Accumulation of epigallocatechin quinone dimers during tea fermentation and formation of theasinensins

Article abstract of DOI:10.1021/np020245k

Production and accumulation of catechin dimer quinones during tea fermentation were chemically confirmed for the first time by trapping as phenazine derivatives. Direct treatment of the fermented tea leaves with o-phenylenediamine yielded five phenazine derivatives (8-12) of o-quinones of an epigallocatechin dimer and its galloyl esters (13-16), in which two flavan units were linked at the B-rings through a C-C bond. Atrop isomerism of the biphenyl bonds was shown to be the R configuration, suggesting that the o-quinone dimers were generated by stereoselective coupling of monomeric quinones. The total concentration of the phenazine derivatives in the o-phenylenediamine-treated tea leaves was higher than that of theaflavins. In contrast, phenazine derivatives of monomeric quinones of epigallocatechin were not isolated. When the fermented tea leaves were heated, the quinone dimers were converted to theasinensins, which are constituents of black tea, suggesting that theasinensins are generated by reduction of the quinone dimers during the heating and drying steps in black tea manufacturing.

Full text of DOI:10.1021/np020245k

1582
J . Nat. Prod. 2002, 65, 1582-1587
Accu m u la tion of Ep iga lloca tech in Qu in on e Dim er s d u r in g Tea F er m en ta tion
a n d F or m a tion of Th ea sin en sin s
Takashi Tanaka,† Chie Mine,† Sayaka Watarumi,† Toshihiro Fujioka,‡ Kunihide Mihashi,‡ Ying-J un Zhang,† and
Isao Kouno*,†
Department of Molecular Medicinal Sciences, Graduate School of Biomedical Sciences, Nagasaki University,
Bunkyo Machi 1-14, Nagasaki 852-8521, J apan, and Faculty of Pharmaceutical Sciences, Fukuoka University,
8-19-1 Nanakuma, J onan-ku, Fukuoka 814-0180, J apan
Received May 24, 2002
Production and accumulation of catechin dimer quinones during tea fermentation were chemically
confirmed for the first time by trapping as phenazine derivatives. Direct treatment of the fermented tea
leaves with o-phenylenediamine yielded five phenazine derivatives (8-12) of o-quinones of an epigallo-
catechin dimer and its galloyl esters (13-16), in which two flavan units were linked at the B-rings through
a C-C bond. Atrop isomerism of the biphenyl bonds was shown to be the R configuration, suggesting
that the o-quinone dimers were generated by stereoselective coupling of monomeric quinones. The total
concentration of the phenazine derivatives in the o-phenylenediamine-treated tea leaves was higher than
that of theaflavins. In contrast, phenazine derivatives of monomeric quinones of epigallocatechin were
not isolated. When the fermented tea leaves were heated, the quinone dimers were converted to
theasinensins, which are constituents of black tea, suggesting that theasinensins are generated by
reduction of the quinone dimers during the heating and drying steps in black tea manufacturing.
Polyphenol oxidation during black tea manufacturing
remains unclear despite many studies by a number of
groups. When fresh tea leaves [Camellia sinensis (L.) O.
Kuntze, Theaceae] are crushed at the initial stage of black
tea manufacturing, four major catechins in the leaves, (-)-
epicatechin (1), (-)-epigallocatechin (2), and their galloyl
esters (3 and 4), are enzymatically oxidized, and the
resulting quinones undergo complex chemical changes.1
Theaflavins (5),2-4 reddish-orange pigments of black tea,
are catechin dimers having a characteristic benzotropolone
moiety produced by condensation of a pair of quinones (1a -
4a ) derived from dihydroxy and trihydroxy B-rings of
catechins (1-4) (Figure 1). In contrast, the remaining
components of the color of black tea infusions, the so-called
thearubigins,1,5 are a heterogeneous mixture of compounds
with large molecular size. Although they may be partly
produced by further oxidation of theaflavins,6,7 most of the
chemical studies up to the present have presumed that
these metabolites are oligomers and polymers of the
quinones.1 In the late 1950s, Roberts, who designated the
major black tea pigments, pointed out that 2 and its gallate
are important as the principal substrates in the oxidation
during tea fermentation,8 because they comprise about 70%
of the green tea catechins and their trihydroxy benzene
rings have the lowest oxidation-reduction potential among
oxidizable polyphenols in tea leaves.9 Furthermore, besides
production of theaflavins, he postulated a mechanism by
which 2 or 4 is oxidatively coupled to form dimeric quinone
intermediates (6), and further polymerization beyond the
dimer stage appears unlikely.8,10 Following that, a few short
reports supporting his hypothesis appeared within several
years.11,12 However, later studies on structural elucidation
of theaflavins in the early 1960s overshadowed his pioneer-
ing work.2-4
isolating the metabolite 7 (Figure 2), corresponding to a
hydrated form of the Roberts’ quinone (6).14 Although the
yield of 7 was significantly lower than that of theaflavins,
we believed that it might represent only a part of unstable
quinone metabolites generated during catechin oxidation,
based on our previous findings which suggested the pres-
ence of various unknown quinones in the fermentation
mixture.6,7 Since formation of o-quinones at the initial stage
of the oxidation is generally accepted, we attempted to trap
the unstable quinones by addition of o-phenylenediamine,
which condenses with o-quinones to form stable phenazine
derivatives.
