α-Glucosidase inhibition of 6-hydroxyflavones. Part 3: Synthesis and evaluation of 2,3,4-trihydroxybenzoyl-containing flavonoid analogs and 6-aminoflavones as α-glucosidase inhibitors

Article abstract of DOI:10.1016/j.bmc.2004.12.010

The SAR studies suggested that the C-ring of baicalein (1) was not necessary for the activity, and validated the importance of 2,3,4- trihydroxybenzoyl structure of 1. Thus, a series of 2,3,4-trihydroxybenzoyl- containing flavonoid analogs were investigat

Full text of DOI:10.1016/j.bmc.2004.12.010

Bioorganic & MedicinalChemistry 13 (2005) 1661–1671
a-Glucosidase inhibition of 6-hydroxyflavones. Part 3: Synthesis
and evaluation of 2,3,4-trihydroxybenzoyl-containing
flavonoid analogs and 6-aminoflavones as a-glucosidase inhibitors
Hong Gao and Jun Kawabata*
Laboratory of Food Biochemistry, Division of Applied Bioscience, Graduate School of Agriculture, Hokkaido University,
Kita-ku, Sapporo 060-8589, Japan
Received 28 October 2004; accepted 7 December 2004
Available online 30 December 2004
Abstract—The SAR studies suggested that the C-ring of baicalein (1) was not necessary for the activity, and validated the impor-
tance of 2,3,4-trihydroxybenzoylstructure of 1. Thus, a series of 2,3,4-trihydroxybenzoyl-containing flavonoid analogs were inves-
tigated for the a-glucosidase inhibitory activity. The results indicated that 5,6,7-trihydroxy-2-phenyl-4-quinolone (2) and
5,6,7-trihydroxyflavanone (4) showed the comparable activity to 1, while 3,5,6,7-tetrahydroxyflavone (7), 5,6,7-trihydroxyisoflavone
(8), and 6-hydroxygenistein (9) showed moderate a-glucosidase inhibitory activity. In addition, it was found that 6-amino-5,7-
dihydroxyflavone (16) was a more potent and specific rat intestinal a-glucosidase inhibitor than 1, and showed the comparable activ-
ity to acarbose. This is the first report on mammalian intestinal a-glucosidase inhibitory activity of 6-aminoflavones. Kinetic studies
revealed that 16 inhibited both sucrose- and maltose-hydrolyzing activities of rat intestinal a-glucosidase uncompetitively.
ꢀ 2004 Elsevier Ltd. All rights reserved.
1. Introduction
due to the difference of molecular recognition in the tar-
get binding site of these enzymes.
Mammalian a-glucosidases (a-D-glucoside glucohydro-
lase EC. 3.2.1.20) inhibitors, which interfere with enzy-
matic action in the brush-border of the small intestine,
could slow the liberation of D-glucose from oligosaccha-
rides and disaccharides, resulting in delaying glucose
absorption and decreasing postprandial plasma glucose
levels.1 There are reports of established a-glucosidase
inhibitors such as acarbose2 and voglibose3 from micro-
organisms, and nojirimycin,4 and 1-deoxynojirimycin5
from plants and their effects on blood glucose levels
after food uptake.6–9 On the other hand, it is well known
that polyphenols efficiently interact with proteins and
lead to inhibit enzyme activities.10 Recent studies have
indicated that phenonic compounds were potent class
of a-glucosidase inhibitors.11–14 But, most of these
yeastÕs a-glucosidase (class I) inhibitors did not show
any activity against mammalian a-glucosidase (class II)
In the course of our screening study on rat intestinal a-
glucosidase-inhibiting principles from plant sources, we
previously reported that baicalein (5,6,7-trihydroxyflav-
one, 1) and the related 6-hydroxyflavonoids were a new
class of a-glucosidase inhibitors.15,16 As a continuing
study, we evaluated the influences of A- and B-rings sub-
stitution for 1 on the a-glucosidase inhibitory activ-
ity.17,18 From initialSAR studies, it was suggested that
5,6,7-trihydroxylsubsituents and presumabyl 4-car-
bonylgroup of 1 (2,3,4-trihydroxybenzoylmoiety) were
important to exert the inhibitory activity as a minimal
structure, although the actual molecular interaction with
the rat intestinal a-glucosidase is unknown.
Cushman et al. have reported the inhibitory effects of
hydroxyflavones and aminoflavones on protein-tyrosine
kinases.19,20 It is not obvious, however, if the amino sub-
stituents offer any overall advantage over hydroxyl
group, but both are approximately the same size and
can function as hydrogen donors as well a hydrogen
bond acceptors. We have focused on the 6-hydroxyl
group of 1 and been interested in the a-glucosidase
inhibitory activity of aminoflavones.
Keywords: a-Glucosidase inhibitor; Structure–activity relationship
(SAR) study; 5,6,7-Trihydroxyflavone; 6-Hydroxyflavones; 6-Amino-
flavones.
*
Corresponding author. Tel./fax: +81 11 706 2496; e-mail: junk@
chem.agr.hokudai.ac.jp
0968-0896/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2004.12.010

1662
H. Gao, J. Kawabata / Bioorg. Med. Chem. 13 (2005) 1661–1671
In the present study, we studied the influence of the C-
ring of 1 on the a-glucosidase inhibitory activity in order
to clarify further structural requirements of 1 for inhibi-
tion of this enzyme. And furthermore, we prepared a
series of 2,3,4-trihydroxybenzoyl-containing flavonoid
analogs of 1 as well as 6-aminoflavones, and evaluated
their a-glucosidase inhibitory effect.
the corresponding 3-lithioflavone with trimethyl borate
followed by oxidation with hydrogen peroxide in acetic
acid to give 3-hydroxy-5,6,7-trimethoxyflavone,26 which
was demethylated to give 3,5,6,7-tetrahydroxyflavone 7.
The synthesis of compounds 8 and 9 were shown in
Scheme 2. Treatment of 3,4,5-trimethoxyphenolwith
phenylacetyl chloride in the presence of BF3–Et2O
complex
benzylketone
gave
6-hydroxy-2,3,4-trimethoxyphenyl
8a.27 This was treated with N,N-
2. Results and discussion
2.1. Synthetic approach
carbonyldiimidazole/formic acid in THF to give 5,6,7-
trimethoxyisoflavone 8b.28 Demethylation of 8b with
BBr3 provided the corresponding 5,6,7-trihydroxyisof-
lavone 8. 40,5,6,7-Tetrahydroxyisoflavone 9 was ob-
tained by the same procedure described for 8 using
p-methoxyphenylacetyl chloride as the starting material.
Reaction of 3,4,5-trimethoxyphenolwith cinnamoyl
chloride in the presence of BF3–Et2O complex gave 60-
hydroxy-20,30,40-trimethoxychalcone 10a,27 which was
cyclized to the corresponding 5,6,7-trimethoxyflavanone
Compound 1 was prepared as previously described.21
The synthesis of compound 2 is shown in Scheme 1.
Reaction of 3,4,5-trimethoxyaniline with benzoyl chlo-
ride in the presence of triethylamine gave N-(3,4,5-
trimethoxyphenyl)benzamide
2a.22
Friedel–Crafts
acylation with acetyl chloride in the presence of SnCl4
gave N-(2-acetyl-3,4,5-trimethoxyphenyl)benzamide 2b.
Cyclization of 2b in the presence of t-BuOK in THF
gave a quinolone 2c, which was demethoxylated with
excess BBr3 afforded 5,6,7-trihydroxy-2-phenyl-4-quino-
lone 2.22
29
10b with KF in methanolunder reflux. By the proce-
dure described for 2, 10b was demethylated to give
5,6,7-trihydroxyflavanone, which was spontaneously
converted to 20,30,40,60-terahydroxychalcone 10 in the
reaction condition. 1,2,3-Trihydroxyxanthone 14 was
obtained as previously described.30 Condensation of
3,4,5-trimethoxyphenoland 2,5-dihydroxy-4-methoxy-
benzoic acid by anhydrous zinc chloride in phosphorus
oxychloride gave 7-hydroxy-1,2,3,6-tetramethoxyxanth-
one 15a, which was demethylated with hydroiodic acid
to give 1,2,3,6,7-pentahydroxyxanthone 15 (Scheme
3).30,31
Methylation of 1 gave 5,6,7-trimethoxyflavone,23 which
was treated with excess LawessonÕs reagent in dry tolu-
ene to convert into the corresponding 5,6,7-trimeth-
oxy-4-thioflavone.24 As the method described for 2,
demethylation of 5,6,7-trimethoxy-4-thioflavone gave
5,6,7-trihydroxy-4-thioflavone 3. Compounds 4 and 11
were obtained by the hydrogenation of 1 in the presence
of palladium black.25 A 3-hydroxy substituent was
introduced to 5,6,7-trimethoxyflavone by quenching
The synthesis of compounds 16, 17, and 21 was shown
in Scheme 4. Nitration of 5,7-dihydroxyflavone with
Scheme 1. Reagents and conditions: (1) Et3N, THF, 85%; (2) CH3COCl, SnCl4, 1,2-dichloroethane, 83%; (3) t-BuOK, THF, 80 ꢁC, 64%; (4) BBr3/
CH2Cl2, 52%.
