Reaction of the black tea pigment theaflavin during enzymatic oxidation of tea catechins

Article abstract of DOI:10.1021/np900618v

Degradation of the black tea pigment theaflavin was examined in detail. Enzymatic oxidation of a mixture of epigallocatechin and epicatechin initially produced theaflavin, while prolonged reaction decreased the product. Addition of ethanol to the reaction mixture at the point when theaflavin began to decrease afforded four new products, together with theanaphthoquinone, a known oxidation product of theaflavin. The structures of the new products were determined by spectroscopic methods. One of the products was an ethanol adduct of a theanaphthoquinone precursor, and this reacted with theaflavin to give two further products. A product generated by coupling of theaflavin with epicatechin quinone was also obtained. The structures of the products indicate that oxidation and coupling with quinones are key reactions in the degradation of theaflavins. The degradation of theaflavin probably contributes to production of thearubigins.

Full text of DOI:10.1021/np900618v

J. Nat. Prod. 2010, 73, 33–39
33
Reaction of the Black Tea Pigment Theaflavin during Enzymatic Oxidation of Tea Catechins
Yan Li, Akane Shibahara, Yosuke Matsuo, Takashi Tanaka,* and Isao Kouno
Laboratory of Natural Product Chemistry, Graduate School of Biomedical Sciences, Nagasaki UniVersity,
Bunkyo-Machi 1-14, Nagasaki 852-8521, Japan
ReceiVed October 7, 2009
Degradation of the black tea pigment theaflavin was examined in detail. Enzymatic oxidation of a mixture of
epigallocatechin and epicatechin initially produced theaflavin, while prolonged reaction decreased the product. Addition
of ethanol to the reaction mixture at the point when theaflavin began to decrease afforded four new products, together
with theanaphthoquinone, a known oxidation product of theaflavin. The structures of the new products were determined
by spectroscopic methods. One of the products was an ethanol adduct of a theanaphthoquinone precursor, and this
reacted with theaflavin to give two further products. A product generated by coupling of theaflavin with epicatechin
quinone was also obtained. The structures of the products indicate that oxidation and coupling with quinones are key
reactions in the degradation of theaflavins. The degradation of theaflavin probably contributes to production of thearubigins.
Black tea, which is rich in ployphenols,1 is associated with
antioxidative,2 anticancer,3,4 and anti-inflammatory5,6 activities.
Theaflavin (1) and its galloyl esters7 are characteristic pigments
found only in black tea and contribute to its color and flavor.8,9
The pigments are produced by oxidative condensation between
catechol-type catechins [epicatechin (2) and its galloyl ester] and
pyrogallol-type catechins [epigallocatechin (3) and its galloyl ester]
during black tea production. However, the theaflavin content of
the black tea leaf is usually 0.8-2.8% of the dried leaf depending
on the fermentation conditions.10-13 In addition to theaflavins,
numerous other oxidation products are produced, and a heteroge-
neous mixture of pigments with a reddish-brown color is referred
to as thearubigins. Thearubigins constitute up to 60% of the solids
in black tea infusions10,14 and are therefore very important with
respect to color and taste, as well as biological activities. However,
the definition of thearubigins remains ambiguous, and little is known
about their chemical structures despite substantial efforts by many
research groups.14,15 Previously, we demonstrated that theaflavins
are further oxidized to give products that include dimerization
products.16-19 In addition to the individual oxidation of 2, 3, and
their galloyl esters,19-25 oxidation of theaflavins is probably
involved in the formation of thearubigin components. During the
enzymatic oxidation of tea catechins, a number of unstable products
are produced and the subsequent decomposition of these products
are key reactions in the production of thearubigins.22,24,25 Due to
difficulties in separating the unstable intermediates by conventional
chromatography, they were converted into more stable compounds
in our previous study.24,26 During the in vitro oxidation study
described here, we found that addition of EtOH, which was used
to terminate the enzymatic reaction, affords stable EtOH adducts
derived from theaflavin oxidation products. The structure of the
derivatives indicated the nature of the reaction mechanisms occur-
ring during the initial stage of theaflavin oxidation.
