Novel anticancer agents, kayeassamins C-I from the flower of Kayea assamica of Myanmar

Article abstract of DOI:10.1016/j.bmc.2008.07.091

A CHCl₃-soluble fraction of 70% EtOH extract of the flower of Kayea assamica from Myanmar exhibited 100% preferential cytotoxicity (PC₁₀₀) against human pancreatic cancer PANC-1 cells under nutrient-deprived conditions at 1 μg/mL. Bioassay-guided fractionation and isolation afforded nine new coumarins, kayeassamins A (8), B (9), and C-I (1-7), together with nine known coumarins (10-18). The structures of these compounds were identified by extensive spectroscopic techniques as well as by comparison with published data. Absolute configuration at C-1′ of 1 was established as S-configuration by the modified Mosher method. All the isolates were evaluated for their in vitro preferential cytotoxicity using novel anti-austerity strategy. Among them, the novel coumarins, kayeassamins A (8), B (9), D (2), E (3), and G (5) exhibited the most potent preferential cytotoxicity (PC₁₀₀ 1 μM) in a concentration- and time- dependent manner and induced apoptosis-like morphological changes of PANC-1 cells within 24 h of treatment. Based on the observed cytotoxicity, structure-activity relationships have been established.

Full text of DOI:10.1016/j.bmc.2008.07.091

Bioorganic & Medicinal Chemistry 16 (2008) 8653–8660
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Novel anticancer agents, kayeassamins CꢀI from the flower of Kayea assamica
of Myanmar
Nwet Nwet Win a, Suresh Awale a, Hiroyasu Esumi b, Yasuhiro Tezuka a, Shigetoshi Kadota a,
*
a Institute of Natural Medicine, University of Toyama, 2630-Sugitani, Toyama 930-0194, Japan
b National Cancer Center Research Institute East, 6-5-1 Kashiwa, Chiba 277-8577, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
A CHCl3-soluble fraction of 70% EtOH extract of the flower of Kayea assamica from Myanmar exhibited 100%
Received 2 July 2008
Revised 30 July 2008
Accepted 31 July 2008
Available online 6 August 2008
preferential cytotoxicity (PC100) against human pancreatic cancer PANC-1 cells under nutrient-deprived
conditions at 1 lg/mL. Bioassay-guided fractionation and isolation afforded nine new coumarins, kayeassa-
mins A (8), B (9), and CꢀI (1ꢀ7), together with nine known coumarins (10ꢀ18). The structures of these com-
pounds were identified by extensive spectroscopic techniques as well as by comparison with published
data. Absolute configuration at C-10 of 1 was established as S-configuration by the modified Mosher method.
All the isolates were evaluated for their in vitro preferential cytotoxicity using novel anti-austerity strategy.
Among them, the novel coumarins, kayeassamins A (8), B (9), D (2), E (3), and G (5) exhibited the most potent
Keywords:
Anti-austerity strategy
Human pancreatic PANC-1 cells
Anti-cancer agent
Kayeassamin
preferential cytotoxicity (PC100 1 lM) in a concentration- and time- dependent manner and induced apop-
tosis-like morphological changes of PANC-1 cells within 24 h of treatment. Based on the observed cytotox-
icity, structure-activity relationships have been established.
Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction
CHCl3-soluble fraction of 70% EtOH extract of the flower of Kayea
assamica King & Prain (Clusiaceae) collected from Myanmar,
Pancreatic cancer is the fifth leading cause of cancer death with
a median survival of <6 months and a relative 5-year survival rate
of 5.5%. In 2005, it was estimated that 22,926 men and women died
of pancreatic cancer in Japan.1,2 It is most intrinsically resistant to
the conventional anticancer drugs in clinical use such as 5-fluoro-
uracil, taxol, doxorubicin, cisplatin and camptothecin, and is a ma-
jor cause of treatment failure in pancreatic cancer. Therefore, the
development of effective adjunct strategies is urgently necessary.
Gemcitabine currently represents the standard chemotherapeutic
drug for metastatic and advanced disease, but it only leads to a
modest improvement in quality of life and survival.3 Pancreatic tu-
mors are known to have poor angiogenesis that leads to insuffi-
cient nutrition supply to the aggressively proliferating cancer
cells. However, pancreatic cancer cells have inherent tolerance to
survive under low nutrient conditions. Therefore, the ability of can-
cer cells to tolerate nutrient starvation (austerity) is regarded as
another critical factor for tumor progression under hypovascular
conditions. Hence, an agent that can retard the cancer cells’ toler-
ance to nutrient starvation (anti-austerity agent) was considered
as a novel target in anticancer drug discovery.4–7
showed potent cytotoxicity to PANC-1 cells preferentially in nutri-
ent-deprived conditions at 1 g/mL. K. assamica is a slow growing,
l
tall, evergreen tree and blooms from early in October to May. It is
locally known as ‘Tharapi’ and has been used to reduce extreme
hotness in the body, dizziness, dry skin, and fever. In India, the
fruits of this species are used as a fish poison, and the aqueous ex-
tract of the stem bark is used as remedy for treating fevers. The
pollen is used for sores, fistulas, fever, and malaria.13 Previous
investigation of this plant reported cytotoxic and antimalarial
alkylated coumarins from the bark, root bark, and fruit.14 In the
present study, we carried out bioassay-guided fractionation and
isolation to identify preferentially cytotoxic anticancer agents,
which afforded nine novel coumarins, kayeassamins A–I, together
with nine known ones. Among the new compounds, kayeassamins
A (8) and B (9) were reported in our preliminary communication.15
In this paper, we report the isolation and identification of seven no-
vel anticancer agents, kayeassamins C–I (1–7), together with the
preferential cytotoxic activity of the isolated compounds.
2. Results and discussions
In our continuing program to discover new anticancer agents
based on a novel ‘anti-austerity strategy’,8–12 we found that the
2.1. Isolation and identification
The 70% EtOH extract of the flowers of K. assamica showed 100%
preferential cytotoxicity (PC100) against PANC-1 cancer cells under
* Corresponding author. Tel.: +81 76 434 7625; fax: +81 76 434 5059.
E-mail address: kadota@inm.u-toyama.ac.jp (S. Kadota).
