New analogues of the potent cytotoxic saponin OSW-1

Article abstract of DOI:10.1021/jm0613572

Saponin OSW-1 (5e-G2; 3β,16β,17α-trihydroxycholest-5-en-22- one 16-O-{O-[2-O-(4-methoxybenzoyl)-β-D-xylopyranosyl]-(1→3)-2-O- acetyl-α-arabinopyranoside}) analogues: with modified side chain (5a/d-G2), 22-deoxo-23,24,25,26,27-pentanor- (14), 22-deoxo-23-o

Full text of DOI:10.1021/jm0613572

J. Med. Chem. 2007, 50, 3667-3673
3667
New Analogues of the Potent Cytotoxic Saponin OSW-1
Agnieszka Wojtkielewicz,*,† Maciej Długosz,‡ Jadwiga Maj,† Jacek W. Morzycki,*,† Michał Nowakowski,‡ Joanna Renkiewicz,†
Miroslav Strnad,§ Jana Swaczynova´,§ Agnieszka Z. Wilczewska,† and Jacek Wo´jcik‡
Institute of Chemistry, UniVersity of Białystok, al. Pilsudskiego 11/4, 15-443 Bialystok, Poland, Institute of Biochemistry and Biophysics, Polish
Academy of Sciences, ul. Pawinskiego 5a, 02-106 Warsaw, Poland, and Laboratory of Growth Regulators, Palacky UniVersity and Institute of
Experimental Botany ASCR, Slechtitelu 11, 78371 Olomouc, Czech Republic
ReceiVed NoVember 24, 2006
Saponin OSW-1 (5e-G2; 3â,16â,17R-trihydroxycholest-5-en-22-one 16-O-{O-[2-O-(4-methoxybenzoyl)-
â-D-xylopyranosyl]-(1f3)-2-O-acetyl-R-arabinopyranoside}) analogues: with modified side chain (5a/d-
G2), 22-deoxo-23,24,25,26,27-pentanor- (14), 22-deoxo-23-oxa- (17), glycosylated with various monosac-
charides (5e-G4/G6/G8), and OSW-1 structural isomer (10) were obtained. The analogues were synthesized
using a previously published method for the synthesis of OSW-1. The structures of analogues were fully
confirmed by spectroscopic methods, and the S-chirality at C-22 of the structural isomer was established by
conformational analysis combined with the NMR spectrometry. The cytotoxicity of the analogues toward
several types of malignant tumor cells was examined and compared with that of OSW-1. The results suggest
that modification of the steroidal aglycone may lead to compounds with high cytotoxicity.
1. Introduction
could tolerate certain modifications without significant loss of
its antiproliferative potency. This includes replacement of the
ketone group in the side chain by an ester group. Attempts to
synthesize structurally simpler OSW-1 analogues with full
antitumor activity led to the discovery of a potent analogue
containing a monosaccharide sugar part.26 However, much
remains to be discovered about the structural elements of saponin
OSW-1 that are associated with its high activity, so we
synthesized two series of novel analogues of the compound with
modified side chains and various sugar moieties, and evaluated
their toxicity toward a battery of cancer cell lines.
The value of screening extracts of natural product in the
search for novel, complex lead structures for drug discovery
has been demonstrated in many studies.1-5 Indeed, most
currently used cancer drugs are synthesized by simple modifica-
tions of natural products.6 Recently, a new group of saponins
was isolated from the bulbs of Ornithogalum saundersiae, a
perennial grown in southern Africa, where it is cultivated as a
cut flower and garden plant.7 These saponins, each of which
contains a novel 16â,17R-dihydroxycholest-22-one aglycone
unit glycosylated at the 16-OH group with an acylated disac-
charide, proved to be strongly cytotoxic, with very similar
cytotoxicity profiles to those of cephalostatins.8-10 The most
abundant saponin in the plant, OSW-1, is weakly toxic toward
normal cells but inhibits the growth of various types of
malignant tumor cells and is 10-100 times more potent than
clinically applied anticancer agents, such as adriamycin, cispl-
atin, camptothecin, and taxol.
The action mechanism of OSW-1 has been recently shown
to damage the mitochondrial membrane and cristae in human
leukemia and pancreatic cancer cells, leading to losses of
transmembrane potential, increases in cytosolic calcium contents,
and activation of calcium-dependent apoptotic pathways.11 This
mechanism differs from those of all other anticancer compounds
examined to date. Thus, OSW-1 has high apparent potential
for effectively treating some cancers that are strongly resistant
to currently available drugs, and it clearly warrants detailed
further investigation.
2. Results and Discussion
We have already published preliminary reports on the
synthesis of OSW-1 analogues,27,28 using a procedure based on
the previously described synthesis of saponin OSW-1.17 One
of the intermediates in the OSW-1 synthesis route, 17R-
hydroxylactone 1, was successfully used to synthesize analogues
with modified side chains. Various aglycones with side chains
of various sizes and shapes (linear or branched) in hemiketal
forms (2a-f) were obtained by reacting 1 with corresponding
alkyllithium reagents in high yields (Scheme 1). All aglycones
were coupled with the OSW-1 disaccharide trichloroacetimidate
(CCl3C(NH)O∼G1) prepared according to the procedure de-
scribed in the literature.16 The reactions were catalyzed by
trimethylsilyl triflate. Apart from the desired 16â-glycosides
4a-f, variable amounts of isomeric hemiketal glycosides 3a-f
were formed (Scheme 2). In the case of 2a, the glycosylation
product 3a was not found in the reaction mixture, in contrast
to the reaction with 2f, which did not afford the desired 16â-
glycoside 4f. The observed difference in the results of the
attempted glycosylation may be attributable to steric hindrance
present in the cyclic aglycone. During glycosylation, the glycosyl
donor may be attacked by either the oxygen atom at C16 (for
16â-O-glycosides) or that at C22 (for 22-O-glycosides). Because
there is no spectral evidence for the presence of an open-chain
form in equilibrium with the cyclic aglycone in any cases
studied, the desired 16â-O-glycosides were probably formed
by the direct attack of the ring oxygen on the sugar donor. The
relative rates of the competing reactions determine the propor-
tions of the products. However, there is no simple explanation
Several research groups have described the synthesis of
OSW-1 aglycone.12-15 Numerous methods of saponin OSW-1
semisynthesis are also known, including coupling of the
aglycone with the sugar moiety.14,16-18 Various analogues of
OSW-1 have also been prepared and tested for cytotoxicity
recently.19-25 It was found that the C17 side chain of OSW-1
* To whom correspondence should be addressed. Tel.: +48 85 7457604
(A.W.). Fax: +48 85 7457581 (A.W.). E-mail: jeremy@uwb.edu.pl.
Tel.: +48 85 7457585 (J.W.M.). Fax: +48 85 7457581 (J.W.M.). E-mail:
morzycki@uwb.edu.pl (J.W.M.).
† University of Białystok.
‡ Polish Academy of Sciences.
§ Palacky University and Institute of Experimental Botany ASCR.
10.1021/jm0613572 CCC: $37.00 © 2007 American Chemical Society
Published on Web 07/03/2007

3668 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 15
Wojtkielewicz et al.