Figure 1. Color change of reaction mixture and HPLC after the
reaction started. The color of the reaction mixture at 0 min was
about 30 s after addition of Japanese pear homogenate. The
chromatogram at 0 min was a control solution, in which water was
added instead of Japanese pear homogenate.
completely consumed, the color had dramatically changed to dark
green. The change in color was probably caused by the formation
of a quinhydrone-type π-π complex between the benzotropolone
moiety of 1 and the ortho-quinone unit (2a) derived from 2.18
Finally, the color of the mixture became brownish-yellow. At this
stage, the concentration of 1 was low and theanaphthoquinone (4)
had accumulated.16 Interestingly, when the color had changed to
dark green, the HPLC chromatogram (Figure 1C) exhibited minor
peaks that were not observed at other reaction stages. To identify
the unknown products, a large-scale experiment was performed and
four new products, 5-8, in addition to five known compounds,
were isolated (Figure 2). The known compounds were identified
as 1, 4, theasinensin C,27 and desgalloyltheasinensins F and G.28
Product 5 was detected after the reaction mixture was treated
with EtOH (at 45.3 min in the HPLC, Figure 1), and this peak was
not observed in the chromatogram obtained by direct analysis of
the same reaction mixture without addition of EtOH. The product
was obtained as a brown, amorphous powder that showed UV
maxima at 257 and 348 nm. The MALDITOFMS spectrum showed
the [M + Na]+ peak at m/z 613. The appearance of 31 signals in
the 13C NMR spectrum and the results of elemental analysis
suggested that the molecular formula is C31H26O12. Among the
carbon signals (Table 1), 18 signals were assigned to the A- and
Results and Discussion
An aqueous solution of 2 and 3 was mixed with a Japanese pear
homogenate18,21 and stirred vigorously at room temperature.
Aliquots of the reaction mixture were mixed with EtOH to terminate
the enzymatic oxidation and then analyzed by HPLC (Figure 1).
The color of the reaction mixture changed from colorless to reddish-
yellow in proportion to the increase in concentration of 1. Due to
the occurrence of an oxidative dimerization reaction of 3,24 the
concentration of 3 decreased faster than that of 2. When 3 was
* To whom correspondence should be addressed. Tel: 81-95-819-2433.
Fax: 81-95-819-2477. E-mail: t-tanaka@nagasaki-u.ac.jp.
10.1021/np900618v  2010 American Chemical Society and American Society of Pharmacognosy
Published on Web 12/16/2009

34 Journal of Natural Products, 2010, Vol. 73, No. 1
Li et al.
Figure 2. Structures of compounds 1-8.
H-2 and H-c was observed in the 1H-1H COSY spectrum,
indicating the presence of the same partial structure as that of 4. In
1
turn, both H-2 and H-c showed long-range H-1H coupling with
an aromatic proton, H-e, and H-e was correlated with the remaining
aromatic proton H-g and with H-2′ of another C-ring. Furthermore,
H-g showed HMBC correlation with an additional carboxylic
carbon, C-k, which also correlated with the methylene protons of
the ethyl group. These spectroscopic observations allowed us to
construct the structure of this product as given in formula 5. The
HMBC correlations illustrated in Figure 3 supported the structure.
This product was formed by addition of EtOH to a carbonyl group
of theaflavin quinone (1a) and subsequent rearrangement (Scheme
1). The intermediate 1a was also presumed to be a precursor of 4.
Isolation of 5 in this experiment evidenced generation of 1a and
4a.
Product 6 showed an [M + H]+ peak at 1153 and an [M + Na]+
peak at m/z 1175, indicating that this product is a tetrameric
catechin. The 1H and 13C NMR spectra (Table 2) showed four sets
of signals due to flavan A- and C-rings. In addition, signals
attributable to two aromatic rings, an ethoxy group, three carbonyl
carbons, one carboxylic carbon, one double bond, three aliphatic
quaternary carbons, and an aliphatic methine were observed. In the
HMBC spectrum (Figure 4), correlations of the C-ring H-2′′′ and
aromatic proton TF-g-H indicated a catechol-type aromatic ring
(TF-f-TF-k). A phenolic hydroxy proton resonating at δ 11.79 was
correlated with aromatic carbons TF-h, i, and j; thus, this proton
was assigned as TF-i-OH, and the chemical shift indicated hydrogen
bonding with a carbonyl group (TF-a) attached to TF-j. The TF-j
Figure 3. Selected HMBC and 1H-1H COSY correlations for
compound 5.
C-rings of the flavan-3-ol skeleton and two (δ 62.2 and 14.2) were
attributed to the ethyl group on the basis of HSQC and HMBC
experiments. The remaining 11 signals were related to the theanaph-
thoquinone moiety of 4, except for an additional carboxylic carbon
at δ 169.4 (C-k). In the HMBC spectrum, two conjugated carbonyl
carbons [δ 179.7 (C-a), 180.9 (C-b)] were correlated with an
aromatic proton at δ 6.85 (H-c), which also correlated with C-2
(C-ring) (δ 75.5). In addition, long-range allylic coupling between

Oxidation of TheaflaVin
Journal of Natural Products, 2010, Vol. 73, No. 1 35
Table 1. 1H and 13C NMR Data for Theanaphthoquinone (4)
and Compound 5 (in acetone-d6)
aromatic protons TNQ-e and TNQ-g showed HMBC correlations
to C-2′. In addition, correlation peaks of the ester carbonyl carbon
(δ 170.3), the methylene protons of the ethoxy group, and the
aromatic proton at TNQ-g were similar to those observed for 5.