0968-0896/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2008.07.091

8654
N. N. Win et al. / Bioorg. Med. Chem. 16 (2008) 8653–8660
Table 2
nutrient-deprived conditions at 1 lg/mL. Thus, it was separated
into the CHCl3-soluble and -insoluble fractions. Among them, only
13C NMR data (CDCl3, 100 MHz) for kayeassamins CꢀG (1ꢀ5)
the CHCl3-soluble fraction exhibited preferential cytotoxicity at
Carbon
1
2
3
4
5
1 lg/mL. This was therefore subjected to a series of chromato-
2
3
4
4a
5
6
7
8
159.7
109.0
156.3
100.7
156.2
104.2
166.3
113.8
157.7
77.5
28.2
10.5
210.4
47.1
16.6
27.2
11.7
21.9
121.2
137.4
16.3
39.8
26.7
159.8
109.1
156.4
100.6
156.2
104.3
166.5
113.9
157.7
77.7
160.8
108.9
156.8
100.6
156.2
103.9
166.7
114.8
157.5
78.4
159.2
108.8
155.9
100.3
155.4
103.8
165.9
113.5
157.2
77.4
27.7
10.1
209.9
46.6
16.2
159.9
108.9
156.4
100.7
156.2
104.3
166.6
114.2
157.7
77.9
graphic separation which led to the isolation of nine new couma-
rins, kayeassamins CꢀI (1ꢀ7), A (8)15 and B (9),15 and nine
known ones [mammea A/AA cyclo D (10),16,17 mammea A/BC
(11),18 mammea B/AC (12),19,20 mammea A/AC (13),21,22 mammea
A/AC cyclo D (14),21,22 a mixture of theraphins B and C (15 and
16),14 mammea B/AC cyclo F (17),17,19 and deacetylmammea E/
BA cyclo D (18)23]. Their structures were determined by means
of extensive spectroscopic techniques and by comparison of their
8a
10
20
28.1
10.5
205.5
53.7
27.5
10.5
27.9
10.5
205.5
53.7
published data.
30
23
100
200
300
400
500
1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
205.2
46.8
Kayeassamin C (1) was obtained as pale yellow oil with [a]
D
ꢀ60.87°. The molecular formula of 1 was established to be
27H36O6 by HREIMS. The IR spectrum exhibited the absorption
of hydroxyl group (3480 cmꢀ1),
,b-unsaturated lactone
25.4
17.9a
13.8
25.4
C
22.6a
22.6a
21.9
26.8
11.3
22.3
120.9
133.2
25.4
22.6a
22.6a
22.0
a
(1720 cmꢀ1), and chelated acyl group (1610 cmꢀ1). The UV spec-
trum showed absorption maxima at 220 and 329 nm in EtOH, sim-
ilar to those of 5,7-dihydroxy coumarins.20 The 1H NMR spectrum
of 1 displayed signals of an oxygenated methine (dH 4.78, H-10),
three olefinic methines (dH 5.06, H-7000; 5.22, H-2000; 6.14, H-3),
one alkyl methine (dH 3.78, H-200), five methylenes (dH 1.44, 1.85,
H2-400; 1.76, 1.94, H2-20; 1.99, H2-5000; 2.06, H2-6000; 3.42, H2-1000),
two primary and one secondary methyls (dH 0.96, H3-500; 1.02,
H3-30; 1.23, H3-300), three vinyl methyls (dH 1.57, H3-10000; 1.65,
H3-9000; 1.81, H3-4000), and three hydroxyl protons (dH 14.27, 7-OH;
10.51, 5-OH; 4.18, 10-OH) (Table 1). Meanwhile, 13C NMR spectrum
of 1 indicated 27 carbons, including those for a ketone carbonyl
carbon, a lactone carbonyl carbon, six aromatics, one oxygenated
methine, three olefinic methines, one alkylated methine, five
methylenes, six methyls, and three quaternary olefinic carbons
(Table 2). Furthermore, COSY and HMBC correlations (Fig. 1a) indi-
cated the presence of a hydroxypropyl, a 2-methyl-1-oxobutyl, and
a geranyl substituent. These data match well to those of surangin C,
previously isolated from the Mammea longifolia24 and the yellow
batai (Peltophorum dasyrachis).25 However, they differ from each
other due to the presence of an additional hydroxyl proton at dH
10.51 (5-OH) in 1, a characteristic of 5,7-dihydroxy-6-acylcouma-
rin.20 A bathochromic shift observed after the addition of alkali
in its UV spectrum (Table 4) further suggested the location of the
acyl group to be at C-6.20 Therefore, the 2-methyl-1-oxobutyl unit
was fixed at C-6, which was confirmed by HMBC correlations of the
21.9
121.2
137.3
16.3
39.8
26.7
124.1
131.5
25.7
121.4
133.0
25.8
121.4
133.5
25.8
17.9a
17.5
17.9
124.2
131.5
25.7
17.7
17.7
a
Overlapping resonances within the same column.
chelated hydroxyl proton at dH 14.27 (7-OH) with C-6, C-7, and C-8.
In addition, the HMBC correlations of H-10 with C-3 and C-4a, of H-
20 with C-4, and of H-1000 with C-7, C-8, and C-8a indicated the loca-
tion of the hydroxypropyl group at C-4 and the geranyl unit at C-8.
Based on this evidence, the planar structure of kayeassamin C (1)
was established as 5,7-dihydroxy-4-(1-hydroxypropyl)-6-(2-
methyl-1-oxobutyl)-8-(3,7-dimethyl-2,6-octadienyl)-2H-benzopy-
ran-2-one. The absolute configuration at C-10 of 1 was determined
by the modified Mosher method.26,27 The MTPA esters of 1 were
obtained by treating 1 with (R)- and (S)-MTPA chloride and their
1H NMR resonances were assigned based on COSY correlations.