Scheme 2. Reagents and conditions: (1) PhCH2COCl(or ( p-OMe)PhCOCl), BF3–Et2O; (2) N,N0-carbonyldiimidazole, formic acid, THF; (3) BBr3/
CH2Cl2.

H. Gao, J. Kawabata / Bioorg. Med. Chem. 13 (2005) 1661–1671
1663
Scheme 3. Reagents and conditions: (1) ZnCl2, POCl3, 70–80 ꢁC, 2 h, 25%; (2) HI (d = 1.5), reflux, 2 h, 26%.
Scheme 4. Reagents and conditions: (1) 70% HNO3, AcOH, 65 ꢁC, 1 h, 70%; (2) MeI, K2CO3, acetone, 8 h, 85%; (3) H2, 10% Pd–C, THF, 10 h, 92%;
(4) HI (d = 1.6), reflux, 3 h, 55%; (5) H2, 10% Pd–C, THF, 10 h, 79%; (6) 70% HNO3, AcOH, 65 ꢁC, 10 h, 52%; (7) Zn, AcOH, 1 h, 71%.
70% HNO3 in acetic acid gave 8-nitro-5,7-dihydroxy-
flavone 16a, which was hydrogenated to give 8-amino-
5,7-dihydroxyflavone 17.32,33 Methylation of 16a with
MeI followed by hydrogenation in the presence of 10%
Pd–C afforded 8-amino-5,7-dimethoxyflavone 16c,
which was rearranged to 6-amino-5,7-dihydroxyflavone
16 by demethylation in the Wessely–Moser reaction con-
dition.32 Treatment of 5,7-dihydroxyflavone with excess
70% HNO3 in acetic acid gave 5,7-dihydroxy-6,8-dini-
troflavone 21a, which was reduced with Zn–acetic acid
system to give 21.34 Treatment of 2,6-dihydroxy-3-nitro-
acetophenone with sodium benzoate and benzoic
anhydride gave 5-hydroxy-6-nitroflavone.35 Reduction
of 5-hydroxy-6-nitroflavone with Zn–acetic acid system
gave 6-amino-5-hydroxyflavone 18.34 The same proce-
dure as described above can smoothly produce 6-ami-
a dramatic drop in the activity. Reduction of the double
bond at C2/C3 of 1 led to a flavanone 4 (IC50 = 79 lM),
which was slightly less active than 1 or 2. These results
suggested that a c-pyrone structure of the C-ring of 1
was not essentialbut advantageous to the activity. This
rationale is supported by the activity of compound 5
(IC50 = 170 lM) and 6 (IC50 = 201 lM), which have nei-
ther a B- nor C-ring, as described in the previous
study.18 From these results, it is fair to conclude that
the 2,3,4-trihydroxybenzoylstructure of 1 is essential
in eliciting a-glucosidase inhibitory activity.
Having identified the importance of 2,3,4-tri-
hydroxybenzoylmoiety of 1, we decided to evaluate
the activity of a series of 2,3,4-trihydroxybenzoyl-con-
taining flavonol( 7), isoflavones (8, 9), and chalcones
(10, 11), together with anthraqinone (12) and xanthones
(13–15). Unfortunately, all of these compounds tended
to decrease the potency compared to 1 (Fig. 1). As for
a flavonol 7 (IC50 = 153 lM), an addition of a hydroxyl
group to position 3 of 1 partially reduced the activity.
no-7-hydroxyflavone
19
from
4,6-dihydroxy-3-
nitroacetophenone as a starting-material. Structures of
the compounds 1–4, 7–11, and 14–21 were confirmed
by spectroscopic data (NMR, Mass) and satisfactory
analytical values were obtained for all the compounds.
Furthermore, isoflavone
8 (IC50 = 640 lM) or 9
2.2. Biological activity tests
(IC50 = 624 lM), which has B-ring (phenylgroup) at po-
sition 3, was 12-fold less active than 1. Rossi et al.36 have
described a relationship between molecular structure
and activity toward enzymes of baicalein (1), and re-
vealed that 1 could act as a template for attractive
hydrogen-bonding interactions with the amino acids so
As shown in Figure 1, a replacement of the O-1 for 1 by
nitrogen atom did not affect the activity [2
a
(IC50 = 45 lM) and 1 (IC50 = 52 lM)], whereas a sul-
fur-substitution of the C-4 carbonyloxygen ( 3) led to

1664
H. Gao, J. Kawabata / Bioorg. Med. Chem. 13 (2005) 1661–1671
Figure 1. Structures and IC50 values of compounds 1–21 for inhibition of rat intestinal a-glucosidase (sucrose, 56 mM; phosphate buffer, 100 mM;
pH 6.3). No inhibition, less than 30% inhibition at the concentration of 1000 lM.
as to reduce or eliminate enzyme activity. Based on their
study, the results might be explained by the interaction
tency and the structure of flavone itself was taken part in
the a-glucosidase inhibitory activity.
between
1
with a-glucosidase. This 2,3,4-tri-
hydroxybenzoylmoiety of 1 can be considered to bind
to the enzyme by hydrogen-bonding interactions. As
predicated previously, the phenyl group (B-ring) at posi-
tion 2 for 1 seems not to greatly affect the activity. How-
ever, a bulky substitution at position 3 may hinder the
approach of a 2,3,4-trihydroxybenzoylmoiety to the
binding site in this enzyme. On the other hand, com-
pounds 9–15, which do not have flavone skeleton
although possessing the 2,3,4-trihydroxybenzoyl moiety,
showed almost inactive. Hence, it was suggested that the
structure near the C-ring of 1 was important for the po-
A replacement of the hydroxyl group (–OH) at C-6 in 1
with an amino group (–NH2) afforded 16. Interestingly,
16 (IC50 = 2.4 lM) showed a dramatic improvement in
potency, whereas 8-amino-5,7-dihydroxyflavone 17 was
inactive. And also, 19 (IC50 = 135 lM) was found to
show a moderate activity even though 6,7-dihydroxyf-
lavone17 was inactive. Hence, it was suggested that the
amino group at position 6 of the flavones gave an addi-
tionalfavorabel effect on a-glucosidase inhibition. Hav-
ing established that the potency could be improved
through this modification, we sought to prepare a series

H. Gao, J. Kawabata / Bioorg. Med. Chem. 13 (2005) 1661–1671
1665
of 6-aminoflavones and tested their activity (Fig. 1).
Compound 18, which lacks a hydroxyl group at position
7, was found to be inactive. The introduction of an ami-
no group to position 8 in 19 afforded 20, which showed a
dramatic loss of potency. To our surprise, 21
(IC50 = 82 lM), which possessed an additionalamino
group substitution on position 8 of 16, decreased the
activity 34-fold compared to 16 in spite of the presence
of 6-amino-5,7-dihydroxyflavone structure. These re-
sults suggested that the 6-amino group and 7-hydroxyl
group of 16 were crucialfor the activity; 5-hydroxylsub-
stitution was favorable to the activity, whereas the sub-
stitution of 8-amino group was unfavorable.
lyzing activity with Ki value of 1.4 and 2.2 lM, respec-
tively. Next, we compared the a-glucosidase inhibitory
activity and the inhibition type against the enzyme be-
tween 16 and acarbose, a commercialmedicine as an
a-glucosidase inhibitor (Table 2). The inhibition type
of 16 (uncompetitive inhibition) against sucrase or mal-
tase was different from that of acarbose (competitive
inhibition), while the sucrase or maltase inhibitory activ-
ity of 16 (sucrase: IC50 = 2.4 lM, Ki = 1.4 lM; maltase:
IC50 = 4.4 lM, Ki = 2.2 lM) was comparable to that of
acarbose (sucrase: IC50 = 1.8 lM, Ki = 0.8 lM; maltase:
IC50 = 2.1 lM, Ki = 0.7 lM).
In summary, we have reported rat intestinal a-glucosi-
dase inhibitory activity of 2,3,4-trihydroxybenzoyl-con-
taining flavonoid analogs of 1 and 6-aminoflavones.
