Synthesis of analogues of a potent antitumor saponin OSW-1

Article abstract of DOI:10.1016/j.bmcl.2004.03.102

A series of side chain analogues (5a-e), a 22-glycosylated isomer (10), and 16β-O-L-arabinosyl (13a) or 16β-O-D-xylosyl (13b) analogues of OSW-1 were synthesized. All analogues were found to be less cytotoxic against breast and endometrial cancer cell lines than the natural product.

Full text of DOI:10.1016/j.bmcl.2004.03.102

Bioorganic & Medicinal Chemistry Letters 14 (2004) 3323–3326
Synthesis of analogues of a potent antitumor saponin OSW-1
a
b
Jacek W. Morzycki,a, Agnieszka Wojtkielewicz and Sławomir Wołczynski
*
ꢀ
aInstitute of Chemistry, University of Bialystok, al. Pilsudskiego 11/4, 15-443 Bialystok, Poland
bDepartment of Gynecological Endocrinology, Medical University of Bialystok, Kilinskiego 1, 15-230 Bialystok, Poland
Received 3 December 2003; revised 8 March 2004; accepted 19 March 2004
Abstract—A series of side chain analogues (5a–e), a 22-glycosylated isomer (10), and 16b-O-
(13b) analogues of OSW-1 were synthesized. All analogues were found to be less cytotoxic against breast and endometrial cancer cell
L
-arabinosyl (13a) or 16b-O-D-xylosyl
lines than the natural product.
ꢀ 2004 Elsevier Ltd. All rights reserved.
OSW-1 belongs to a family of saponins isolated 11 years
ago by Sashida et al. from the bulbs of Ornithogalum
saundersiae.1 The saponins appeared to be strongly
cytotoxic against a broad spectrum of malignant tumor
cells, such as leukemia HL-60, mouse mastrocarcinoma,
human pulmonary adenocarcinoma, large cell carci-
noma, and squamous cell carcinoma including adria-
mycin-resistant and camptothecin-resistant cell lines
with IC50 between 0.1 and 0.7 nM for the most active
OSW-1 (5f).2 This extraordinary cytotoxicity of OSW-1
encouraged several research groups to undertake efforts
of its synthesis.3 The first was Fuchs group which syn-
thesized OSW-1 aglycone, but it proved to be biologi-
Several analogues bearing the disaccharide moiety of
OSW-1 attached to nonsteroidal aglycones and steroid
glycosides with the sugar part at various positions (other
than 16b) were prepared and tested for cytotoxicity.10
All of them were many times less active than OSW-1.
Also any alteration in the sugar part, for example,
removal of the ester groups on the disaccharide moiety
or change of linkage of
L-arabinopyranose and D-
xylopyranose derivatives (into 1 fi 4), resulted in dra-
matic decrease of activity. It seems that an a glycoside
bond between 16b-hydroxysteroid and acylated sugar
moieties is a pharmacophore requirement.
cally
inactive.4
Fuchs
speculated
that
the
For further study of a structure–activity relationship
analogues with structures more close to OSW-1 were
designed, among them compounds with different size of
a side chain. The method for their synthesis was direct
glycosylation of a steroid aglycone in its cyclic form.