The chemical shift differences (DSR = dS ꢀ dR) of the individual
protons of 1a and 1b are shown in Figure 2. In the 1H NMR spec-
trum of the (S)-MPTA ester (1a), H2-20 and H3-30 appeared shielded,
Table 1
1H NMR data (CDCl3, 400 MHz) for kayeassamins CꢀG (1ꢀ5), J values (in Hz) in Parentheses
Proton
1
2
3
4
5
3
6.14 s
6.09 s
6.01 s
6.14 s
6.03 s
10
20
30
200
4.78 t (6.4)
1.76 m; 1.94 m
1.02 t (7.3)
4.75 t (7.8)
1.78 m; 1.92 m
1.02 t (7.6)
2.99 dd (6.6, 15.7);
3.07 dd (6.6, 15.7)
2.23 m
4.61 t (7.8)
1.74 m; 1.98 m
1.00a t (7.6)
3.00 m
4.78 t (7.3)
1.76 m; 1.92 m
1.03 t (7.3)
4.69 t (7.3)
1.78 m; 1.97 m
1.02 t (7.3)
3.78 dd (6.6, 13.2)
3.82 dd (6.6, 12.9)
2.96 dd (6.6, 15.6)
3.05 dd (6.6, 15.6)
2.21 m
300
1.23 d (6.6)
1.44 m; 1.85 m
0.96 t (7.3)
3.42 t (6.5)
5.22 t (6.5)
1.81 s
1.99 m
2.06 m
5.06 t (6.8)
1.65 s
1.71 m
1.23 d (6.6)
1.45 m; 1.85 m
0.96 t (7.3)
3.42 t (7.0)
5.22 t (7.0)
1.71 s
400
1.00a d (4.5)
1.00a d (4.5)
3.40 t (7.3)
5.20 t (7.3)
1.81 s
1.99 m
2.04 m
5.06 t (7.1)
1.64 s
1.00a t (7.6)
0.99a
0.99a
m
m
500
1000
3.34 m
5.15 t (6.8)
1.67 s
3.39 t (7.0)
5.20 t (7.0)
1.70 s
2000
4000
5000
1.79 s
1.81 s
1.80 s
6000
7000
9000
10000
10-OH
5-OH
7-OH
1.57 s
1.57 s
4.18 br s
10.51 br s
14.27 s
5.38 br s
11.41 br s
14.32 s
3.71 br s
10.39 br s
14.28 s
10.58 br s
14.42 s
10.88 br s
14.39 s
a
Overlapping resonances within the same column.

N. N. Win et al. / Bioorg. Med. Chem. 16 (2008) 8653–8660
8655
HO
OH
HO
OH
HO
OH
HO
OH
O
O
O
O
HO
O
O
HO
O
O
HO
O
O
HO
O
O
(a)
(b)
(c)
(d)
HO
OH
O
O
O
HO
O
HO
O
O
HO
O
O
HO
O
O
O
(f)
(e)
(g)
Figure 1. COSY (bold lines) and HMBC (1H ? 13C) (arrows) correlations in 1(a), 2(b), 3(c), 4(d), 5(e), 6(f), and 7 (g).
whereas H-3 was deshielded, in comparison to the corresponding
signals of (R)-MTPA ester (1b). Thus, H2-20 and H3-30 of 1a were
more affected by the phenyl ring of the MTPA part that indicated
cated the presence of a hydroxypropyl, an oxobutyl, and an isopre-
nyl substituent. However, besides the signal due to a chelated
hydroxyl proton at dH 14.32 (7-OH), the 1H NMR spectrum of 3
showed an additional signal ascribable to a hydroxyl proton at dH
11.41 (5-OH). Therefore, kayeassamin E (3) was assumed to be 6-
acylcoumarin, a regioisomer of theraphin A. This conclusion was
further supported by the bathochromic shift from 328 to 380 nm
after addition of alkali in UV spectra (Table 4). Kayeassamin E (3)
showed negative optical rotation as in theraphins AꢀD14 and
siamenols AꢀD.28 Therefore the absolute configuration at C-10 in
3 should be S. Accordingly, the structure of 3 was assigned as
(ꢀ)-5,7-dihydroxy-4-(1S-hydroxypropyl)-6-(1-oxobutyl)-8-(3-
methylbut-2-enyl)-2H-benzopyran-2-one.
the absolute configuration of C-10 in 1 to be S.
23
Kayeassamin D (2) was isolated as a pale yellow oil with [a]
D
ꢀ47.90°. The IR and UV data of 2 (Table 4) showed similar patterns
to those of 1, indicating 2 to be a 5,7-dihydroxy-6-acylcoumarin.
HREIMS of 2 also showed the same molecular formula C27H36O6
as 1. Extensive analyses of 1H and 13C NMR data (Tables 1 and 2)
and COSY and HMBC correlations (Fig. 1b) indicated that 2 differed
only on a 3-methyl-1-oxobutyl side chain at C-6. Kayeassamin D
(2) showed negative optical rotation, similar to that of 1, suggest-
ing the absolute configuration at C-10 to be S. Therefore, the
structure of kayeassamin D (2) was established as (ꢀ)-5,7-dihy-
droxy-4-(1S-hydroxypropyl)-6-(3-methyl-1-oxobutyl)-8-(3,7-di-
Kayeassamins F (4) and G (5) were obtained as pale yellow oils
23
with [
a]
ꢀ9.49° and ꢀ33.23°, respectively. They showed the
D
methyl-2,6-octadienyl)-2H-benzopyran-2-one.
same molecular formula C22H28O6 in HREIMS. Moreover, they
exhibited similar UV patterns as 3 (Table 4), ascribable to 6-acyl-
coumarin. The 1H and 13C NMR data (Tables 1 and 2) were similar
to those of theraphins B and C.14 However, the appearance of an
additional hydroxyl broad singlet (dH 10.39 in 4, dH 10.88 in 5) in
the 1H NMR spectra of 4 and 5 suggested that they were regioisom-
ers of theraphins B and C, respectively. The absolute configuration
at C-10 of 4 and 5 was also assumed to be S based on their negative
optical activity. Therefore, the structures of 4 and 5 were con-
23
Kayeassamin E (3) was isolated as a pale yellow oil with [a]
D
ꢀ9.79°. Its molecular formula C21H26O6 was established by HRE-
IMS. The 1H and 13C NMR data of 3 (Tables 1 and 2) resembled
those of theraphin A,14 an isolate from the same species, and indi-
-0.05
RO
H
-0.05,-0.04
+0.09
cluded as shown in Chart 1.29,30
O
OH
23
Kayeassamin H (6) was isolated as a pale yellow oil with [a]
D
+0.19
O
ꢀ31.95°. Its molecular formula C21H24O5 was established by HRE-
IMS. The 1H and 13C NMR data of 6 closely resembled those of ther-
aphin A14 and showed the signals of an oxobutyl, and an isoprenyl
substituent together with a singlet olefinic proton ascribable to H-
3 at dH 5.88 and a chelated hydroxyl proton at dH 14.20 (8-OH).
