Structure of polymeric polyphenols of cinnamon bark deduced from condensation products of cinnamaldehyde with catechin and procyanidins

Article abstract of DOI:10.1021/jf800921r

Structures of two condensation products obtained by the reaction of cinnamaldehyde with (+)-catechin were determined by spectroscopic methods. One had two phenylpropanoid units at the C-6 and C-8 positions of the catechin skeleton. The other product had a dimeric structure with two catechin and two phenylpropanoid units. Matrix-assisted laser desorption time-of-flight mass spectrometric analysis of the reaction products of cinnamaldehyde with procyanidin B1 suggested that procyanidins were oligomerized in a manner similar to the reaction with catechin. Furthermore, ¹³C NMR spectral comparison of the condensation products with the polymeric procyanidins obtained from commercial cinnamon bark strongly suggested that the procyanidins in the cinnamon bark also were polymerized by reaction with cinnamaldehyde.

Full text of DOI:10.1021/jf800921r

5864 J. Agric. Food Chem. 2008, 56, 5864–5870
Structure of Polymeric Polyphenols of Cinnamon
Bark Deduced from Condensation Products of
Cinnamaldehyde with Catechin and Procyanidins
TAKASHI TANAKA,* YOSUKE MATSUO, YUKO YAMADA, AND ISAO KOUNO
Graduate School of Biomedical Sciences, Department of Molecular Medicinal Sciences, Nagasaki
University, 1-14 Bunkyo-machi, Nagasaki 852-8521, Japan
Structures of two condensation products obtained by the reaction of cinnamaldehyde with (+)-catechin
were determined by spectroscopic methods. One had two phenylpropanoid units at the C-6 and C-8
positions of the catechin skeleton. The other product had a dimeric structure with two catechin and
two phenylpropanoid units. Matrix-assisted laser desorption time-of-flight mass spectrometric analysis
of the reaction products of cinnamaldehyde with procyanidin B1 suggested that procyanidins were
oligomerized in a manner similar to the reaction with catechin. Furthermore, 13C NMR spectral
comparison of the condensation products with the polymeric procyanidins obtained from commercial
cinnamon bark strongly suggested that the procyanidins in the cinnamon bark also were polymerized
by reaction with cinnamaldehyde.
KEYWORDS: Catechin; cinnamaldehyde; cinnamon bark; polyphenol; procyanidin
INTRODUCTION
found that a similar color change also was observed when a
mixture of (+)-catechin (1) and cinnamaldehyde (2) was heated
(100 °C) or irradiated with sun light. Even at room temperature,
the color changed to red very slowly to give a complex mixture
of products after 6 days. This study reports the chemical reaction
of cinnamaldehyde with catechin and procyanidin B1. Further-
more, incorporation of cinnamaldehyde to polymeric procya-
nidins of commercial cinnamon bark was deduced by 13C NMR
spectroscopic comparison with the in vitro reaction products.
Cinnamon bark [Cinnamomum zeylanicum, Cinnamomum
cassia (Chinese cinnamon), Cinnamomum burmannii (Indone-
sian cinnamon), and Cinnamomum sieboldii (Japanese cin-
namon)] is an important spice used worldwide as well as one
of the most essential crude drugs in Oriental medicine. Cinna-
maldehyde (2) is a major component of the essential oil that
characterizes its flavor and aroma and shows various biological
activities (1–8). In addition, cinnamon bark contains catechins
and procyanidins, which have an astringent taste and antioxi-
dative activities, and the structures of the procyanidins have
been reported to have procyanidin B- and A-type linkages (9–13).
The presence of polymeric polyphenols also was shown by
normal-phase thin layer chromatography (TLC) and reversed-
phase high-performance liquid chromatography (HPLC). The
polymers remained at the TLC origin and were detected as a
broad hump on the baseline by HPLC analysis. By analogy to
the structures of procyanidin dimers to hexamers previously
isolated from the same plant (11, 12), the structures of the
polymer were deduced to be relatively simple procyanidin
polymers with B- and A-type linkages. However, polyphenol
fractions of cinnamon bark are dark reddish brown, which
cannot be explained by the simple procyanidin structures. When
fresh bark of Japanese cinnamon (C. sieboldii) was peeled from
a branch, the color of the wood surface immediately changed
from white to reddish brown. Since this dynamic change of color
was not observed when the branch was heated in advance, the
reaction was apparently catalyzed by enzymes. However, we
MATERIALS AND METHODS
Materials. (+)-Catechin was isolated from commercial gambir
(extract of the leaves and twigs of Uncaria gambir) and recrystallized
from H2O (14). Procyanidin B1 was produced by degradation of
polymeric procyanidin of Areca catechu with (+)-catechin according
to the method previously reported (15). Cinnamaldehyde was purchased
from Nacalai Tesque, Inc. Cinnamon bark (dried bark of C. cassia)
was purchased from Uchida Wakanyaku Co., Ltd.
