23-Oxa-analogues of OSW-1: Efficient synthesis and extremely potent antitumor activity

Article abstract of DOI:10.1002/anie.200454237

An aldol condensation is the key to the eight-step linear synthesis of 23-oxa analogues of OSW-1 (e.g. 2) in more than 20% overall yield from the industrially produced steroid 1. Compound 2 is up to 2000 times more potent than cisplatin as an inhibitor of

Full text of DOI:10.1002/anie.200454237

Communications
Medicinal Chemistry
23-Oxa-Analogues of OSW-1: Efficient Synthesis
and Extremely Potent Antitumor Activity**
Bingfeng Shi, Hao Wu, Biao Yu,* and Jiarui Wu
Saponin OSW-1, which contains a novel 16b,17a-dihydroxy-
cholest-22-one aglycone unit with an acylated disaccharide at
the 16-OH group, was discovered by Sashida, Mimaki, and co-
workers in 1992 in the bulbs of Ornithogalum saudersiae, a
perennial garden plant of the lily family widely cultivated in
southern Africa.[1] Tremendous attention has been given to
this compound since its antitumor activity was tested in 1995:
OSW-1 has a low toxicity for normal cells but inhibits the
growth of a variety of malignant tumor cells and is 10–100
times more potent than clinically applied anticancer agents
such as mitomycin C, adriamycin, cisplatin, camptothecin,
and taxol.[2] Further research on O. saudersiae and taxonomi-
cally related plants has revealed a family of 16b,17a-
dihydroxycholest-22-one saponins.[3] Considerable effort has
been directed toward the synthesis of OSW-1, the first and
most abundant member of the family having been isolated.[4–7]
The aglycone was first synthesized by Fuchs and Guo in
1998.[4] Shortly afterwards, we coupled the aglycone with the
disaccharide moiety to complete the total synthesis of OSW-
1.[5] Jin and Yu recently greatly improved this synthesis.[6]
However, studies on the structure–activity relationships
(SAR) of this novel type of anticancer agent have so far
been limited;[3,8] the ascertained structural requirements for
the exceptionally strong antitumor activity of OSW-1 are the
presence of Ac and MBz groups on the disaccharide moiety
and b orientation of the 16-O-sugar unit.[3,8b] Detailed bio-
logical, toxicological, and pharmacokinetic studies await the
outcome of attempts to overcome a new synthetic challenge:
the preparation of a variety of analogues, derivatives, and
hopefully a lead compound in multigram quantities. Herein
we report a novel and efficient approach to the construction
of the 16b,17a-dihydroxycholest-22-one structure. One of the
[*] B. Shi, Prof. B. Yu
State Key Laboratory of Bioorganic
and Natural Products Chemistry
Shanghai Institute of Organic Chemistry
Chinese Academy of Sciences
354 Fenglin Road, Shanghai 200032 (P.R. China)
Fax: (+86)21-6416-6128
E-mail: byu@mail.sioc.ac.cn
H. Wu, Prof. J. Wu
Institute of Biochemistry and Cell Biology
Shanghai Institute for Biological Sciences
Chinese Academy of Sciences
Shanghai 200031 (P.R. China)
[**] This work was supported by the National Natural Science
Foundation of China (Grant nos. 20372070, 20321202, and
29925203) and the Chinese Academy of Sciences (Grant no. KGCX2-
SW-209).
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
4324
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/anie.200454237
Angew. Chem. Int. Ed. 2004, 43, 4324 –4327