However, significant differences were observed in the 1H NMR
spectrum of 6 for the signals having been assigned to the hydroxy-
propyl unit at C-4 in theraphin A. The COSY and HMQC spectra led
to the assignment of the partial structure C(10)H3–C(5)H(O)–C(4)H2
(Fig. 1f) in 6, which was confirmed by the HMBC spectrum (Fig. 1f).
This, together with the consideration of its molecular formula,
indicated that 6 should be a pyranocoumarin, similar to those re-
ported from Mammea siamensis.31
HO
O
1a R = (
S
)-MTPA
R
1b R = ( )-MTPAa
Figure 2. Difference in the DSR (dS ꢀ dR) values for the (S)- and (R)-MTPA esters of 1
in CDCl3.

8656
N. N. Win et al. / Bioorg. Med. Chem. 16 (2008) 8653–8660
10''
9''
8''
7''
1'
3'
6''
5''
5
HO
OH
3'
4
O
1''
O6
1'
2''
3a
6a
4
HO
R1
6b
2''
6
4''
3
3''
3''
O
1'
7
1''
4
2
2
4''
8
HO
9a O
9
O
HO
O
O
6
8
R2
2
1'''
2'''
3'''
HO
O
O
8
3'''
1'''
R1
R2
5'''
O
4'''
4'''
2'''
10'''
8'''
4'''
3'''
6
7
4''
6'''
1'''
2''
1
2
5''
1''
O
O
9'''
2'''
2'''
5'''
7'''
3''
4''
2''
2''
5''
OH
1''
1''
3''
OH
R2
5'''
R
R1
HO
1'''
3'''
3
4
4''
O
O
3''
4'''
O
O
O
O
O
R1
R2
R
5
8
9
O
10
14
11
13
O
O
O
O
O
O
HO
O
O
O
OH
O
OH
15
16
O
O
HO
O
O
O
O
HO
O
O
O
12
17
18
HO
Chart 1. Structures of isolated coumarins (1–18) from the flower of K. assamica.
However, the HMBC and UV spectral data32 indicated the acyl
group to be at C-7. Therefore, the structure of kayeassamin H (6)
was concluded as 8-hydroxy-5-methyl-7-(1-oxobutyl)-9-(3-meth-
ylbut-2-enyl)-4,5-dihydropyrano [4,3,2-de]chromen-2-one.
2.2. In vitro preferential cytotoxicity
The effect of isolated coumarins on PANC-1 cell viability under
nutrient-deprived conditions was determined in vitro and their
preferential cytotoxicity (PC100) is summarized in Table 5. Among
the compounds tested, kayeassamins A (8), B (9), D (2), E (3), and
Kayeassamin I (7) was isolated as a pale yellow oil and its
molecular formula was deduced to be C26H32O6 by HREIMS. Its
UV spectrum showed absorption at 320 nm (Table 4), suggesting
it to be a 7-hydroxycoumarin derivative.33 The 1H NMR data of 7
(Table 3) displayed an oxygenated methine (dH 5.43, H-10), four
olefinic methines (dH 5.07, H-700; 5.54, H-300; 6.61, H-3; 6.79, H-
400), five methylenes (dH 1.50, 1.97, H2-20; 1.79, H2-3000; 1.90,
H2-500; 2.09, H2-600; 3.27, H2-2000), five methyls (dH 1.05, H3-4000;
1.13, H3-30; 1.51, 200-Me; 1.55, H3-1000; 1.64, H3-900), and a hydro-
gen-bonded hydroxyl proton (dH 14.48, 7-OH). On the other
hand, the 13C NMR spectrum of 7 exhibited 26 carbon signals
including those of a ketone carbonyl carbon, a lactone carbonyl
carbon, six aromatics, an oxygenated methine, an oxygenated
quaternary carbon, five methylenes, five methyls, four olefinic
methines, and two quaternary olefinic carbons. These data clo-
sely resembled those of deacetylmammea E/BA cyclo D (18),23
a known isolate in the present work, except for the presence
of additional signals due to a methyl (dH 1.51), an olefinic
methine (dH 5.07), two methylenes (dH 1.90, 2.09) and 1-oxobu-
tyl unit. The HMBC correlations of H-300, H-400, H-500, and 200-Me
with C-200 and of H-600 with C-700 and C-800 indicated the presence
of a 2-methyl-2-(4-methylpent-3-enyl)-2H-pyran ring in 7 in-
stead of the 2,2-dimethylpyran ring in 18. Thus, the structure
of kayeassamin I (7) was concluded as 7-hydroxy-200-methyl-
200-(4-methylpent-3-enyl)-4-(1-hydroxypropyl)-8-(1-oxobutyl)-2H,
G (5) were found to be the most potent with PC100 of 1
was comparable to that of the positive control, arctigenin.8 The
order of potency of the other isolates was 4, 12 (2 M) > 1
(4 M) > 11 (8 M) > 13 (16 M) > 10, 18 (32 M) > 6, 7, 15–17
(64 M) > 14 (> 256 M). Upon careful inspection of the structure
lM, which
l
l
l
l
l
l
l
and activity, compounds possessing an isoprenyl or a geranyl sub-
stituent at C-8 and a hydroxypropyl substituent at C-4 showed the
most potent activity. On C-6, presence of an oxobutyl or a 3-
methyl-1-oxobutyl group was found to be favored than a 2-
methyl-1-oxobutyl group (e.g., 2, 9 > 1; 3, 5 > 4). Interchange of
geranyl and oxobutyl units at C-8 and C-6 did not alter the activity
in the presence of a hydroxypropyl unit at C-4 (8ꢁ9). However
interchange of isoprenyl and any acyl groups between C-6 and C-
8 led to a dramatic reduction in activity even in the presence of a
hydroxypropyl group at C-4 (15 << 4; 16 << 5). Replacement of
the hydroxypropyl group at C-4 by a phenyl group led to significant
decrease in activity (13 < 3). Similarly, presence of any additional
cyclic ring in the coumarin nucleus led to the dramatic loss of
activity (3 >> 6; 8 >> 7; 12 >> 17; 13 >> 14).