These results suggested that C-ring of 1 was not neces-
sary for the activity, and validated the importance of
2,3,4-trihydroxybenzoylstructure of 1. By SAR-studies,
16 was found to be a more potent and specific rat
Having identified a potent series of rat intestinal a-glu-
cosidase inhibitors, we evaluated the selectivity of 1, 2,
4, and 16 for several selected sugar hydrolases inhibition
(Table 1). As indicated above, each of these compounds
showed the high activity against rat intestinalsucrase.
Compound 1 (IC50 = 500 lM) was a mild rat intestinal
maltase inhibitor, while 16 (IC50 = 4.4 lM) strongly
inhibited this enzyme. However, 2 or 4 showed no inhib-
itory activity against maltase. It is interesting to note
that all of the compounds 1, 2, 4, and 16 did not inhibit
the other glycosidases, ex. isomaltase, b-glucosidase, a-
mannosidase, a-galactosidase, b-galactosidase, and por-
cine pancreatic a-amylase. Therefore, compounds 1, 2,
4, and 16 would belong to specific rat intestinal sucrase
inhibitors. In addition, the result indicated that 16 was
also a potent, specific rat intestinal maltase inhibitor.
Among the tested compounds 1–21, 16 showed the
strongest inhibitory activity against rat intestinal a-glu-
cosidase; thus, we examined the type of inhibition of this
enzyme by 16. Lineweaver–Burk plot of a-glucosidase
kinetics is shown in Figure 2. The kinetic result demon-
strated that the mechanism of a-glucosidase inhibition
of 16 was uncompetitive on sucrose- and maltose-hydro-
Table 1. Selectivity of compounds 1, 2, 4, and 16 for the inhibitory
effect on eight sugar hydrolases
IC50 (lM)
1
2
4
16
Sucrase (rat intestine)
Maltase (rat intestine)
52
45
79
2.4
4.4
NI
NI
NI
NI
NI
NI
500
NI
NI
NI
NI
NI
NI
NIa
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
Isomaltase (rat intestine)
b-Glucosidase (almonds)
a-Mannosidase (jack bean)
a-Galactosidase (green coffee been)
b-Galactosidase (bovine liver)
a-Amylase (porcine pancreas)
Figure 2. Lineweaver–Burk plot analysis of the inhibition kinetics of
rat intestinal a-glucosidase inhibitory effects by 6-amino-5,7-dihydr-
oxyflavone (16).
a No inhibition, less than 30% inhibition at the concentration of
1000 lM.
Table 2. IC50, Ki values, and inhibition type of 16 and acarbose for rat small intestinal disaccharidase
Substrate
Km(mM)
IC50 (lM)
Acarbose
Ki (lM)
Acarbose
Inhibition type
16
16
16
Acarbose
Sucrose
Maltose
27
4
2.4
4.4
1.8
2.1
1.4
2.2
0.8
0.7
Uncompetitive
Uncompetitive
Competitive
Competitive

1666
H. Gao, J. Kawabata / Bioorg. Med. Chem. 13 (2005) 1661–1671
intestinal a-glucosidase inhibitor than 1. In addition, 16
showed the comparable a-glucosidase inhibitory activity
to ararbose. Therefore, it is concluded that 16 would be
a lead compound suitable for designing new potent a-
glucosidase inhibitors of this kind.
3.1.2.
N-(2-Acetyl-3,4,5-trimethoxyphenyl)benzamide
(2b). To an ice-cooled solution of 2a (5.25 g, 18.3 mmol)
in dry 1,2-dichloroethane (60 mL) under argon gas was
added dropwise SnCl4 (4.3 mL, 36.6 mmol) and acetyl
chloride (1.41 mL, 20 mmol). After stirring at room tem-
perature for 1.5 h, the solution was poured into ice-
cooled water, extracted with ethyl acetate, washed with
brine, dried over Na2SO4, and evaporated. The residue
was purified by silica gel column chromatography using
hexane–ethylacetate (1:1) as the eul ent to yiedl 2b as an
oil(4.98 g, 83%): EIHRMS m/z 329.1255 (calcd for
C18H19O5N, 329.1264); 1H NMR d (chloroform-d)
ppm (J in Hz): 2.66 (3H, s, Ac-2), 3.85, 3.99, 4.00 (each
3H, s, OMe-3, 4, and 5), 7.51–7.55 (3H, m, H-30, 40, and
50), 8.02 (2H, dd, J = 6.8, 1.3, H-20 and 60).
3. Experimental
3.1. Chemistry
NMR spectra were recorded with a Bruker AMX500
instrument (1H, 500 MHz; 13C, 125 MHz). Chemical
shift data were calculated from the residual solvent sig-
nals of dH 3.30 and dC 49.0 ppm in methanol-d4, dH 2.04
and dC 29.8 ppm in acetone-d6, dH 7.24 and dC 77.0 ppm
in chloroform-d, and dH 2.49 and dC 39.5 ppm in di-
methylsufloxide- d6. Field desorption (FD), FD-high
resolution (HR), electron ionization (EI), and EIHR
mass spectra (MS) were obtained with a JeolJMS-
SX102A instrument. Melting point data were measured
with a hot-stage apparatus and are uncorrected. The
following experimental conditions were used for
chromatography: ordinary-phase silica gel column
chromatography, WakogelC-300 (Wako Pure Chem.
Co., Osaka, Japan); reverse-phase ODS column chro-
matography, Cosmosil75C 18-OPN (Nacalai Tesque,
Kyoto, Japan); TLC, precoated TLC plates with silica
3.1.3. 5,6,7-Trimethoxy-2-phenyl-4-quinolone (2c). To a
stirred solution of 2b (2.5 g, 7.6 mmol) in dry THF
(50 mL) under argon gas was added t-BuOK (4.21 g,
38 mmol). The reaction mixture was then refluxed for
20 h. After cooling to room temperature, the resultant
mixture was added to a saturated aqueous solution of
NH4Cland extracted with ethylacetate. The extract
was washed with brine, dried over Na2SO4, and evapo-
rated. The residue was purified by silica gel column chro-
matography using hexane–ethylacetate (4:6) as the
eluent to yield 2c as a white solid (1.51 g, 64%): mp
230 ꢁC; EIHRMS m/z 311.1140 (calcd for C18H17O4N,
1
gel60 F
malphase) and RP-18 F
(Merck, 0.25 mm or 0.5 mm thickness, nor-
(Merck, 0.2 mm thickness,
311.1158); H NMR d (chloroform-d) ppm (J in Hz):
254
3.93, 3.97, 4.14 (each 3H, s, OMe-5, 6, and 7), 7.00
(1H, s, H-3), 7.39 (1H, s, H-8), 7.43 (2H, d, J = 6.6, H-
20 and 60), 7.37–7.40 (3H, m, H-30, 40, and 50).
254s
reverse phase). Spots were detected by a UV lamp
(254 nm). Preparative HPLC was conducted with an
InertsilPREP-ODS coulmn (20.0 · 250 mm, GL-Sci-
ence). Detailed analytical conditions are mentioned in
each section.
3.1.4. 5,6,7-Trihydroxy-2-phenyl-4-quinolone (2). To a
stirred solution of 2c (0.35 g, 1.13 mmol) in dry CH2Cl2
(10 mL) at ꢀ80 ꢁC was added BBr3 (1 M CH2Cl2 solu-
tion, 10 mL, 10 mmol). The reaction mixture was then
stirred for 1 h at ꢀ80 ꢁC and 12 h at room temperature.
The reaction was stopped by adding ice-cooled water
(100 mL). To this mixture was added 1-butanol
(200 mL). The organic phase was separated, washed
with brine, dried over anhydrous Na2SO4, and evapo-
rated. The residue was crystallized from hexane–ethyl
acetate to give 2 as a yellow solid (0.16 g, 52%): mp
292 ꢁC; FABHRMS (negative) m/z 268.0617 ([MꢀH]ꢀ,
All chemicals used were of reagent grade and were pur-
chased from Wako Pure Chem Co. (Osaka, Japan).
20,30,40-Trihydroxyacetophenone (5), 2,3,4-trihydroxy-
benzaldehyde (6), and mangeferin (13) were supplied
by Sigma–Aldrich Japan Co. (Tokyo, Japan). 1,2,3-Tri-
hydroxyanthraquinone (12) was supplied by Tokyokasei
Co. (Tokyo, Japan). Acetone and tetrahydrofuran were
dried by storing over 3A molecular sieves. All solvents
were distilled before use.