General Procedures. Infrared (IR) and ultraviolet (UV) spectra were
obtained with Jasco FT/IR-410 and Jasco V-560 spectrophotometers
(Jasco Co.), and optical rotations were measured with a Jasco DIP-
370 digital polarimeter. 1H and 13C NMR, 1H-1H COSY, NOESY,
HSQC, and HMBC spectra were recorded in acetone-d6 at 27 °C with
a Varian Unity plus 500 spectrometer (Varian Inc.) operating at 500
1
MHz for H and 125 MHz for 13C, respectively. Coupling constants
are expressed in hertz, and chemical shifts are given on a δ (ppm)
scale. HMQC, HMBC, and NOESY experiments were performed using
standard Varian pulse sequences. The matrix-assisted laser desorption
time-of-flight mass spectrometry (MALDI-TOF MS) images were
recorded on a Voyager-DE Pro spectrometer (Applied Biosystems),
and 2, 5-dihydrohydroxybenzoic acid (10 mg/mL in 50% acetone
containing 0.05% trifluoroacetic acid) was used as the matrix. Fast atom
bombardment (FAB) MS was recorded on a JMS 700N spectrometer
* Author to whom correspondence should be addressed. E-mail:
t-tanaka@nagasaki-u.ac.jp.
10.1021/jf800921r CCC: $40.75  2008 American Chemical Society
Published on Web 06/18/2008

Polyphenols of Cinnamon Bark
J. Agric. Food Chem., Vol. 56, No. 14, 2008 5865
Figure 1. Structures of 1-4, 3a, and 3b.
(JEOL Ltd.), and m-nitrobenzyl alcohol or glycerol was used as a
matrix. Column chromatography was performed using 25-100 µm
Sephadex LH-20 column (Amersham Pharmacia Biotech AB), Diaion
HP20SS, MCI-gel CHP20 (Mitsubishi Chemical Co.), Chromatorex
ODS (Fuji Silysia Chemical Ltd.), and silica gel 60 (Merck).
TLC was performed on 0.2 mm precoated Kieselgel 60 F254 plates
(Merck) with toluene/ethyl formate/formic acid (2:7:1, v/v) and CHCl3/
MeOH/H2O (90:10:1, v/v). Spots were detected by UV illumination
and by spraying with 2% ethanolic FeCl3 or 10% sulfuric acid reagent
followed by heating. Detection of catechins and procyanidins was
performed by spraying 1% ethanolic vanillin followed by concentrated
HCl. Analytical HPLC was performed on a 250 mm × 4.6 mm i.d.
Cosmosil 5C18-AR II column (Nacalai Tesque, Inc.) with gradient
elution from 4 to 30% (39 min) and 30 to 75% (15 min) of CH3CN in
50 mM H3PO4 at a flow rate of 0.8 mL/min and detection with an
MD-910 photodiode array detector (Jasco).
119.0, 118.8, 118.7 (C-9′, 9′′, B-ring C-6), 115.7, 115.6 (B-ring C-5),
115.0, 114.9 (B-ring C-2), 104.3, 104.2, 104.14, 104.09 (C-6, 8), 102.7,
102.6 (C-4a), 83.0, 82.9, 82.8 (C-2), 77.4, 77.34, 77.27, 77.2, 77.04,
76.99, 76.97 (C-7′, 7′′), 67.4, 67.3 (C-3), 28.5, 28.30, 28.27, 28.09 (C-
4). The NMR signals were assigned with the aid of 1H-1H COSY,
HSQC, and HMBC spectroscopic analysis. After measurement of NMR
spectra, the color of the solution changed to red. The sample was
analyzed by HPLC to show two large peaks at 51.9 and 52.2 min
attributable to isomers of 3, which were accompanied by a small peak
at 31.8 min attributable to the pigment (3a or 3b) produced from 3.
The UV spectrum of the peak showed absorption maxima at 467, 399,
307, 279, and 271 nm.
Reaction of (+)-Catechin and Cinnamaldehyde upon Heating.
An acetone solution (50 mL) containing 1 (2.0 g) and 2 (5.0 g) was
concentrated to dryness in vacuo, and the resulting syrup was heated
at 100 °C for 50 min. The mixture was separated by Sephadex LH-20
column chromatography (30 cm × 3.0 cm i.d.) with EtOH to give four
fractions. The first fraction contained 2. The second fraction was further
separated by MCI-gel CHP20 (30 cm × 3.0 cm i.d.) with H2O that
contained increasing proportions of MeOH (10% stepwise elution, each
200 mL), to give 1 (990 mg) and a crude crop of 4. Purification by
column chromatography on Sephadex LH-20 (20 cm × 2.5 cm i.d.)
with H2O-MeOH (40-100%, 10% stepwise elution, each 100 mL) and
Chromatorex ODS (25 cm × 3.0 cm i.d.) with H2O/MeOH (20-100%,
5% stepwise elution, each 100 mL) afforded 4 (16.9 mg).