Resu lts a n d Discu ssion
Fresh tea leaves were crushed and spread on a glass tray
in a thin layer at 25 °C. After 2.5 h, an aliquot of the
fermented leaves was extracted with EtOH and another
aliquot was extracted with EtOH containing 1% acetic acid
and 0.2% o-phenylenediamine, the quinone-trapping re-
agent,15 and the extracts were separately analyzed by
reversed-phase HPLC (Figure 3). The chromatogram of the
ethanol extract (Figure 3a) showed peaks arising from
caffeine, catechins (1-4), flavonol glycosides, and theafla-
vins. In contrast, in the chromatogram of the extract
treated with o-phenylenediamine (Figure 3b), peaks due
to many additional products appeared. Photodiode array
detection indicated that the major products (8-12) showed
similar UV spectra with absorption maxima at 376 nm.
Follow-up experiments with larger amounts of the leaves
led to the isolation of the major products. The 80% acetone
extract of the o-phenylenediamine-treated leaves was
concentrated and partitioned between H2O and ethyl
acetate. The ethyl acetate layer was subjected to a combi-
nation of column chromatography over MCI-gel CHP20P,
Sephadex LH-20, Chromatorex ODS, and Toyopearl HW-
40F to give four compounds (8-11). We failed to isolate
another product 12; however, the products 8 and 12 were
easily obtained, along with theaflavin, by oxidation of a
mixture of 1 and 2 with banana homogenate,6,14 followed
by treatment with o-phenylenediamine. Banana homoge-
In the course of a chemical study on the oxidation of
catechins in plants,6,7,13,14 we have recently succeeded in
* Corresponding author. Tel: +81-95-847-1111. Fax: +81-95-848-4387.
E-mail: ikouno@net.nagasaki-u.ac.jp.
† Nagasaki University.
‡ Fukuoka University.
10.1021/np020245k CCC: $22.00
© 2002 American Chemical Society and American Society of Pharmacognosy
Published on Web 09/13/2002

Epigallocatechin Quinone Dimers
J ournal of Natural Products, 2002, Vol. 65, No. 11 1583
F igu r e 1. Proposed mechanism for formation of theaflavins and quinone dimers.
F igu r e 2. Structures of epigallocatechin oxidation products and phenazine derivatives.
1
nate was used because it efficiently catalyzes the oxidation
of 1 and 2 to theaflavin without production of interfering
minor products.14
proton singlet signals in the H NMR spectra (δ 7.06 for 9,
δ 7.07 for 10, δ 7.12 and 7.06 for 11). Hydrolysis of the
galloyl groups of 9-11 by treatment with tannase yielded
8 and gallic acid, confirming that the products 9-11 were
galloyl esters of 8. In the HMBC spectrum of 10 (Figure
4), two one-proton singlets at δ 8.24 and 7.12 were
attributable to the B-ring protons H-6′ and H-6′′′, respec-
tively, because of their long-range coupling with the C-2
and C-2′′ carbons. The correlations between the H-6′′′, H-2′′,
H-3′′, and galloyl protons indicated the presence of a 3-O-
galloyl epigallocatechin unit similar to that of theasinensin
A (15).16 The 13C NMR chemical shifts of the B-ring (C-
The 1H NMR spectra of 8-11 were related to each other,
showing two sets of signals arising from A- and C-rings of
2,3-cis-flavan-3-ols similar to those of 1-4. In addition, the
[M + H]+ peak in the FABMS (m/z 681 for 8, m/z 833 for
9 and 10, m/z 985 for 11) indicated that each of these
products contains two flavan-3-ol units. The difference of
their molecular weights (152 mass units) coincided with
the mass of a galloyl group. The presence of the galloyl
groups in 9-11 was also shown by the appearance of two-

1584 J ournal of Natural Products, 2002, Vol. 65, No. 11
Tanaka et al.
F igu r e 3. HPLC profiles of fermented tea leaves (UV max. absorbance obtained by photodiodearray detector). (a) Chromatogram of the ethanol
extract of fermented tea leaves. (b) Chromatogram of the ethanol extract of fermented tea leaves treated with o-phenylenediamine. The reagent
was detected as a large peak at the solvent front. (c) Chromatogram of the ethanol extract of fermented tea leaves heated at 90 °C for 10 min.
Hence, 9 and 10 were concluded to be 3-O-gallate and the
3,3′′-di-O-gallate of 8, respectively.
The molecular weight of 12 was the same as that of 8
1
by FABMS, and the H and 13C NMR spectral comparison
indicated that these products were closely related to each
other. However, the chemical shifts of H-6′ (δ 7.81), H-3′′
(δ 3.71), and H-4′′ (δ 2.32 and 1.89) of 12 were observed at
higher field compared to those of 8 [δ 8.21 (H-6′), 4.10 (H-
3′′), 2.53 and 2.11 (H-4′′)]. The HMBC spectrum of 12
(Figure 4) revealed that this product is a positional isomer
2
of 8 by the appearance of a strong J correlation between
H-6′ (δ 7.81) and a hydroxyl-bearing carbon (δ 152.2, C-5′)
and the absence of correlation between H-6′ and C-3′ (δ
144.4). The location of the hydroxyl group on the hydrox-
F igu r e 4. Important HMBC correlations of 10 and 12.
yphenazine moiety was further confirmed by a differential
NOE experiment of the octamethyl ether 12a , which
showed a strong NOE between H-6′ (δ 7.64) and a methoxyl
proton (δ 4.26). The high-field shifts of H-3′′ and H-4′′ of
12 could be explained by the anisotropic effect of the
phenazine ring, which hung over these protons. Therefore,
the structure of 12 was determined to be as shown in
Figure 2.
