Highly enantioselective synthesis of angelmarin

Article abstract of DOI:10.1039/b913400j

Angelmarin (1), a novel anti-cancer agent, was efficiently synthesized through a highly enantioselective epoxidation and a copper cyanide-mediated esterification of the hindered alcohol as the key steps in 53% overall yield.

Full text of DOI:10.1039/b913400j

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/obc | Organic & Biomolecular Chemistry
Highly enantioselective synthesis of angelmarin†
Hang Jiang and Yasumasa Hamada*
Received 6th July 2009, Accepted 23rd July 2009
First published as an Advance Article on the web 4th August 2009
DOI: 10.1039/b913400j
Angelmarin (1), a novel anti-cancer agent, was efficiently
synthesized through a highly enantioselective epoxidation
and a copper cyanide-mediated esterification of the hindered
alcohol as the key steps in 53% overall yield.
During the course of a unique screening program using the anti-
austerity strategy,1-3 Kadota, Esumi, and co-workers identified
a new coumarin natural product, angelmarin (1),4 from an
extract of Angelica pubescens, which preferentially displays strong
activity against PANC-1 cancer cells under a nutrient starvation
environment (Fig. 1). Its core structure is columbianetin (2) with
a 2-substituted benzofuran moiety, which occurs widely in many
natural products of plant origin.5 Based on its interesting anti-
tumor activity as well as the scarcity of an efficient method for
the construction of optically active 2-substituted benzofurans,6
we decided to attempt the synthesis of this natural product.
A recent publication of the synthesis of (+)-1 by Coster and
Magolan7 prompted us to disclose our efforts on the synthesis of
1. Our synthesis of angelmarin contains a highly enantioselective
epoxidation and an efficient copper-mediated esterification as the
key steps. Our method is superior to the reported one with regard
to its efficiency and enantioselectivity.
Scheme 1
Scheme 2
◦
According to the precedent,8 7 was heated at 130 C for 1.5 h
to give a separable mixture of the desired 4 and its isomer 8 in an
86/14 ratio. After the chromatographic separation of the mixture,
we examined the enantioselective construction of the benzofuran
moiety using pure 4. The authentic racemic columbianetin was
obtained according to Franke’s method.10 The enantioselective
epoxidation of 4 and its derivative is shown in Table 1.
First, we investigated the Sharpless asymmetric epoxidation
using titanium tetraisopropoxide and tartrates (entries 1–3).
Although we extensively surveyed the reaction conditions, we
were unable to obtain a satisfactory result. The enantioselectivity
was moderate (~69% ee) and the yield was low due to the
production of the side product 11, which was formed by the tita-
nium tetraisopropoxide-induced ring opening reaction of the
product epoxide. After some experiments, we succeeded in the
suppression of this side reaction by changing the oxidant from
TBHP to tritylhydroperoxide (TPHP), a bulkyl peroxide, but the
reactivity was slower than that of TBHP and the chemical yield
Fig. 1 Angelmarin and columbianetin.
The obvious route to 2 is the enantioselective epoxidation of
7-hydroxy-8-prenylcoumarin (osthenol, 4), which is derived from
the commercially available 7-hydroxycoumarin (5) through the
Claisen rearrangement (Scheme 1). The precursor 7 for the Claisen
rearrangement is conventionally derived from the 2-(2-methyl)-
but-3-ynylation of 5 and subsequent hydrogenation using the
Lindlar catalyst.8 However, 7 could be conveniently prepared from
5 by the palladium-catalyzed direct allylation using the mixed
carbonate 6 in high yield (Scheme 2).9
Yayoi-cho, Inage-ku, Chiba 263-8522, Japan. E-mail: hamada@p.chiba-
u.ac.jp; Fax: +81-43-290-2987; Tel: +81-43-290-2987
† Electronic supplementary information (ESI) available: Experimental and
1
NMR data of compounds; copies of H NMR and 13C NMR spectra of
compounds. See DOI: 10.1039/b913400j
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 4173–4176 | 4173
©

View Article Online
Table 1 Enantioselective epoxidation
Entry
Substrate
Reagents (eq.)