Reaction of (+)-Catechin and Cinnamaldehyde at Room Tem-
perature. (+)-Catechin (1, 1.0 g) and cinnamaldehyde (2, 2.0 g)
(Figure 1) were dissolved in acetone (30 mL), the organic solvent was
evaporated under reduced pressure, and the resulting syrup was kept
at room temperature for 6 days. The mixture was separated by silica
gel column chromatography (35 cm × 3.0 cm i.d.) with CHCl3/MeOH/
H2O (90:10:1, v/v). The first fraction positive to the FeCl3 reagent was
further purified by silica gel column chromatography (30 cm × 2.0
cm i.d.) with CHCl3/MeOH/H2O (90:10:1, v/v) to yield 3 (64.1 mg) as
a precipitate powder from MeOH. Further elution of the first silica gel
column with CHCl3/MeOH/H2O (40:10:1 and 14:6:1, v/v) gave a
complex mixture of products.
Compound 4. Yellow amorphous powder, [R]20D -110.6° (c 0.12,
MeOH); FAB MS m/z 809 [M + H]+, 831 [M + Na]+; HR-FAB MS
m/z 809.2595 (C48H41O12 requires 809.2598); UV λmax (nm) (log ꢀ):
283sh (3.99), 236sh (4.48); IR νmax (cm-1): 3340, 2915, 1619, 1520,
Compound 3. Yellow powder (precipitated from MeOH), [R]15
D
0° (c 0.25, acetone); MALDI-TOF MS m/z 519 [M + H]+, 517 [M -
H]+ (base peak, corresponding to the M+ of 3a or 3b); HR-FAB MS
m/z 517.1658 ([M - H]+ C33H25O6 requires 517.1651); UV λmax (nm)
(log ꢀ): 389 (3.42), 276 (4.29); IR νmax (cm-1): 3390, 2973, 1699, 1599,
1527, 1448, 1365, 1280, 1126; 1H NMR (500 MHz, in acetone-d6)
(major isomer) δ: 7.47-7.29 (10H, m, benzene rings), 6.87 (1H, d, J
) 2.0 Hz, B-ring H-2), 6.69-6.78 (4H, m, H-9′, 9′′, B-ring H-5, 6),
5.89 (2H, br s, H-7′, 7′′), 5.73 (2H, dd, J ) 4, 10.0 Hz, H-8′, 8′′), 4.63
(1H, d, J ) 7.8 Hz, H-2), 3.93 (1H, ddd, J ) 5.3, 7.8, 8.3 Hz, H-3),
2.79 (1H, dd, J ) 5.3, 16.6 Hz, H-4), 2.50 (1H, dd, J ) 8.3, 16.6 Hz,
H-4). After measurement of the 1H NMR spectrum, the concentration
of isomers of 3 that were initially minor components was increased to
give complex signals. 13C NMR (125 MHz, in acetone-d6) (mixture of
four isomers) δ: 152.5 (C-5), 151.5 (C-8a), 148.5, 148.4 (C-7), 145.7,
145.6 (B-ring C-3, 4), 142.0, 141.9 (C-1′, 1′′), 131.3, 131.2 (B-ring
C-1), 129.3, 129.2 (C-3′, 5′, 3′′, 5′′), 128.9, 128.8 (C-4′, 4′′), 127.51,
127.48, 127.40, 127.37 (C-2′, 6′, 2′′, 6′′), 121.14, 121.10, 121.05, 121.0,
120.9, 120.7 (C-8′, C-8′′), 119.7, 119.6, 119.5, 119.4, 119.32, 119.28,
1
1447; H and 13C NMR data: see Table 1.
Reaction of Procyanidin B1 with Cinnamaldehyde. Procyanidin
B1 (5) (20 mg) and 2 (20 mg) were dissolved in acetone (0.5 mL), and
the organic solvent was evaporated under reduced pressure. The mixture
was kept at room temperature for 8 days. The mixture was dissolved
in a small amount of EtOH and subjected to Sephadex LH-20 column
chromatography (15 cm × 1.0 cm i.d.) with EtOH. The nonpolar
compounds including 2 were washed out with EtOH (30 mL).
Subsequent elution of the column with 80% EtOH (20 mL) and then
EtOH/H2O/acetone (2:1:1, v/v) afforded a mixture of phenolic products
(20.1 mg). MALDI-TOF MS of the products is shown in Figure 5.
The products (2 mg) were dissolved in 60% EtOH solution (0.5 mL)
that contained 4% mercaptoethanol and 0.1% HCl and heated at 70 °C
for 7 h. HPLC analysis showed a small peak of epicatechin (tR 22.3
min). In contrast, thiol degradation of 5 (2 mg) under similar conditions
yielded epicatechin and epicatechin 4-hydroxyethylthioether (5a) (tR
28.2 min) (17).