Treatment by kayeassamins D (2), E (3), and G (5) at 1 lM trig-
gered apoptosis-like morphological change in PANC-1 cells within
24 h (Fig. 4). The dying cells severed attachments to other cells and
rounded up with clearly visible blebs. Furthermore, 2, 3 and 5
showed preferential cytotoxicity in a concentration- and time-
dependent manner (Fig. 3). Complete preferential cell death was
observed within 12 h of treatment with 2, and within 24 h of treat-
8H-benzo[1,2-b:3,4-b0]dipyran-2-one. Compound
7
exhibited
negative optical rotation, similar to that of 18,23 and therefore
the absolute configuration at C-10 was assumed to be S.

N. N. Win et al. / Bioorg. Med. Chem. 16 (2008) 8653–8660
8657
Table 3
1H and 13C NMR data for kayeassamins H (6) and I (7), J values (in Hz) in parentheses
3. Conclusion
Position
6
Position
7
The CHCl3-soluble fraction of the flower of K. assamica from
Myanmar exhibited 100% preferential cytotoxicity (PC100
) at
dH
dC
dH
dC
1
l
g/mL against the human pancreatic PANC-1 cancer cell line
2
3
3a
4
5
6a
6b
7
8
9
160.3
106.3
148.2
34.6
73.8
155.9
99.4
110.5
165.8
106.7
155.7
20.7
206.7
46.8
18.1
2
3
4
4a
5
6
7
8
159.6
107.0
160.6
101.0
155.9
105.8
162.9
104.5
157.3
71.8
30.7
10.2
83.0
124.9
116.5
41.6
under nutrient-deprived medium (NDM). Bioassay-guided frac-
tion and isolation led to the isolation of novel coumarins, kaye-
assamins A–I, together with nine known ones. All of the
isolated compounds were evaluated for their preferential cytotox-
icity against PANC-1 cells under NDM. Among them, five novel
coumarins, kayeassamins A (8), B (9), D (2), E (3), and G (5), dis-
5.88 s
6.61 s
2.84 m
4.41 m
8a
played the potent activity (PC100, 1 lM). The structure–activity
10
5.43 d (8.1)
1.50 m; 1.97 m
1.13 t (7.4)
relationship of all the isolated coumarins demonstrated that a
hydroxypropyl substituent at C-4, an oxobutyl or a 3-methyl-1-
oxobutyl moiety at C-6, and a geranyl group at C-8 in 5,7-dihydr-
oxycoumarin are important structural features in exhibiting the
preferential cytotoxicity. Kayeassamins D (2), E (3), and G (5) in-
duced apoptosis-like morphological changes in PANC-1 cells and
exhibited preferential cytotoxicity in a concentration- and time-
dependent manner.
9a
10
20
1.59 d (6.4)
30
100
200
300
400
1000
2000
3000
4000
5000
8-OH
200
300
400
500
600
700
800
900
1000
1000
2000
3000
4000
200-Me
7-OH
3.06 t (6.9)
1.72 m
1.01 t (7.5)
3.44 d (7.3)
5.23 t (7.3)
5.54 d (10.0)
6.79 d (10.0)
1.90 m
2.09 m
5.07 t (7.1)
13.9
21.2
121.1
132.8
25.8
23.0
123.1
132.6
25.6
17.7
206.4
46.7
18.1
13.8
27.3
1.67 s
1.83 s
14.20 s
1.64 s
1.55 s
18.0
4. Experimental
3.27 t (7.1)
1.79 m
1.05 m
1.51 s
14.48 br s
4.1. General experimental procedures
Optical rotations were recorded on a JASCO DIP-140 digital
polarimeter. IR spectra were measured with a Shimadzu IR-408
spectrophotometer in CHCl3. UV spectra were obtained using a
Shimadzu UV-160A UV–visible recording spectrophotometer.
NMR spectra were taken on a JEOL JNM-LA400 spectrometer with
tetramethylsilane (TMS) as an internal standard, and chemical
shifts are expressed in d values. EIMS and HREIMS measurements
were carried out on a JEOL JMS-700T spectrometer. Column chro-
matography was performed with normal-phase silica gel (Silica
Table 4
UV data of kayeassamins CꢀI (1ꢀ7)
Compound
In EtOH
kmax nm (loge)
In 0.1 N KOH in EtOH
1
2
3
4
5
6
7
220 (4.85), 329 (4.05)
207 (3.66), 329 (3.87)
223 (4.23), 328 (4.42)
224 (4.62), 328 (4.82)
224 (4.62), 328 (4.80)
211 (4.85), 380 (3.67)
211 (4.85), 380 (3.49)
213 (4.80), 380 (3.92)
213 (4.84), 380 (4.26)
213 (4.84), 380 (4.20)
Gel 60N, Spherical, neutral, 40–50 lm, Kanto Chemical Co., Inc.)
and reversed-phase silica gel (Cosmosil 75C18-OPN, Nacalai Tes-
que Inc.). Analytical and preparative TLC were carried out on pre-
coated silica gel 60F254 and RP-18F254 plates (Merck, 0.25 or 0.50
mm thickness). HPLC was performed with a Supelco Discovery C-
207 (4.40), 246 (3.80), 281 (4.60) 212 (4.78), 323 (3.85), 391 (4.29)
220 (4.20), 271 (4.43), 320 (4.27) 212 (4.86), 383 (4.24)
18 column (25 cm ꢂ 10 mm,
5 lm) using Tosoh UV-8000
detector.
Table 5
4.2. Plant material
Preferential cytotoxicity of compounds (118) on human pancreatic PANC-1 cancer
cells in nutrient-deprived conditions
Flowers of Kayea assamica King & Prain were collected from
Taunggyi, Shan State, Myanmar in November, 2004. The plant
was identified by Associate Professor Tin Maung Ohn (Department
of Botany, University of Yangon, Myanmar). A voucher specimen
(TMPW 25183) was deposited at the Museum for Materia Medica,
Analytical Research Center for Ethnomedicines, Institute of Natural
Medicine, University of Toyama, Japan.
Compound
PC100 [lM]
1
4
1
2
2, 3, 5, 8, 9
4, 12
6, 7, 15 + 16, 17
10, 18
64
32
8
16
> 256
1
11
13
14
Arctigenin*
4.3. Extraction and isolation
*
Positive control.