Conditions
Yielda
% eeb
1
2
3
4
5
6
4
4
4
4
9
9
method A, DMT-TBHP
method A, DET-CHP
method A, DET-TPHP
method Bf, ketone 12 (0.3)
method Bf, ketone 12 (0.3)
method Bh, ketone 13 (0.15)
-20 ◦C, 87 h
0 ◦C, 17.5 h
-20 ◦C, 89 h
-10 ◦C, 4 h
0 ◦C, 3 h
31c
27d
38e
28g
25g
88g
68
61
69
21
77
97
0 ◦C, 23 h
a Isolated yield. b Enantiomeric excess of 2. c Yield of 11 (R = t-Bu): 56%. d Yield of 11 (R = Cumenyl): 67%. e Yield of 11 (R = trityl): trace. f Oxone (1.38
eq.), K2CO3 (5.8 eq.), CH3CN–DMM–buffer (1/2/2). Buffer: 0.05 M solution of Na2B4O7 10H2O in 4 ¥ 10-4 M aqueous Na2(EDTA). g Yield of two steps.
h Oxone (1.63 eq.), K2CO3 (2.4 eq.), CH3CN–DMM–buffer (1.8/3.6/1). Buffer: 0.1 M aqueous KOH/0.1 M aqueous KH2PO4/H2O = 5.6/50/44.4,
pH 6.
was still unsatisfactory. Next, we turned our attention to the Shi
epoxidation11 using the chiral dioxirane generated from a ketone
and Oxone. The epoxidation of 4 using the ketone 12 under
the standard Shi conditions at -10 ◦C for 4 h directly afforded
the cyclized 2 in 28% yield (entry 4). The enantioselectivity was
disappointedly low (21% ee). After some preliminary experiments,
we found that the protection of the phenolic function with a silyl
group significantly enhanced the enantioselectivity (entry 5). The
protection of 4 with tert-butylchlorodimethylsilane and imidazole
in dimethylformamide at rt for 1.5 h gave the silylated 9 in a
quantitative yield. The thus-obtained 9 was subjected to the Shi
epoxidation using a catalytic amount of tetra-n-butylammonium
hydrogen sulfate and the chiral ketone 12 (0.3 equiv.) in the
presence of an excess amount of Oxone (1.38 equiv.) and potas-
sium carbonate (5.8 equiv.) in a mixed solvent of acetonitrile,
dimethoxymethane (DMM), and buffer12 at 0 ◦C for 3 h, yielding
a mixture of the target epoxide 10 and the starting material with a
25/75 ratio. Treatment of the crude epoxidized product with tetra-
n-butylammonium fluoride (TBAF) in tetrahydrofuran (THF) at
columbianetin 2 with a 97% ee in 88% yield (two steps). In contrast
to the original Shi procedure, lower catalyst loadings (15 mol%)
and a smooth reaction were possible without any loss of reaction
efficiency. The optically pure 2 could be obtained by one recrystal-
lization from ethyl acetate. To demonstrate the preparative utility
of this new asymmetric synthesis, the epoxidation–fluoride anion-
induced cyclization process was performed on a 10 gram-scale to
afford columbianetin (+)-2 with almost the same efficiency of 97%
ee in 85% yield.‡
As an alternative route for the preparation of the optically
active 2, we briefly investigated the Sharpless enantioselective
dihydroxylation of 4 and its derivatives as shown in Table 2.
Interestingly, the enantioselectivity was strongly dependent on
the phenolic protecting group employed. Sulfonyl groups caused
negative effects concerning the enantioselectivity (entries 2 and
5). In the case of the triflate 14, the reaction directly produced
the cyclized 2 in 16.4% yield, which apparently arose from the
intramolecular aromatic nucleophilic substitution of the diol 18
(R = Tf). The benzyl and acetyl groups largely enhanced the
enantioselectivity (entries 3 and 4). Although the route using the
asymmetric dihydroxylation was attractive, we selected the more
efficient pathway above using the asymmetric epoxidation.
The final task remaining in the synthesis of angelmarin (1)
was the esterification of the tertiary alcohol of 2 with p-
hydroxycinnamic acid. We carried out model experiments using
(rac)-2 and tiglic acid as shown in Table 3. The reaction using
either the usual condensing agents or the Mitsunobu procedure did
0
◦C for 30 min afforded 2 with a 77% ee in 25% yield. Since
the enantioselectivity and chemical yield by the Shi epoxidation
method were still unsatisfactory, we surveyed other ketones for this
transformation. In 2005, Vidal-Ferran and co-workers reported a
Shi-type ketone 13 for epoxidations, which is robust and equally
effective when compared to the ketone 12.13 Surprisingly, for
the epoxidation of 9 the Vidal-Ferran’s ketone 13 was superior
to 12 and afforded, after treatment of the epoxide with TBAF,
4174 | Org. Biomol. Chem., 2009, 7, 4173–4176
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
Table 2 Enantioselective dihydroxylationa
Entry
Substrate
Time (h)
Yield (%)b
% ee
1
2
3
4 (R = H)
4
10
24
42
8
19
19c
69
20
68
50
0
80
88
6
14 (R = Tf)
15 (R = Bn)
16 (R = Ac)
17 (R = Ns)
4
5d
a AD-mix-a (140 mg/0.1 mmol). b Isolated yield. c The cyclized product
2 was obtained in 16.4%. d The reaction was carried out in 50% aqueous
t-BuOH/CH2Cl2 (1/1).