5866 J. Agric. Food Chem., Vol. 56, No. 14, 2008
Table 1. 1H and 13C NMR Spectroscopic Data for 3a
Tanaka et al.
position
1H
4.55 (d, 7.6)
3.89 (m)
2.58 (dd, 16.1, 8.2)
2.89 (dd, 16.1, 5.3)
13C
HMBC (H to C)
position
1H
4.81 (d, 7.3)
4.02 (m)
2.63 (dd, 16.1, 8.0)
2.92 (dd, 16.1, 5.3)
13C
HMBC (H to C)
2
3
4
82.8
68.4
28.7
3, 4, 8a, B-1, B-2, B-6
2′′
3′′
4′′
82.6
68.0
28.2
3′′, 4′′, 8a′′, B′′-1, B′′-2, B′′-6
2′′, 3′′, 4a′′, 5′′, 8a′′
2, 3, 4a, 5, 8a
4a, 5, 7, 8
4a
5
6
7
8
8a
B-1
B-2
B-3
B-4
B-5
B-6
1′
101.4
156.0
95.8
4a′′
5′′
6′′
7′′
101.5
157.7
95.6
5.99 (s)
6.03 (s)
4a′′, 5′′, 7′′ 8′′
152.9
155.2
100.4
153.9
132.1
6.78 (d, 1.9)
8′′
102.8
154.6
132.2
6.88 (d, 1.8)
8a′′
B′′-1
2, B-1, B-3, B-4, B-6
115.3
145.3
145.4
115.6
119.6
138.5
127.0
129.2
127.6
131.1
133.8
33.8
B′′-2
B′′-3
B′′-4
B′′-5
B′′-6
1′′′
115.1
145.5
145.6
115.8
119.3
140.2
128.2
129.2
129.1
77.4
2′′, B′′-3, B′′-4, B′′-6
b
b
b
b
6.59 (d, 8.2)
6.64 (dd, 8.2, 1.9)
B-1, B-3, B-4
2, B-2, B-4
6.69 (d, 8.1)
6.73 (dd, 8.1, 1.8)
B′′-1, B′′-3, B′′-4, B′′-6
2′′, B′′-2, B′′-4
2′, 6′
3′, 5′
4′
7.29 (2H, m)
4′, 7′
2′′′, 6′′′
3′′′, 5′′′
4′′′
7.31 (2H, m)
7.37 (2H, br t, 7.3)
7.30 (m)
4.88 (d, 11.0)
2.38 (ddd, 11.0, 1.5, 2.3)
5.47 (d, 2.3)
4′′′, 7′′′
7.24 (2H, br t, 7.5)
7.14 (dt, 7.5, 1.4)
6.12 (dd, 15.9, 1.3)
6.28 (dd, 15.9, 5.9)
3.27 (ddd, 5.9, 1.5, 1.3)
3.97 (d, 4.6)
1′, 2′, 4′, 6′
2′, 6′
1′, 2′, 6′, 8′, 9′
8, 1′, 7′, 9′, 8′′′
7, 8, 8a, 7′, 8′, 7′′′, 8′′′, 9′′′
2, 3, 4
1′′′, 2′′′, 4′′′, 6′′′
2′′′, 6′′′
7′
7′′′
1′′′, 2′′′, 6′′′, 8′′′, 9′′′
8, 9′, 7′′′
8′
8′′′
41.4
62.7
9′
9′′′
7′′, 8′′, 8a′′, 7′′′
2′′, 3′′, 4′′
3-OH
5-OH
3′′-OH
5′′-OH
4.13 (d, 4.6)
8.65 (br s)
8.38 (br s)
4a, 5
4a′′, 5′′
a Measured in acetone-d6, δ in ppm, J in Hz. b Assignments may be interchanged in each column.
urea (3:2 v/v, adjusted to pH 2 with concd HCl) (16). The first fraction,
which contained polyphenols with large molecular weights, was
collected and concentrated until the acetone was removed. The aqueous
solution was directly applied to a Diaion HP20SS column (20 cm ×
2.0 cm i.d.) with water to remove urea and HCl, and then the
polyphenols were eluted with 80% MeOH to yield polymeric polyphe-
nols (661.7 mg) as a brown amorphous powder. For 13C NMR, see
Figure 7.
Thiol Degradation of Polymeric Polyphenols. A solution of the
polymeric polyphenols in 70% EtOH (25 mg/mL, 0.2 mL) was mixed
with a 60% EtOH solution (0.8 mL) that contained 5% mercaptoethanol
and 0.1% HCl and heated at 70 °C for 7 h. HPLC analysis showed
two major peaks. The retention times and UV peak absorptions
coincided with those of epicatechin (tR 22.3 min) and epicatechin
4-hydroxyethylthioether (tR 28.2 min), in that order of elution (17).
The large broad hump on the baseline between tR 40 and 50 min
observed in the chromatogram of the original polymers decreased but
did not completely disappear. Thiol degradation of the reaction products
of 5 and 2 was performed in a manner similar to that described for
cinnamon polymeric polyphenols.
RESULTS AND DISCUSSION
A mixture of catechin (1) and cinnamaldehyde (2) was kept
without solvents at room temperature for 6 days. The reaction
products were separated by silica gel column chromatography
to give an unstable colorless product 3 as an insoluble precipitate
from MeOH. Among the reaction products, 3 showed the highest
Rf value on normal-phase TLC. The product was dark green
with the FeCl3 reagent, which suggested that 3 was derived from
catechin. In the 1H NMR spectrum, signals arising from catechin
B- and C-rings were similar to those of 1; however, the absence
of signals due to the A-ring indicated that cinnamaldehyde was
attached to the catechin C-6 and C-8 positions. With the aid of
Figure 2. Important NOESY correlations (arrows) observed for 4.