The location of the remaining two conjugated carbonyl carbons at
δ 187.8 (TNQ-a) and 198.0 (TNQ-b) was concluded to be between
TNQ-i and TNQ-c, because this TNQ unit was suggested to
originate from compound 5. From these spectroscopic observations
and the molecular weight evidenced by HRFABMS, the molecular
formula of 6 was deduced to be C60H48O24, which indicates that
the degree of unsaturation was 37. This implies that an ether ring
was formed between the oxygen-bearing quaternary carbon TNQ-c
and the C-ring C-3 (δ 70.1) hydroxy group. C-3 (δ 70.1) resonated
at a much lower field than the other C-ring C-3 carbons (∆δ >+4.0).
In addition, an accompanying upfield shift of C-4 (δ 24.3) also
supported the conclusion of ether ring formation.
theanaphthoquinone (4)
5
position
1H
13C
1H
13C
2
3
4
5.38, br s
4.39, br d (2.7)
2.96, dd (4.6, 16.5)
2.82, br d (16.5)
75.4
64.8
28.9
5.48, s
4.46, m
2.98, br d (16.5)
2.84, br d (16.5)
75.5
64.1
28.7
4a
5
6
7
8
8a
2′
3′
4′
99.7
157.7aa,b
96.7b
157.6a
95.6b
156.2
78.9
99.7
157.8a
96.7b
157.6a
95.7b
156.1
78.8
6.06, d, (2.3)
6.00, d (2.3)
6.07, d (2.3)
6.07, d (2.3)
5.16, br s
4.43, m
2.71, dd (3.7, 16.6)
2.89, dd (4.6, 16.6)
5.26, s
4.43, m
2.92, dd (4.6, 16.7)
2.75, dd (3.2, 16.7)
66.3
28.8
66.3
28.8
The stereochemistry was determined by a NOESY experiment;
the methine proton TF-e showed a NOE with H-2, H-2′′′, and TNQ-
e. H-2′′′ was correlated with H-2′′. Furthermore, the C-ring H-3′
exhibited a correlation peak with A-ring H-8 and H-8′′. These
NOESY correlations and molecular modeling indicated that the
configuration at TF-e and TNQ-d could be concluded to be as shown
in formula 6. The possible mechanism of formation 6 is illustrated
in Scheme 2, in which initial attack of the benzotropolone of 1 to
the enone of 5 is followed by ring formation between TNQ-c and
TF-b (route A) and subsequent Michael addition of C-3-OH to
TNQ-c.
4a′
99.4
157.5a
96.7b
157.6a
95.6b
156.1
180.6
183.4
126.9
152.2
118.1
152.9
120.1
166.9
115.2
133.9
99.3
157.6a
96.7b
157.6a
95.7b
156.1
179.7
180.9
127.5
152.6
125.8
148.9
128.1
138.3
128.6
134.5
169.4
62.2
5′
6′
6.06, d (2.3)
6.00, d (2.3)
6.05, s
6.03, s
7′
8′
8a′
a
b
c
d
e
f
g
h
6.79, d (1.0)
7.33, d (1.0)
7.44, d (1.0)
6.85, d, 1.0
7.98, d, 1.0
7.72, d, 1.0
1
The H and 13C NMR spectra of compound 7 resembled those
of 6 (Table 2), and the FABMS data exhibited an [M + H]+ peak
at m/z 1155, indicating that the molecular weight of 7 was 2 mass
units larger than that of 6. The HMBC correlations of 7 (Figure 4)
revealed the presence of partial structures similar to those of 6.
C-2′ of one of the four sets of flavan-A- and C-rings is attached to
an aromatic ring bearing an ethoxycarbonyl group of the TNQ
subunit. C-2 of a second A/C-ring showed a HMBC correlation
with a quaternary carbon (TNQ-d), which further correlated with
the aromatic proton of TNQ-e and an aliphatic methine proton of
TF-e. Chemical shifts of C-3 and C-4 of this C-ring resonated at
higher and lower fields, respectively, compared with those of 6,
indicating the absence of the additional etherocyclic linkage in 6.
C-2′′ and C-2′′′ of the remaining two sets of A/C-rings are
connected to C-d and C-f of the TF subunit. The connection between
the TNQ and TF subunits was shown to be similar to that of 6 by
observation of similar HMBC correlations of H-e in the TF subunit.