Table 3 Reaction conditions for esterification
Scheme 3
In conclusion, we have established the efficient asymmetric
synthesis of angelmarin (1), which ensures the supply of the
optically pure 1. Using this method, many analogs of 1 will
be available for evaluation of their biological activities. Further
investigations into the formation of 2-substituted benzofurans and
the structure–activity relationship of angelmarin and its analogs
are underway.
Entry
X (eq.)
Conditions
Yield (%)a
1
2
3
4
5
6
7
8
OH (2)
OH (1)
Cl (1)
DCC–DMAP, CH2Cl2, rt, 24 h
DEAD–PPh3, THF, rt, 24 h
pyridine, rt, 10 h
nr
nr
nr
nr
Cl (2)
n-BuLi, THF, -78 ◦C to rt, 24 h
NaH, THF, 0 ◦C to rt, 24 h
AgCN (15 eq.), PhCH3, 50 ◦C, 72 h
AgCN (3 eq.), PhCH3, 50 ◦C, 26 h
CuCN (3 eq.), PhCH3, 50 ◦C, 26 h
Cl (1.5)
Cl (10)
Cl (2)
nr
100
91b
91b
Acknowledgements
Cl (2)
This work was financially supported in part by Special Funds
for Education and Research (Development of SPECT Probes
for Pharmaceutical Innovation) from the Ministry of Education,
Culture, Sports, Science and Technology, Japan.
a Isolated yield. b A small amount of 2 was recovered.
not take place and the starting material was completely recovered
(entries 1 and 2). The acid chloride in combination with bases
resulted in no reaction (entries 3–5). Finally, we found that the
addition of silver cyanide or cuprous cyanide as a base to the
solution of the acid chloride effectively provided the corresponding
ester 19 in high yield (entries 6–8).14 The cheap cuprous cyanide
method is preferable based on the cost.§
As illustrated in Scheme 3, the required acid chloride 22 for the
final construction of 1 was obtained from 20 by (1) the protection
of the phenolic hydroxyl group and carboxylic function with tert-
butylchlorodimethylsilane and imidazole, (2) the selective removal
of the silyl ester with lithium hydroxide in aqueous THF, and
(3) the treatment of the carboxylic acid 21 with refluxing thionyl
chloride. The esterification of 2 with the acid chloride 22 proceeded
smoothly in the presence of an excess amount of cuprous cyanide
(5 equiv.) in toluene at 50 ◦C for 17 h to afford the ester 23 in
87.4% yield. Final deprotection of 23 with TBAF (1.2 equiv.)
in THF furnished angelmarin (1) in a quantitative yield.15 This
asymmetric synthesis of 1 is the most efficient method so far
and the overall yield is 53% in six steps from the commercially
available 5.
Notes and references
‡ Typical procedure for the synthesis of (+)-2 from 9: Alkene 9 (10.0 g,
29.0 mmol) and ketone 13 (1.32 g, 4.35 mmol) were dissolved in acetonitrile
(192 mL) and dimethoxymethane (384 mL). A buffer (pH 6, made from
0.1 M aqueous KOH/0.1 M aqueous KH2PO4/H2O = 5.6/50/44.4)
solution (105 mL) and tetra-n-butylammonium hydrogen sulfate (0.39 g,
1.16 mmol) were slowly added with stirring, and the mixture was cooled to
0 ◦C. The flask was equipped with two dropping funnels; one of them was
filled with a solution of Oxone (29.1 g, 47.3 mmol) in the buffer solution
(pH 6, 183 mL), and the other one with a solution of K2CO3 (9.63 g,
69.7 mmol) in water (183 mL). The two solutions were added dropwise
over 4 h. The reaction mixture was stirred at 0 ◦C for another 43.5 h. The
reaction mixture was then quenched by the addition of water (500 mL)
and n-hexane (150 mL) and extracted with n-hexane (300 mL ¥ 2). The
combined organic extracts were washed with brine (300 mL), dried over
sodium sulfate, filtered, and concentrated in vacuo. The crude material
was used without purification. The above residue was dissolved in THF
(100 mL), and TBAF (1 M in THF, 34.8 mL) was added at 0 ◦C with
stirring. The reaction mixture was gradually warmed to room temperature.