Scheme 1
quaternary carbon C-22 signal appeared at about 115 ppm for
compounds 3, while 16â-glycosides 4 showed a characteristic
signal for ketones at ∼218 ppm. The last step of the synthesis
of the saponin OSW-1 analogues 5a-e consisted of the
simultaneous removal of the protective groups from both the
steroid and the sugar moieties by hydrolysis in the presence of
p-TsOH under controlled conditions (2 h at 80 °C). The best
overall yield from 17R-hydroxylactone 1 was obtained for
saponin 5c (9%).
Scheme 2a
To obtain an isomer of OSW-1 with a carbonyl group at C-16
and a sugar moiety attached at C-22 (compound 10), the 16â,-
17R,22-triol 6 was glycosylated (Scheme 3). Compound 6 was
obtained by a LiAlH4 reduction of OSW-1 aglycone in its
hemiketal form (compound 2e). The stereoselectivity of the
reaction was not satisfactory s a mixture of epimers at C-22
was formed although the mixture was homogeneous according
to TLC analysis. Previous studies on the benzylation13 of this
compound indicated that the hydroxyl group at C-22 is the most
reactive one. However, the reaction of the 16â,17R,22-triol 6
with glycosyl trichloroacetimidate (CCl3C(NH)O∼G1) was not
regioselective. Both products, glycosylated at C-16 (7) and at
C-22 (8), were formed. They were quite easy to separate because
compound 7 proved to be significantly less polar than its
regioisomer 8. Both TLC and NMR analysis indicated that
compound 8 was homogeneous. One of the epimers at C-22
appeared to react more quickly with a glycosyl donor than the
other. The regioisomers (7 and 8) were subjected to oxidation
with pyridinium dichromate. The isomeric ketones showed very
similar polarity. One of them proved to be identical to the
protected OSW-1 (4e-G1) previously obtained. The isomeric
product 9 was subjected to simultaneous desilylation and
cycloreversion with p-TsOH to afford the saponin OSW-1
isomer 10. The structure of compond 10 was fully confirmed
by spectral analysis (IR, NMR, MS).
We also took effort to assign the configuration at the C-22
stereogenic center. The combination of NMR data with the
exhaustive conformational analysis for the analogue in question
allowed us to solve the problem. Initially, almost all proton and
carbon signals were assigned by different types of NMR
measurements. The chemical shifts observed in two solvents
(Table 1) were in good agreement with those measured earlier
for similar compounds.7,29,30 The correlation signals observed
in the 2D homonuclear spectra of compound 10 as well as those
observed in their 2D heteronuclear spectra revealing long range
connectivities allowed us to fully confirm the structure. Thus,
numerous correlations between sugar units, between sugar and
OMBz moieties, and from the aliphatic chain to either ring D
or sugar were observed. This in turn allowed us to determine
the positions of all groups and residues. However, the analysis
provided no clear, unambiguous evidence for the chirality at
C-22. Therefore, conformational analysis of both possible
isomers was carried out.
a TMSOTf, trimethylsilyl triflate; MBz, 4-methoxybenzoyl; TES, trieth-
ylsilyl; p-TsOH, 4-toluenesulfonic acid.
for the obtained results because both the structure and the
conformation of the side chain affect the course of reactions.
Products 3 and 4 were readily distinguished by analysis of
1
1
their H and 13C NMR spectra. The most characteristic H
signals of the 16â-glycosides 4 are a quartet at about 3.1 ppm
deriving from the proton at C-20, a doublet (δ ∼1.1 ppm) arising
from the 21-methyl protons, and a peak at around 4.8 ppm
corresponding to the anomeric proton. For compounds 3, the
anomeric proton signal was shifted downfield to ∼5.4 ppm, the
20-H quartet appeared at ∼2.4 ppm, and there was a charac-
teristic upfield shift of acetate protons from their normal
resonance at ∼2 ppm to ∼1.7 ppm. In the 13C NMR spectra, a
Because the signal assignment obtained for 10 in chloroform
was more complete than that obtained in the mixed solvent
(chloroform/pyridine), we performed the conformational analysis
using the data acquired from the former. The number of papers
concerning full conformational analysis of saponins is rather
limited.31,32 In most other cases, structural analyses have been
restricted to evaluations of the isolated conformations.30 Here
we attempted to analyze the conformational properties of
saponin 10 by means of an exhaustive conformational search.
The final set of 500 structures obtained for each isomer (R and
S) by the procedure described in detail in the Experimental
Section consists of molecular structures corresponding to
Scheme 3

Analogues of the Potent Cytotoxic Saponin OSW-1
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 15 3669
Table 1. Proton and Carbon NMR Chemical Shifts (in ppm) Measured
for 10 in Solutiona
in CDCl3
1H
in CDCl3/C5D5N 1:1 (v/v)
atom
13C
1H
13C
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
1′
2′
OAc
3′
4′
5′
1′′
2′′
3′′
4′′
5′′
1.120; 1.602
1.514; 1.859
3.543
31.44
31.42
71.62
42.11
NAb
121.08
31.04
NAb
1.001; 1.745
1.456; 1.790
3.412
NAb
35.02
74.52
NAb
144.80
124.00
NAb
45.74
52.59
39.95
NAb
2.234; 2.298
2.190; 2.226
5.340
5.229
1.581; 1.862
1.479; 1.782
NAb
NAb
1.058
49.17
NAb
NAb
1.491; 1.608
1.259
20.50
32.03
NAb
43.85
37.84
218.81
NAb
14.06
19.38
35.52
7.59
82.83
37.05
35.11
28.16
22.58
22.58
100.58
70.53
20.64
80.84
55.59
63.83
101.87
74.29
74.30
69.67
64.59
NAb
1.390; 1.443
1.533; 1.748
34.38
49.62
47.29
NAb
1.886
1.682; 2.109
1.848
1.628; 2.132
NAb
86.27
17.20
22.75
38.94
10.89
86.00
39.96
38.36
31.50
25.56
26.25
104.70
73.66
NAb
83.34
70.89
68.06
105.30
77.16
77.93
73.01
68.74
166.50
135.30
116.90
166.80
58.79
168.60
0.743
1.013
1.876
0.942
0.652
0.923
1.797
0.833
Figure 1. Sets of conformers of compound 10 obtained by the
procedure described in the text. Molecules are superposed with regard
to heavy atoms of the steroid moiety. The lowest energy conformer is
shown in bold: (A) 58 conformers of the 22S isomer; (B) 29 conformers
of the 22R isomer. Lowest energy conformers with a cutoff of 5 kcal/
mol: (C) 30 conformers of the 22S isomer; (D) 15 conformers of the
22R isomer.
4.961
4.884
1.114; 1.856
0.970; 1.123
1.497
0.859
0.859
4.571
5.037
1.796
3.800
4.048
3.523; 3.736
4.682
4.934
3.718
1.177; 1.489
0.886; 1.063
1.408
0.764
0.771
4.439
observed for meta protons. These findings indicate that one
might be able to establish the chirality at C-22 by direct
quantitative comparison of the conformationally averaged
distances measured by NMR with those obtained from confor-
mational analysis. However, this approach did not work because
all distances obtained from NOE for the saponin were system-
atically shorter than those obtained from generated structures.