1
calcd for C15H10O4N, 268.0610); H NMR d (DMSO-
3.1.1. N-(3,4,5-Trimethoxyphenyl)benzamide (2a). To a
solution of 3,4,5-trimethoxyaniline (3.6 g, 19.6 mmol)
in dry THF (50 mL) was added Et3N (4.1 mL,
29.4 mmol). The solution was stirred at 0 ꢁC for
10 min and treated dropwise by benzoylcholride
(3.4 mL, 29.4 mmol). The reaction mixture was stirred
at room temperature for 1 h, then quenched by adding
water, and extracted with ethylacetate. The organic
layer was washed with brine, dried over anhydrous
Na2SO4, and evaporated. The residue was purified by
silica gel column chromatography using hexane–ethyl
acetate (8:2) as the eluent to yield 2a as a gray solid
(4.78 g, 85%): mp 141 ꢁC; EIHRMS m/z 287.1165 (calcd
d6) ppm (J in Hz): 6.16 (1H, s, H-3), 6.62 (1H, s, H-8),
7.55–7.56 (3H, m, H-30, 40, and 50), 7.77 (2H, d,
J = 3.1, H-20 and 60), 9.98, 11.71, 14.58 (each 1H, s,
OH-5, 6, and 7).
3.1.5. 5,6,7-Trimethoxyflavone (3a). A mixture of 1
(2.7 g, 10 mmol), 10 mL of methyl iodide and anhydrous
potassium carbonate (6.9 g, 50 mmol) in dry acetone
(100 mL) was well stirred under reflux for 8 h. The reac-
tion mixture was cooled and filtered. The filtrate was
evaporated and the residue was crystallized from acetic
acid–water to give 3a (2.56 g, 82%) as white plates: mp
145–147 ꢁC; EIHRMS m/z 312.1029 (calcd for
1
1
for C16H17O4N, 287.1158); H NMR d (chloroform-d)
C18H16O5, 312.0999); H NMR d (chloroform-d) ppm:
ppm: 4.14 (9H, s, OMe-3, 4, and 5), 7.24, 7.26 (each
1H, s, H-2 and 6), 7.81–7.85 (3H, m, H-30, 40, and 50),
8.16–8.20 (2H, m, H-20 and 60).
3.99 (9H, s, OMe-5, 6, and 7), 6.67 (1H, s, H-3), 6.82
(1H, s, H-8), 7.51–7.52 (3H, m, H-30, 40, and 50), 7.87–
7.89 (2H, m, H-20 and 60).

H. Gao, J. Kawabata / Bioorg. Med. Chem. 13 (2005) 1661–1671
1667
3.1.6. 5,6,7-Trimethoxy-4-thioflavone (3b). A mixture of
3a (1.56 g, 5 mmol), LawessonÕs reagent (1.23 g,
3 mmol), and anhydrous toluene (40 mL) was refluxed
for 3 h. The solvent was evaporated and the residue
was purified by silica gel column chromatography using
hexane–ethylacetate (7:3) as the eul ent to give 3b as a
purple solid (1.18 g, 72%): mp 116–117 ꢁC; EIHRMS
m/z 328.0756 (calcd for C18H16O4S, 328.0770); 1H
NMR d (chloroform-d) ppm: 3.93, 3.95, 4.01 (each
3H, s, OMe-5, 6, and 7), 6.83 (1H, s, H-3), 7.50–7.52
(3H, m, H-30, 40, and 50), 7.59 (1H, s, H-8), 7.92–7.93
(2H, m, H-20 and 60).
aqueous saturated sodium bicarbonate solution
(50 mL). The aqueous mixture was extracted with ethyl
acetate (100 mL · 3). The combined organic extracts
were dried over anhydrous Na2SO4 and concentrated.
The residue was purified by silica gel column chroma-
tography using chloroform–methanol (10:0.2) as the elu-
ent to give 7a (0.84 g, 32%) as yellow powders: mp
168 ꢁC; EIHRMS m/z 328.0938 (calcd for C18H16O6,
1
328.0947); H NMR d (chloroform-d) ppm (J in Hz):
3.91, 3.97, 4.02 (each 3H, s, OMe-5, 6, and 7), 6.79
(1H, s, H-8), 7.43 (1H, t, J = 7.4, H-40), 7.50 (2H, t,
J = 7.4, H-30 and 50), 8.18 (2H, d, J = 7.4, H-20 and 60).
3.1.7. 5,6,7-Trihydroxy-4-thioflavone (3). Using the pro-
cedure described for 2, 3b (164 mg, 0.5 mmol) was deme-
thylated to 3. The solid was further purified by
preparative HPLC (mobile phase, water–methanol–
formic acid (20:80:0.1); flow rate, 5 mL/min; tR 25.2 min;
detection, UV 254 nm) to give 3 (108 mg, 76%) as green
crystals: mp 198 ꢁC; EIHRMS m/z 286.0291 (calcd for
3.1.10. 3,5,6,7-Tetrahydroxyflavone (7). Using the proce-
dure described for 2, 7a (130 mg, 0.4 mmol) was deme-
thylated to 4. Crystallization from hexane and ethyl
acetate gave 7 (74 mg, 65%) as yellow powders: mp
246 ꢁC; EIHRMS m/z 286.0472 (calcd for C15H10O6,
1
286.0477); H NMR d (methanol-d4) ppm: 6.53 (1H, s,
H-8), 7.43–7.49 (3H, m, H-30, 40, and 50), 8.18 (2H, d,
J = 7.1, H-20 and 60).
1
C15H10O4S, 286.0300); H NMR d (DMSO-d6) ppm (J
in Hz): 6.69 (1H, s, H-3), 7.55 (1H, s, H-8), 7.56–7.60
(3H, m, H-30, 40, and 50), 8.12 (2H, d, J = 8.1, H-20
and 60).
3.1.11. 6-Hydroxy-2,3,4-trimethoxyphenyl benzyl ketone
(8a).
A mixture of 3,4,5-trimethoxyphenol(3.7 g,
20 mmol), phenylacetyl chloride (3.4 g, 22 mmol), and
BF3–Et2O complex solution (20 mL) was heated under
reflux for 15 min and quenched with excess of water.
The resulting mixture was extracted with ethyl acetate
(200 mL · 3) and the combined organic extracts were
washed with brine, dried over anhydrous Na2SO4, and
concentrated. The residue was purified by silica gel col-
umn chromatography using hexane–ethylacetate (7:3)
as the eluent to give 8a (5.98 g, 90%) as yellow pale
oil: EIHRMS m/z 302.1171 (calcd for C17H18O5,
3.1.8. 5,6,7-Trihydroxyflavanone (4) and 20,30,40,60-tetra-
hydroxydihydrochalcone (11). A solution of 1 (108 mg,
0.4 mmol) in water–methanol–acetic acid (1:10:1,
24 mL) was hydrogenated for 36 h in the presence of
palladium black (10 mg) at room temperature. The cat-
alyst was removed by filtration and the filtrate was con-
centrated. The residue was dissolved in a small amount
of methanoland subjected to preparative HPLC (mobile
phase, water–methanol–formic acid (30:70:0.1); flow
rate, 4.5 mL/min; detection, UV 254 nm) to give 24 mg
of 4 (tR 17.2 min) as yellow powders: mp 228–230 ꢁC;
EIHRMS m/z 272.0703 (calcd for C15H12O5,
272.0685); 1H NMR d (methanol-d4) ppm (J in Hz):
2.75 (1H, dd, J = 17.2, 3.2, H-3eq), 3.06 (1H, dd,
J = 17.2, 12.8, H-3ax), 5.41 (1H, dd, J = 12.8, 2.9, H-
2), 6.00 (1H, s, H-8), 7.33 (1H, d, J = 4.6, H-40), 7.35–
7.40 (2H, m, H-30 and 50), 7.48 (1H, d, J = 7.4, H-20
and 60); and 32 mg of 11 (tR 19.7 min) as a yellow solid:
mp 140 ꢁC; EIHRMS m/z 274.0870 (calcd for C15H14O5,
274.0841); 1H NMR d (methanol-d4) ppm (J in Hz):
2.86–2.91 (2H, m, H-a), 3.27–3.34 (2H, m, H-b), 5.88
(1H, s, H-50), 7.18 (2H, d, J = 1.5, H-2 and 6), 7.24–
7.28 (3H, m, H-3, 4, and 5).
1
302.1154); H NMR d (chloroform-d) ppm (J in Hz):
3.78, 3.82, 3.88 (each 3H, s, OMe-2, 3, and 4), 4.37
(2H, s, H-8), 6.24 (1H, s, H-5), 7.23–7.26 (3H, m, H-
30, 40, and 50), 7.27–7.35 (2H, m, H-20 and 60).