Angewandte
Chemie
23-oxa-analogues of OSW-1 (1–3) prepared by this method is
a more potent inhibitor of the growth of tumor cells than
OSW-1.
Encouragingly, Gros and Doller found that b-face attack of
tert-butyl propionate by a lithiumenolate leads to a 16 a-
acetoxy-17-oxo-androstane.[11]
Scheme 2. Introduction of the 16a-OH group. a) CuBr2 (3.0 equiv),
CH3OH, reflux, 91%; b) TBSCl, imidazole, DMF, RT, overnight, 96%;
c) NaOH (1.3 equiv), DMF/CH2Cl2/H2O (3:2:1, v/v/v), RT, overnight,
95%. TBS=tert-butyldimethylsilyl, DMF=dimethylformamide.
A major challenge in the synthesis of OSW-1 is the
elaboration of the 16b,17a-dihydroxycholest-22-one struc-
ture. Fuchs and Guo assembled the side chain on 5-androsten-
3b-ol-17-one by a Wittig olefination and ene reaction
sequence. The 16b,17a-diol was then introduced by dihydrox-
ylation of the 16,17-ene group (1 equiv OsO4) followed by
inversion of the resulting 16a-OH group.[4] Jin and Yu
developed a novel protocol for the synthesis of OSW-1 that
involves 1,4-addition of an a-alkoxy vinyl cuprate to a 17(20)-
en-16-one. The 17a-OH group was introduced to the resulting
16,17-enolate by treatment with Davis reagent.[6] Morzycki
et al. used intramolecular ring opening of the 16a,17a-
epoxide by the 22-carbonyl function as a starting point for
further elaboration.[7a,b] We reanalyzed the characteristic
16b,17a-dihydroxycholest-22-one structure of OSW-1 sapo-
nins and envisioned the synthetic target as an a-methyl-b,g-
diol-one. This evaluation suggested that aldol condensation
(between an a-hydroxy ketone and a propionate) would be
the most direct approach to synthesis of the
The required 16a-hydroxy-5-androsten-17-one 7, with the
3-OH group protected by a TBS ether, was readily prepared
fromthe industrially produced acetate 4 by using a modified
version of a literature procedure (Scheme 2).[12] Bromination
of ketone 4 by treatment with CuBr2 (3 equiv) in methanol
under reflux provided 16a-bromide 5 in 91% yield (full
cleavage of the 3-OAc group). Protection of the 3-OH group
by treatment with TBSCl in the presence of imidazole in
DMF gave 6. Epimerization of the 16a-bromide and its
16b isomer took place readily. SN2 displacement of the 16b-
bromide by a hydroxide ion in DMF afforded the 16a-alcohol
7 in 95% yield.
The crucial aldol condensation of 16a-hydroxy-17-one 7
with propionate enolates was examined carefully (Scheme 3).
The lithium E enolate of ethyl propionate was generated by
desired product.[9] We hoped to achieve stereo-
control by using a rigid 16a-hydroxy-17-oxo
steroid as the ketone substrate, which would
lead to the desired natural stereochemistry at
C17 and C20 if the bulky 16a-OR group could
force the enolate to approach fromthe b face (by
nonchelation control,[10] Scheme 1, left). The
chair transition state (Zimmerman–Traxler
Scheme 3. Aldol condensation of 16a-hydroxy-17-one 7 with propionate enolates. Conditions:
A) 1. iPr2NH, nBuLi, ꢀ788C, 15 min; 2. HMPA, THF, ꢀ788C, then ethyl propionate, 0.5 h; 3. 7,
ꢀ788C, 63% (for 8a). B) 1. iPr2NH, nBuLi, ꢀ788C, 15 min; 2. propionates, THF, ꢀ788C, 0.5 h;
3. 7, ꢀ788C, 75% (8a), 12% (9a), 78% (8b), 81% (8c). HMPA=hexamethyl phosphoramide,
THF=tetrahydrofuran.
model)[10a] favors the E enolate over the Z en-
olate since the E conformation avoids steric
interaction of the 18-methyl group with the
methyl group of the propionate enolate (right).
the method described by Ireland et al.[13] When 7 was treated
with this enolate (A, Scheme 3), only the expected 17a-
hydroxy-21a-methyl product 8a was produced (in 63%
yield). Without the use of stereocontrol to generate exclu-
sively the E enolate of ethyl propionate (B, Scheme 3), the
condensation reaction predominantly led to the desired 21a-
methyl product 8a (75% yield), but its 21b-methyl isomer 9a
Scheme 1. Proposed construction of the 21a-methyl-16a,17a-diol
steroid structure by stereocontrolled aldol condensation.
was also isolated in 12% yield. Condensation with isobutyl
Angew. Chem. Int. Ed. 2004, 43, 4324 –4327
www.angewandte.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4325