Extraction of Polyphenols from Cinnamon Bark. Commercial
cinnamon bark (100 g) was extracted 3 times with 60% acetone (1 L)
at 50 °C for 2 h. The extract was concentrated under reduced pressure
until the organic solvent was removed, and precipitates formed in the
resulting aqueous solution were dissolved by the addition of a small
amount of MeOH. The solution was applied to Sephadex LH-20 column
chromatography (32 cm × 4.0 cm i.d.) with H2O that contained
increasing proportions of MeOH (20% stepwise elution, each 300 mL)
and finally with 50% aqueous acetone. The fractions that contained
catechins and procyanidins, which were positive to FeCl3 and vanillin/
HCl reagents on TLC, were collected and concentrated until the organic
solvent was completely removed. Contaminating nonpolar compounds
were completely removed by partitioning with Et2O 3 times, and the
aqueous layer was lyophilized to give a polyphenol fraction (1.83 g).
The lyophilized powder was subjected to gel permeation chromatog-
raphy on Sephadex LH-20 (32 cm × 5.0 cm i.d.) with acetone/7 M
1
the H-1H COSY spectrum, the presence of two monosubsti-
tuted benzene rings, two oxygenated methines (H-7′ and H-7′′),
and two sets of mutually coupled (J ) 10 Hz) cis-olefinic
protons (H-8′, H-9′, H-8′′, and H-9′′) was revealed. The
oxygenated methine protons H-7′ and H-7′′ correlated with the
H-8′ and H-8′′, respectively, as well as the phenyl proton signals
(H-7′ with H-2′ and H-6′ and H-7′′ with H-2′′ and H-6′′). In

Polyphenols of Cinnamon Bark
J. Agric. Food Chem., Vol. 56, No. 14, 2008 5867
Figure 3. Production mechanisms of 3 and 4.
1
addition, H-7′ and H-7′′ showed H-13C long-range coupling
Similar condensation reactions of 1 and 2 occurred when the
mixture was heated at 100 °C for 50 min. In this experiment,
separation of the products by Sephadex LH-20 column chro-
matography yielded a new product 4 as a yellow amorphous
powder. FAB MS showed the [M + H]+ peak at m/z 809 and
the [M + Na]+ peak at m/z 831, which suggested that this
compound was composed of two molecules of 1 and two of 2.
From the results of HR-FAB MS, the molecular formula was
with C-5 and C-7 of the catechin A-ring in the HMBC spectrum.
This observation indicated that C-7′ and C-7′′ were attached to
catechin C-5 and C-7 through ether linkages. Furthermore, the
signals attributable to H-9′ and H-9′′ correlated with catechin
A-ring carbon signals (C-5, C-6, C-7, C-8, and C-8a). These
spectroscopic data indicated that 3 was a condensation product
produced by coupling between A-ring C-6 and C-8 of 1 and
two molecules of 2 (Figure 1). The product 3 was first isolated
1
deduced to be C48H40O12. The H and 13C NMR spectra also
1
as a precipitate from methanol and exhibited reasonable H
indicated the presence of two catechin skeletons and two
phenylpropanoid units (Table 1). Substitution at the C-8 and
C-8′′ positions of the two catechin units was revealed by HMBC
spectral analysis (Table 1), and assignment of the adjacent C-8a
and C-8a′′ signals was confirmed by correlations from H-2 and
H-2′′, respectively. These two carbons did not correlate with
the A-ring singlet proton signals (H-6 and H-6′′), which showed
all other 2J and 3J correlations with the remaining A-ring
carbons. Low-field shifts of the carbon signals attributable to
C-8 and C-8′′ (δ 100.4 and 102.8, respectively) also supported
substitution at these positions. C-8, C-7, and C-8a showed cross-
peaks with an aliphatic methine proton H-9′, which showed
correlation peaks with a pair of trans-olefinic protons H-8′ and
H-7′ (J7′,8′ ) 15.9 Hz) in the 1H-1H COSY spectrum.