3.1.12. 5,6,7-Trimethoxyisoflavone (8b). Formic acid
(9.2 g, 0.1 mol) was gradually added to a solution of
N,N0-carbonyldiimidazole (16.2 g, 0.1 mol) in THF
(100 mL) at 0 ꢁC. The resulting solution was stirred
for 1 h. Compound 8a (4.53 g, 15 mmol) in dry THF
(50 mL) was then added to it and stirred for 12 h at
0 ꢁC. The solution was then heated at 100 ꢁC for
30 min. The mixture was concentrated. The residue
was treated with ice and the solid obtained was purified
by silica gel column chromatography using hexane–ethyl
acetate (7:3) as the eluent to give 8b as yellow powders
(3.46 g, 74%): mp 156–158 ꢁC; EIHRMS m/z 312.0984
(calcd for C18H16O5, 312.0848); 1H NMR d (chloro-
form-d) ppm (J in Hz): 3.92, 3.96, 3.97 (each 3H, s,
OMe-5, 6, and 7), 6.70 (1H, s, H-8), 7.38 (1H, d,
J = 7.2, H-40), 7.41–7.44 (2H, m, H-30 and 50), 7.54
(2H, d, J = 6.9, H-20 and 60), 7.84 (1H, s, H-2).
3.1.9. 3-Hydroxy-5,6,7-methoxyflavone (7a). n-Butyl-
lithium (1.6 M solution in hexane, 5 mL, 8 mmol) was
added to diisopropylamine (1.2 mL, 8 mmol) in dry
THF (40 mL) at ꢀ80 ꢁC under argon gas. The LDA
solution was warmed to 0 ꢁC for 30 min and re-cooled
to ꢀ80 ꢁC. Compound 3a (2.5 g, 8 mmol) in dry THF
(80 mL) was added dropwise to the LDA solution and
stirred for 1 h. After an addition of trimethylborate
(0.92 mL, 8 mmol) in THF (20 mL), the resulting solu-
tion was then stirred at ꢀ80 ꢁC for 30 min. The reaction
mixture was acidified with 0.8 g of acetic acid, stirred for
15 min, and then oxidized with 30% hydrogen peroxide
(1.2 mL, 8 mmol). The solution was allowed to warm
slowly to room temperature and was then shaken with
3.1.13. 6-Hydroxy-2,3,4-trimethoxyphenyl p-methoxy-
benzyl ketone (9a). By the same method as 8a, 9a was
obtained from 3,4,5-trimethoxyphenol(3.7 g, 20 mmo)l
and p-methoxyphenylacetyl chloride (4.1 g, 22 mmol).
Chromatography on silica gel using hexane–ethyl ace-
tate (3:7) as the eluent gave 9a (4.78 g, 72%) as a yellow

1668
H. Gao, J. Kawabata / Bioorg. Med. Chem. 13 (2005) 1661–1671
pale solid: mp 95–96 ꢁC; EIHRMS m/z 332.1262 (calcd
for C18H20O6, 332.1260); 1H NMR d (chloroform-d)
ppm (J in Hz): 3.77, 3.78, 3.86, 3.96 (each 3H, s,
OMe-2, 3, 40, and 4), 4.29 (2H, s, H-8), 6.22 (1H, s,
H-5), 6.87 (2H, d, J = 8.4, H-30 and 50), 7.13 (2H, d,
J = 8.4, H-20 and 60).
phenol, 23%) as pale yellow powders: mp 157–158 ꢁC;
EIHRMS m/z 314.1176 (calcd for C18H18O5,
314.1154); H NMR d (chloroform-d) ppm (J in Hz):
2.78 (1H, dd, J = 16.7, 2.7, H-3eq), 3.00 (1H, dd,
J = 16.7, 13.3, H-3ax), 3.81, 3.86, 3.93 (each 3H, OMe-
5, 6, and 7), 5.39 (1H, dd, J = 13.3, 2.7, H-2), 6.34
(1H, s, H-8), 7.36–7.38 (2H, m, H-20 and 60), 7.40–7.44
(3H, m, H-30, 40, and 50).
1
3.1.14. 40,5,6,7-Tetramethoxyisoflavone (9b). By the
same method as 8b, 9b was obtained from 9a (3.32 g,
10 mmol). The crude solid was purified with silica gel
column chromatography using hexane–ethyl acetate
(7:3) as the eluent to 9b (2.46 g, 72%) as a yellow solid:
mp 175 ꢁC; EIHRMS m/z 342.1086 (calcd for C19H18O6,
Using the procedure described for 2, 5,6,7-trimethoxy-
flavanone (314 mg, 1 mmol) was demethylated to give
5,6,7-trihydroxyflavanone, which was spontaneously
converted to 20,30,40,60-tetrahydroxychalcone by ring-
opening under the reaction condition. The product
was further purified by preparative HPLC (mobile
phase, water–methanol–formic acid (30:70:0.1); flow
rate, 4.5 mL/min; detection, UV 254 nm) to give 75 mg
of 10 (tR 19.4 min, 28%) as yellow powders: mp
175 ꢁC; EIHRMS m/z 272.0643 (calcd for C15H12O5,
272.0685); 1H NMR d (methanol-d4) ppm (J in Hz):
5.92 (1H, s, H-50), 7.38–7.40 (3H, m, H-3, 4, and 5),
7.63 (2H, d, J = 1.7, H-2 and 6), 7.73 (1H, d, J = 15.5,
H-a), 8.25 (1H, d, J = 15.5, H-b).
1
342.1103); H NMR d (chloroform-d) ppm (J in Hz):
3.81, 3.89, 3.94, 3.94 (each 3H, s, OMe-40, 5, 6, and 7),
6.67 (1H, s, H-8), 6.94 (2H, d, J = 8.6, H-30 and 50),
7.45 (2H, d, J = 8.6, H-20 and 60), 7.79 (1H, s, H-2).
3.1.15. 5,6,7-Trihydroxyisoflavone (8) and 40,5,6,7-tetra-
hydroxyisoflavone (9). Using the procedure described for
2, 8b, and 9b was demethylated to give 8 (81%) and 9
(79%), respectively. Compound 8: yellow powders; mp
284 ꢁC; EIHRMS m/z 270.0569 (calcd for C15H10O5,
1
270.0528); H NMR d (DMSO-d6) ppm (J in Hz): 6.51
(1H, d, J = 1.7, H-8), 7.41–7.44 (3H, m, H-30, 40, and
50), 7.56 (2H, d, J = 8.1, H-20 and 60), 8.39 (1H, d,
J = 0.7, H-2); 13C NMR d (DMSO-d6) ppm: 94.5 (C-
8), 105.6 (C-10), 122.4 (C-10), 128.7 (C-6), 129.0 (C-20
and 60), 129.9 (C-30 and 50), 130.2 (C-40), 132.0 (C-3),
148.3 (C-5), 150.9 (C-9), 154.5 (C-7), 155.7 (C-2),
180.9 (C-4); 9: yellow powders; mp 260 ꢁC; EIHRMS
m/z 286.0493 (calcd for C15H10O6, 286.0477); 1H
NMR d (DMSO-d6) ppm (J in Hz): 6.47 (1H, d,
J = 1.7, H-8), 6.81 (2H, d, J = 1.7, H-30 and 50), 7.36
(2H, dd, J = 8.4, 1.7, H-20 and 60), 8.38 (1H, d,
J = 1.0, H-2); 13C NMR d (DMSO-d6) ppm: 93.5 (C-
8), 104.8 (C-10), 115.0 (C-30 and 50), 121.5 (C-3), 121.5
(C-10), 129.2 (C-6), 130.2 (C-20 and 60), 147.4 (C-5),
150.0 (C-9), 153.6 (C-7), 153.9 (C-2), 157.3 (C-40),
180.4 (C-4).
3.1.17. 1,2,3-Trihydroxyxanthone (14). A mixture of sali-
cylic acid (2.07 g, 15 mmol), 20,40,60-trihydroxyacetoph-
enone (2.52 g, 15 mmol), anhydrous ZnCl2 (6 g), and
POCl3 (20 mL) were heated at 70–80 ꢁC for 2 h. The
reaction mixture was cooled, poured into ice-cooled
water, and extracted with ethylacetate (100 mL · 3).
The extracts were combined, washed with water, dried
over Na2SO4, and concentrated. The residue was puri-
fied by silica gel column chromatography using hex-
ane–ethylacetate (7:3) as the uelent to give 1,3-
dihydroxy-2-acetylxanthone (1.74 g, 43%) as pale yellow
powders: mp 252–254 ꢁC; EIHRMS m/z 270.0529 (calcd
for C15H10O5, 270.0528); 1H NMR d (chloroform-d)
ppm: 2.79 (3H, s, Ac-2), 6.36 (1H, s, H-4), 7.40–7.41
(1H, m, H-5), 7.42–7.44 (1H, m, H-7), 7.73 (1H, m, H-
6), 8.22 (1H, d, J = 7.6, H-8), 14.33 (1H, s, OH-3),
15.27 (1H, s, OH-1).