Communications
and dodecyl propionate under conditions B gave predom-
16-OH group of the minor aldol product 9a (conditions
similar to those for 8!10, Scheme 4).
inantly the E enolates as a result of the greater size of the
isobutyl and dodecyl groups compared to the ethyl group;[14]
the desired 17a-hydroxy-21a-methyl compounds (8b and 8c)
were the only isolated products (78% and 81% yield,
respectively).
Once the desired 17a-hydroxy-21a-methyl-22-one archi-
tecture had been constructed, synthesis of OSW-1 and its
congeners was straightforward. However, we first turned our
attention to the synthesis of the 23-oxa analogues of OSW-1,
1–3, which we obtained from 8a–c in four steps by employing
transformations similar to those used in our previously
described OSW-1 synthesis.[5] The results are shown in
Scheme 4. Oxidation of the 16a-OH group of 8b by treatment
with TPAP/NMO[16] gave the 16-keto compound 10b in 91%
yield. Reduction of the resulting 16-ketone with NaBH4 in the
presence of CeCl3 afforded 16b-ol 11b stereoselectively in
81% yield. This reaction had to be quenched at ꢀ408C, as
higher temperatures resulted in the production of lactone 14.
The lactonization also took place in CDCl3 during NMR
measurement. Coupling of 16b,17a-diol 11b with disacchar-
ide imidate 12[5] in the presence of TMSOTf (0.1 equiv) and 4-
Š MS gave glycoside 13b in a satisfactory yield (64%),
without production of a significant amount of lactone 14.
Finally, deprotection of the silyl groups by treatment with
[PdCl2(MeCN)2][17] afforded the desired product 1 in 86%
yield. The congeners 2 and 3, which have a shorter or a longer
side chain than 1, were similarly prepared from 8a and 8c,
respectively.
The stereochemistry of the aldol products was confirmed
by analysis of their reaction products 14 and 15. Lactone 14
was produced in the course of inversion of the 16a-OH group
of 8 (10!11, Scheme 4) and was analyzed by NMR spectros-
copy. NOE interactions were observed between the a-
oriented protons at position 16 and on the 21-methyl group,
and between the b-oriented protons at position 20 and on the
18-methyl group. Single-crystal X-ray diffraction analysis was
performed on 15,[15] which was prepared by oxidation of the
23-Oxa-OSW-1 (1) strongly suppressed the growth of the
three types of malignant tumor cells tested. The concentration
of 1 required for 50% inhibition (IC50) was 0.031–3.1 mm,
which indicates that this compound is as potent as OSW-1 and
is 8–60 times more potent than the anticancer drug cisplatin in
the assay we used (Table 1). The short congener 2 was slightly
Table 1: Cytotoxic activities of 1–3 and the controls OSW-1 and cisplatin
against tumor cells.[a]
Tumor cells[b]
IC50 [mm]
1
2
3
OSW-1
cisplatin
AGS
7404
MCF-7
3.10
0.031
0.31
11.82
0.30
1.48
0.40
0.004
0.052
1.42
0.10
0.27
24.1
8.37
18.7
[a] The standard MTT assay was used (MTT=3-(4,5-dimethylthiazol-2-
yl)-2,5-diphenyltetrazolium bromide). The IC50 values determined for
cisplatin against the three cell lines used in our assay are consistent with
those determined by others.[15] [b] AGS=human stomach cancer cell
line, 7404=human liver carcinoma cell line, MCF-7=human breast
cancer cell line.
less potent than 1. Interestingly, the longer congener 3 was
considerably more potent than OSW-1 (as much as 25 times
more potent against the human liver carcinoma cell line
7404).
In summary, a novel and efficient approach to the
construction of the 16b,17a-dihydroxycholest-22-one steroi-
dal architecture characteristic of the saponin OSW-1 family
was developed that uses an aldol condensation as the key
reaction. Members of this family of compounds have excep-
tionally potent antitumor activities. The 23-oxa analogues of
OSW-1 1–3 were synthesized fromthe industrially produced
steroid 4 in 26, 30, and 20% overall yield, respectively, by a
Scheme 4. Completion of the synthesis of 23-oxa-OSW-1 (1) and its
congeners 2 and 3. a) TPAP, NMO, 4-Š MS, CH2Cl2, RT, overnight.
b) NaBH4, CeCl3·7H2O, THF, ꢀ108C, then ꢀ408C, MeOH, 15 min.
c) TMSOTf (0.1 equiv), 4-Š MS, CH2Cl2, ꢀ208C, 2 h.
d) [PdCl2(MeCN)2], acetone/water (20:1, v/v), RT, overnight. MBz=4-
methylbenzoyl, MS=molecular sieves, NMO=N-methylmorpholine
N-oxide, TES=triethylsilyl, TMSOTf=trimethylsilyl trifluoromethane-
sulfonate, TPAP=tetrapropylammonium perruthenate.
4326
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 4324 –4327