The flowers of K. assamica (1 kg) were extracted with 70% EtOH
under sonication (2 L, 3ꢂ 90 min) at room temperature and the sol-
vent was evaporated under reduced pressure to give 40 g of ex-
tract. Sonication of this extract with CHCl3 (100 mL, 3ꢂ 90 min)
afforded 14 g and 25 g each of CHCl3-soluble and -insoluble
fractions.
ment with 3 and 5, at 1 lM. This observation indicated that the
geranyl moiety at C-8 might be an important feature for enhance-
ment of activity. The sensitivities of the tested compounds were
further increased when treated at the concentrations of 4 or
16 lM; more than 50% cells were killed within 3 h of exposure
The CHCl3-soluble fraction (14 g) was chromatographed over
silica gel using increasing polarity of EtOAc in n-hexane to afford
seven fractions [fr. 1: EtOAc/n-hexane (20:80) eluate, 4.88 g; fr.
2: EtOAc/n-hexane (30:70) eluate, 2.87 g; fr. 3: EtOAc/n-hexane
(40:60) eluate, 2.34 g; fr. 4: EtOAc/n-hexane (50:50) eluate,
and the preferential cytotoxicity was observed within 9 h expo-
sure. These observations indicate that kayeassamins D (2), E (3)
and G (5) are potent anticancer agents, and might be useful for
the treatment of pancreatic cancer in real clinical situations.

8658
N. N. Win et al. / Bioorg. Med. Chem. 16 (2008) 8653–8660
Figure 4. Morphological change of human pancreatic cancer PANC-1 cells under nutrient-deprived medium after 24 h exposure with 1 lM of 2, 3 or 5.
Si gel column chromatography (CC) on fr. 2 using CH2Cl2/n-hex-
ane (2:1), CH2Cl2, and CH2Cl2/MeOH (40:1) afforded three subfrac-
tions [fr. 2-1: CH2Cl2/n-hexane (2:1) eluate, 671 mg; fr. 2-2: CH2Cl2
eluate, 188 mg; fr. 2-3: CH2Cl2/MeOH (40:1) eluate, 1.17 g]. Si gel
CC of subfraction 2-1 (671 mg) with n-hexane/Me2CO to give four
fractions [fr. 2-1-1: n-hexane/Me2CO (25:1) eluate, 125 mg; fr. 2-1-
2: n-hexane/Me2CO (15:1) eluate, 120 mg; fr. 2-1-3: n-hexane/
Me2CO (7:1) eluate, 54 mg; fr. 2-1-4: n-hexane/Me2CO (3:1) eluate,
225 mg]. Subfraction 2-1-1 was subjected to normal-phase pre-
parative TLC with benzene/Me2CO (40:1) to give mammea A/BC
(11, 46 mg) and mammea B/AC (12, 22 mg). Mammea A/AC (13,
60 mg) and mammea A/AC cyclo D (14, 40 mg) were obtained from
subfractions 2-1-2 and 2-1-3 by normal-phase pTLC with benzene/
Me2CO (20:1).
Fraction 3 (2.34 g) was rechromatographed on reversed-phase
Si gel with MeCN/H2O to afford three subfractions [fr. 3-1:
MeCN/H2O (7:3) eluate, 524 mg; fr. 3-2: MeCN/H2O (4:1) eluate,
270 mg; fr. 3-3: MeCN, 687 mg]. Si gel CC of fr. 3-1 using n-hex-
ane/Me2CO to give two subfractions [fr. 3-1-1: n-hexane/Me2CO
(20:1) eluate, 306 mg; fr. 3-1-2: n-hexane/Me2CO (20:1) eluate,
180 mg]. Purification of fr. 3-1-1 by normal-phase pTLC using ben-
zene/Me2CO (4:1), followed by reversed-phase pTLC using Me2CO/
H2O (3:1), to give kayeassamin E (3, 2 mg), a mixture of theraphins
B and C (15 and 16, 2 mg), mammea B/AC cyclo F (17, 2 mg), and
deacetylmammea E/BA cyclo D (18, 2 mg). Subfraction 3-1-2 was
subjected to reversed-phase pTLC using MeCN/H2O (7:3) to give
a mixture (78 mg) and kayeassamin B (9, 26 mg). Purification of
the mixture with semipreparative HPLC using MeCN/H2O (7:3) sol-
vent system [column: Discovery C-18, Supelco; flow rate: 5 mL/
min] afforded kayeassamins F (4, 15.3 mg; tR 38 min) and G (5,
3 mg; tR 40 min). Subfraction 3-2 was subjected to a series of nor-
mal-phase pTLC using benzene/Me2CO (5:1) and n-hexane/EtOAc
(2:1) afforded kayeassamins A (8, 10.6 mg), H (6, 3.7 mg), and I
(7, 25.2 mg). Subfraction 3-3 was purified by semipreparative HPLC
with MeCNꢀH2O (7:3) [column: Discovery C-18, Supelco; flow
rate: 5 mL/min] to afford kayeassamins A (8, 90 mg; tR 33 min), C
(1, 9.6 mg; tR 35 min), and D (2, 2 mg; tR 38 min).
100
75
1
50
4
16 μM
25
0
3
6
9
12
24
-25
Time (h)
100
1
75
50
25
0
4
16 μM
3
6
9
12
24
-25
Time (h)
100
75
50
25
0
1
4
16 μM
3
6
9
12
24
-25
Time (h)
Figure 3. Survival of human pancreatic cancer PANC-1 cells under nutrient-
deprived conditions within 024 h by 116 lM of 2 (A), 3 (B) and 5 (C). Data are
means SEM, n = 3.
0.49 g; fr. 5: EtOAc/n-hexane (60:40) eluate, 0.63 g; fr. 6: EtOAc/n-
hexane (80:20) eluate, 0.26 g; fr. 7: EtOAc eluate, 2.16 g]. Addition
of MeOH to fr. 1 (4.88 g) crystallized mammea A/AA cyclo D (10,
4 g).

N. N. Win et al. / Bioorg. Med. Chem. 16 (2008) 8653–8660
8659
4.3.1. Kayeassamin C (1)
ing to the procedure described by Izuishi et al.4 PANC-1 cells were
seeded in 96-well plates (2 ꢂ 104 cells per well) and incubated in
fresh Dulbecco’s modified Eagle’s medium (DMEM; Nissui Pharma-
ceuticals; Tokyo, Japan) at 37ꢀC under 5% CO2 and 95% air for 24 h.