3.1.16. 20,30,40,60-Tetrahydroxychalcone (10). A mixture
of 3,4,5-trimethoxyphenol(3.7 g, 20 mmo)l and cinna-
moyl chloride (3.7 g, 22 mmol) and BF3–Et2O complex
(20 mL) was refluxed for 15 min, cooled, and quenched
with excess of water. The resulting mixture was ex-
tracted with ethylacetate (200 mL · 3) and the com-
bined organic extracts were washed with brine, dried
over anhydrous Na2SO4, and concentrated to provide
crude 60-hydroxy-20,30,40-trimethoxychalcone (5.15g,
crude yield 82%), which was used without further
purification.
1,3-Dihydroxy-2-acetylxanthone (0.755 g, 2.79 mmol)
was dissolved in a mixture of 6 mL of 4% NaOH and
6 mL of pyridine and the mixture was cooled in an ice
bath. Hydrogen peroxide (6 mL, 20%) was added drop-
wise with stirring at 5 min. The reaction mixture was left
for 1 h. Acidification with 10 mL of 1 M HCl(10 mmo)l
yielded 14 as a yellow solid, which was crystallized from
ethanolto give 14 (442 mg, 65%) as yellow powders: mp
268–270 ꢁC; EIHRMS m/z 244.0345 (calcd for
1
C13H8O5, 244.0371); H NMR d (acetone-d6) ppm (J
in Hz): 6.54 (1H, s, H-4), 7.44 (1H, t, J = 7.9, H-7),
7.53 (1H, d, J = 7.9, H-5), 7.83 (1H, t, J = 7.9, H-6),
8.20 (1H, dd, J = 7.9, 1.7, H-8), 12.75 (1H, s, OH-1).
Crude 60-hydroxy-20,30,40-trimethoxychalcone (3.77 g)
was added to a stirred solution of KF (1 g) in methanol
(300 mL) and the reaction mixture was refluxed for 24 h.
The reaction mixture was diluted with water and ex-
tracted with chloroform (300 mL · 3). The combined
organic extracts were washed with brine, dried over
anhydrous Na2SO4, and concentrated. The residue was
purified by silica gel column chromatography using hex-
ane–ethylacetate (7:3) as the eulent to give 1.06 g of
5,6,7-trimethoxyflavanone (yield from 3,4,5-trimethoxy-
3.1.18. 7-Hydroxy-1,2,3,6-tetramethoxyxanthone (15a).
A mixture of 2,5-dihydroxy-4-methoxybenzoic acid
(1.98 g, 10 mmol), 3,4,5-trimethoxyphenol (3.13 g,
17 mmol), anhydrous ZnCl2 (15 g), and POCl3 (26 mL)
were heated at 70–80 ꢁC for 2 h. Treatment following
the usualprocedure gave 15a (0.83 g, 25%) as yellow
powders: mp 186–190 ꢁC; EIHRMS m/z 332.0941 (calcd

H. Gao, J. Kawabata / Bioorg. Med. Chem. 13 (2005) 1661–1671
1669
for C17H16O7, 332.0954); 1H NMR d (chloroform-d)
ppm: 3.93, 3.94, 3.96, 4.00 (each 3H, OMe-1, 2, 3, and
6), 6.38 (1H, s, H-4), 6.97 (1H, s, H-5), 7.53 (1H, s, H-8).
residue was crystallized from hexane–ethyl acetate to
give 16 (51 mg, 55%) as yellow powders: mp 263 ꢁC;
EIHRMS m/z 269.0713 (calcd for C15H11O4N,
1
269.0688); H NMR d (DMSO-d6) ppm (J in Hz): 6.62
3.1.19. 1,2,3,6,7-Pentahydroxyxanthone (15). A mixture
of 15a (166 mg, 0.5 mmol), acetic anhydride (5 mL),
and hydroiodic acid (d = 1.6, 2 mL) was heated under
reflux at 150 ꢁC for 2 h. The cooled mixture was poured
into a saturated solution of sodium hydrogen sulfite.
The mixture was extracted with ethylacetate
(50 mL · 3). The combined organic extracts were
washed with brine, dried over anhydrous Na2SO4, and
concentrated. The residue was dissolved in a small
amount of methanoland subjected to preparative
HPLC (mobile phase, water–methanol–formic acid
(30:70:0.1); flow rate, 4.5 mL/min; detection, UV
254 nm) to give 36 mg of 15 (tR 11.4 min, 26%) as yellow
powders: mp 284 ꢁC; EIHRMS m/z 276.0267 (calcd for
C13H8O7, 276.0269); 1H NMR d (methanol-d4) ppm:
6.33 (1H, s, H-4), 6.92 (1H, s, H-5), 7.44 (1H, s, H-8).
(1H, s, H-8), 6.92 (1H, s, H-3), 7.54–7.59 (3H, m, H-
30, 40, and 50), 8.05 (2H, d, J = 7.1, H-20 and 60), 12.65
(1H, s, OH-5); 13C NMR d (DMSO-d6) ppm: 94.0 (C-
8), 104.2 (C-10), 104.5 (C-3), 126.3 (C-20 and 60), 129.1
(C-30 and 50), 129.4 (C-6), 131.0 (C-10), 131.8 (C-40),
146.8 (C-5), 149.9 (C-9), 154.0 (C-7), 162.8 (C-2),
182.1 (C-4).
3.1.24. 8-Amino-5,7-dihydroxyflavone (17). By the same
method described for 16c, 16a was hydrogenated to give
17 (yield, 79%) as yellow powders: mp 244–246 ꢁC;
EIHRMS m/z 269.0663 (calcd for C15H11O4N,
1
269.0688); H NMR d (DMSO-d6) ppm (J in Hz): 6.29
(1H, s, H-6), 6.90 (1H, s, H-3), 7.54–7.62 (3H, m, H-
30, 40, and 50), 8.21 (2H, d, J = 7.6, H-20 and 60), 12.07
(1H, s, OH-5); 13C NMR d (DMSO-d6) ppm: 98.4 (C-
6), 103.7 (C-10), 104.5 (C-3), 116.5 (C-8), 126.7 (C-20
and 60), 129.0 (C-30 and 50), 130.9 (C-10), 131.9 (C-40),
143.3 (C-7), 151.3 (C-5), 151.4 (C-9), 162.9 (C-2),
182.4 (C-4).
3.1.20. 5,7-Dihydroxy-8-nitroflavone (16a). Nitric acid
(d = 1.42, 1 mL) in acetic acid (10 mL) was slowly added
to a solution of 5,7-dihydroxyflavone (2.54 g, 10 mmol)
in acetic acid (20 mL). The mixture was stirred for 1 h at
65 ꢁC and the mixture was then poured into ice-cooled
water (500 mL). The precipitated solid were filtered,
washed with water, and crystallized from ethanol to give
16a (2.1 g, 70%) as yellow powders: mp 224 ꢁC;
EIHRMS m/z 299.0449 (calcd for C15H9O6N,
3.1.25. 5,7-Dihydroxy-6,8-dinitroflavone (21a). Nitric
acid (d = 1.42, 5 mL) in acetic acid (10 mL) was slowly
added to a solution of 5,7-dihydroxyflavone (2.54 g,
10 mmol) in acetic acid (20 mL). The mixture was stirred
for 10 h at 65 ꢁC and the mixture was then poured into
ice-cooled water (500 mL). The precipitated solid were
filtered, washed with water, and crystallized from etha-
nolto give 21a (1.8 g, 52%) as pale yellow plates: mp
231 ꢁC; EIHRMS m/z 344.0305 (calcd for C15H8O8N2,
1
299.0430); H NMR d (chloroform-d) ppm (J in Hz):
6.38 (1H, s, H-6), 7.21 (1H, s, H-3), 7.61–7.64 (3H, m,
H-30, 40, and 50), 7.99 (2H, dd, J = 8.3, 1.4, H-20 and 60).
1
3.1.21. 5,7-Dimethoxy-8-nitroflavone (16b). By the
method described for 3a, 16a was methylated to give
16b (yield, 85%) as pale yellow powders: mp 207–
208 ꢁC; EIHRMS m/z 327.0751 (calcd for C17H13O6N,
344.0280); H NMR d (chloroform-d) ppm (J in Hz):
6.91 (1H, s, H-3), 7.54–7.59 (3H, m, H-30, 40, and 50),
7.88 (2H, d, J = 7.4, H-20 and 60).
1
327.0743); H NMR d (chloroform-d) ppm (J in Hz):
3.1.26. 6,8-Diamino-5,7-dihydroxyflavone (21). Zinc dust
(100 mg) and acetic acid (20 mL) were added succes-
sively to a stirred solution of 21a (200 mg) in CH2Cl2
(80 mL) at room temperature. After 1 h, the reaction
mixture was filtered and 200 mL of CH2Cl2 was added
to the filtrate. The solution was washed with water,
dried over anhydrous Na2SO4, passed through a plug
of silica gel, and concentrated. The residue was crystal-
lized from hexane–ethyl acetate to give 21 (117 mg, 71%)
as pale purple powders: mp 182 ꢁC; EIHRMS m/z
4.01, 4.08 (each 3H, s, OMe-5 and 7), 6.81 (1H, s, H-
3), 6.92 (1H, s, H-6), 7.56–7.59 (3H, m, H-30, 40, and
50), 7.87 (2H, dd, J = 8.1, 1.7, H-20 and 60).