One of the remarkable differences between 7 and 6 was the presence
of an enolic group, which showed correlation peaks with the
i
j
k
OCH2
CH3
4.40, 2H, m
1.35, 3H, t, 7.1
14.2
a,b Assignments may be interchanged in each column.
and TF-k carbons showed correlations with the methine proton TF-e
(δ 6.06), and, in turn, the TF-e proton correlated to double-bond
carbons (TF-c and TF-d) and an oxygen-bearing quaternary carbon
(TF-b). The olefinic proton TF-c was correlated to the C-ring C-2′′.
1H-1H long-range coupling between the olefinic protons TF-c and
1
H-2′′ was also observed in the H-1H COSY spectrum. Further-
more, a hydroxy group (δ 6.44), assignable to TF-b-OH, showed
correlations to TF-a and TF-b. These observations suggest the
presence of a theaflavin unit in which TF-b and TF-e carbons are
connected to the remaining structural component. The TF-b-OH
also showed an HMBC cross-peak with an oxygen-bearing qua-
ternary carbon TNQ-c (δ 98.1), and a TF-e methine proton was
correlated to the TNQ-d carbon. The TNQ-d carbon also showed
a correlation peak with an aromatic proton, TNQ-e. The aromatic
ring (TNQ-e-TNQ-j) was attached to the C-ring C-2′, because the
3
4
methine proton of TF-e and H-2 through J (C-c) and J (C-b)
correlations. This enolic group was assigned to C-b and C-c of the
TNQ subunit. Another difference was observed in the TF subunit:
the 13C NMR chemical shift of TF-i of 7 (δ 144.7) was observed
Scheme 1. Proposed Mechanism for Formation of 4 and 5 from 1

36 Journal of Natural Products, 2010, Vol. 73, No. 1
Li et al.
1
Table 2. 13C and H NMR Data for 6 and 7 (in acetone-d6)
assigned to C-b in the TF unit and C-a was connected to C-c of
the TNQ subunit. Compound 7 gradually decomposed during NMR
experiments to give a mixture of decomposition products. In the
spectra of the mixture, a set of 1H NMR peaks coincided with those
of compound 6. The mechanism of formation of 6 from 1 and 7
was deduced to be route B in Scheme 2. The compound 7 is related
to bistheaflavin B (9), previously reported as an oxidation product
of 1.17
6
7
atom no.
13C
1H
TNQ unit
13C
1H
2
3
4
76.0
70.1
24.3
4.94, d (2.5)
4.47, m
2.79, dd (1.9, 17.4)
2.60, dd (5.0, 17.4)
81.4
62.8
31.5
3.99, br s
3.99, br s
2.37, dd (4.2, 16.7)
2.46, br d (16.7)
1
In contrast to the above three products, the H and 13C NMR
4a
5
6
7
8
8a
2′
3′
4′
88.0
157.9a
96.5
99.0
157.3a
95.6
spectra of 8 (Table 3) did not show signals of the ethyl group; thus
8 was not an ethanol adduct of reaction products. This product was
detected by HPLC when 1 began to oxidize (Figure 1). The
MALDITOFMS of 8 showed an [M + Na]+ peak at m/z 875,
indicating this product is a flavan-3-ol trimer. The HMBC spectrum
revealed the presence of a TF subunit similar to that of 6 (Figure
5). The chemical shift (δ 12.91) of the phenolic hydroxy proton
connected to TF-i showed hydrogen bonding with the TF-a carbonyl
group. In addition, the HMBC correlations of the C-ring H-2′′′ and
the aromatic proton H-6′′′ indicate the presence of an epicatechin
subunit. The connection between these subunits was obvious from
the HMBC correlation of H-e to C-1′′′, C-2′′′, and C-3′′′, and H-c
to C-3′′′ (Figure 5). The NOESY spectrum showed correlations
between H-e and all three C-ring H-2 protons. However, this
observation is not sufficient to determine the configuration at C-e
and C-b. Thus, the structure of 8 was tentatively presumed to have
the same configuration as that of 6. A possible mechanism to explain
the genesis of 8 is illustrated in Scheme 3. Compound 8 repre-
sents the condensation product of 1 and epicatechin quinone (2a).
The results indicate that oxidation and coupling with quinones
are key reactions in the degradation of theaflavins. Our previous
study suggested that epicatechin quinone (2a) was first generated
in the reaction mixture because of higher enzyme specificity to 2.
Subsequently, 2a reacts with 3 to yield 1.10,18 The quinone 2a acts
as an oxidant and first oxidizes 3 due to the low redox potential of
3. The resulting quinone of 3 was used for producing theasin-
ensins;20 therefore, the concentration of 3 decreased faster than 2.