The nutrient-deprived medium (NDM) was prepared following the
procedure described by Izuishi et al.4 After the cells were washed
with PBS (Nissui Pharmaceuticals), the medium was changed to
either DMEM or NDM and serial dilutions of the test samples were
added. For general preferential cytotoxicity assay, the cells were
23
Pale yellow oil; [
a
]
ꢀ60.87° (c 0.18, CHCl3); UV, see Table 4;
D
IR (CHCl3) mmax 3480, 1720, 1610, 1510, 1420, 1210, 1030,
930 cmꢀ1 1H and 13C NMR, see Tables 1 and 2; HREIMS m/z
;
456.2546 [M]+ (calcd for C27H36O6, 456.2512).
4.3.1.1. Preparation of (S)- and (R)-MTPA ester derivatives of
1. Two equal portions of 1 (each 2.5 mg) were dissolved in pyri-
dine (250 lL) and (R)-MTPA-Cl (7.5 lL) or (S)-MTPA-Cl (7.5 lL)
was added, respectively. The reaction mixtures were maintained
at room temperature for 12 h. The reaction products were purified
by normal-phase pTLC eluting n-hexane/EtOAc (3:1) afforded
1.8 mg each of 1a and 1b. 1H NMR data of 1a (CDCl3, 400 MHz):
dH 0.87 (H3-30), 1.72;1.95 (H2-20), 6.21 (H-3), 6.81 (H-10); EIMS
m/z 672 ([M]+). 1H NMR data of 1b (CDCl3, 400 MHz): dH 0.92
(H3-30), 1.77;1.99 (H2-20), 6.02 (H-3), 6.72 (H-10); EIMS m/z 672
([M]+).
incubated for 24 h, then washed with PBS, and 100 lL of DMEM
containing 10% WST-8 cell counting kit (Dojindo; Kumamoto, Ja-
pan) solution was added to the wells. After 2 h incubation, the
absorbance at 450 nm was measured. The crude extracts were
tested at 1, 10, 50, 100, and 200
lg/mL concentrations, while the
pure isolates were tested ranging from 1
l
M to 256 M.
l
4.5. Time- and concentration-dependent in vitro preferential
cytotoxicity
4.3.2. Kayeassamin D (2)
23
Pale yellow oil; [
a
]
ꢀ47.90° (c 0.18, CHCl3); UV, see Table 4;
For time- and concentration-dependent experiments, the cells
were incubated with the tested compounds for 3, 6, 9, 12, and
24 h. After incubation, morphological changes in PANC-1 cells were
observed and photographs were taken under 20ꢂ magnification
using phase-contrast microscopy (Olympus D-340L/C-840L Digital
Camera, Tokyo, Japan). The cells were then washed with PBS, and
D
IR (CHCl3) mmax 1720, 1600, 1510, 1420, 1210, 1030, 930 cmꢀ1 1H
;
and 13C NMR, see Tables 1 and 2; HREIMS m/z 456.2513 [M]+ (calcd
for C27H36O6, 456.2512).
4.3.3. Kayeassamin E (3)
23
Pale yellow oil; [
(CHCl3)
13C NMR, see Tables 1 and 2; HREIMS m/z 374.1744 [M]+ (calcd for
21H26O6, 374.1729).
a
]
D ꢀ9.79° (c 0.18, CHCl3); UV, see Table 4; IR
100 lL of DMEM containing 10% WST-8 cell counting kit (Dojindo;
m
max 1720, 1600, 1510, 1420, 1210, 1030, 930 cmꢀ1; 1H and
Kumamoto, Japan) solution was added to the wells. After 2 h incu-
bation, the absorbance at 450 nm was measured. Cell viability was
calculated from the mean values of data from three wells by using
the following equation:
C
4.3.4. Kayeassamin F (4)
ð%Þ Cell viability ¼ ½fAbsðtest sampleÞ ꢀ AbsðblankÞg=fAbsðcontrolÞ
ꢀ AbsðblankÞgꢃ ꢂ 100
23
Pale yellow oil; [
(CHCl3) mmax 1720, 1610, 1520, 1220, 1040, 930 cmꢀ1
NMR, see Tables 1 and 2; HREIMS m/z 388.1884 [M]+ (calcd for
22H28O6, 388.1886).
a
]
D ꢀ9.49° (c 0.52, CHCl3); UV, see Table 4; IR
;
1H and 13C
C
Acknowledgments
4.3.5. Kayeassamin G (5)
23
Pale yellow oil; [
a
]
D ꢀ33.23° (c 0.2, CHCl3); UV, see Table 4; IR
This work was supported in part by a grant from the Ministry of
Health and Welfare for the Second-Term Comprehensive 10-year
Strategy for Cancer Control and Grants-in-Aid for Cancer Research
from the Ministry of Health and Welfare, Japan.
(CHCl3) m
max 1720, 1600, 1520, 1410, 1200, 1040, 925 cmꢀ1; 1H and
13C NMR, see Tables 1 and 2; HREIMS m/z 388.1924 [M]+ (calcd for
C22H28O6, 388.1886).
4.3.6. Kayeassamin H (6)
References and notes
23
Pale yellow oil; [
a
]
ꢀ31.95° (c 0.06, CHCl3); UV, see Table 4;
1H
and 13C NMR, see Table 3; HREIMS m/z 356.1669 [M]+ (calcd for
21H24O5, 356.1624).
D
IR (CHCl3) mmax 1730, 1600, 1520, 1430, 1380, 1220, 930 cmꢀ1
;
1. National Cancer Center. Cancer statistics in Japan 2007. http://
ganjoho.ncc.go.jp/public/statistics/backnumber/2007_en.html.
2. Tsukuma, H.; Ajiki, W.; Ioka, A.; Oshima, A. Jpn. J. Clin. Oncol. 2006, 36, 602–607.
3. Heinemann, V. Oncology 2001, 60, 8–18.
C
4. Izuishi, K.; Kato, K.; Ogura, T.; Kinoshita, T.; Esumi, H. Cancer Res. 2000, 60,
6201–6207.
5. Koito, K.; Namieno, T.; Nagakawa, T.; Morita, K. Am. J. Roentgenol. 1997, 169,
1263–1267.