Observation of HMBC correlations between the H-7′ and the
benzene carbons indicated that this double bond was attached
to a benzene ring (Table 1). The 1H-1H COSY spectrum also
showed cross-peaks between H-9′ and H-8′′′, and in turn, H-8′′′
was coupled with two oxygen-bearing methine protons, H-7′′′
and H-9′′′. The connection of these four aliphatic methine
carbons, C-9′, C-7′′′, C-8′′′, and C-9′′′, also was confirmed by
their mutual HMBC correlations (Table 1). Furthermore, H-9′′′
exhibited HMBC correlations with C-8′′, C-7′′, and C-8a′′,
which indicated that C-9′′ was attached to the C-8′′ of one of
the catechin units. The location of a benzene ring at C-7′′′ was
apparent from their mutual HMBC correlations, as shown in
Table 1. On the basis of the molecular formula (C48H40O12),
the index of hydrogen deficiency was calculated to be 29. This
indicated the presence of two additional rings in the molecule
besides two catechin units, two benzene rings, and a double
bond. Taking the structure of 3 into account, it was strongly
suggested that two oxygen-bearing methine carbons C-7′′′ and
C-9′′′ formed ether rings with the oxygen atoms attached at C-7
NMR signals; however, this compound was unstable, and signals
arising from several isomers gradually increased in the spectrum
within several hours after dissolving NMR solvent (acetone-
d6). Chemical shifts of the newly observed signals were similar
to that of 1, which indicated that the structures of the isomers
were closely related to 3 and suggested that the isomers differed
in their configurations at C-7′ and C-7′′ because these oxygen-
bearing methines were positioned between benzene and the
C-8dC-9 and C-8′dC-9′ double bonds, respectively, and were
expected to be easily isomerized. The MALDI-TOF MS of 3
showed a small [M + H]+ peak at m/z 519 along with a large
[M - H]+ peak. This strongly suggested that 3a or 3b ox-
idatively was generated from 3 upon irradiation of the laser
beam during the MS measurement. Actually, 3 was unstable
when it was exposed to air for a few days to give a small amount
of red pigment, which was detected as a red spot on TLC.
Although isolation failed, HPLC revealed that the pigment had
UV-vis absorption maxima at 467, 399, 307, 279, and 271 nm.
From the structure of 3, it was deduced that there was a
benzopyrylium ion moiety in the molecule. This also was
supported by the production of related pigments by reactions
between catechin and 4-hydroxy-3,5-dimethoxycinnamaldehyde
(18, 19). It strongly was suggested that the dynamic color change
observed when fresh bark of Japanese cinnamon was peeled
off was explained by the formation of pigments related to 3a
or 3b by enzymatic oxidation of 3 and related compounds. The
yield of 3 was very low, and the composition of the remaining
products was too complex to separate into individual compo-
nents. The 13C NMR and MALDI-TOF MS analyses of the
mixture showed that the products were also condensation
products of 1 and 2 with a larger molecular size (data not
shown).

5868 J. Agric. Food Chem., Vol. 56, No. 14, 2008
Tanaka et al.
Figure 4. Thiol degradation of procyanidin B1 (5).
Figure 5. MALDI-TOF MS of reaction products of 2 and 5.
and C-7′′, respectively. Configuration of the methine carbons
were determined by the mutual 1H-1H coupling constants and
NOESY correlations (Figure 2). The large coupling constant
between H-7′′′ and H-8′′′ (11.0 Hz) indicated that the dihedral
angle between these protons was almost 180°. H-8′′′ also
coupled with H-9′, the coupling constant of which was 1.5 Hz.
Since H-9′ showed NOESY correlations with both H-7′′′ and
H-8′′′, H-9′ was spatially placed between H-7′′′ and H-8′′′. The
coupling constant of H-8′′′ with another methine proton H-9′′′
was 2.3 Hz, and H-9′′′ showed NOE to H-8′′′, H-8′, and B′′-
ring protons (B′′-2 and B′′-6). The NOE between H-9′′′ and
B′′-ring protons indicated that these two were oriented to the
same side of the molecule. The absolute configuration at catechin
C-2′′ was R because (+)-catechin was used in this experiment.
Therefore, these observations allowed us to construct the
structure of 4, including the absolute configuration shown in
Figure 1.
The production mechanism of 3 and 4 is illustrated in Figure
3. It is known that the nucleophilicity of the C-8 position of
the catechin A-ring is stronger than that of C-6 (20); therefore,
C-C bond formation between 2 and 1 first occurs at the C-8
position of 1 to give 1a. Then, simultaneous dehydration and
benzopyran formation produces 1b. Similar condensation at the
C-6 position of 1b yielded 3, and reaction between 1b and 1a
generated the dimer 4.
examined by thiol degradation, MALDI-TOF MS (Figure 5),
and 13C NMR spectroscopy (Figure 7). The thiol degradation
of the starting material 5 yielded epicatechin 4-hydroxyethyl
thioether (5a) arising from the extension (upper) unit of 5 and
1 from the terminal (lower) unit (Figure 4). However, the
reaction mixture produced by condensation between 5 and 2
did not yield 5a. Only a small peak attributable to 1 was detected
by HPLC analysis, which indicated that 2 also reacted with
procyanidins. MALDI-TOF MS (Figure 5) showed some
prominent peaks arising from the products, and their plausible
structures, deduced from 3 and 4, are shown in Figure 6. The
results suggested that procyanidins were oligomerized by
condensation with 2. The 13C NMR spectrum of the product
mixture (Figure 7B) was closely related to that of 5 (Figure
7C), and the difference in the two spectra was the appearance
of large peaks at δ 128 and small peaks in the range of δ
100-105. These peaks were related to those of the benzene
moiety and the A-ring C-8 of 4, respectively (Table 1).