3.1.22. 8-Amino-5,7-dimethoxyflavone (16c). A solution
of 16b (200 mg) in THF (100 mL) was hydrogenated
for 10 h in the presence of 10% palladium on charcoal
(200 mg) at room temperature. The catalyst was re-
moved by filtration and the filtrate was concentrated
to give 16c (yield 92%) as yellow powders: mp 215 ꢁC;
EIHRMS m/z 297.0968 (calcd for C17H15O4N,
1
284.0820 (calcd for C15H12O4N2, 284.0798); H NMR
d (DMSO-d6) ppm (J in Hz): 6.91 (1H, s, H-3), 7.54–
7.61 (3H, m, H-30, 40, and 50), 8.21 (2H, d, J = 7.1, H-
20 and 60).
1
297.1002); H NMR d (chloroform-d) ppm: 3.96, 4.00
(each 3H, s, OMe-5 and 7), 6.65 (1H, s, H-6), 7.26
(1H, s, H-3), 7.50–7.52 (3H, m, H-30, 40, and 50), 7.86–
7.88 (2H, m, H-20 and 60).
3.1.27. 5-Hydroxy-6-nitroflavone (18a). A mixture of 2,6-
dihydroxy-3-nitroacetophenone (3 g), sodium benzoate
(4.25 g), benzoic anhydride (15 g), and pyridine
(1.0 mL) was heated at 150–160 ꢁC for 9 h under argon
gas. After cooling to room temperature, the reaction
mixture was poured into water (100 mL) and extracted
with chloroform (100 mL · 3). The combined organic
phase was washed with water, dried over anhydrous
Na2SO4, and concentrated. The residue was refluxed
with alcoholic potassium hydroxide (5%, 100 mL) for
3.1.23. 6-Amino-5,7-dihydroxyflavone (16). Compound
16c (100 mg) was heated with hydroiodic acid (d = 1.6,
10 mL) in a sealed tube at 175–180 ꢁC for 3 h, and the
reaction mixture was poured into 10% sodium hydrogen
sulfate. The solution was neutralized by an addition of
1 M NaOH and the mixture was extracted with 1-buta-
nol(100 mL · 3). The extracts were washed with water,
dried over anhydrous Na2SO4, and concentrated. The

1670
H. Gao, J. Kawabata / Bioorg. Med. Chem. 13 (2005) 1661–1671
3 h under argon gas. After acidified with 1 M HCl, the
mixture was extracted with chloroform (100 mL · 3).
The extract was washed with water, dried over anhy-
drous Na2SO4, and concentrated. The residue was sub-
jected to silica gel column chromatography using
prepared from rat intestinalacetone powder (Sigma–
Aldrich Japan Co., Tokyo, Japan), was used as the small
intestinal a-glucosidases, sucrase, maltase, and isomalt-
ase. The reaction mixture consisted of crude enzyme
solution (as sucrase, 0.2 mL; as maltase, 0.05 mL; as iso-
maltase, 0.2 mL), substrate (sucrose: 56 mM, 0.2 mL;
maltose: 5 mM, 0.35 mL; isomaltose: 7 mM, 0.2 mL,
respectively), in 0.1 M potassium phosphate buffer
(pH 6.3), and the test sample in 50% aqueous dimethyl
sulfoxide (0.1 mL). After incubation for 15 min at
37 ꢁC, the reaction was stopped by adding 0.75 mL of
2 M Tris–HClbuffer (pH 7.0). The reaction mixture
was passed through a short column of basic alumina
(ICN Alumina B, grade, ICN Biomedical GmbH, Esc-
hwege, Germany) to remove phenolic compounds,
which might interfere with glucose measurement.17
The amount of liberated glucose was measured by the
glucose oxidase method using a commercial test kit
(Glucose B test Wako, Wako Pure Chem. Co., Osaka,
Japan). For kinetic analyses of sucrase and maltase by
6-amino-5,7-dihydroxyflavone (16), the enzyme and 16
(10 and 20 lM for sucrase; 20 and 40 lM for isomaltase)
were incubated with increasing concentrations of su-
crose (7–56 mM) or maltose (1.25–15 mM), respectively.
The concentration of inhibitors required for inhibiting
50% of the a-glucosidase activity under the assay condi-
tions was defined as the IC50 value. The IC50 value was
measured graphically by a plot of percent inhibition ver-
sus log of the test compound.
hexane–ethylacetate (8:2) as the eulent to give
18a
(1.6 g, 40%) as pale yellow powders: mp 210 ꢁC;
EIHRMS m/z 283.0443 (calcd for C15H9O5N,
1
283.0481); H NMR d (chloroform-d) ppm (J in Hz):
6.86 (1H, s, H-3), 7.10 (1H, d, J = 9.3, H-8), 7.57–7.60
(3H, m, H-30, 40, and 50), 7.94 (2H, dd, J = 6.9, 1.2, H-
20 and 60), 8.42 (1H, d, J = 9.3, H-7).
3.1.28. 6-Amino-5-hydroxyflavone (18). By the same pro-
cedure as described for 21, 18a (566 mg, 2 mmol) was re-
duced by Zn–AcOH to give 18 (330 mg, 65%) as yellow
powders: mp 183–185 ꢁC; EIHRMS m/z 253.0743 (calcd
for C15H11O3N, 253.0739); 1H NMR d (DMSO-d6) ppm
(J in Hz): 6.98 (1H, s, H-3), 7.07 (1H, d, J = 8.9, H-7),
7.15 (1H, d, J = 8.9, H-8), 7.56–7.61 (3H, m, H-30, 40,
and 50), 8.09 (2H, d, J = 6.6, H-20 and 60).
3.1.29. 6-Amino-7-hydroxyflavone (19). According to the
method using preparation of 18a, 7-hydroxy-6-nitroflav-
one (19a) was obtained by the reaction of 4,6-dihy-
droxy-3-nitroacetophenone with sodium benzoate and
benzoic anhydride to give 19a (yield 35%) as pale yellow
powders: mp 234 ꢁC; EIHRMS m/z 283.0489 (calcd for
1
C15H9O5N, 283.0481); H NMR d (chloroform-d) ppm
(J in Hz): 6.79 (1H, s, H-3), 7.26, 7.27 (each 1H, s, H-
5 and 8), 7.54–7.59 (3H, m, H-30, 40, and 50), 7.90 (2H,
d, J = 6.9, H-20 and 60).
3.2.2. b-Glucosidase, a-mannosidase, a-galactosidase, and
b-galactosidase inhibitory activity determination. The
inhibitory activity of b-glucosidase (Sigma, G0395;
almonds), a-mannosidase (Sigma, M7257; Jack bean),
a-galactosidase (Sigma, G8507; Green coffee bean), or
b-galactosidase (Sigma, G1875; Bovine liver) was per-
formed with a slight modification of the literature proce-
dure.38 The enzymatic hydrolysis of substrates was
monitored by the amount of o- or p-nitrophenolreleased
in the reaction mixture: b-glucosidase (6 lg/mL) and
p-nitrophenyl b-D-glucopyranoside (4 mM) in 50 mM
sodium citrate buffer (pH 5.0); a-mannosidase (10 lg/
mL) and p-nitrophenyl a-D-mannopyranoside (4 mM)
in 50 mM sodium citrate buffer (pH 4.5); a-galactosidase
(3.8 lg/mL) and o-nitrophenyl a-D-galactopyranoside
(4 mM) in 50 mM potassium phosphate buffer
(pH 6.5); b-galactosidase (1.4 mg/mL) and o-nitrophenyl
b-D-galactopyranoside (40 mM) in 40 mM potassium
phosphate buffer (pH 7.3). After the mixture of enzyme
solution (0.2 mL) and substrate (0.2 mL) was preincu-
bated for 5 min at 37 ꢁC, 0.2 mL of test compounds in
50% DMSO was added and then incubated for 15 min
at 37 ꢁC. After stopping the reaction by addition of
1 M Na2CO3 (0.4 mL), the amount of o-, or p-nitrophen-
ol in the mixture was measured spectrophotometrically
at 405 nm.