Angewandte
Chemie
linear eight-step procedure. The 23-oxa-OSW-1 (1) showed
similar antitumor potency to that of OSW-1, while the longer
congener 3 was even more potent. Progress toward an
understanding of the SARs and antitumor mechanisms of
OSW-1 saponins should be greatly accelerated by the devel-
opment described herein. The results of our ongoing research
will be reported in due course.
Received: March 11, 2004
Revised: June 2, 2004 [Z54237]
Keywords: aldol reaction · antitumor agents · OSW-1 analogues ·
.
saponins · synthetic methods
[1] S. Kubo, Y. Mimaki, M. Terao, Y. Sashida, T. Nikaido, T.
Ohmoto, Phytochemistry 1992, 31, 3969.
[2] a) Y. Mimaki, M. Kuroda, A. Kameyama, Y. Sashida, T. Hirano,
K. Oka, R. Maekawa, T. Wada, K. Sugita, J. A. Beutler, Bioorg.
Med. Chem. Lett. 1997, 7, 633; b) A. M. Rouhi, Chem. Eng. News
1995, 11, 28.
[3] a) M. Kuroda, Y. Mimaki, A. Yokosuka, F. Hasegawa, Y.
Sashida, J. Nat. Prod. 2002, 65, 1417; b) M. Kuroda, Y.
Mimaki, A. Yokosuka, Y. Sashida, Chem. Pharm. Bull. 2001,
49, 1042; c) M. Kuroda, Y. Mimaki, A. Yokosuka, Y. Sashida,
J. A. Beutler, J. Nat. Prod. 2001, 64, 88.
[4] C. Guo, P. L. Fuchs, Tetrahedron Lett. 1998, 39, 1099.
[5] S. Deng, B. Yu, Y. Lou, Y. Hui, J. Org. Chem. 1999, 64, 202.
[6] a) W. Yu, Z. Jin, J. Am. Chem. Soc. 2001, 123, 3369; b) W. Yu, Z.
Jin, J. Am. Chem. Soc. 2002, 124, 6576.
[7] a) J. W. Morzycki, A. Gryszkiewicz, I. Jastrzebska, Tetrahedron
Lett. 2000, 41, 3751; b) J. W. Morzycki, A. Wojtkielewicz,
Carbohydr. Res. 2002, 337, 1269; c) Q. Xu, X. Peng, W. Tian,
Tetrahedron Lett. 2003, 44, 9375.
[8] a) C. Guo, T. G. LaCour, P. L. Fuchs, Bioorg. Med. Chem. Lett.
1999, 9, 419; b) X. Ma, B. Yu, Y. Hui, Z. Miao, J. Ding, Bioorg.
Med. Chem. Lett. 2001, 11, 2153; c) L. Deng, H. Wu, B. Yu, M.
Jiang, J. Wu, Bioorg. Med. Chem. Lett. 2004, 14, 2781.
[9] M. T. Reetz, Angew. Chem. 1984, 96, 542; Angew. Chem. Int. Ed.
Engl. 1984, 23, 556.
[10] a) H. E. Zimmerman, M. D. Traxler, J. Am. Chem. Soc. 1957, 79,
1920; b) W. A. Kleschick, C. T. Buse, C. H. Heathcock, J. Am.
Chem. Soc. 1977, 99, 247; c) D. A. Evans, Top. Stereochem. 1982,
13, 1.
[11] D. Doller, E. G. Gros, Synth. Commun. 1990, 20, 3115.
[12] M. Numazawa, M. Nagaoka, Y. Osawa, J. Org. Chem. 1982, 47,
4024.
[13] a) R. E. Ireland, R. H. Mueller, J. Am. Chem. Soc. 1976, 98,
2891; b) R. E. Ireland, P. Wipf, J. D. Armstrong III, J. Org. Chem.
1991, 56, 650.
[14] J. Otera, Y. Fujita, S. Fukuzumi, Synlett 1994, 213.
[15] See the Supporting Information.
[16] L. A. Paquette, H. Wang, Z. Su, M. Zhao, J. Am. Chem. Soc.
1998, 121, 5213.
[17] B. H. Lipshutz, D. Pollart, J. Monforte, H. Kotsuki, Tetrahedron
Lett. 1985, 26, 705.
Angew. Chem. Int. Ed. 2004, 43, 4324 –4327
www.angewandte.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4327