After being stirred for 4 h, the mixture was diluted with ethyl acetate,
washed with brine, dried over sodium sulfate, filtered, and concentrated
in vacuo. The residue was purified by silica gel column chromatography
(ethyl acetate/n-hexane = 1/1) to give columbianetin (2, 6.07 g, 84.9%) as
a colorless solid. Recrystallization from ethyl acetate provided the optically
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 4173–4176 | 4175
©

View Article Online
pure 2 (100% ee judged by HPLC): mp 161–162 ◦C; [a]D +183.55 (c 1.0,
S. Murahashi, Bull. Chem. Soc. Jpn, 1985, 58, 3282; (c) T. Hosokawa,
T. Kono, T. Shinohara and S. Murahashi, J. Organomet. Chem., 1989,
370, C13; (d) Y. Uozumi, K. Kato and T. Hayashi, J. Am. Chem. Soc.,
1997, 119, 5063; (e) Y. Uozumi, H. Kyota, K. Kato, M. Ogasawara and
T. Hayashi, J. Org. Chem., 1999, 64, 1620; (f) T. Nemoto, T. Ohshima
and M. Shibasaki, Tetrahedron Lett., 2000, 41, 9569; (g) T. Nemoto,
T. Ohshima and M. Shibasaki, Tetrahedron, 2003, 59, 6889; (h) Y.
Natori, H. Tsutsui, N. Sato, S. Nakamura, H. Nambu, M. Shiro and S.
Hashimoto, J. Org. Chem., 2009, 74, 4418.
25
CHCl3) (lit.4 [a]D +198 (c 0.025, CHCl3)); 1H NMR (400 MHz, CDCl3)
25
d 1.24 (s, 3H), 1.37 (s, 3H) 3.32 (m, 2H), 4.80 (t, J = 8.8 Hz, 1H), 6.21
(d, J = 9.6 Hz, 1H), 6.75 (d, J = 8.4 Hz, 1H), 7.27 (d, J = 8.0 Hz, 1H),
7.63 (d, J = 9.6 Hz, 1H): 13C NMR (100 MHz, CDCl3) d 23.9, 26.0, 27.6,
71.8, 91.3, 106.7, 112.3, 113.1, 114.0, 128.7, 144.0, 151.3, 161.0, 163.7;
IR (KBr, cm-1) 3502, 2969, 1698, 1609, 1490, 1453, 1406, 1337, 1316,
1254, 1146, 1118, 1065, 1011, 946, 826, 779, 759; HRMS (FAB) calcd
for C14H15O4: 247.0970 [M + H]+. Found: 247.0956. HPLC analysis using
CHIRALPAC AD and n-hexane/i-PrOH (75:25, 0.5 mL min-1), Retention
time for major: 15.6 min; minor: 22.9 min. 97% ee.
7 J. Magolan and M. J. Coster, J. Org. Chem., 2009, 74, 5083.
8 R. D. H. Murray, M. M. Ballantyne and K. P. Mathai, Tetrahedron,
1971, 27, 1247.
§ Typical procedure for the copper cyanide-mediated esterification: To a
stirred suspension of 2 (400 mg, 1.62 mmol, 100% ee) and CuCN (727 mg,
8.12 mmol) in toluene (5 mL) at room temperature was added a solution
of the acid chloride 22 (4.87 mmol) in toluene (10 mL). After being
stirred for 18.5 h at 50 ◦C, the mixture was filtered through a Celite pad
and the insoluble material was washed with ethyl acetate. The combined
filtrates were washed with brine, dried over sodium sulfate, filtered, and
concentrated in vacuo. The residue was purified by silica gel column
chromatography (n-hexane/ethyl acetate = 3/1) to give 23 (719.5 mg,
87.4%) as a brown oil.
9 For
the
3-(3-methyl)-but-1-enylation
of
phenols,
see:
T. Kaiho, T. Yokoyama, H. Mori, J. Fujiwara, T. Nobori,
H. Odaka, J. Kamiya, M. Maruyama, and T. Sugawara, JP 06128238,
1994; [Chem. Abs., 1995, 123, 55900].