1′′′-C-6′′′) were also consistent with those of 15. The other
B-ring proton H-6′ was correlated with six aromatic carbon
signals at δ 135.6 (C-4′), 116.8 (C-2′), 143.8 (C-1′), 144.2
(C-5′), 151.7 (C-3′), and 112.4 (C-2′′′) through 2J , J , and
3
4J couplings. Since the presence of a o-phenylenediamine
unit was apparent from the aromatic multiplets at δ 7.94
(2H, m) and 8.25 (2H, m), the aromatic ring of C-1′-C-6′
and the phenylenediamine unit should form a hydroxy-
phenazine moiety.15 The structure of this hydroxyphena-
zine moiety was confirmed as follows: compound 8 was
methylated with CH2N2 to give octamethyl ether 8a , and
the differential NOE experiment of 8a showed no NOE
between H-6′ and any methoxyl protons. Therefore, the
position of the hydroxyl group was concluded to be at C-3′.
On the basis of these spectroscopic findings, the structures
of 8 and 10 were determined to be as shown in Figure 2.
The 1H NMR spectrum of 9 showed the low-field shift of
the H-3 [δ 5.49 (br d, J ) 4.4 Hz)] instead of H-3′′ as in 10.
The location of the galloyl groups in 11 was determined to
be at the C-3 and C-3′′ positions because of the low-field
shifts of H-3 and H-3′′ (δ 5.52 and 5.44, respectively).
The configuration of the biphenyl bond of 8-12 was
shown to be R by comparison of the CD spectrum of 8 and
12 with those of theasinensin C (13), having an R biphenyl
bond, and theasinensin E (16), having an S biphenyl
bond.17,18 The CD spectra of 8 and 12 were very similar to
each other and showed positive Cotton effects (CE) at 244
and 242 nm and negative ones at 211 and 216 nm,
respectively. These Cotton effects were similar to those of
13 (positive CE at 240 nm and negative CE at 219 nm)
and different from those of 16 (negative CE at 235 nm and
positive CE at 227 nm). This result indicated that dimer-
ization of epigallocatechin quinones (2a and 4a ) was highly
stereoselective. It was noteworthy that the hydrated quino-
ne dimer 7, previously obtained from our model fermenta-
tion experiments as a minor metabolite,14 was derived from

Epigallocatechin Quinone Dimers
J ournal of Natural Products, 2002, Vol. 65, No. 11 1585
are expressed in Hz, and chemical shifts are given on a δ (ppm)
scale with TMS as an internal standard. MS were recorded
on a J EOL J MS DX-303 spectrometer, and glycerol or m-
nitrobenzyl alcohol was used as a matrix for FABMS measure-
ment. Column chromatography was performed with MCI-gel
CHP 20P (75-150 µm, Mitsubishi Chemical Co.), Sephadex
LH-20 (25-100 µm, Pharmacia Fine Chemical Co. Ltd.),
Toyopearl HW-40F (Tosoh Corp.), and Chromatorex ODS (Fuji
Silysia Chemical Ltd.). TLC was performed on precoated
Kieselgel 60 F254 plates (0.2 mm thick, Merck) with benzene-
ethyl formate-formic acid (1:7:1, v/v) or CHCl3-MeOH-H2O
(14:6:1, v/v), and spots were detected by UV illumination and
by spraying with 2% ethanolic FeCl3 or 10% sulfuric acid
reagent followed by heating. Compounds 1 and 2 were isolated
from commercial green tea according to Nonaka et al. and
recrystallized from H2O.20 Tannase (a crude enzyme prepared
from Aspergillus sp.) was provided from Sankyo Co., Ltd.
P la n t Ma ter ia l. Fresh tea leaves (Camellia sinensis var.
sinensis) were collected in May at Nagasaki Agricultural and
Forestry Experimental Station, Higashisonogi tea branch.
HP LC An a lysis of P h en a zin e Der iva tives. Tea leaves
(80 g), left withering for 12 h, were crumpled and crushed.
The leaves was spread on a tray in a thin layer at 24 °C and
mixed well at intervals with occasional spraying of H2O to
avoid drying. After 2.5 h, a portion (10 g) of the leaves was
extracted with EtOH (50 mL), a portion (10 g) was treated
with 0.2% o-phenylenediamine solution in 1% AcOH-EtOH
(50 mL), and another portion (10 g) was heated at 90 °C for
10 min in a steamer and extracted with EtOH (50 mL).
Aliquots (2 mL) of extracts were separately passed through a
Sep-pak cartridge (Waters Associates) with 90% EtOH, and
the volume was adjusted to 5.0 mL. Each of the solutions was
analyzed by reversed-phase HPLC performed on a Cosmosil
5C18-AR II (Nacalai Tesque Inc.) column (250 × 4.6 mm i.d.)
with gradient elution from 4% to 30% (39 min) and 30% to
75% (15 min) of CH3CN in 50 mM H3PO4 (flow rate, 0.8 mL/
min; detection, J ASCO photodiode array detector MD-910).
Quantification of the total theaflavins and phenazine deriva-
tives was performed by incorprating the peak area onto the
calibration curve of the theaflavin (at 375 nm) and phenazine
derivative 8 at 370 nm, respectively.
the quinone with opposite configuration, which corre-
sponded to a quinone of theasinensin E (16).