The oxidation of 1 to theanaphthoquinone (4) by 2a began when 3
was depleted (Scheme 1), and this reaction is the major route for
degradation of 1. The formation of 5 supported this reaction
mechanism. In addition to oxidation, our results demonstrated that
addition of the benzotropolone ring to the quinones occurs during
decomposition of 1 (Schemes 2 and 3). The final product 4 of this
experiment has not previously been found in black tea. Probably,
4 undergoes further coupling reactions with coexisting compounds.
It is evidenced by the production of 9 and related compounds, 6
and 7, isolated in this study. The major part of the black tea
polyphenols, including thearubigins, still remained to be identified.14
The results of the present study indicate the reaction mechanism
for theaflavin decomposition, which probably participates in the
production of the unknown products.
5.80, d (2.3)
5.70, d (2.3)
5.97, d (2.3)
5.97, d (2.3)
157.7a
96.1
157.6a
96.1
155.0a
78.4
65.9
156.0a
78.6
66.6
5.15, br s
4.82, br s
2.89, 2H, br s
5.03, br s
4.39, br s
2.56, dd (3.6, 16.3)
2.84, m
29.2
28.0
4a′
99.8
157.6a
96.6b
157.3a
96.0b
155.0a
187.8
198.0
98.1
99.5
157.4a
96.4b
157.5a
96.5b
156.3a
180.3
143.0
125.1
50.9
129.2
144.5
125.9
135.1
128.5
149.5
170.4
61.8
5′
6′
6.04, d (2.3)c
5.91, d (2.3)c
5.91, d (2.3)c
5.93, d (2.3)c
7′
8′
8a′
TNQ-a
TNQ-b
TNQ-c
TNQ-d
TNQ-e
TNQ-f
TNQ-g
TNQ-h
TNQ-i
TNQ-j
COO
CH2
55.2
129.0
145.5
125.8
136.9
126.8
141.7
170.3
61.6
8.24, s
7.62, s
8.56, s
7.53, s
4.22, 2H, m
4.30, 2H, q (7.1)
1.37, 3H, t (7.1)
CH3
14.2
1.24, 3H, t (7.1)
TF unit
14.3
2′′
3′′
4′′
76.0
65.6
28.0
4.80, br s
4.49, br s
2.74, dd (4.4, 16.0)
2.27, dd (6.2, 16.0)
80.4
62.7
29.2
4.08, br s
3.77, br s
2.65, 2H, m
4a′′
5′′
99.2
157.6a
95.0b
157.3a
96.6b
156.3a
76.0
99.4
157.4a
96.8b
157.5a
95.6b
156.7a
75.4
6′′
5.67, d (2.3)
5.88, d (2.3)
6.09, d (2.3)
6.03, d (2.3)
7′′
8′′
8a′′
2′′′
3′′′
4′′′
5.63, br s
4.88, br s
3.08, dd (4.1, 16.7)
2.97, br d (16.7)
5.89, br s
4.38, br s
2.89, m
2.90
63.3
31.0
65.7
31.0
4a′′′
100.5
157.3a
97.4
100.1
157.5a
95.6
5′′′
6′′′
6.25, d (2.3)
6.42, d (2.3)
6.03, d, (2.3)
6.08, d, (2.3)
7′′′
157.1a
96.8
157.6a
97.0
8′′′
8a′′′
156.4a
186.3
65.7
123.6
165.8
41.5
129.9
125.2
146.8
153.2
116.2
132.6
157.1a
86.3
189.4
126.9
159.3
44.7
129.8
116.4
145.5
144.7
113.9
127.4
TF-a
TF-b
TF-c
TF-d
TF-e
TF-f
TF-g
TF-h
TF-i
TF-j
TF-k
TF-b-OH
TF-i-OH
Experimental Section
General Experimental Procedures. IR and UV spectra were
obtained using JASCO FT/IR-410 and JASCO V-560 spectropho-
tometers, and optical rotations were measured using a JASCO DIP-
370 digital polarimeter (JASCO Co., Tokyo, Japan). Elemental
analysis was conducted with a Perkin-Elmer 2400R analyzer (Perkin-
Elmer Inc., Waltham, MA). 1H and 13C NMR, 1H-1H COSY,
NOESY, HSQC, and HMBC spectra were recorded in a mixture of
acetone-d6 and D2O, using a Varian Unity plus 500 spectrometer
5.75, d (1.6)
6.06, br s
7.86, s
6.21, br s
6.03, br s
7.19, s
1
operating at 500 MHz for H and 125 MHz for 13C (Varian, Palo
6.44, s
11.79, s
Alto, CA). Coupling constants are expressed in Hz, and chemical
shifts are given on a δ (ppm) scale. HMQC, HMBC, and NOESY
experiments were performed using standard Varian pulse sequences.