By the same procedure as described for 21, 19a was
reduced by Zn–AcOH to give 19 (yield 71%) as yellow
powders: mp 209 ꢁC; EIHRMS m/z 253.0712 (calcd for
C15H11O3N, 253.0739); 1H NMR d (DMSO-d6) ppm
(J in Hz): 6.77 (1H, s, H-3), 6.92 (1H, s, H-5), 7.12
(1H, s, H-8), 7.54–7.59 (3H, m, H-30, 40, and 50), 7.90–
8.02 (2H, m, H-20 and 60).
3.1.30. 6,8-Diamino-7-hydroxyflavone (20). Using the
procedure for preparation of 21a, 7-hydroxyflavone
was nitrated to give 7-hydroxy-6,8-dinitroflavone 20a
(yield 72%) as yellow powders: mp 289–293 ꢁC;
EIHRMS m/z 328.0309 (calcd for C15H8O7N2,
1
328.0331); H NMR d (chloroform-d) ppm (J in Hz):
7.01 (1H, s, H-3), 7.56 (1H, s, H-5), 7.55–7.61 (3H, m,
H-30, 40, and 50), 7.96 (2H, d, J = 7.9, H-20 and 60).
By the same procedure as described for 21, 20a was re-
duced using Zn–AcOH to give 20 (yield 68%) as yellow
powders: 223 ꢁC; EIHRMS m/z 268.0876 (calcd for
1
C15H12O3N2, 268.0849); H NMR d (DMSO-d6) ppm
(J in Hz): 6.59 (1H, s, H-3), 6.76 (1H, s, H-5), 7.52–
7.54 (3H, m, H-30, 40, and 50), 8.15 (2H, d, J = 8.6, H-
20 and 60).
3.2.3. Porcine pancreatic a-amylase inhibitory activity
determination. The inhibitory effect of the compounds
on a-amylase activity was assessed by the usual meth-
od.39 Briefly, starch azure (Sigma; 2.0 mg) used as a sub-
strate was suspended in 50 mM Tris–HClbuffer
(pH 6.9) containing 10 mM CaCl2, and boiled for
3.2. Biology
3.2.1. a-Glucosidase inhibitory activity determination.
The a-glucosidase inhibitory activity was measured as
described previously.17,37 The crude enzyme solution

H. Gao, J. Kawabata / Bioorg. Med. Chem. 13 (2005) 1661–1671
1671
Deesamer, S.; Yibchok-anun, S. Bioorg. Med. Chem. Lett.
2004, 14, 2893–2896.
15. Nishioka, T.; Kawabata, J.; Aoyama, Y. J. Nat. Prod.
1998, 61, 1413–1415.
16. Kawabata, J.; Mizuhata, T.; Sato, E. Biosci. Biotechnol.
Biochem. 2003, 67(2), 445–447.
17. Hong, G.; Nishioka, T.; Kawabata, J.; Kasai, T. Biosci.
Biotechnol. Biochem. 2004, 68(2), 369–375.
18. Hong, G.; Kawabata, J. Biosci. Biotechnol. Biochem. 2004,
68(9), 281–287.
5 min at 100 ꢁC. Then, the starch azure solution was
preincubated for 10 min at 37 ꢁC. The test samples in
50% DMSO, and 0.2 mL of porcine pancreatic a-amy-
lase solution (Sigma, A-6255; 2.0 U/mL; 50 mM Tris–
HClbuffer containing 10 mM CaCl 2, pH 6.9) were
added into each assay. The reaction was carried out at
37 ꢁC for 10 min and stopped by adding 0.5 mL of
50% acetic acid. The reaction mixture was then centri-
fuged at 2000 rpm for 5 min at 4 ꢁC. The absorbance
of the supernatant at 595 nm was measured.
19. Cushman, M.; Nagarathnam, D.; Geahlen, R. L. J. Nat.
Prod. 1991, 54, 1345–1352.
20. Cushman, M.; Zhu, H.; Geahlen, R.; Kraker, A. J. J.
Med. Chem. 1994, 37, 3352–3362.
21. Agasimundin, Y. S.; Siddappa, s. J. Chem. Soc., Perkin
Trans. 1 1973, 503–505.
22. Handjeri, M.; Mariotte, A. M.; Boumendjel, A. Chem.
Pharm. Bull. 2001, 49, 1352–1355.
23. Zheng, X.; Meng, W. D.; Xu, Y. Y.; Cao, J. G.; Qing, F.
L. Bioorg. Med. Chem. Lett. 2003, 13, 881–884.
24. Varma, R. S.; Kumar, D. Org. Lett. 1999, 1, 697–700.
25. Gleye, C.; Lewin, G.; Laurens, A.; Jullian, J. C.; Loiseau,
P.; Bories, C.; Hocquemiller, R. J. Nat. Prod. 2003, 66,
690–692.
Acknowledgements
We are gratefulto Mr. Kenji Watanabe and Dr. Eri
Fukushi, GC–MS & NMR Laboratory, Faculty of
Agriculture, Hokkaido University, for their skillful mass
spectralmeasurements.
References and notes
1. Lebovitz, H. Clin. Diabetes 1995, 13, 99–103.
2. Schmidit, D.; Frommer, W.; Junge, B.; Muller, L.;
Wingender, W.; Truscheit, E.; Schafer, D. Naturwissens-
chaften 1977, 64, 535–536.
26. McGarry, L.; Detty, M. R. J. Org. Chem. 1990, 55, 4349–
4356.
27. Huang, W. H.; Chen, P. Y.; Yang, C. H.; Lee, A. R.
Chem. Pharm. Bull. 2003, 51, 339–340.
3. Matsuo, T.; Odaka, H.; Ikeda, H. Am. J. Clin. Nutr. 1992,
55, 314s–317s.
28. Krishnamurty, H. G.; Prasad, J. S. Tetrahedron Lett. 1977,
35, 3071–3072.
4. Asano, N.; Tomioka, E.; Kizu, H.; Matsui, K. Carbohydr.
Res. 1994, 253, 235–245.
29. Rao, Y. K.; Rao, C. V.; Kishore, P. H.; Gunasekar, D. J.
Nat. Prod. 2001, 64, 368–369.
5. Asano, N.; Kizu, H.; Oseki, K.; Tomioka, E.; Matsui, K.
J. Med. Chem. 1995, 38, 2349–2356.
30. Agasimundin, Y. S.; Rajagopal, S. J. Org. Chem. 1971, 36,
845–846.
6. Hoffmann, J.; Spengler, M. Diabetes Care 1994, 17, 561–
566.
31. Shaw, S. C.; Azad, R.; Mandal, S. P.; Gandhi, R. S. J.
Indian Chem. Soc. 1988, LXV, 107–109.
7. Shinozaki, K.; Suzuki, M.; Ikebuchi, M.; Hirose, J.;
Harano, Y. Metabolism 1996, 6, 731–737.
8. Yoshikuni, Y. Agric. Biol. Chem. 1988, 52, 121–128.
9. Joubert, P. H.; Bam, W. J.; Manyane, N. Eur. J. Clin
Pharmacol. 1986, 30, 253–2555.
32. Larget, R.; Lockhart, B.; Renard, P.; Largeron, M.
Bioorg. Med. Chem. Lett. 2000, 10, 835–838.
33. Hu, C. Q.; Chen, K.; Shi, Q. J. Nat. Prod. 1994, 57, 42–51.
34. Morrris, G. A.; Mullah, K. B.; Sutherland, S. Tetrahedron
1986, 42, 3303–3309.
10. Haslam, E. Practical Polyphenols; Cambridge University
Press: Cambridge, 1998; pp 168–174.
35. Seshadri, S.; Trivedi, P. L. J. Org. Chem. 1958, 23, 1735–
1738.
11. Lee, D. S.; Lee, S. H. FEBS Lett. 2001, 501, 84–
86.
12. Niwa, T.; Doi, U.; Osawa, T. J. Agric. Food Chem. 2003,
51, 90–94.
13. Babu, K. S.; Tiwari, A. K.; Srinivas, P. V.; Ali, A. Z.;
Raju, B. C.; Rao, M. Bioorg. Med. Chem. Lett. 2004, 14,
3841–3845.
36. Rossi, M.; Mayer, R.; Constantinou, P.; Caruso, F.;
Castelbuono, D.; Brien, M. O.; Narasimhan, V. J. Nat.
Prod. 2001, 64, 26–31.
37. Toda, M.; Kawabata, J.; Kasai, T. Biosci. Biotechnol.
Biochem. 2000, 64(2), 294–298.
38. Brocker, E. R.; Benn, M. H. Phytochemistry 1983, 22,
770–772.
14. Adisakwattana, S.; Sookkongwaree, K.; Roengsumran,
S.; Petsom, A.; Ngamrojnavanich, N.; Chavasiri, W.;
39. Hansawasdi, C.; Kawabata, J.; Kasai, T. Biosci. Biotech-
nol. Biochem. 2001, 65(9), 2087–2089.