Our results clearly confirmed that the quinone dimer (6)
derived from 2 and 4 is present in fairly high concentra-
tions in tea leaves during fermentation as predicted by
Roberts in 1957.10 In the case of the experiment shown in
Figure 3b, the total concentration of these phenazine
derivatives was estimated to be about 6.4 µmol/g of fresh
leaves, which is higher than that of theaflavins (about 5.4
µmol/g). Although we have previously confirmed the gen-
eration of the monomeric quinones (1a and 2a ) by reaction
with gluthatione, no phenazine derivatives of monomeric
quinones were isolated in this experiment. The reason may
be that the rate of the dimerization reaction was too fast
to trap the monomeric quinones with o-phenylenediamine.
The quinone dimer (6) was also generated from 2 alone on
treatment with banana homogenate.13 However, the reac-
tion was accelerated in the presence of 1, because 1 is
oxidized by enzymes much faster than 2 and the resulting
quinone 1a readily oxidizes 2 to the quinone 2a .9,14 It is
known that theaflavins can be chemically synthesized from
1 and 2 by treatment with potassium ferricyanide.2 Inter-
estingly, addition of o-phenylenediamine to the reaction
mixture yielded 8 together with a small amount of 12.
These results indicated that stereoselective dimerization
of the quinone 2a is a nonenzymatic process.
The HPLC analysis of the fermented tea leaves showed
broad peaks attributable to the Roberts’ quinones (peaks
Q in Figure 3a), which disappeared after treatment with
o-phenylenediamine (Figure 3b). An attempt to isolate the
quinones from the fermented tea leaves by the usual
column chromatography was not successful because the
quinones disappeared during separation, and theasinensins
were isolated instead, which were not detected originally
before the separation. The result suggested that theasin-
ensins were produced by reduction of these unstable
quinone dimers. In the fermented tea leaves, the quinone
dimers were also decomposed on heating (90 °C, 10 min)
to give theasinensins as shown in Figure 3c, and the
phenazine derivatives were no longer generated from the
fermented leaves after heating. This result explained why
the commercial black tea did not produce the phenazine
derivatives on treatment with o-phenylenediamine. The
Roberts’ quinones were probably converted to theasinensins
in the later stage of black tea manufacturing, especially
at the stage of heating and drying, where enzymes were
inactivated. The oxidation products, which should be
formed concomitantly with the reduction of the quinones
to theasinensins, could not be isolated at this stage. A
dismutation between two molecules of the quinone inter-
mediates was possible because the intermediates could
function as suitable hydrogen donors and acceptors.5,8
Reaction with coexisting substances, including other
polyphenols, might occur. Our results suggested that the
mechanism of theasinensin production in fermented tea
leaves was different from that of radical oxidation of
epigallocatechin and its gallate.19
Isola tion . Fresh tea leaves (450 g) were macerated with
H2O (200 mL) using a Waring blender and spread onto trays
in thin layers at 28 °C. After 3 h, the leaves were treated with
o-phenylenediamine (1%) in EtOH-AcOH (9:1, v/v, 500 mL)
for 8 h and then extracted with 80% acetone. The extract was
concentrated and partitioned between H2O and ethyl acetate.
The ethyl acetate extract (30.7 g) was subjected to column
chromatography over MCI-gel CHP20P with H2O containing
increasing proportions of MeOH to give four fractions. The
second fraction, eluted with 30-60% MeOH, was further
separated by Sephadex LH-20 column chromatography with
EtOH into four fractions (2-1-2-4). Fraction 2-2 was applied
to a column of Chromatorex ODS (H2O-MeOH) and then
Toyopearl HW-40F (H2O-MeOH) to give 8 (100.2 mg). Similar
separation of the fraction 2-3 yielded 9 (28.3 mg) and 10 (45.5
mg). Fraction 2-4 was also chromatographed over Chromatorex
ODS followed by Toyopearl HW-40F to give 11 (102.7 mg).
Syn th esis of 8 a n d 12 fr om 1 a n d 2. Fresh banana fruit
(50 g) was homogenized with H2O (100 mL) in a Waring
blender, and the homogenate was filtered through four layers
of gauze. The filtered homogenate (40 mL) was mixed with a
solution of 1 (229 mg) and 2 (245 mg) in H2O (12 mL) and
vigorously stirred for 1.5 h at room temperature. Then, an
EtOH solution (200 mL) containing 0.2% o-phenylenediamine
and 1% AcOH was added, and the resulting mixture was
filtered. HPLC analysis of the filtrate showed that 2 was
absent and the major products were 1, 8, 12, and theaflavin.
The filtrate was concentrated and separated by a combination
of column chromatography over Chromatorex ODS (H2O-
MeOH) and Sephadex LH-20 (80% MeOH) to yield 1 (142 mg),
12 (25.8 mg), 8 (83.5 mg), and theaflavin (132.2 mg).
Exp er im en ta l Section
Gen er a l P r oced u r es. UV spectra were obtained with a
J ASCO V-560 UV/vis spectrophotometer. Optical rotations
were measured with a J ASCO DIP-370 digital polarimeter.