FAB and HR-FABMS data were recorded on a JEOL JMS-700N
spectrometer (JEOL Ltd., Tokyo, Japan) with glycerol or m-
nitrobenzyl alcohol matrixes. MALDITOFMS data were recorded
on a Voyager-DE Pro spectrometer (Applied Biosystems, Foster
City, CA), and 2,5-dihydrohydroxybenzoic acid (10 mg/mL in 50%
a-c Assignments may be interchanged in each column.
at a higher field than that of 6 (δ 153.2). This suggests that this
aromatic ring was not conjugated with a carbonyl group, and thus,
the oxygen-bearing quaternary carbon at δ 86.3 was attributed to
C-a. Accordingly, a conjugated carbonyl carbon at δ 189.4 was

Oxidation of TheaflaVin
Journal of Natural Products, 2010, Vol. 73, No. 1 37
Figure 4. Selected HMBC correlations for 6 and 7.
Scheme 2. Proposed Mechanism for Formation of 6 and 7
1
Table 3. 13C and H NMR Data for 8 (in acetone-d6)
TLC was performed on 0.2 mm thick precoated Kieselgel 60 F254
plates (Merck, Darmstadt, Germany) using toluene-ethyl formate-
formic acid (1:7:1, v/v). Spots were detected by UV illumination,
sprayed with 2% methanolic FeCl3 or 10% H2SO4 reagent, and then
heated. Analytical reversed-phase HPLC was performed on a
Cosmosil 5C18-AR II column (Nacalai Tesque Inc., Tokyo, Japan;
4.6 mm i.d. × 250 mm) using an elution gradient of 4-30% (39
min) and 30-75% (15 min) CH3CN in 50 mM H3PO4 (flow rate
0.8 mL/min; detection uisng a JASCO MD-910 photodiode array
detector). Preparative HPLC was performed on a Cosmosil 5C18-
AR-II column (Nacalai Tesque Inc.; 10 mm i.d. × 250 mm) using
a linear elution gradient of 20-70% CH3CN in 0.5% TFA. Japanese
pear was purchased at a local market, and epicatechin (2) and
position 13C
1H
75.8 5.38, br s
64.8 3.98, br s
position 13C
1H
2
3
4
8
8a
96.1c 6.01, d, (1.6)
155.8
29.9 2.61, dd (4.0, 16.7) 2′
75.0 5.68, br s
67.0 4.27, br s
29.3 2.82, br d (16.5)
3.06, dd (4.4, 16.5)
99.1
2.67, dd (16.7)
3′
4′
4a
5
6
7
8
8a
1′
2′
3′
4′
5′
6′
2′′
3′′
4′′
99.7a
157.3b
96.1c 5.87, d (1.6)
157.0b
4a′
5′
157.6b
95.3c 5.86, d (1.6)
6′
96.0c 6.00, d (1.6)
157.5b
155.0b
128.1
130.0
7′
8′
95.3c 5.94, d (1.6)
8a′
156.9b
121.7
a
b
c
d
e
f
g
h
197.8
85.2
132.8 6.43, s
146.8
41.4 5.80, br s
127.8
143.7d
144.4d
116.2 7.19, s
77.2 4.94, br s
65.1 4.60, br s
28.3 2.43, dd (6.5, 15.8)
2.79, dd (4.7, 15.8)
99.8a
123.3 7.51, d (1.0)
145.6
4a′′
5′′
i
j
k
155.4
113.3
134.0
157.8b
6′′
96.4c 6.02, d (1.6)
157.7b
7′′
C-i-OH
12.91, s
a-d Assignments may be interchanged.
acetone containing 0.05% TFA) was used as the matrix. Column
chromatography was performed using Sephadex LH-20 (25-100
µm, GE Healthcare Bio-Sciences AB, Uppsala, Sweden), Diaion
HP20SS (Mitsubishi Chemical, Tokyo, Japan), Bondapak C18 125A
(37-35 µm, Waters Co., Milford, MA), and Chromatorex ODS
(100-200 mesh; Fuji Silysia Chemical, Kasugai, Japan) columns.
Figure 5. Selected HMBC correlations for 8.

38 Journal of Natural Products, 2010, Vol. 73, No. 1
Scheme 3. Proposed Mechanism for Formation of 8
Li et al.
epigallocatechin (3) were extracted and separated from commercial
green tea and purified by crystallization from H2O.
(5.32), 272 (4.07); IR νmax cm-1 3353, 1631, 1607, 1513, 1506; anal.
C 59.85%, H 4.98%, calcd for C44H36O18 ·2H2O, C 59.46%, H 4.54%;
1H NMR (500 MHz, acetone-d6) and 13C NMR (125 MHz, acetone-d6)
data, see Table 3.
