Science Technology; a
series of monographs; Academic Press: Orlando, FL, 1986.
(30) Thomas, P. Radiation preservation of foods of plant origins Part
VI. Mushrooms, tomatoes, minor fruits and vegetables, dried
fruits and nuts. CRC Crit. ReV. Food. Sci. Nutr. 1988, 26, 313-
358.
(31) Oufedjikh, H.; Lacroix, M.; Mahrouz, M.; Amiot, M. J. Effect
of γ irradiation on phenolic compounds and phenylalanine
ammonia-lyase activity during storage in relation to peel injury
from peel of Citrus clementina Hort. ex. Tanaka. J. Agric. Food
Chem. 2000, 48, 559-565.
(20) Gottlieb, O. R. In The flaVonoids: FlaVonols; Harborne, J. B.,
Mabry, T. J., Eds.; Chapman and Hall: London, 1975; pp 296-
375.
(21) Das, N. P. Studies on flavonoid metabolism: Absorption and
metabolism of (+)-catechin in man. Biochem. Pharmacol. 1971,
20, 3435-3445.
(22) Winter, J.; Moore, L. H.; Dowell, V. R.; Bokkenheuser, V. D.
C-ring cleavage of flavonoids by human intestinal bacteria. Appl.
EnViron. Microbiol. 1989, 55, 1203-1208.
Received for review July 29, 2002. Revised manuscript received
November 4, 2002. Accepted November 5, 2002. This work was
supported in part by the “Ministe`re de l’Education Nationale, de la
Recherche et de la Technologie” and by the “Conseil Re´gional du
Limousin” (France).
(23) Schneider, H.; Blaut, M. Anaerobic degradation of flavonoids
by Eubacterium ramulus. Arch. Microbiol. 2000, 173, 71-75.
JF020836G