1H and 13C NMR spectra were measured in a mixture of
acetone-d6 and D2O (19:1, v/v) with a Varian Gemini 300 (300
MHz for 1H and 75 MHz for 13C) or a Varian Unity plus 500
1
1
(500 MHz for H and 125 MHz for 13C) spectrometer. The H-
1H COSY, HSQC, and HMBC spectra were recorded at 27 °C
with a Varian Unity plus 500 spectrometer. Coupling constants
Com p ou n d 8: reddish brown amorphous powder; [R]26
D
-6.0° (c 0.8, MeOH); UV (EtOH) λmax (log ꢀ) 275 (4.79), 376

1586 J ournal of Natural Products, 2002, Vol. 65, No. 11
Tanaka et al.
(3.89) nm; CD (EtOH) ∆ꢀ (nm) +11.5 (275), -3.6 (255), +0.82
(244), -32.2 (211); 1H NMR (acetone-d6, 300 MHz) (see Figure
2 for numbering scheme) δ 8.28 (2H, m, H-9′, H-12′), 8.21 (1H,
s, H-6′), 7.96 (2H, m, H-10′, H-11′), 7.04 (1H, s, H-6′′′), 6.01,
6.00, 5.88, 5.83 (each 1H, br s, H-6, H-8, H-6′′, H-8′′), 4.98 (1H,
br s, H-2), 4.55 (1H, br s, H-2"), 4.30 (1H, br s, H-3), 4.10 (1H,
br s, H-3′′), 2.76 (1H, br d, J 4a,4b ) 16.5 Hz, H-4a), 2.53 (1H,
br d, J 4′′a,4′′b ) 16.5 Hz, H-4′′a), 2.42 (1H, dd, J 3,4b ) 4.4 Hz,
2), 76.4 (C-2′′), 68.7 (C-3′′), 64.0 (C-3), 29.5 (C-4), 26.9 (C-4′′);
HRFABMS m/z 833.1815 (M + H)+ (calcd for C43H33N2O16
833.1830).
,
Com p ou n d 11: reddish brown amorphous powder; [R]26
D
-24.3° (c 0.7, MeOH); UV (EtOH) λmax (log ꢀ) 274 (4.84), 376
(3.88) nm; 1H NMR (acetone-d6, 300 MHz) δ 8.41 (1H, s, H-6′),
8.27 (2H, m, H-9′, H-12′), 7.94 (2H, m, H-10′, H-11′), 7.18 (1H,
s, H-6′′′), 7.12, 7.06 (each 2H, s, galloyl-H), 6.07, 6.06, 5.93,
5.91 (each 1H, d, J ) 2.2 Hz, H-6, H-8, H-6′′, H-8′′), 5.52 (1H,
br s, H-3), 5.44 (1H, br s, H-3′′), 5.19 (1H, br s, H-2), 4.80 (1H,
br s, H-2′′), 3.12 (1H, br d, J 4a,4b ) 18.1 Hz, H-4a), 2.80 (1H,
br d, J 4a,4b ) 18.1 Hz, H-4′′a), 2.60 (1H, dd, J 3,4b ) 4.4 Hz, J 4a,4b
) 18.1 Hz, H-4b), 2.35 (1H, dd, J 3′′,4′′b ) 4.4 Hz, J 4′′a,4′′b ) 18.1
Hz, H-4′′b); 13C NMR (acetone-d6, 75.5 MHz) δ 166.5, 166.3
(galloyl C-7), 157.8, 157.5 (2C), 157.2, 157.1, 156.8 (C-5, C-7,
C-8a, C-5′′, C-7′′, C-8a′′), 152.0 (C-3′), 146.5 (C-5′′′), 145.9, 145.8
(galloyl C-3, C-5), 144.5, 142.4 (C-7′, C-8′), 144.2, 143.5, 143.4
(C-1′, C-5′, C-3′′′), 138.8, 138.7 (galloyl C-4), 135.8 (C-4′), 133.7
(C-4′′′), 132.1, 131.4, 130.3, 129.5 (C-9′, C-10′, C-11′, C-12′),
128.5 (C-1′′′), 121.8, 121.2 (galloyl C-1), 119.1 (C-6′), 117.0 (C-
2′), 112.1 (C-2′′′), 109.9, 109.6 (galloyl C-2, C-6), 108.3 (C-6′′′),
98.4 (2C) (C-4a, C-4a′′), 96.7, 96.3, 95.8, 95.6 (C-6, C-8, C-6′′,
C-8′′), 76.5, 76.3 (C-2, C-2′′), 68.7, 67.8 (C-3′′, C-3), 26.8 (2C)
(C-4, C-4′′); HRFABMS m/z 985.1932 (M + H)+ (calcd for
J 4a,4b ) 16.7 Hz, H-4b), 2.11 (1H, dd, J 3′′,4′′b ) 3.8 Hz, J 4′′a,4′′b
)
16.5 Hz, H-4′′b); 13C NMR (acetone-d6, 75.5 MHz) δ 157.6,
157.5, 157.4, 157.2, 157.1, 157.0 (C-5, C-7, C-8a, C-5′′, C-7′′,
C-8a′′), 151.3 (C-3′), 146.0 (C-5′′′), 144.7, 142.0 (C-7′, C-8′),
144.5 (C-3′′′), 144.2 (C-5′), 144.0 (C-1′), 135.5 (C-4′), 133.3 (C-
4′′′), 131.5, 131.1, 130.3, 130.1 (C-9′, C-10′, C-11′, C-12′), 129.2
(C-1′′′), 119.4 (C-6′), 117.3 (C-2′), 112.0 (C-2′′′), 108.1 (C-6′′′),
99.4, 99.1 (C-4a, C-4a′′), 96.2, 96.0, 95.7, 95.5 (C-6, C-8, C-6′′,
C-8′′), 77.7, 77.3 (C-2, C-2′′), 64.9, 63.8 (C-3, C-3′′), 29.5-30.5
(C-4, C-4′′, overlapped with solvent signals); HRFABMS m/z
681.1720 (M + H)+ (calcd for C36H29N2O12, 681.1721).