Enzymatic Oxidation of 2 and 3. Japanese pear (100 g) was
homogenized in H2O (100 mL) and filtered through four layers of
gauze. The homogenate (20 mL) was mixed with an aqueous solution
(5 mL) of 2 (50 mg) and 3 (50 mg) and vigorously stirred at room
temperature. Aliquots (1.0 mL) of the mixture were taken at 15,
30, and 60 min after addition of the homogenate, mixed with EtOH
(3 mL), and analyzed by HPLC. A solution of 2 (10 mg) and 3 (10
mg) in 75% EtOH (20 mL) was used for the control solution.
Large-Scale Enzymatic Oxidation of 2 and 3. Japanese pear (500
g) was homogenized in H2O (500 mL) and filtered through four
layers of gauze. The homogenate (1000 mL) was mixed with an
aqueous solution (300 mL) of 2 (2.0 g) and 3 (2.0 g) and vigorously
stirred for 30 min at room temperature. The mixture was poured
into EtOH (3 L) and gently stirred for 30 min, and insoluble material
was removed by filtration. The filtrate was concentrated by evapora-
tion until the EtOH was completely removed. The resulting aqueous
solution was subjected to Diaion HP20SS column chromatography
(4.0 cm i.d. × 26 cm) using H2O containing an increasing proportion
of MeOH (10% stepwise elution from 0% to 100%, each 300 mL).
The eluate was monitored by TLC and separated into 11 fractions.
The first fraction mainly contained sugars and was not examined
further. Fractions 2 (103 mg), 3 (59 mg), and 4 (173 mg) were
separately subjected to Sephadex LH-20 column chromatography
(40-60% MeOH) to give theasinensin C (45 mg from Fr. 2) and
desgalloyl theasinensins F (31 mg, from Fr. 3) and G (80 mg, from
Fr. 4). Fraction 7 (722 mg) was successively separated by Sephadex
LH-20 (60% MeOH) and preparative HPLC using Cosmosil 5C18
AR II to yield product 8 (11.0 mg). Fraction 8 (358 mg) was applied
to a column of Chromatorex ODS (30-100% MeOH) and then
Bondapak C18 (20-60% MeOH) to yield products 6 (62 mg) and 7
(48 mg). Fraction 10 (1.6 g) was subjected to Sephadex LH-20
column chromatography (80-100% MeOH) to afford 4 (448 mg)
and 1 (953 mg). Fraction 11 (571 mg) was separated by Sephadex
LH-20 column chromatography (80-100% MeOH) to yield product
5 (206 mg).
Acknowledgment. The authors are grateful to Mr. K. Inada and
Mr. N. Yamaguchi for NMR and MS measurements. This work was
supported by a Grant-in-aid for Scientific Research No. 18510189 from
the Japan Society for the Promotion of Science.
1
Supporting Information Available: H and 13C NMR spectra for
products 5-8; 1H-1H COSY, HSQC, and HMBC spectra of 5, 6, and
8; and NOESY spectrum of 6. This material is available free of charge
via the Internet at http://pubs.acs.org.
References and Notes
(1) Shahidi, F.; Naczk M. Phenolics in Food and Nutraceuticals; CRC
Press: Boca Raton, 2004; pp 241-312.
(2) Luczaj, W.; Skrzydlewska, E. PreV. Med. 2005, 40, 910–918.
(3) Way, T.-Z.; Lee, H.-H.; Kao, M.-C.; Lin, J.-K. Eur. J. Cancer 2004,
40, 2165–2174.
(4) Krishnan, R.; Maru, G. B. J. Agric. Food Chem. 2004, 52, 4261–
4269.
(5) Maity, S.; Ukil, A.; Karmakar, A.; Datta, N.; Chaudhuri, T.; Veda-
siromoni, J. R.; Ganguly, D. K.; Das, P. K. Eur. J. Pharmacol. 2003,
470, 103–112.
(6) Sang, S.; Lambert, J. D.; Tian, S.; Hong, J.; Hou, Z.; Ryu, J.-H.; Stark,
R. E.; Rosen, R. T.; Huang, M.-T.; Yang, C. S.; Ho, C.-T. Bioorg.
Med. Chem. 2004, 12, 459–467.
(7) Takino, Y.; Imagawa, H.; Horikawa, H.; Tanaka, A. Agric. Biol. Chem.
1964, 28, 64–71.
(8) Hashimoto, F.; Nonaka, G.; Nishioka, I. Chem. Pharm. Bull. 1992,
40, 1383–1389.
(9) Wang, D.; Kurasawa, E.; Yamaguchi, Y.; Kubota, K.; Kobayashi, A.
J. Agric. Food Chem. 2001, 49, 1900–1903.
(10) Haslam, E. Practical Polyphenolics, From Structure to Molecular
Recognition and Physiological Action; Cambridge University Press,
1998; pp 335-373.