The required steroid aglycones were prepared by
addition of alkyllithium to the easily available11 hy-
droxylactone 1 (Scheme 1). Five aglycones (2a–f) were
obtained with 3 (using DIBAL-H), 6 (linear and
branched), 7, 8, or 10 carbon atom side chain. Each of
the aglycones was coupled with the OSW-1 disacchar-
ide. An alternative synthesis of the side chain ana-
logues of OSW-1 consisting of addition of alkyllithium
to the 16b-glycoside 22-aldehyde (compound 4a in
Scheme 2) failed due to the undesired retro-aldol
reaction leading to a formation of 17-ketone, a product
devoid of side chain. All aglycones existed in a cyclic
hemiketal form. The products were stereochemically
pure, but configuration of the new stereogenic center at
C-22 could not be elucidated from the spectra. Pre-
sumably, the thermodynamically favored 22R isomers
22-oxocarbenium ions might be the active intermediate
for the anticancer activity of both OSW-1 and cepha-
lostatins.5 Then Yu and Hui described complete syn-
thesis of OSW-1.6 The Chinese chemists employed
essentially the same method for the OSW-1 aglycone,
but for the first time described synthesis of a sugar part
and a glycosylation procedure. A different approach to
the synthesis of both parts of OSW-1, which were cou-
pled by the same trichloroacetimidate method, was re-
ported by the American chemists Yu and Jin.7 A new
strategy for synthesizing the aglycone of OSW-1 by
using the intact skeleton of diosgenin was recently re-
ported by Tian et al.8 One more synthesis of OSW-1 was
elaborated in our laboratory.9 Yu and Hui also proved
that both parts of OSW-1, the aglycone and the sugar
part, are equally important for the biological activity.
* Corresponding author. Tel.: +48-85-745-76-04; fax: +48-85-745-75-
81; e-mail: morzycki@uwb.edu.pl
0960-894X/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.03.102

3324
J. W. Morzycki et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3323–3326
OH
O
R
2a: R = H
b: R = n-C3H7
c: R = i-C3H7
d: R = n-C4H9
e: R = n-C5H11
f: R = i-C5H11
g: R = n-C7H15
O
O
O
H
O
H
RLi
1
OCH3
OCH3
Scheme 1.
OH
R
2a: R = H
b: R = n-C3H7
c: R = i-C3H7
d: R = n-C4H9
e: R = n-C5H11
f: R = i-C5H11
O
O
Disacch*:
Disacch:
O
H
AcO
OOTES
OOH
O
TESO
OMBz
TESO
OCH3
g: R = n -CH15
7
O
O
NH
AcO
Cl3C-C-O~Disacch*
TMSOTf
HO
OMBz
HO
O
O
R
ODisacch*
R
R
ODisacch
O
H
O
ODisacch*
O
H
O
H
H2O, p-TsOH
HO
5
3
4
OSW-1 (5f)
Scheme 2.
OH
O
OH
PDC
ODisacch*
ODisacch*
O
O
O
H
H
7
4f
OR
OH
NH
O
H
Cl3C-C-O~Disacch*
TMSOTf
ODisacch*
6
OCH3
PDC
OH
O
O
H
H
8
9: R = Disacch*
H2O, p-TsOH
10: R = Disacch,
5
(3β-OH, ∆ )
Scheme 3.
were formed. Glycosylation reactions were carried out
with the disaccharide trichloroacetimidate under the
standard conditions (TMSOTf was used as a pro-
moter).12 The conversion was about 20% and usually
two types of products were formed (Scheme 2), 22-O-
glycosides (3) and 16b-O-glycosides (4). The configu-
ration of glycosides at the anomeric position was found
to be a. The ratio of products depended strongly on
size of a side chain. In the case of compound 2b with a
short side chain the only product was 16b-O-glycoside,
whereas the reaction of aglycone with the longest side
chain (compound 2g) afforded exclusively 22-O-glycoside.

J. W. Morzycki et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3323–3326
3325
O
O
OH
OMonosacch*
OMonosacch
O
O
H
H
H
O
O
H
H2O, p-TsOH
12
13
HO
TMSOTf
2f
OMonosacch*
O
OCH3
NH
Cl3C-C-O~Monosacch*
O
OR'
RO
OR'
O
a:
b:
a:
b:
L
-arabinopyranosyl (4S; R = Ac, R' = TES)
-xylopyranosyl (4R; R = MBz, R' = TES)
-arabinopyranosyl (4S; R = Ac, R' = H)
-xylopyranosyl (4R; R = MBz, R' = H)
Monosacch* -
+
D
Monosacch
-
L
11
D
Scheme 4.