It seems that the main source of this systematic discrepancy
comes from the indirect effect,33 which clearly disturbs not only
the intensity of NOE of methylene protons usually used for
calibration but also the intensity of all peaks measured for the
multispin system of saponins. In this situation, only qualitative
analysis was possible. Figure 2 shows the distances between
ortho protons of the OMBz ring and 7R H, 9R H, 14R H, and
15R H observed in the conformers of both isomers of 10. The
figure shows that quite substantial proportions of generated
conformers of the isomer S have distances of about 5 Å, which
is not the case for the isomer R. Furthermore, only one
conformer of the S isomer was found in which the MBz ring is
positioned over the â surface of the steroid; in all other
conformers of this isomer, the MBz ring is over the R surface.
The opposite situation obtains with the R isomer, where only
two conformers were identified in which the MBz ring lies over
the R surface. It is important to note that the relative population
of the identified conformers cannot be accurately predicted, but
these findings strongly suggest that the conformational prefer-
ences of the C-22 epimers are distinct and support the assign-
ment of compound 10 as the C-22 S epimer.
5.031
NAb
3.676
3.937
3.412; 3.925
4.624
4.971
3.697
3.722
3.310; 4.023
3.823
3.395; 4.136
1′′′
2′′′; 6′′′
3′′′; 5′′′
4′′′
4′′′OMe
7′′′
7.972
6.915
132.22
113.86
NAb
7.870
6.789
3.864
55.67
3.741
NAb
a Numbers followed by ′, ′′, and ′′′ refer to carbon atoms in arabinose,
xylose, and the MBz group, respectively. b NA ) not assigned.
possible minima in the conformational space of the studied
molecules. As some of the conformers were very similar to each
other, to simplify analysis, the values of dihedral angles
measured in generated structures were chosen as a criterion
describing their similarity. Two conformers were assumed
identical if the largest difference observed between the values
of all of their matching dihedral angles was less than 2 degrees
and only one of them was further considered. This procedure
led to a total number of 29 different conformers of R and 58
conformers of S; see Figure 1. For clarity, the numbers of lowest
energy conformers of R and S after applying a cutoff of 5 kcal/
mol were reduced to 15 and 30, respectively, see Figure 1.
Inspection of Figure 1 indicates that there is a distinctive
difference in conformational freedom between the two isomers
of 10. In the S isomer, OMBz groups in the sugar units are
generally positioned over the R-surface of the steroid moiety,
whereas in the R isomer, these moieties are positioned over the
â-surface. In ROESYa spectra, four cross-peaks have been
observed between ortho protons of the OMBz group and four
protons of the R-surface, namely, the 7R, 9R, 14R, and 15R
protons. Clearly weaker cross-peaks of the same type were
In a recent study21 it was claimed that the size of the side
chain and even the presence of a carbonyl group were not very
important for cytotoxicity of saponins. To check this hypothesis,
an analogue 14 lacking the carbonyl group with a short side
a Abbreviations: NA, not assigned; NT, not tested; DEPT, distortionless
enhancement by polarization transfer; DQCOSY, double quantum filtered
correlated spectroscopy; GHSQC, gradient heteronuclear single quantum
coherence; HMBC, heteronuclear multiple-bond correlation; ROESY,
rotating-frame Overhauser effect spectroscopy; TOCSY, total correlation
spectroscopy.

3670 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 15
Wojtkielewicz et al.
Scheme 5
of OSW-1, but different glycosyl donors were used (G3, G5,
G7, or G9). Generally, yields for coupling aglycone 2e with
monosaccharide trichloroacetimidates (D-xylopyranose and L-
arabinopyranose derivatives) were higher than those obtained
with CCl3C(NH)O∼G1. However, the reaction with the glucose
derivative was rather sluggish, and the desired 16â-glucoside
was not obtained (the hemiketal glucoside 3e-G9 was the major
product). In some reactions, orthoacetates (e.g., 4e-G10), which
showed a characteristic signal in the 13C NMR spectrum at about
122 ppm, were isolated as byproducts. Both hemiketal glyco-
sides and orthoacetates proved to be acid-sensitive. During
attempts to deprotect them under acidic conditions (p-TsOH),
these products underwent fragmentation to the starting sugar
and furan derivative of the steroid aglycone. The formation of
this compound with a heteroaromatic ring E under acidic
conditions from OSW-1 aglycone via a Ferrier type of rear-
rangement was reported several times in the literature.12,34,35
The anticancer activity of the new OSW-1 analogues was
evaluated in vitro using eight cancer cell lines of different
histopathological origins and normal mouse fibroblast NIH 3T3
cells. The results are summarized in Table 2. The cancer cell
lines exhibited distinct sensitivity to OSW-1 (5e-G2) and its
analogues, with CEM, K 562, and A 549 cell lines being the
most sensitive and G 361 human melanoma cells being the least
sensitive. In contrast, no compounds tested showed cytotoxicity
to the normal mouse NIH 3T3 fibroblasts. Among the tested
analogues, the most active appeared to be compound 5d-G2,
which showed similar antitumor potency to that of OSW-1.
Against the ARN 8 cell line, it was even 10 times more active
than natural saponin. However, shortening of the alkyl side chain
(5a-G2) led to a slight loss of activity. The results demonstrate
that small variations in the structure, for example, in the size
of the cholestane side chain, do not affect antitumor activity
significantly.21-23
Figure 2. Comparison of distances between ortho protons of the OMBz
ring and 7R H, 9R H, 14R H, and 15R H observed in the conformers
of the two isomers of 10 shown in Figure 1: (A) isomer 22S; (B) isomer
22R.
Scheme 4
chain was designed. Lactol 11, readily available from lactone
1 by DIBAL-H reduction, was treated with hydrazine and then
with potassium t-butoxide in DMSO. The Wolff-Kishner
reduction afforded 16â,17R-diol 12, which was glycosylated
with CCl3C(NH)O∼G1. Removal of protective groups in 13
with simultaneous cycloreversion resulted in the desired ana-
logue 14 (Scheme 4).
Further studies22,23 have shown that compounds containing
an oxygen atom instead of a carbon atom at position 23 are
equally or even more potent than saponin OSW-1. Combining
this finding with the previous mentioned claim that the carbonyl
group is not essential for biological activity,21 the simplified
OSW-1 analogue 17 lacking the 22-carbonyl group and with
oxygen in place of C-23 was designed, as a good candidate for
a cytotoxic agent. The required ether aglycone 15 (which we
previously synthesized from lactol 11 via reduction followed
by a selective Williamson reaction28) was subjected to glyco-
sylation with the OSW-1 disaccharide. After routine removal
of the protective groups, the desired 23-oxa-22-deoxo-OSW-1
analogue (17) was obtained in 20% overall yield (four steps
from lactol 11; Scheme 5).
The monosaccharide analogues of OSW-1 (5e-G4, 5e-G8)
appeared to be lethal to 50% of the tumor cells at concentrations
of 0.2-7.7 µM, that is, they are about 1000 times less active
than OSW-1 (5e-G2). These findings clearly indicate that the
disaccharide moiety is essential for the antitumor activities of
OSW-1.