6. Esumi, H.; Izuishi, K.; Kato, K.; Hashimoto, K.; Kurashima, Y.; Kishimoto, A.;
Ogura, T.; Ozawa, T. J. Biol. Chem. 2002, 277, 32791–32798.
7. Esumi, H.; Lu, J.; Kurashima, Y.; Hanaoka, T. Cancer Sci. 2004, 95, 685–690.
8. Awale, S.; Lu, J.; Kalauni, S. K.; Kurashima, Y.; Tezuka, Y.; Kadota, S.; Esumi, H.
Cancer Res. 2006, 66, 1751–1757.
9. Awale, S.; Nakashima, E. M. N.; Kalauni, S. K.; Tezuka, Y.; Kurashima, Y.; Lu, J.;
Esumi, H.; Kadota, S. Bioorg. Med. Chem. Lett. 2006, 16, 581–583.
10. Win, N. N.; Awale, S.; Esumi, H.; Tezuka, Y.; Kadota, S. J. Nat. Prod. 2007, 70,
1582–1587.
4.3.7. Kayeassamin I (7)
23
Pale yellow oil; [
a
]
ꢀ35.52° (c 0.90, CHCl3); UV, see Table 4;
D
IR (CHCl3) m C
max 1720, 1610, 1410, 1210, 1020, 930 cmꢀ1; 1H and 13
NMR, see Table 3; HREIMS m/z 440.2191 [M]+ (calcd for C26H32O6,
440.2199).
4.3.8. Mammea B/AC (12)
13C NMR (CDCl3) dC 207.7 (C-100), 163.9 (C-5), 160.3 (C-2), 160.1
(C-7), 159.7 (C-4), 156.6 (C-8a), 138.5 (C-3000), 120.2 (C-2000), 110.1
(C-3), 106.7 (C-6), 104.5 (C-8), 103.5 (C-4a), 46.6 (C-200), 38.6 (C-
10), 25.9 (C-4000), 22.8 (C-20), 22.0 (C-1000), 18.0 (C-5000), 17.8 (C-300),
13.9 (C-400), 14.0 (C-30).
11. Win, N. N.; Awale, S.; Esumi, H.; Tezuka, Y.; Kadota, S. Chem. Pharm. Bull. 2008,
56, 491–496.
12. Awale, S.; Li, F.; Onozuka, H.; Esumi, H.; Tezuka, Y.; Kadota, S. Bioorg. Med.
Chem. 2008, 16, 181–189.
13. Bordoloi, M.; Mohan, S.; Barua, N.; Dutta, S. C.; Mathur, R. K.; Ghosh, A. C.
Phytochemistry 1997, 44, 939–942.
14. Lee, K. H.; Chai, H. B.; Tamez, P. A.; Pezzuto, J. M.; Cordell, G. A.; Win, K. K.; Tin-
Wa, M. Phytochemistry 2003, 64, 535–541.
15. Win, N. N.; Awale, S.; Esumi, H.; Tezuka, Y.; Kadota, S. Bioorg. Med. Chem. Lett.
2008, 18, 4688–4691.
16. Crombie, L.; Games, D. E.; McCormick, A. J. Chem. Soc. (C) 1967, 2553–2559.
4.4. In vitro preferential cytotoxicity
The in vitro preferential cytotoxicities of the crude extract and
the isolated compounds were investigated using PANC-1 human
pancreatic cancer cells under nutrient-deprived conditions accord-

8660
N. N. Win et al. / Bioorg. Med. Chem. 16 (2008) 8653–8660
17. Prachyawarakorn, V.; Mahidol, C.; Ruchirawat, S. Pharm. Biol. 2000, 38, 58–
62.
26. Ohtani, I.; Kusumi, T.; Kashman, Y.; Kakisawa, H. J. Am. Chem. Soc. 1991, 113,
4092–4096.
18. Morel, C.; Dartiguelongue, C.; Youhana, T.; Oger, J. M.; Séraphin, D.; Duval, O.;
Richomme, P.; Bruneton, J. Heterocycles 1999, 51, 2183–2191.
19. Crombie, L.; Games, D. G.; Haskins, N. J.; Reed, G. F. J. Chem. Soc. Perkin Trans. 1
1972, 2248–2254.
27. Rieser, M. J.; Hui, Y. H.; Rupprecht, J. K.; Kozlowski, J. F.; Wood, K. V.;
McLaughlin, J. L.; Hanson, P. R.; Zhuang, Z.; Hoye, T. R. J. Am. Chem. Soc. 1992,
114, 10203–10213.
28. Prachyawarakorn, V.; Mahidol, C.; Ruchirawat, S. Chem. Pharm. Bull. 2006, 54,
20. Crombie, L.; Jones, R. C. F.; Palmer, C. J. J. Chem. Soc. Perkin Trans. 1 1987, 317–
331.
884–886.
29. Crombie, L.; Games, D. E.; Reed, G. F. J. Chem. Soc. Perkin Trans. 1 1972, 2255–
21. Thebtaranonth, C.; Imraporn, S.; Padungkul, N. Phytochemistry 1981, 20, 2305–
2306.
22. Morel, C.; Guilet, D.; Oger, J. M.; Séraphin, D.; Sévenet, T.; Wiart, C.; Hadi, A. H.
A.; Richomme, P.; Bruneton, J. Phytochemistry 1999, 50, 1243–1247.
23. Mahidol, C.; Prawat, H.; Kaweetripob, W.; Ruchirawat, S. Nat. Prod. Commun.
2007, 2, 557–564.
2260.
30. These compounds were reported as a mixture of semi-synthetic product by
Crombie et al. (1972). However, no physical and spectroscopic data in pure
form were available.
31. Prachyawarakorn, V.; Mahidol, C.; Ruchirawat, S. Phytochemistry 2006, 67, 924–
928.
24. Mahandru, M. M.; Ravindran, V. K. Phytochemistry 1986, 25, 555–556.
25. Yagi, N.; Ohkubo, K.; Okuno, Y.; Oda, Y.; Miyazawa, M. J. Oleo Sci. 2006, 55, 173–
180.
32. Crombie, L.; Jones, R. C. F.; Palmer, C. J. J. Chem. Soc. Perkin Trans. 1 1987, 333–
343.
33. Kalyanmay, S.; Bagchi, P. J. Org. Chem. 1958, 24, 316–319.