Meth yla tion of 8. A solution of 8 (40 mg) in EtOH (2 mL)
was treated with CH2N2 in Et2O for 12 h at 0 °C. After
evaporation of the solvent, the residue was separated by silica
gel column chromatography with hexane-acetone (3:2, v/v) to
give octamethyl ether (9.4 mg) as a yellow amorphous pow-
der: [R]25 +45.1° (c 0.5, MeOH); 1H NMR (CDCl3, 300 MHz)
C
50H37N2O20, 985.1939).
D
δ 8.49 (1H, d, J 6′,2 ) 0.8 Hz, H-6′), 8.29 (2H, m, H-9′, 12′), 7.86
(2H, m, H-10′, 11′), 7.31 (1H, s, H-6′′′), 6.17, 6.08, 6.06, 5.95
(each 1H, d, J ) 2 Hz, H-6, H-8, H-6′′, H-8′′), 4.88, 4.52 (each
1H, br s, H-2, H-2′′), 4.30, 3.99, 3.93, 3.91, 3.76, 3.74, 3.69,
3.62 (each 3H, s, OCH3), 2.87, 2.67 (each 1H, br d, J ) 17.0
Hz, H-4a, H-4′′a), 2.48, 2.25 (each 1H, dd, J ) 4.4, 17.0 Hz,
H-4b, H-4′′b); HRFABMS m/z 793.2972 (M + H)+ (calcd for
Hyd r olysis of 9-11. A solution of each compound (3-5 mg)
in H2O (0.5 mL) was treated with tannase at 35 °C for 3 h.
The reaction mixture was concentrated, and the residue was
treated with EtOH. The insoluble material was filtered off,
and the filtrate was analyzed by HPLC. The chromatograms
showed peaks corresponding to gallic acid (8.6 min) and 8 (31.1
min).
C
44H45N2O12, 793.2974).
Com p ou n d 12: reddish brown amorphous powder; [R]26
D
Com p ou n d 9: reddish brown amorphous powder; [R]26
+18.1° (c 0.2, MeOH); CD (EtOH) ∆ꢀ (nm) +14.5 (274), +0.1
(255), +1.80 (242), -22.8 (216); 1H NMR (acetone-d6, 300 MHz)
δ 8.21, 8.13 (each 1H, dd, J ) 7, 2 Hz, H-9′, H-12′), 7.87 (2H,
m, H-10′, H-11′), 7.81 (1H, s, H-6′), 7.08 (1H, s, H-6′′′), 6.01,
5.94, 5.85, 5.81 (each 1H, d, J ) 2.4 Hz, H-6, H-8, H-6′′, H-8′′),
5.11 (1H, br s, H-2), 4.54 (1H, br s, H-2′′), 4.40 (1H, br s, H-3),
3.71 (1H, br s, H-3′′), 2.78 (1H, br d, J 4a,4b ) 17.0 Hz, H-4a),
2.52 (1H, dd, J 3,4b ) 4.7 Hz, J 4a,4b 17.0 Hz, H-4b), 2.32 (1H, br
d, J 4′′a,4′′b ) 17.0 Hz, H-4′′a), 1.89 (1H, dd, J 3′′,4′′b ) 4.1 Hz, J 4′′a,4′′b
) 17.0 Hz, H-4′′b); 13C NMR (acetone-d6, 75.5 MHz) δ 157.6,
157.5, 157.3, 157.2, 157.1, 157.0 (C-5, C-7, C-8a, C-5′′, C-7′′,
C-8a′′), 152.2 (C-5′), 145.7 (C-5′′′), 145.2 (C-3′′′), 144.6 (C-1′),
144.4 (C-3′), 144.2, 141.7 (C-7′, C-8′), 135.9 (C-4′), 133.4 (C-
4′′′), 131.4, 131.3, 130.6, 129.9 (C-9′, C-10′, C-11′, C-12′), 129.3
(C-1′′′), 124.1 (C-2′), 114.8 (C-2′′′), 111.5 (C-6′), 108.5 (C-6′′′),
99.3, 99.2 (C-4a, C-4a′′), 96.4, 96.0, 95.9, 95.8 (C-6, C-8, C-6′′,
C-8′′), 78.0, 77.3 (C-2, C-2′′), 65.1, 65.0 (C-3, C-3′′), 30.2, 29.0
(C-4, C-4′′); HRFABMS m/z 681.1719 (M + H)+ (calcd for
D
-15.4° (c 0.6, MeOH); UV(EtOH) λmax (log ꢀ) 274 (4.83), 376
(3.89) nm; 1H NMR (acetone-d6, 300 MHz) δ 8.38 (1H, s, H-6′),
8.26 (2H, m, H-9′, H-12′), 7.95 (2H, m, H-10′, H-11′), 7.11 (1H,
s, H-6′′′), 7.06 (2H, s, galloyl-H), 6.08, 6.04, 5.89, 5.87 (each
1H, d, J ) 2.2 Hz, H-6, H-8, H-6′′, H-8′′), 5.49 (1H, br d, J 3′′,4′′b
) 4.4 Hz, H-3), 5.20 (1H, br s, H-2), 4.56 (1H, br s, H-2′′), 4.19
(1H, br s, H-3′′), 3.09 (1H, br d, J 4a,4b ) 17.5 Hz, H-4a), 2.57
(1H, br d, J 4a,4b ) 17.5 Hz, H-4′′a), 2.56 (1H, dd, J 3,4b ) 4.4