10 F. Bohlmann and H. Franke, Chem. Ber., 1971, 104, 3229.
11 (a) Y. Tu, Z.-X. Wang and Y. Shi, J. Am. Chem. Soc., 1996, 118, 9806;
(b) Z.-X. Wang, Y. Tu, M. Frohn, J.-R. Zhang and Y. Shi, J. Am. Chem.
Soc., 1997, 119, 11224.
12 Buffer: 0.05 M solution of Na2B4O7 10 H2O in 4 ¥ 10-4 M aqueous
Na2(EDTA).
1 K. Izuishi, K. Kato, T. Ogura, T. Kinoshita and H. Esumi, Cancer Res.,
2000, 60, 6201.
2 J. Lu, S. Kunimoto, Y. Yamazaki, M. Kaminishi and H. Esumi, Cancer
Sci., 2004, 95, 547.
13 (a) N. Nieto, P. Molas, J. Benet-Buchholz and A. Vidal-Ferran, J. Org.
Chem., 2005, 70, 10143; (b) B. Wang, X.-Y. Wu, O. A. Wong, B.
Nettles, M.-X. Zhao, D. Chen and Y. Shi, J. Org. Chem., 2009, 74,
3986.
3 S. Awale, J. Lu, S. K. Kalauni, Y. Kurashima, Y. Tezuka, S. Kadota and
H. Esumi, Cancer Res., 2006, 66, 1751.
4 S. Awale, E. M. N. Nakashima, S. K. Kalauni, Y. Tezuka, Y. Kurashima,
J. Lu, H. Esumi and S. Kadota, Bioorg. Med. Chem. Lett., 2006, 16,
581.
5 For reviews of naturally occurring coumarins, see: (a) R. D. H. Murray,
in Progress in the Chemistry of Organic Natural Products, eds. W. Herz,
G. W. Kirby, W. Steglich, and Ch. Tamm, Springer-Verlag, Wien New
York, 1991, 58, 83–316; (b) R. D. H. Murray, in Progress in the
Chemistry of Organic Natural Products, eds. W. Herz, G. W. Kirby,
R. E. Moore, W. Steglich, and Ch. Tamm, Springer-Verlag, Wien New
York, 1997, 72, 2–119.
6 For the enantioselective synthesis of the optically active 2-substituted
benzofurans, see: (a) T. Hosokawa, T. Uno, S. Inui and S. Murahashi,
J. Am. Chem. Soc., 1981, 103, 2318; (b) T. Hosokawa, Y. Imada and
14 For silver cyanide-mediated esterification, see: (a) S. Takimoto, J.
Inanaga, T. Katsuki and M. Yamaguchi, Bull. Chem. Soc. Jpn, 1976,
49, 2335; (b) S. Hara, K. Makino and Y. Hamada, Tetrahedron Lett,
2006, 47, 1081.
25
25
15 1 as amorphous solids: [a]D + 237.8 (c 1.025, CHCl3), [a]D + 217.5
(c 0.028, CHCl3) (lit.4 [a]D + 218.7 (c 0.025, CHCl3)); 1H NMR
25
(400 MHz, CDCl3) d 1.60 (s, 3H), 1.65 (s, 3H), 3.39 (m, 2H), 5.20
(t, J = 9.6 Hz, 1H), 6.15 (d, J = 16.0 Hz, 1H), 6.24 (d, J = 9.2 Hz, 1H),
6.79 (d, J = 8.8 Hz, 1H), 6.83 (d, J = 8.8 Hz, 2H), 7.30 (m, 3H), 7.35
(d, J = 15.6 Hz, 1H), 7.67 (d, J = 9.2 Hz, 1H); 13C NMR (100 MHz,
CDCl3) d 21.1, 22.1, 27.6, 82.2, 89.1, 106.9, 112.1, 113.0, 113.6, 115.9,
116.4, 126.9, 128.9, 129.9, 144.2, 144.4, 151.2, 158.0, 161.4, 164.1, 166.4;
IR (KBr, cm-1) 3303, 1696, 1602, 1513, 1489, 1455, 1406, 1387, 1369,
1328, 1260, 1200, 1167, 1118, 1065, 976, 871, 828, 753; HRMSFAB
calcd for C23H21O6 393.1338 [M + H]+, found 393.1321.
4176 | Org. Biomol. Chem., 2009, 7, 4173–4176
This journal is
The Royal Society of Chemistry 2009
©