Table 1. Growth inhibition ratio (%) and IC50 for the cancer cell necrosis
Compound conc. (mol/L)
10À9
10À8
10À7
10À6
10À5
OSW-1 (5f)
MCF-7
87%
100
100
100
100
100
100
100
100
100
100
100
100
100
MDA–MB-231
Ishikawa cells 24 h
95%
Compound 5b
MCF-7
0
0
0
0
0
0
0
0
0
5
7
12
15
18
MDA–MB-231
Ishikawa cells 24 h
10
Compound 5d
MCF-7
0
0
0
0
0
0
0
0
0
15
12
17
27
23
28
MDA–MB-231
Ishikawa cells 24 h
Compound 5e
MCF-7
0
0
0
0
0
0
0
0
0
10
14
12
18
26
17
MDA–MB-231
Ishikawa cells 24 h
OSW-1 (5f)
MCF-7
MDA–MB-231
Ishikawa cells 24 h
IC50 for necrosis
7 · 10À6
5 · 10À7
3 · 10À6
Other glycosylation reactions afforded both products,
which were relatively easy to separate by flash chro-
matography and to distinguish by spectroscopic meth-
ods. The most characteristic was signal of a 20-H
proton (q) deshielded by a neighboring carbonyl group
in 16b-O-glycosides. It was easily achieved by treatment
with p-TsOH in dioxane-aqueous solution at 80 °C. All
novel compounds 5a–g (except for earlier prepared
OSW-1 5f) were fully characterized (selected spectral
and analytical data for 5e are given in Ref. 13) and
subjected to biological evaluation.
1
to d ꢀ 3:1 ppm in the H NMR spectra of compounds
4 (in 3 it appeared at ꢀ2.4 ppm). The signal of arabi-
nopyranose anomeric proton (ꢀ4.8 ppm in 4) was
shifted to d ꢀ 5:4 ppm in compounds 3. Also charac-
teristic was shape of a 16a-H proton at ꢀ4.2 ppm
(t, J ¼ 7:5 Hz) in compounds 3. Acetate protons ap-
peared at d ꢀ 2:0 ppm in 16b-O-glycosides (4), whereas
at ꢀ1.7 ppm in the 22-O-glycosylated isomers (3). An
obvious difference in the 13C NMR spectra was ob-
served in the chemical shift of C-22 (d ꢀ 218 ppm for 4
and ꢀ115 ppm for 3).
A synthesis of OSW-1 isomer with a sugar moiety at-
tached to oxygen atom at C-22 was also attempted
(Scheme 3). The OSW-1 aglycone (2f) was reduced with
lithium aluminum hydride to afford 16b,17a,22n-triol
6, which was glycosylated with 1 equiv of the disac-
charide trichloroacetimidate. A regioselective glycosyl-
ation at O-22 was expected on the basis of the previous
study on benzylation of this compound.9 However, the
reaction was not regioselective and glycosylation
occurred equally at both secondary positions (O-16 and
O-22). The products 7 and 8 were separated and oxi-
dized with pyridinium dichromate to the corresponding
ketones. One of them appeared to be identical in all
A final step of saponin synthesis (compounds 5) was
simultaneous deprotection of functional groups of both
steroid (cycloreversion) and sugar (desilylation) moieties

3326
J. W. Morzycki et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3323–3326
2. (a) Rouhi, A. M. Chem. Eng. News 1995 (September 11),
28; (b) Advances in Experimental Medicine and Biology;
Saponins Used in Traditional and Modern Medicine;
Waller, G. R., Yamasaki, K., Eds.; Plenum: New York,
1996; Vol. 404.
3. Gryszkiewicz-Wojtkielewicz, A.; Jastrzebska, I.; Mor-
zycki, J. W.; Romanowska, D. B. Curr. Org. Chem.
2003, 7, 1257.
respects with the previously obtained protected saponin
OSW-1 (4f). Interestingly, the 22-O-glycoside 9 proved
to be stereochemically pure compound, though the
starting triol 6 was a mixture of epimers at C-22.