Hz, J 4a,4b ) 17.5 Hz, H-4b), 2.11 (1H, dd, J 3′′,4′′b ) 4.4 Hz, J 4′′a,4′′b
) 17.5 Hz, H-4′′b); 13C NMR (acetone-d6, 75.5 MHz) δ 166.5
(galloyl C-7), 157.7, 157.5, 157.4, 157.2 (2C), 156.8 (C-5, C-7,
C-8a, C-5′′, C-7′′, C-8a′′), 151.6 (C-3′), 146.3 (C-5′′′), 145.8
(galloyl C-3, C-5), 144.6, 142.4 (C-7′, C-8′), 144.0 (C-5′, C-3′′′),
143.1 (C-1′), 138.8 (galloyl C-4), 135.8 (C-4′), 133.5 (C-4′′′),
131.1, 131.3, 130.3, 129.4 (C-9′, C-10′, C-11′, C-12′), 129.2 (C-
1′′′), 121.2 (galloyl C-1), 118.7 (C-6′), 117.7 (C-2′), 112.0 (C-
2′′′), 109.5 (galloyl C-2, C-6), 108.4 (C-6′′′), 99.1, 98.3 (C-4a,
C-4a′′), 96.6, 96.5, 95.7, 95.6 (C-6, C-8, C-6′′, C-8′′), 77.2, 76.6
(C-2, C-2′′), 67.9, 64.8 (C-3′′, C-3), 28.8, 26.8 (C-4, C-4′′);
C
36H29N2O12, 681.1721).
Meth yla tion of 12. A solution of 12 (15 mg) in EtOH (1
HRFABMS m/z 833.1826 (M + H)+ (calcd for C43H33N2O16
833.1830).
,
mL) was treated with CH2N2 in Et2O for 12 h at 0 °C. After
evaporation of the solvent, the residue was separated by silica
gel column chromatography with hexane-acetone (3:2, v/v) to
give octamethyl ether 12a (5.0 mg) as a yellow amorphous
Com p ou n d 10: reddish brown amorphous powder; [R]26
D
-32.9° (c 0.5, MeOH); UV (EtOH) λmax (log ꢀ) 275 (4.83), 375
(3.89) nm; 1H NMR (acetone-d6, 500 MHz) δ 8.25 (2H, m, H-9′,
H-12′), 8.24 (1H, d, J 2,6′ ) 0.9 Hz, H-6′), 7.94 (2H, m, H-10′,
H-11′), 7.12 (1H, s, H-6′′′), 7.07 (2H, s, galloyl-H), 6.03, 6.01
(each 1H, d, J ) 2.2 Hz, H-6, H-8), 5.93, 5.89 (each 1H, d, J )
2.2 Hz, H-6′′, H-8′′), 5.39 (1H, br d, J 3′′,4′′b ) 4.4 Hz, H-3′′), 4.99
(1H, br s, H-2), 4.80 (1H, br s, H-2′′), 4.35 (1H, br d, J 3,4b ) 4.4
Hz, H-3), 2.80 (1H, br d, J 4a,4b ) 16.7 Hz, H-4a), 2.77 (1H, br
d, J 4′′a,4′′b ) 17.6 Hz, H-4′′a), 2.47 (1H, dd, J 3,4b ) 4.4 Hz, J 4a,4b
) 16.7 Hz, H-4b), 2.29 (1H, dd, J 3′′,4′′b ) 4.4 Hz, J 4′′a,4′′b ) 17.6
Hz, H-4′′b); 13C NMR (acetone-d6, 125.7 MHz) 166.2 (galloyl
C-7), 157.7,157.6, 157.5, 157.3, 157.2, 157.1 (C-5, C-7, C-8a,
C-5′′, C-7′′, C-8a′′), 151.7 (C-3′), 146.3 (C-5′′′), 145.9 (galloyl
C-3, C-5), 145.0, 142.1 (C-7′, C-8′), 144.5 (C-3′′′), 144.2 (C-5′),
143.8 (C-1′), 138.8 (galloyl C-4), 135.6 (C-4′), 133.6 (C-4′′′),
131.6, 131.2, 130.4, 130.2 (C-9′, C-10′, C-11′, C-12′), 128.6 (C-
1′′′), 121.9 (galloyl C-1), 119.9 (C-6′), 116.8 (C-2′), 112.4 (C-
2′′′), 110.0 (galloyl C-2, C-6), 108.1 (C-6′′′), 99.4 (C-4a), 98.6
(C-4a′′), 96.5, 96.4, 96.0, 95.8 (C-6, C-8, C-6′′, C-8′′), 77.7 (C-
powder, [R]25 +12.4° (c 0.3, CHCl3); 1H NMR (CDCl3, 300
D
MHz) δ 8.44, 8.49 (each 1H, dd, J ) 8.2, 1.9 Hz, H-9′, 12′),
7.82 (2H, m, H-10′, 11′), 7.64 (1H, s, H-6′), 7.30 (1H, s, H-6′′′),
6.16, 6.09, 6.08, 5.98 (each 1H, d, J ) 2 Hz, H-6, H-8, H-6′′,
H-8′′), 4.97, 4.62 (each 1H, br s, H-2, H-2′′), 4.36 (1H, br s,
H-3), 4.26, 4.02, 3.93, 3.77, 3.73, 3.70, 3.62, 3.45 (each 3H, s,
OCH3), 2.83 (1H, d, J ) 16.8 Hz, H-4a), 2.55 (1H, dd, J ) 4.4,
16.8 Hz, H-4b), 2.48 (1H, d, J ) 16.8 Hz, H-4′′a), 1.91 (1H, dd,
J ) 4.8, 16.8 Hz, H-4′′b); HRFABMS m/z 793.2976 (M + H)+
(calcd for C44H45N2O12, 793.2974).