Compound 10, the structural isomer of saponin OSW-1, was
also tested using the panel of cancer and normal cells. The
cytotoxicity (TSC50) values of 10 varied between 0.28-
14.4 µM and were about 1000 times lower than that of OSW-
1, proving that the position of the disaccharide moiety is also
important.
Contrary to previously published assertions,21 the presence
of a carbonyl group at C22 also seems to be a pharmacophore
requirement. Compound 14 showed much lower activity in
anticancer tests than OSW-1. Similarly, the OSW-1 ether
analogue 17 appeared to be much less active than the very potent
ester analogue (23-oxa analogue of OSW-1) described recently.22
A number of OSW-1 analogues with different sugar moieties
were obtained. We reasoned that truncation of the disaccharide
residue in OSW-1 into a monosaccharide derivative with the
acetyl and p-methoxybenzoyl groups retained at their positions
would not significantly affect its antiproliferative activity. The
synthetic procedure was essentially the same as for synthesis
3. Experimental Section
3.1. General Remarks. Melting points were determined using
a Kofler apparatus of the Boetius type. NMR spectra were recorded
with a Bruker AC 200F or Varian UNITY 500plus (equipped with

Analogues of the Potent Cytotoxic Saponin OSW-1
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 15 3671
Table 2. Antitumor Activities of the New OSW-1 Analogues (TCS50 in nM) in the Calcein AM Cytotoxicity Assaya
cell line (TCS50 in nM)
compound
5e-G2
5a-G2
5d-G2
14
CEM
MCF7
K 562
ARN 8
G 361
HeLa
HOS
A 549
NIH 3T3
0.3 ( 0.03
0.5 ( 0.02
0.2 ( 0.01
1300 ( 90
280 ( 10
<200
200 ( 20
1000 ( 80
62 ( 10
50 ( 2
0.8 ( 0.05
1.4 ( 0.4
0.7 ( 0.05
5100 ( 73
430 ( 17
<400
430 ( 20
1800 ( 35
230 ( 9
4 ( 0.4
3 ( 0.2
0.6 ( 0.05
6800 ( 66
500 ( 7.5
<400
1950 ( 61
1500 ( 27
NTb
1000 ( 31
3100 ( 26
1400 ( 52
9300 ( 101
14400 ( 99
1900 ( 78
2200 ( 91
3500 ( 85
670 ( 20
8 ( 0.4
40 ( 2
0.5 ( 0.01
0.9 ( 0.03
0.5 ( 0.025
3200 ( 85
480 ( 39
<200
4200 ( 96
7200 ( 120
NTb
>50 000
>50 000
>50 000
>50 000
>50 000
>50 000
>50 000
>50 000
NTb
90 ( 4
12 ( 0.5
8 ( 0.15
6400 ( 87
1000 ( 54
350 ( 25
570 ( 31
2300 ( 100
670 ( 31
140 ( 11
40 ( 1.5
42 ( 6
18200 ( 130
3400 ( 66
860 ( 32
1300 ( 15
3500 ( 63
140 ( 12
17400 ( 202
11700 ( 150
1800 ( 63
4200 ( 92
7200 ( 103
330 ( 49
10
17
5e-G4
5e-G8
daunorubicin
a The values shown are the mean ( SDs obtained in three experiments. Daunorubicin was used as a positive control. b NT ) not tested.
a Performa II gradient generator unit, WFG, Ultrashims, high
stability temperature unit and a 5 mm 1H/13C/15N-PFG triple
resonance inverse probe head) spectrometer using CDCl3/C5ND5
solutions with TMS as the internal standard. Only selected signals
in the 1H NMR spectra are reported. Infrared spectra were recorded
on a Nicolet series II Magna-IR 550 FT-IR spectrometer from
chloroform solutions. Mass spectra were obtained at 70 eV with
an AMD-604 spectrometer. The reaction products were isolated
by column chromatography performed on 70-230 mesh silica gel
(J. T. Baker). TLC was carried out using commercially available
plates (Merck, Silica gel 60 F254).
6â-Methoxy-16â,17R-dihydroxy-3R,5-cyclo-27-nor-5R-
cholestan-22-one 16-O-{O-[2-O-(4-methoxybenzoyl)-3,4-di-O-
triethylsilyl-â-D-xylopyranosyl]-(1f3)-2-O-acetyl-4-O-triethylsilyl-
R-L-arabinopyranoside} (4c-G1). Yield 9%; an oily product eluted
with benzene-ethyl acetate (98:2). IR (CHCl3) 3467, 1737, 1717,
1607, 1459, 1256, 1101, 1018 cm-1. 1H NMR (200 ΜΗz, CDCl3)
δ 8.06 (2 H, d, J ) 8.8 Hz), 6.92 (2 H, d, J ) 8.8 Hz), 4.92 (1 H,
t, J ) 5.0 Hz), 4.81 (1 H, m), 4.75 (1 H, m), 4.37 (1 H, br s), 4.23
(1 H, m), 3.95 (1 H, m), 3.88 (3 H, s), 3.63-3.80 (3 H, m), 3.32
(3 H, s), 3.11 (1 H, q, J ) 7.3 Hz), 2.76 (1 H, m), 1.95 (3 H, s),
1.13 (3 H, d, J ) 7.3 Hz), 1.02 (3 H, s), 0.40 (1 H, dd, J ) 8.0, 5.1
Hz).
Compound 1, glycosyl trichloroacetimidates (CCl3C(NH)O∼G),
lactol 11, and ether 15 were obtained according to procedures
described, respectively, in refs 34, 16, 35, and 28.
6â-Methoxy-3R,5-cyclo-27-nor-5R-furostane-17R,22R-diol 22-
O-{O-[2-O-(4-methoxybenzoyl)-3,4-di-O-triethylsilyl-â-D-xylopy-
ranosyl]-(1f3)-2-O-acetyl-4-O-triethylsilyl-R-L-arabinopyrano-
side} (3c-G1). Yield 9%; an oily product eluted with benzene-
ethyl acetate (96:4). IR (CHCl3) 3510, 1727, 1607, 1511, 1256,
3.2. Chemical Synthesis. Representative Procedure for Syn-
thesizing Aglycones 2a-2f. A solution of n-butyllithium in
anhydrous ether was prepared from lithium (13 mmol) and n-butyl
bromide (13 mmol). The reagent was added, dropwise, over 1 h to
a stirred solution of lactone 1 (1.3 mmol) in 100 mL of anhydrous
ether at room temperature under argon. The reaction mixture was
stirred for an additional hour, then quenched with saturated aqueous
NH4Cl, and the reaction product was extracted with ether. Evapora-
tion of the solvent from dried (anhydrous MgSO4) extract afforded
crude product 2c, which was purified by silica gel column
chromatography.