Furthermore, the spectrum of the reaction products resembled
the 13C NMR spectrum of the polymeric polyphenols separated
from commercial cinnamon bark. The separation of the poly-
meric polyphenols was performed by a combination of absorp-
tion chromatography on Sephadex LH-20 with ethanol and gel
permeation chromatography on the same Sephadex LH-20 with
7 M urea/acetone as the elution solvent (16). Thiol degradation
of the polymeric polyphenols yielded 5a and epicatechin as the
major products; however, the ratio of thioether (representing
extension units)/epicatechin (representing terminal units) was ca.
2:1, which indicated that the apparent average degree of polym-
erization was only 3. Although epicatechin trimers (procyanidin
To examine as to whether similar reactions occur with
procyanidins, next procyanidin B1 (5) (Figure 4) and 2 were
mixed and left to stand at room temperature for 8 days. After
removal of the starting material by Sephadex LH-20 column
chromatography, the mixture of the reaction products was

Polyphenols of Cinnamon Bark
J. Agric. Food Chem., Vol. 56, No. 14, 2008 5869
Figure 6. Possible structures of products detected by MALDI-TOF MS.
Figure 7. 13C NMR spectra of polymeric polyphenols of cinnamon bark (A), mixture of reaction products of procyanidin B1 and cinnamaldehyde (B), and
procyanidin B1 (C). Arrows indicates the peaks due to benzene rings of cinnamaldehyde moieties.
C-1, MW 866) can be detected as a clear spot on TLC and a sharp
peak in HPLC, the polymeric polyphenols obtained from this
experiment did not show the presence of such molecules. These
observations strongly suggested that procyanidin oligomers and
polymers of the commercial cinnamon bark were further polym-
erized by condensation with cinnamaldehyde.
condensation products. Similar reactions probably occur during
the drying process in production of commercial cinnamon bark.
Actually, polymeric procyanidins obtained from commercial
cinnamon bark showed 13C NMR signals similar to those of
the products generated by reaction of 5 and 2. In addition, it
was strongly suggested that oxidation of the products yielded
benzopyrylium pigments, which explains the dynamic color
change observed when the fresh bark of Japanese cinnamon
In this study, we demonstrated that catechins and procyanidins
reacted with cinnamaldehyde at room temperature to generate

5870 J. Agric. Food Chem., Vol. 56, No. 14, 2008
Tanaka et al.
(4) Shahverdi, A. R.; Monsef-Esfahani, H. R.; Tavasoli, F.; Zaheri,
A.; Mirjani, R. trans-Cinnamaldehyde from Cinnamomum zey-
lanicum bark essential oil reduces the clindamycin resistance of
Clostridium difficile in vitro. J. Food Sci. 2007, 72, 55–58.
(5) Youn, H. S.; Lee, J. K.; Choi, Y. J.; Saitoh, S. I.; Miyake, K.;
Hwang, D. H.; Lee, J. Y. Cinnamaldehyde suppresses toll-like
receptor 4 activation mediated through the inhibition of receptor
oligomerization. Biochem. Pharmacol. 2008, 75, 494–502.
(6) Domadia, P.; Swarup, S.; Bhunia, A.; Sivaraman, J.; Dasgupta,
D. Inhibition of bacterial cell division protein FtsZ by cinnama-
ldehyde. Biochem. Pharmacol. 2007, 74, 831–840.
(7) Shan, B.; Cai, Y. Z.; Brooks, J. D.; Corke, H. Antibacterial
properties and major bioactive components of cinnamon stick
(Cinnamomum burmannii): Activity against foodborne pathogenic
bacteria. J. Agric. Food Chem. 2007, 55, 5484–5490.
(8) Yanaga, A.; Goto, H.; Nakagawa, T.; Hikiami, H.; Shibahara, N.;
Shimada, Y. Cinnamaldehyde induces endothelium-dependent and
-independent vasorelaxant action on isolated rat aorta. Biol. Pharm.
Bull. 2006, 29, 2415–2418.
(9) Nonaka, G.; Morimoto, S.; Nishioka, I. Tannins and related
compounds. Part 13. Isolation and structures of trimeric, tet-
rameric, and pentameric proanthicyanidins from cinnamon.
J. Chem. Soc., Perkin Trans. 1 1983, 2139–2145.
(10) Morimoto, S.; Nonaka, G.; Nishioka, I. Tannins and related
compounds. XXXV. Proanthocyanidins with a doubly linked unit
from the root bark of Cinnamomum sieboldii Meisner. Chem.
Pharm. Bull. 1985, 33, 4338–4345.
(11) Morimoto, S.; Nonaka, G.; Nishioka, I. Tannins and related com-
pounds. XXXVIII. Isolation and characterization of flavan-3-ol
glucosides and procyanidin oligomers from cassia bark (Cinnamo-
mum cassia Blume). Chem. Pharm. Bull. 1986, 34, 633–642.
(12) Anderson, R. A.; Broadhurst, C. L.; Polansky, M. M.; Schmidt,
W. F.; Khan, A.; Flanagan, V. P.; Schoene, N. W.; Graves, D. J.
Isolation and characterization of polyphenol type-A polymers from
cinnamon with insulin-like biological activity. J. Agric. Food
Chem. 2004, 52, 65–70.