F er r icya n id e Oxid a tion of 1 a n d 2. To a solution of 1
and 2 (each 5 mg) in H2O (0.5 mL) was added a solution of
K3[Fe(CN)6] (15 µg) and NaHCO3 (4 µg) in H2O (0.1 mL) and
left to stand at 0 °C for 15 min. The mixture was acidified
with 0.1 M HCl (0.4 mL), and a portion (400 µL) of the mixture
was applied to a MCI-gel CHP 20P short column (8 mm i.d. ×
50 mm). After the column was washed with H2O (10 mL), a
solution of o-phenylenediamine (2 mg) in EtOH-acetic acid

Epigallocatechin Quinone Dimers
J ournal of Natural Products, 2002, Vol. 65, No. 11 1587
(6) Tanaka, T.; Betsumiya, Y.; Mine, C.; Kouno, I. Chem. Commun. 2000,
1365-1366.
(7) Tanaka, T.; Inoue, K.; Betsumiya, Y.; Mine, C.; Kouno, I. J . Agric.
Food Chem. 2001, 49, 5785-5789.
(8) Roberts, E. A. H. J . Sci. Food Agric. 1959, 10, 167-172.
(9) Roberts, E. A. H. Chem. Ind. 1957, 1354-1355.
(10) Roberts, E. A. H. Chem. Ind. 1957, 1355-1356.
(11) Imagawa, H.; Takino, Y. Agric. Biol. Chem. 1962, 26, 541-542.
(12) Takino, Y.; Imagawa, H.; Aoki, Y.; Ozawa, T. Agric. Biol. Chem. 1963,
27, 562-568.
(13) Tanaka, T.; Kondou, K.; Kouno, I. Phytochemistry 2000, 53, 311-
316.
(14) Tanaka, T.; Mine, C.; Inoue, K.; Matsuda, M.; Kouno, I. J . Agric. Food
Chem. 2002, 50, 2142-2148.
(15) Nonaka, G.; Hayashi, M.; Tanaka, T.; Saijo, R.; Nishioka, I. Chem.
Pharm. Bull. 1990, 38, 861-865.
(16) Nonaka, G.; Kawahara, O.; Nishioka, I. Chem. Pharm. Bull. 1983,
31, 3906-3914.
(17) Hashimoto, F.; Nonaka, G.; Nishioka, I. Chem. Pharm. Bull. 1988,
36, 1676-1684.
(18) Hashimoto, F.; Nonaka, G.; Nishioka, I. Chem. Pharm. Bull. 1992,
40, 1383-1389.
(19) Zhu, N.; Wang, M.; Wei, G.-J .; Lin, J .-K.; Yang, C. S.; Ho, C.-T. Food
Chem. 2001, 73, 345-349.
(20) Nonaka, G.; Kawahara, O.; Nishioka, I. Chem. Pharm. Bull. 1983,
31, 3906-3914.
(3:2, v/v, 0.5 mL) was added to the column, followed by elution
with EtOH. The product eluted out with EtOH was collected
and analyzed by HPLC. The chromatogram showed peaks
arising from 2, 1, 12, 8, and theaflavin.
Ack n ow led gm en t. We thank the staff of Nagasaki Agri-
cultural and Forestry Experimental Station for the collection
of tea leaves and valuable information. This work was sup-
ported by a Grant-in-Aid for Scientific Research No. 12680594
from the J apan Society for the Promotion of Science. The
authors are also grateful to Mr. K. Inada and Mr. N. Yamagu-
chi (Nagasaki University) for NMR and MS measurements.
Refer en ces a n d Notes
(1) Haslam, E. Practical Polyphenolics, from Structure to Molecular
Recognition and Physiological Action; Cambridge University Press:
Cambridge, 1998; pp 335-354.
(2) Takino, Y.; Imagawa, H.; Horikawa, H.; Tanaka, A. Agric. Biol. Chem.
1964, 28, 64-71.
(3) Takino, Y.; Ferretti, A.; Flanagan, M.; Gianturco, M.; Vogel, M.
Tetrahedron Lett. 1965, 4019-4025.
(4) Brown, A. G.; Falshaw, C. P.; Haslam, E.; Holmes, A.; Ollis, W. D.
Tetrahedron Lett. 1996, 1193-1204.
(5) Roberts, E. A. H. Economic importance of flavonoid substances: tea
fermentation. The Chemistry of Flavinoid Compounds; Pergammon
Press: Oxford, 1962; pp 468-512.
NP020245K