1
1098, 1010 cm-1. H NMR (200 ΜΗz, CDCl3) δ 7.96 (2 H, d, J
) 8.8 Hz), 6.89 (2 H, d, J ) 8.8 Hz), 5.33 (1 H, d, J ) 3.3 Hz),
4.97-5.03 (2 H, m), 4.68 (1 H, d, J ) 6.7 Hz), 4.23 (1 H, t, J )
7.5 Hz), 3.93-4.19 (3 H, m), 3.86 (3 H, s), 3.60-3.82 (3 H, m),
3.55 (1 H, dd, J ) 11.8, 3.5 Hz), 3.32 (3 H, s), 3.25 (1 H, m), 2.77
(1H, m,), 2.37 (1 H, q, J ) 7.0 Hz), 1.74 (3 H, s). 13C NMR (50
MHz, CDCl3) δ 169.8 (C), 164.5 (C), 163.2 (C), 131.7 (2 × CH),
122.8 (C), 115.1 (C), 113.4 (2 × CH), 102.0 (CH), 91.8 (CH),
90.1 (CH), 89.8 (C), 82.1 (CH). ESI-MS m/z (%) 1238.2 (MNa+).
Representative Procedure for the Deprotection of Glycosides.
To the solution of a protected glycoside 4c (0.012 mmol) in a
mixture of dioxane (1.5 mL) and water (0.05 mL), p-TsOH × H2O
(0.002 mmol) was added. The reaction mixture was stirred for 2 h
at 80 °C, the solvent was evaporated in vacuo at a temperature
below 50 °C, and the residue was chromatographed on a silica gel
column.
6â-Methoxy-3R,5-cyclo-27-nor-5R-furostane-17R,22R-diol (2c).
Yield 73%; an oily product eluted with hexane-ethyl acetate (77.5:
1
22.5). IR (CHCl3) 3594, 3469, 1091 cm-1. H NMR (200 ΜΗz,
CDCl3) δ 4.17 (1 H, t, J ) 7.5 Hz), 3.32 (3 H, s), 3.15 (1 H, br s),
2.78 (1 H, m), 2.29 (1 H, q, J ) 7.4 Hz), 1.04 (3 H, s), 0.94 (3 H,
d, J ) 7.4 Hz), 0.92 (2 × 3 H, m), 0.66 (1 H, m), 0.45 (1 H, dd,
J ) 7.8, 5.2 Hz). 13C NMR (50 MHz, CDCl3) δ 111.2 (C), 90.5
(C), 90.3 (CH), 82.0 (CH), 56.4 (CH3), 52.2 (CH), 47.5 (CH), 44.5
(C), 43.3 (C), 42.2 (CH), 37.4 (CH2), 35.01 (C), 34.97 (CH2), 33.2
(CH2), 32.0 (CH2), 30.8 (CH2), 29.9 (CH), 25.5 (CH2), 24.8 (CH2),
22.8 (CH2), 22.2 (CH2), 21.3 (CH), 19.2 (CH3), 17.6 (CH3), 13.9
(CH3), 13.0 (CH2), 8.4 (CH3). EI-MS m/z (%) 414 (M - H2O, 21),
385 (47), 382 (9), 269 (100).
3â,16â,17R-Trihydroxy-27-norcholest-5-en-22-one 16-O-{O-
[2-O-(4-methoxybenzoyl)-â-D-xylopyranosyl]-(1f3)-2-O-acetyl-
R-L-arabinopyranoside} (5c-G2). Yield 98%; amorphous solid
eluted with ethyl acetate. IR (CHCl3) 3591, 3453, 1728, 1692, 1606,
1512, 1259, 1170, 1033 cm-1. 1H NMR (500 MHz, CDCl3) δ 8.10
(2 H, d, J ) 8.7 Hz), 6.99 (2 H, d, J ) 8.7 Hz), 5.34 (1 H, m),
4.94 (1 H, dd, J ) 8.0, 7.2 Hz), 4.71 (2 H, m), 4.20 (1 H, br s),
4.17 (1 H, br s), 4.14 (1 H, dd, J ) 11.6, 4.2 Hz), 3.89 (3 H, s),
3.77 (1 H, m), 3.66-3.73 (2 H, m), 3.38-3.58 (4 H, m), 3.23 (1
H, br s), 2.78 (1 H, br s), 2.67 (1 H, q, J ) 7.4 Hz), 1.97 (3 H, s),
1.04 (3 H, s), 1.02 (3 H, d, J ) 7.4 Hz), 0.82 (3 H, t, J ) 7.0 Hz),
0.80 (3 H, s). 13C NMR (125 MHz, CDCl3) δ 218.9 (C), 170.1
(C), 168.3 (C), 166.4 (C), 142.8 (C), 132.5 (2 × CH), 121.7 (CH),
121.4 (C), 114.3 (2 × CH), 102.4 (CH), 99.4 (CH), 88.7 (CH),
85.8 (C), 80.2 (CH). ESI-MS m/z (%) 881.7 (MNa+). Anal.
(C46H66O15) C, H.
Procedure for Oxidation of Compounds 7 and 8. To a solution
of dihydroxy-compound 7 or 8 (0.008 mmol) in dichloromethane
(3 mL), pyridinium dichromate (0.016 mmol) was added, and the
reaction mixture was stirred for 2 h at room temperature. The
solvent was evaporated in vacuo, and then the crude product was
purified by silica gel column chromatography to afford the
corresponding ketone (4e-G1 or 9-G1) in a quantitative yield.
Representative Procedure for Glycosylation of Aglycones 2a-
2f, 6, 12, and 15 with CCl3C(NH)O-G (G1, G3, G5, G7, or G9).
A solution of the glycosyl trichloroacetimidate (CCl3C(NH)O-G1,
0.64 mmol) and steroid aglycone 2c (0.5 mmol) in dry dichlo-
romethane (15 mL) was stirred with 4 Å molecular sieves (1.5 g)
at room temperature for 15 min, and then the reaction mixture was
cooled to -68 °C (ethanol-dry ice bath) and a 0.14 M solution of
TMSOTf in CH2Cl2 (1.3 mL) was slowly added. The reaction
mixture was stirred for an additional 30 min, quenched with
triethylamine, and the molecular sieves were filtered out. The filtrate
was evaporated in vacuo, and the products, protected saponin 4c
and its cyclic isomer 3c, were separated by silica gel column
chromatography.
The above method was also used for the preparation of the
glycosides 4a-4e. The compounds 4a-4e were subjected to
deprotection without full characterization. The crude products of
glycosylation of diol 6 (compounds 7 and 8) were separated and
oxidized to the corresponding ketones (vide infra).

3672 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 15
Wojtkielewicz et al.
The oxidation product of compound 7 (eluted with hexane-ethyl
acetate (92:8)) was shown to be identical in all respects to the
compound 4e-G1 obtained by direct glycosylation of the aglycone
2e, as described in ref 17.
included in the standard program distribution. Obtained structures
were visualized using the VMD program.47
Details of the Simulation Procedure. A single molecular
dynamics trajectory at constant temperature was generated. Simula-
tion started from an arbitrarily chosen conformer of the molecule
of interest. The Berendsen bath coupling48 method, with a time
constant of 0.1 ps, was used to maintain constant temperature
(1000 K) during the simulation.
The molecular dynamics trajectory was updated using a modified
Beeman method49 to integrate the Newtonian equations of motion.