Probably, an opposite epimer reacted faster at O-16.
However, the configuration at C-22 in the isolated
product could not be concluded from its spectra. The
protecting groups were removed in the same way as
described above and the novel OSW-1 analogue 10 was
obtained.13
4. Guo, C.; Fuchs, P. L. Tetrahedron Lett. 1998, 39, 1099.
5. Guo, C.; LaCour, T. G.; Fuchs, P. L. Bioorg. Med. Chem.
Lett. 1999, 9, 419.
6. Deng, S.; Yu, B.; Lou, Y.; Hui, Y. J. Org. Chem. 1999, 64,
202.
7. Yu, W.; Jin, Z. J. Am. Chem. Soc. 2002, 124, 6576.
8. Xu, Q.; Peng, X.; Tian, W. Tetrahedron Lett. 2003, 44,
9375.
The aglycone of OSW-1 in its hemiketal form (2f) was
also treated with the monosaccharide (L-arabinopyr-
anose and D-xylopyranose derivatives) trichloroacetim-
idates (Scheme 4). The conversions were slightly higher
(about 30%) than in the previous glycosylation reac-
tions. Again a mixture of 22-O- (11) and 16b-O-gly-
cosylated (12) products was formed in almost equal
amounts. The former products were subjected to de-
protection in a usual manner to afford new analogues
13a and 13b.13
9. Morzycki, J. W.; Wojtkielewicz, A. Carbohydr. Res. 2002,
337, 1269.
10. (a) Ma, X.; Yu, B.; Hui, Y.; Xiao, D.; Ding, J. Carbohydr.
Res. 2000, 329, 495; (b) Ma, X.; Yu, B.; Hui, Y.; Miao, Z.;
Ding, J. Carbohydr. Res. 2001, 334, 159; (c) Ma, X.; Yu,
B.; Hui, Y.; Miao, Z.; Ding, J. Bioorg. Med. Chem. Lett.
2001, 11, 2153.
11. (a) Morzycki, J. W.; Gryszkiewicz, A.; Jastrzebska, I.
Tetrahedron 2001, 57, 2185; (b) Morzycki, J. W.;
Gryszkiewicz, A. Polish J. Chem. 2001, 75, 983.
12. Schmidt, R. R.; Toepfer, A. Tetrahedron Lett. 1991, 32,
3353.
All described analogues were tested for cytotoxicity
against two breast cancer cell lines (MCF-7 and MDA–
MB-231) and endometrial cancer Ishikawa cell line.
OSW-1 (5f) influenced significantly [3H]thymidine incor-
poration, cell viability, and growth. The saponin induced
necrosis of the cells without apoptosis. The analogues
with linear side chain 5b, 5d, and 5e showed much lower
cytotoxicity that the saponin OSW-1. Other analogues
were not biologically active. Table 1 shows the results of
the growth inhibition tests and IC50 values determined for
the necrosis induced by OSW-1 in three lines of cancer.
IC50’s for the analogues (5b, 5d, 5e) were more than 1000
times higher than these determined for OSW-1.