(11) Hall, M. N.; Robertson, A.; Scotter, C. N. G. Food Chem. 1988, 80,
283–290.
(12) Sakakibara, H.; Honda, Y.; Nakagawa, S.; Ashida, H.; Kanazawa, K.
J. Agric. Food. Chem. 2003, 51, 571–581.
(13) Liang, Y.; Lu, J.; Zhang, L.; Wu, S.; Wu, Y. Food Chem. 2003, 80,
283–290.
(14) Haslam, E. Phytochemistry 2003, 64, 61–73.
(15) Ozawa, T.; Kataoka, M.; Morikawa, K.; Negishi, O. Biosci. Biotech.
Biochem. 1996, 60, 2023–2027.
(16) Tanaka, T.; Betsumiya, Y.; Mine, C.; Kouno, I. Chem. Commun. 2000,
1365–1366.
(17) Tanaka, T.; Inoue, K.; Betsumiya, Y.; Mine, C.; Kouno, I. J. Agric.
Food Chem. 2001, 49, 5785–5789.
(18) Tanaka, T.; Mine, C.; Inoue, K.; Matsuda, M.; Kouno, I. J. Agric.
Food Chem. 2002, 50, 2142–2148.
Product 5: brown, amorphous powder; [R]29 -276 (c 0.06,
D
MeOH); MALDITOFMS m/z 613 [M + Na]+; UV (EtOH) λmax nm
(log ε) 257 (4.34), 348 (4.86); IR νmax cm-1 3389, 1694, 1613, 1516,
1467; anal. C 60.53%, H 4.80%, calcd for C31H26O12 · 1.5H2O, C
1
60.29%, H 4.73%; H NMR (500 MHz, acetone-d6) and 13C NMR
(125 MHz, acetone-d6) data, see Table 1.
Product 6: brown, amorphous powder; [R]29D +295 (c 0.06, MeOH);
FABMS m/z 1153 [M + H]+, 1175 [M + Na]+; UV (EtOH) λmax (log
ε) nm 270 (4.23), 377 (3.67); IR νmax cm-1 3389, 1697, 1632, 1517,
1467; HR-FABMS m/z 1153.2726 [M + H]+ (calcd for C60H48O24;
1153.2712); 1H NMR (500 MHz, acetone-d6) and 13C NMR (125 MHz,
acetone-d6) data, see Table 2.
Product 7: brown, amorphous powder; [R]18 -36 (c 0.05,
D
MeOH); FABMS m/z 1155 [M + H]+; UV (EtOH) λmax nm (log ε)
272 (4.07), 378 (4.61); IR νmax cm-1 3372, 1710, 1675, 1666; anal.
C 58.82%, H 4.84%, calcd for C60H50O24 · 4H2O, C 58.73%, H 4.76%;
1H NMR (500 MHz, acetone-d6) and 13C NMR (125 MHz, acetone-
d6) data, see Table 6.
(19) Tanaka, T.; Mine, C.; Kouno, I. Tetrahedron 2002, 58, 8851–8856.
(20) Li, Y.; Tanaka, T.; Kouno, I. Phytochemistry 2007, 68, 1081–1088.
(21) Kusano, R.; Tanaka, T.; Matsuo, Y.; Kouno, I. Chem. Pharm. Bull.
2007, 55, 1768–1772.
(22) Tanaka, T.; Watarumi, S.; Matsuo, Y.; Kamei, M.; Kouno, I.
Tetrahedron 2003, 59, 7939–7947.
Product 8: brown, amorphous powder; [R]18D +50 (c 0.05, MeOH);
MALDITOFMS m/z 875 [M + Na]+; UV (EtOH) λmax nm (log ε) 212
(23) Tanaka, T.; Matsuo, Y.; Kouno, I. J. Agric. Food Chem. 2005, 53,
7571–7578.

Oxidation of TheaflaVin
Journal of Natural Products, 2010, Vol. 73, No. 1 39
(24) Matsuo, Y.; Tanaka, T.; Kouno, I. Tetrahedron 2006, 62, 4774-
4783.
(25) Matsuo, Y.; Tanaka, T.; Kouno, I. Tetrahedron Lett. 2009, 50, 1348–
1351.
(26) Tanaka, T.; Mine, C.; Watarumi, S.; Fujioka, T.; Mihashi, K.; Zhang,
Y.-J.; Kouno, I. J. Nat. Prod. 2002, 65, 1582–1587.
(27) Nonaka, G.; Kawahara, O.; Nishioka, I. Chem. Pharm. Bull. 1983,
31, 3906–3914.
(28) Hashimoto, F.; Nonaka, G.; Nishioka, I. Chem. Pharm. Bull. 1988,
36, 1676–1684.
NP900618V