The time interval for the dynamics steps was set to 1 fs, and the
total length of the simulated trajectory reached 1 ns. From the
generated trajectory, 500 structures, equally spaced in time, were
collected, with the interval between snapshots set to 2 ps. Each
structure from the generated ensemble was then subjected to
molecular dynamics simulated annealing computation. During a 2
ps simulation, the temperature changed between the starting value
of 1000 K to the final value of 100 K. The linear scaling protocol
was used.
(22S)-6â-Methoxy-17R,22-dihydroxy-3R,5-cyclo-5R-cholestan-
16-one 22-O-{O-[2-O-(4-methoxybenzoyl)-3,4-di-O-triethylsilyl-
â-D-xylopyranosyl]-(1f3)-2-O-acetyl-4-O-triethylsilyl-R-L-ara-
binopyranoside} (9-G1). Product 9-G1 was eluted with hexane-
ethyl acetate (87.5:12.5). IR (CHCl3) 3480, 1753, 1735, 1607, 1511,
1169, 1255. 1094 cm-1. 1H NMR (200 MHz, CDCl3) δ 7.97 (2 H,
d, J ) 8.7 Hz), 6.91 (2 H, d, J ) 8.7 Hz), 5.05 (1 H, dd, J ) 8.9,
7.0 Hz), 4.94 (2 H, m), 4.64 (1 H, d, J ) 6.5 Hz), 4.44 (1 H, d, J
) 6.9 Hz), 4.00 (2 H, m), 3.91 (1 H, s), 3.86 (3 H, s), 3.77 (1 H,
m), 3.61-3.71 (3 H, m), 3.40 (1 H, d, J ) 11.8 Hz), 3.33 (3 H, s),
3.24 (1 H, m), 2.78 (1 H, m), 2.30 (1 H, dd, J ) 17.2, 6.9 Hz),
1.87 (3 H, s). 13C NMR (50 MHz, CDCl3) δ 218.9 (C), 168.9 (C),
164.5 (C), 163.1 (C), 131.9 (2 × CH), 122.9 (C), 113.3 (2 × CH),
101.8 (CH), 101.3 (CH), 83.1 (C), 82.8 (CH), 82.1 (CH). ESI-MS
m/z (%) 1252.5 (MNa+).
Wolff-Kishner Reduction of Lactol 11.13 To a stirred solution
of lactol 11 (0.276 g, 0.73 mmol) in ethanol (15 mL), an 89%
solution of hydrazine hydrate (0.05 mL, 1.1 equiv) and triethylamine
(0.05 mL) was added. The reaction mixture was stirred under reflux
for 16 h. Evaporation of the solvent from the reaction mixture
afforded crude hydrazone. The product, without further purification,
was dissolved in DMSO, and a solution of potassium t-butoxide in
DMSO was added dropwise. The reaction mixture was stirred at
room temperature under argon for 4 h, then it was poured into water,
and the product was extracted with ether. Evaporation of the solvent
from dried (anhydrous MgSO4) extract afforded crude product,
which was purified by silica gel column chromatography. Elution
with hexane-ethyl acetate (85:15) yielded compound 12 (0.133 g;
50%).
Finally, all structures were minimized in Cartesian coordinate
space using limited memory LM-BFGS50,51 nonlinear optimization,
with the value of the rms termination criterion set to 0.01 kcal/
mol/Å. The value of the dielectric constant in all calculations was
1.5; the default value for the MM3 force field. Because all
experimental measurements were conducted in media of low
polarizability, this choice of dielectric constant seems to be justified.
The final set of 500 minimized structures was then further analyzed.
3.4. Biological Tests. The ARN 8 cancer cell line was kindly
provided by Dr. J. P. Blaydes (University of Dundee, Scotland).
Other cancer cell lines were obtained from the American Type
Culture Collection (Manassas, VA). The screening cell lines (T-
lymphoblastic leukemia cell line CEM; breast carcinoma cell line
MCF7, lung carcinoma cell line A 549, chronic myelogenous
leukemia cell line K 562, epitheloid carcinoma cell line HeLa,
melanoma cell line ARN 8, malignant melanoma cell line G 361,
osteosarcoma cell line HOS, and mouse fibroblast NIH 3T3) were
cultured in DMEM medium (Gibco BRL) supplemented with 10%
fetal calf serum, 4 mM glutamine, 100 U/mL penicillin, and
100 µg/mL streptomycin, at 37 °C in a fully humidified atmosphere
containing 5% CO2. Cell suspensions with approximately 1.25 ×
105 cells/mL were distributed in 96-well microtiter plates, and after
3 h of stabilization, the tested OSW-1 analogues were added in
20 µL aliquots in dimethylsulfoxide (DMSO) in serial (usually 6-
and 4-fold) dilutions to the microtiter plate wells. Control cultures
were treated with DMSO alone, and the final concentration of
DMSO in the reaction mixture never exceeded 0.6%. In routine
testing, the highest well concentration was 50 µM, but for some
analogues, this varied. After 72 h of culture, the cells were incubated
with Calcein AM solution (Molecular Probes) for 1 h. The
fluorescence of viable cells was quantified using a Fluoroscan
Ascent fluorometer (Microsystems). The tumor cell survival (TCS)
rate was calculated using the following equation: TCS )
(ODdrug-exposed well/mean ODcontrol wells) × 100%. The TCS50 value,
the drug concentration lethal to 50% of the tumor cells, was
calculated from the obtained dose-response curves.
6â-Methoxy-3R,5-cyclo-20a-homo-5R-pregnan-16â,17R-diol
1
(12). IR (CHCl3) 3510, 3468, 1725, 1092 cm-1. H NMR (200
ΜΗz, CDCl3) δ 3.87 (1 H, dd, J ) 4.7, 7.9 Hz), 3.31 (3 H, s),
2.77 (1 H, m), 2.30 (2 H, m), 1.02 (s, 3 H), 0.98 (3 H, d, J ) 6.7
Hz), 0.96 (3 H, s), 0.93 (3 H, d, J ) 6.9 Hz), 0.64 (1 H, m), 0.42
(1 H, dd, J ) 8.0, 5.1 Hz). 13C NMR (50 MHz, CDCl3) δ 86.3 (C),
82.2 (CH), 80.8 (CH), 56.4 (CH3), 48.3 (CH), 47.5 (CH), 46.9 (C),
43.2 (C), 35.9 (CH2), 35.1 (C), 34.9 (CH2), 33.2 (CH2), 33.1 (CH2),
30.3 (CH), 28.9 (CH), 24.8 (CH2), 22.0 (CH2), 21.4 (CH), 19.2
(CH3), 18.7 (CH3), 17.0 (CH3), 13.3 (CH3), 13.00 (CH2). ESI-MS
m/z (%) 747.6 (2MNa+), 385.3 (MNa+).
3.3. NMR and Conformational Analysis of Compound 10.
For NMR measurements, compound 10 was dissolved in a mixture
of pyridine-d5 (D 99.5%, Aldrich, St. Louis, MO) and CDCl3 (D
99.5%, Cambridge Isotope Laboratories, Inc., Andover, MA), 1:1
(v/v), or in CDCl3 to a concentration of 5 mM. A total of 32 K
data points were collected, and a spectral width of 6 kHz was used
in 1D proton experiments. The 2D experiments were measured
using a proton spectral width of 4.5 kHz collecting 2 K data points.