(13) Peng, X.; Cheng, K.-W.; Ma, J.; Chen, B.; Ho, C.; Lo, C.; Chen, F.;
Wang, M. Cinnamon bark proanthocyanidins as reactive carbonyl
scavengers to prevent the formation of advanced glycation endprod-
ucts. J. Agric. Food Chem. 2008, 56, 1907–1911.
(14) Nonaka, G.; Nishioka, I. Novel biflavonoids, chalcan-flavan dimers
from Gambir. Chem. Pharm. Bull. 1980, 28, 3145–3149.
(15) Tanaka, T.; Yoshitake, N.; Zhao, P.; Matsuo, Y.; Kouno, I.; Nonaka,
G. Production of oligomeric proanthocyanidins by fragmentation of
polymers. Jpn. J. Food Chem. 2007, 14, 134–139.
(16) Yanagida, A.; Shoji, T.; Shibusawa, Y. Separation of proantho-
cyanidins by degree of polymerization by means of size-exclusion
chromatography and related techniques. J. Biochem. Biophys.
Methods 2003, 56, 311–322.
Figure 8. Possible structure of cinnamon polymeric polyphenols.
was peeled off. In that case, oxidation probably was catalyzed
by plant enzymes. This may also explain the dark reddish brown
color of the polyphenol fractions obtained from commercial
cinnamon bark. On the basis of these results, a possible structure
of cinnamon polymeric polyphenols is shown in Figure 8. A
similar reaction of proanthocyanidins with aldehyde is known
to occur in persimmon fruits; in that case, polymerization and
insolubilization of the proanthocyanidins were caused by
reaction with acetaldehyde generated by anaerobic metabolism
(17). Similar insolubilization seemed to occur in cinnamon bark,
and it was demonstrated by the following simple experiment:
after repeated extraction of polyphenols from cinnamon bark
powder, the remaining plant debris was directly subjected to
thiol degradation. HPLC of the reaction mixture showed the
production of epicatechin (from terminal units) and 5a (from
extension units), and the ratio (ca. 1:2 based on the peak area)
was similar to that observed for degradation products from the
extract. Production of polymeric polyphenols may have a
significance in plant defense systems. When the polyphenols,
such as 3, were oxidized by the catalysis of enzymes, the oxygen
molecules were simultaneously reduced to give reactive oxygen
species such as superoxide anion and hydrogen peroxide, which
act as antimicrobial agents. Thus, the results of this study are
important from the viewpoint of not only food chemistry and
pharmacognosy but also plant physiology.
(17) Tanaka, T.; Takahashi, R.; Kouno, I.; Nonaka, G. Chemical evidence
for the de-astringency (insolubilization of tannins) of persimmon fruit.
J. Chem. Soc., Perkin Trans. 1 1994, 3013–3022.
ACKNOWLEDGMENT
(18) de Freitas, V.; Sousa, C.; Silva, A.; Santos-Buelga, C.; Mateus,
N. Synthesis of a new catechin-pyrylium derived pigment.
Tetrahedron Lett. 2004, 45, 9349–9352.
We are grateful to K. Inada and N. Yamaguchi for NMR
and MS measurements.
(19) Sousa, C.; Mateus, N.; Perez-Alonso, J.; Santos-Buelga, C.; de Freitas,
V. Preliminary study of oaklins, a new class of brick-red catechin-
pyrylium pigments resulting from the reaction between catechin and
wood aldehydes. J. Agric. Food Chem. 2005, 53, 9249–9256.
(20) Charlton, A. J.; Davis, A. L.; Jones, D. P.; Lewis, J. R.; Davies,
A. P.; Haslam, E.; Williamson, M. P. The self-association of the
black tea polyphenol theaflavin and its complexation with caffeine.
J. Chem. Soc., Perkin Trans. 2 2000, 317–322.
LITERATURE CITED
(1) Cocchiara, J.; Letizia, C. S.; Lalko, J.; Lapczynski, A.; Api, A. M.
Fragrance material review on cinnamaldehyde. Food Chem.
Toxicol. 2005, 43, 867–923.
(2) Harada, M. Pharmacological studies on Chinese cinnamon. II.
Effects of cinnamaldehyde on the cardiovascular and digestive
systems. Chem. Pharm. Bull. 1975, 23, 941–947.
(3) Ma, Y.-Y.; Huo, H.-R.; Li, C.-H.; Zhao, B.-S.; Li, L. F.; Sui, F.;
Guo, S.-Y.; Jiang, T. L. Effects of cinnamaldehyde on PGE2
release and TRPV4 expression in mouse cerebral microvascular
endothelial cells induced by interleukin-1beta. Biol. Pharm. Bull.
2008, 31, 426–430.
Received for review March 25, 2008. Revised manuscript received May
7, 2008. Accepted May 12, 2008. This work was supported by Grant-
in-aid for Scientific Research 18510189 from the Japan Society for the
Promotion of Science.
JF800921R