13. Selected analytical data for compounds 5e, 10, and 13a.
5e: amorphous solid; IR(CHCl3): 3453, 1728, 1692, 1606,
1512, 1259, 1170, 1033; 1H NMR (200 MHz): 8.09 (d,
J ¼ 8:9, 2H), 6.98 (d, J ¼ 8:9, 2H), 5.35 (m, 1H), 4.94 (dd,
J ¼ 8:0, 7.2, 1H), 4.69 (m, 2H), 4.22 (brs, 1H), 4.17 (m,
1H), 4.15 (m, 1H), 3.88 (s, 3H), 3.75 (m, 1H), 3.66–3.73
(m, 2H), 3.40–3.55 (m, 3H), 2.65 (q, J ¼ 7:4, 1H), 1.95 (s,
3H), 1.03 (s, 3H), 1.01 (d, J ¼ 7:4, 3H), 0.86 (m, 3H), 0.80
(s, 3H); 13C NMR (50 MHz): 218.8 (C), 169.4 (C), 166.0
(C), 164.2 (C), 140.6 (C), 132.2 (2 · CH), 121.4 (CH), 121.3
(C), 114.0 (2 · CH), 102.2 (CH), 99.1 (CH), 88.5 (CH),
85.6 (C), 80.0 (CH); ESI-MS: 895.5 (MNaþ); HR-MS
calcd for C47H68O15Na (MNaþ): 895.4456; found:
895.4472. 10: IR(CHCl3): 3461, 1736, 1606, 1512, 1170,
1259, 1081; 1H NMR (200 MHz): 7.95 (d, J ¼ 8:7, 2H),
6.90 (d, J ¼ 8:7, 2H), 5.34 (m, 1H), 4.97 (m, 3H), 4.64 (d,
J ¼ 6:4, 1H), 4.52 (d, J ¼ 6:8, 1H), 3.96–4.04 (m, 2H),
3.94 (brs, 1H), 3.86 (s, 3H), 3.70–3.82 (m, 4H), 3.32–3.58
(m, 3H), 1.89 (s, 3H), 1.01 (s, 3H), 0.93 (d, J ¼ 7:0, 3H),
0.86 (d, J ¼ 6:2, 6H), 0.75 (s, 3H); 13C NMR (50 MHz):
218.8 (C), 169.7 (C), 166.3 (C), 163.9 (C), 141.0 (C), 132.2
(2 · CH), 121.4 (C), 121.0 (CH), 113.8 (2 · CH), 101.9
(CH), 100.6 (CH), 82.9 (C), 82.8 (CH), 80.7 (CH); ESI-
MS: 895.4 (MNaþ); HR-MS calcd for C47H68O15Na
(MNaþ): 895.4456; found: 895.4469. 13a: mp 199–
202 °C; IR (CHCl3): 3489, 1741, 1692, 1234, 1053. 1H
NMR (200 MHz): 5.33 (d, J ¼ 4:6, 1H), 4.73 (dd, J ¼ 6:2,
4.3, 1H), 4.34 (m, 2H), 3.74–3.96 (m, 4H), 3.58 (dd,
J ¼ 11:6, 3.0, 1H), 3.54 (m, 1H), 3.24 (d, J ¼ 7:6, 1H),
3.01 (q, J ¼ 7:3, 1H), 2.14 (s, 3H), 1.23 (d, J ¼ 7:3, 3H),
1.00 (s, 3H), 0.90 (d, J ¼ 6:2, 6H), 0.82 (s, 3H); 13C NMR
(50 MHz): 218.7 (C), 170.7 (C), 140.7 (C), 121.3 (CH), 99.5
(CH), 89.9 (CH), 85.4 (C); ESI-MS: 629.4 (MNaþ); Anal.
calcd for C34H54O9: C 67.30, H 8.97; found: C 67.17, H
9.02.
Acknowledgements
This work was supported by the State Committee for
Scientific Research (Grant No. 3 P05F 014 23). One of
the authors (A.W.) thanks Foundation for Polish Sci-
ence for a domestic grant for young scholars.
References and notes
1. (a) Kubo, S.; Mimaki, Y.; Terao, M.; Sashida, Y.;
Nikaido, T.; Ohmoto, T. Phytochemistry 1992, 31, 3969;
(b) Mimaki, Y.; Kuroda, M.; Kameyama, A.; Sashida, Y.;
Hirano, T.; Oka, K.; Maekawa, R.; Wada, T.; Sugita, K.;
Beutler, J. A. Bioorg. Med. Chem. Lett. 1997, 7, 633;
(c) Kuroda, M.; Mimaki, Y.; Yokosuka, A.; Sashida, Y.;
Beutler, J. A. J. Nat. Prod. 2001, 64, 88; (d) Kuroda, M.;
Mimaki, Y.; Yokosuka, A.; Sashida, Y. Chem. Pharm.
Bull. 2001, 49, 1042.