TOCSY,36 DQF-COSY,37,38 and ROESY39,40 spectra were measured
with 256 increments. A mixing time of 80 ms was used in TOCSY,
while 250 ms was used in ROESY measurements. The 2D {1H,
13C} GHSQC41,42 and HMBC43 experiments with gradients were
performed in proton decoupled mode with a carbon spectral width
of 25 kHz and 256 increments. The spectra were calibrated against
the residual chemical shift of chloroform in proton spectra (7.26
ppm) and chemical shift of chloroform in 13C spectra (77.0 ppm).
Distances between protons were calculated from the volumes of
the corresponding cross-peaks from ROESY spectra, and cross-
peaks obtained for methylene protons were used for the calibration.
Conformational Analysis. Molecular structures of the 22S and
22R isomer of 10 were built using the TINKER44,45 package (version
4.2).
Acknowledgment. Financial support from the Polish and
Czech Ministry of Science and Higher Education (Grant Nos.
PBZ-KBN-126/T09/2004 and MSM6198959216) is gratefully
acknowledged. We would also like to thank Sees-editing Ltd.
(U.K.) for the excellent editing of this manuscript.
Supporting Information Available: Synthetic procedures,
spectral data for all new compounds (IR, 1H NMR, 13C NMR, MS),
a set of spectra of compound 10 (1H NMR, 13C NMR, DEPT,
DQCOSY, GHSQC, TOCSY, ROESY, IR, MS), and elemental
analyses of all final compounds. This material is available free of
charge via the Internet at http://pubs.acs.org.
Structure manipulations and analysis (computation of energy,
rmsd calculations) were carried out with the set of tools provided
with the TINKER package. Molecular dynamics simulations and
the minimizations of structures over Cartesian coordinates employed
the MM346 force field adopted for the TINKER suite, which was
References
(1) Perry, M. C. The chemotherapy sourcebook; Williams & Wilkins:
Baltimore, MD, 1992.

Analogues of the Potent Cytotoxic Saponin OSW-1
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 15 3673
(2) Mukherjee, A. K.; Basu, S.; Sarkar, N.; Ghosh, A. C. Advances in
cancer therapy with plant based natural products. Curr. Med. Chem.
2001, 8, 1467-1486.
(3) Petit, G. R.; Inoue, M.; Kamano, Y.; Herald, D. L.; Arm, C.; Dufresne,
C.; Christie, N. D.; Schmidt, J. M.; Doubek, D. L.; Krupa, T. S.
Isolation and structure of the powerful cell growth inhibitor cepha-
lostatin 1. J. Am. Chem. Soc. 1988, 110, 2006-2007.
(4) Fukuzawa, S.; Matsunaga, S.; Fusetani, N. Ritterazine A, a highly
cytotoxic dimeric steroidal alkaloid, from the tunicate Ritterella
tokioka. J. Org. Chem. 1994, 59, 6164-6166.
(5) Fan, W. Possible mechanisms of poclitaxel-induced apoptosis.
Biochem. Pharmacol. 1999, 57, 1215-1221.
(6) Reddy, L.; Odhav, B.; Bhoola, K. D. Natural products for cancer
prevention: A global perspective. Pharmacol. Ther. 2003, 99, 1-13.
(7) Kubo, S.; Mimaki, Y.; Terao, M.; Sashida, Y.; Nikaido, T.; Ohmoto,
T. Acylated cholestane glycosides from the bulbs of Ornithogalum
saundersiae. Phytochemistry 1992, 31, 3969-3973.
(8) Mimaki, Y.; Kuroda, M.; Kameyama, A.; Sashida, Y.; Hirano, T.;
Oka, K.; Maekawa, R.; Wada, T.; Suita, K.; Beutler, J. A. Cholestane
glycosides with potent cytostatic activities on various tumor cells
from Ornithogalum saundersiae bulbs. Bioorg. Med. Chem. Lett.
1997, 7, 633-636.
(9) Mimaki, Y. Structures and biological activities of plant glycosides:
cholestane glycosides from Ornithogalum saundersiae, O. thyrsoides
and Galtonia candicans, and their cytotoxic and antitumor activities.
Nat. Prod. Commun. 2006, 1, 247-253.
(10) Gryszkiewicz-Wojtkielewicz, A.; Jastrze¸bska, I.; Morzycki, J. W.;
Romanowska, D. B. Approaches towards the synthesis of cephalost-
atins, ritterazines and saponins from Ornithogalum saundersiaes
New natural products with cytostatic activity. Curr. Org. Chem. 2003,
7, 1257-1277.
(11) Zhou, Y.; Garcia-Prieto, C.; Carney, D. A.; Xu, R.; Pelicano, H.;
Kang, Y.; Yu, W.; Lou, C.; Kondo, S.; Liu, J.; Harris, D. M.; Estrov,
Z.; Keating, M. J.; Jin, Z.; Huang, P. OSW-1: A natural compound
with potent anticancer activity and a novel mechanism of action. J.
Natl. Cancer Inst. 2005, 97, 1781-1785.
(12) Guo, C.; Fuchs, P. L. The first synthesis of the aglycone of the potent
anti-tumor steroidal saponin OSW-1. Tetrahedron Lett. 1998, 39,
1099-1102.
(13) Morzycki, J. W.; Gryszkiewicz, A. Synthesis of the potent antitumor
saponin OSW-1 aglycone. Pol. J. Chem. 2001, 75, 983-989.
(14) Yu, W.; Jin, Z. A new strategy for the stereoselective introduction
of steroid side chain via R-alkoxy vinyl cuprates: Total synthesis of
highly potent antitumor natural product OSW-1. J. Am. Chem. Soc.
2001, 123, 3369-3370.
(15) Xu, Q.; Peng, X.; Tian, W. A new strategy for synthesizing the
steroids with side chains from steroidal sapogenins: Synthesis of
the aglycone of OSW-1 by using the intact skeleton of diosgenin.
Tetrahedron Lett. 2003, 44, 9375-9377.
(16) Deng, S.; Yu, B.; Lou, Y.; Hui, Y. First total synthesis of an
exceptionally potent antitumor saponin, OSW-1. J. Org. Chem. 1999,
64, 202-208.
(17) Morzycki, J. W.; Wojtkielewicz, A. Synthesis of a cholestane
glycoside OSW-1 with potent cytostatic activity. Carbohydr. Res.
2002, 337, 1269-1274.
(18) Yu, W.; Jin, Z. The total synthesis of the natural product OSW-1. J.
Am. Chem. Soc. 2002, 124, 6576-6583.
(19) Shi, B.; Wu, H.; Yu, B.; Wu, J. 23-Oxa-analogues of OSW-1:
Efficient synthesis and extremely potent antitumor activity. Angew.
Chem., Int. Ed. 2004, 43, 4324-4327.
(20) Ma, X.; Yu, B.; Hui, Y.; Miao, Z.; Ding, J. Synthesis of OSW-1
analogues and a dimer and their antitumor activities. Bioorg. Med.
Chem. Lett. 2001, 11, 2153-2156.
(21) Deng, L.; Wu, H.; Yu, B.; Jiang, M.; Wu, J. Synthesis of OSW-1
analogs with modified side chains and their antitumor activities.
Bioorg. Med. Chem. Lett. 2004, 14, 2781-2785.
(22) Shi, B.; Wu, H.; Yu, B.; Wu, J. 23-Oxa-analogues of OSW-1:
Efficient synthesis and extremely potent antitumor activity. Angew.
Chem., Int. Ed. 2004, 43, 4324-4327.
(23) Shi, B.; Tang, P.; Hu, X.; Liu, X. O.; Yu, B. OSW Saponins: Facile
synthesis toward a new type of structures with potent antitumor
activities. J. Org. Chem. 2005, 70, 10354-10367.
(24) Qin, J.; Tian, W.; Lin, W. A highly efficient synthesis of 22-deoxy-
OSW-1 by utilizing the intact skeleton of diosgenin. Tetrahedron
Lett. 2006, 47, 3217-3219.
(25) Morzycki, J. W.; Wojtkielewicz, A. Synthesis of a highly potent
antitumor saponin OSW-1 and its analogues. Phytochem. ReV. 2005,
4, 259-277.
(26) Tang, P.; Mamdani, F.; Hu, X.; Liu, J. O.; Yu, B. Synthesis of OSW
saponin analogs with modified sugar residues and their antiprolif-
erative activities. Bioorg. Med. Chem. Lett. 2007, 17, 1003-1007.
(27) Morzycki, J. W.; Wojtkielewicz, A.; Wołczyn´ski, S. Synthesis of
analogues of a potent antitumor saponin OSW-1. Bioorg. Med. Chem.
Lett. 2004, 14, 3323-3326.
(28) Kruszewska, A.; Wilczewska, A. Z.; Wojtkielewicz, A.; Morzycki,
J. W. Synthesis of 23-oxa-22-deoxo analoues of OSW-1 aglycone.
Pol. J. Chem. 2006, 80, 611-615.
(29) Morzycki, J. W.; Łotowski, Z.; Siergiejczyk, L.; Lipkowski, J.;
Tabaszewska, A.; Wo´jcik, J. Structure of 3â-hydroxy-16-oxo-24-
nor-17-azachol-5-eno-23-nitrile and its 20S epimer. Steroids 1995,
60, 195-203.
(30) Soliman, H. S. M.; Simon, A.; To´th, G.; Duddeck, H. Identification
and structure determination of four triterpene saponins from some
middle-east plants. Magn. Reson. Chem. 2001, 39, 567-576.
(31) Akihisha, T.; Kimura, Y.; Tai, T. Arai, K. Astertarone B, a hydroxy-
triterpenoid ketone from the roots of Aster tataricus L. Chem. Pharm.
Bull. 1999, 47, 1161-1163.
(32) Kapou, A.; Fousteris, M. A.; Nikolaropoulos, S. S.; Zervou, M.;
Grdadolnik, S. S.; Zoumpoulakis, P.; Kyrikou, I.; Mavromoustakos,
T. 2D NMR and conformational analysis of a prototype antitumor
steroidal ester. J. Pharm. Biomed. Anal. 2005, 38, 428-434.
(33) Neuhaus, D.; Williamson, M. P. The steady-state NOE in rigid
multispin systems. The nuclear OVerhauser effect in structural and
conformational analysis, 2nd ed.; John Wiley & Sons: New York,
2000.
(34) Morzycki, J. W.; Gryszkiewicz, A.; Jastrze¸bska, I. Some reactions
of 16R,17R-oxido-steroids: A study related to the synthesis of the
potent antitumor saponin OSW-1 aglycone. Tetrahedron Lett. 2000,
41, 3751-3754.
(35) Morzycki, J. W.; Gryszkiewicz, A.; Jastrze¸bska, I. Neighboring group
participation in epoxide ring cleavage in reactions of some 16R,-
17R-oxidosteroids with lithium hydroperoxide. Tetrahedron 2001,
57, 2185-2193.
(36) Bax, A.; Davis, D. G. MLEV-17 based two-dimensional homonuclear
magnetization transfer spectroscopy. J. Magn. Reson. 1985, 65, 355-
360.
(37) Piantini, U.; Sørensen, O. W.; Ernst, R. R. Multiple quantum filters
for elucidating NMR coupling networks. J. Am. Chem. Soc. 1982,
104, 6800-6801.
(38) Rance, M.; Sørensen, O. W.; Bodenhausen, G.; Wagner, G.; Ernst,
R. R.; Wu¨thrich, K. Improved spectral resolution in COSY H-1-
NMR spectra of proteins via double quantum filtering. Biochem.
Biophys. Res. Commun. 1983, 117, 479-485.
(39) Bothner-By, A. A.; Stephens, R. L.; Lee, J.-M., Warren, C. D.;
Jeanloz, R. W. Structure determination of a tetrasaccharide: Transient
nuclear Overhauser effects in the rotating frame. J. Am. Chem. Soc.
1984, 106, 811-813.
(40) Bax, A.; Davis, D. G. Practical aspects of two-dimensional transverse
NOE spectroscopy. J. Magn. Reson. 1985, 63, 207-213.
(41) Palmer, A. G., III; Cavanagh, J.; Wright, P. E. Sensitivity improve-
ment in proton-detected twodimensional heteronuclear correlation
NMR spectroscopy. J. Magn. Reson. 1991, 93, 151-170.
(42) Kay, L. E.; Keifer, P.; Saarinen, T. Pure absorption gradient enhanced
heteronuclear single quantum correlation spectroscopy with improved
sensitivity. J. Am. Chem. Soc. 1992, 114, 10663-10665.
1
(43) Bax, A.; Summers, M. F. H and 13C assignments from sensitivity-
enhanced detection of heteronuclear multiple-bond connectivity by
2D multiple quantum NMR. J. Am. Chem. Soc. 1986, 108, 2093-
2094.
(44) Ren, P.; Ponder, J. W. Polarizable atomic multipole water model for
molecular mechanics simulation. J. Phys. Chem. B 2003, 107, 5933-
5947.
(45) Ren, P.; Ponder, J. W. Consistent treatment of inter- and intramo-
lecular polarization in molecular mechanics calculations. J. Comput.
Chem. 2002, 23, 1497-1506.
(46) Allinger, N. L.; Yuh, Y. H.; Lii, J.-H. Molecular mechanics. The
MM3 force field for hydrocarbons. J. Am. Chem. Soc. 1989, 111,
8551-8566.
(47) Humphrey, W.; Dalke, A.; Schulten, K. VMDsVisual molecular
dynamics. J. Mol. Graphics 1996, 14, 33-38.
(48) Berendsen, H. J. C.; Postma, J. P. M.; van Gunsteren, W. F.; DiNola,
A.; Haak, J. R. Molecular dynamics with coupling to an external
bath. J. Chem. Phys. 1984, 81, 3684-3690.
(49) Beeman, D. Some multistep methods for use in molecular dynamics
calculations. J. Comput. Phys. 1976, 20, 130-139.
(50) Nocedal, J.; Wright, S. J. Numerical Optimization; Springer-Verlag:
New York, 1999.
(51) Nocedal, J. Updating quasi-Newton matrices with limited storage.
Math. Comp. 1980, 35, 773-782.
JM0613572