An improved synthesis of methyl protodioscin: Tautomerization and direct access to the 3-o-substituted kryptogenin

Article abstract of DOI:10.1246/cl.2008.780

The acid-catalyzed tautomerization of 3-O-substituted kryptogenin 2 was studied, and the effects of acids such as silica gel, TMSOTf, acetic acid, and CDCl₃, are discussed. Additionally, a Zn/KI/HOAc reduction based on this tautomerization was adopted, which afforded 2 directly, and provided a facile procedure for the synthesis of methyl protodioscin and other furostanol saponins in a mild way. Copyright

Full text of DOI:10.1246/cl.2008.780

780
Chemistry Letters Vol.37, No.7 (2008)
An Improved Synthesis of Methyl Protodioscin:
Tautomerization and Direct Access to the 3-O-Substituted Kryptogenin
Qing-Chun Xu, Yang Liu, Jiao Liu, Chun-Xian He, Mao-Cai Yan, and Mao-Sheng Chengꢀ
Key Laboratory of New Drugs Design and Discovery of Liaoning Province, School of Pharmaceutical Engineering,
Shenyang Pharmaceutical University, Shenyang 110016, P. R. China
(Received March 31, 2008; CL-080327; E-mail: mscheng@syphu.edu.cn)
O
2, R2 = H 2a, R1 = TPS, R2 = H
The acid-catalyzed tautomerization of 3-O-substituted
2b, R1 = TBS, R2 = H
2c, R1 = R, R2 = H
kryptogenin 2 was studied, and the effects of acids such as silica
gel, TMSOTf, acetic acid, and CDCl3, are discussed. Addition-
ally, a Zn/KI/HOAc reduction based on this tautomerization
was adopted, which afforded 2 directly, and provided a facile
procedure for the synthesis of methyl protodioscin and other
furostanol saponins in a mild way.
OR2
OBz
O
O
BzO
7, R1 = R, R2 =
BzO
OBz
R1O
R1O
R1O
O
3, R3 = OH 3a, R1 = TPS, R3 = OH
5, R1 = R, R3 = H
O
R3
Steroidal saponins are usually divided into three classes:
spirostan, cholestan, and furostan. Due to their structural com-
plexities and intrinsic instabilities, studies on furostanol sapo-
nins are still limited.1 Methyl protodioscin (1, Figure 1), a repre-
sentative furostanol saponin which may have a novel mechanism
of anti-cancer action according to the NCI’s comparative analy-
sis, has been under study in recent years.2,3 The total synthesis
of 1 has been achieved,4,5 and has aided further studies of this
molecule. Two basic strategies have been applied to install the
ꢀ-chacotriosyl at the 3-OH. In glycosylation via a chacotriosyl
thioglycoside,4 though the right ꢀ-configuration was affirmed,6
to our knowledge, there has been no theory that could illustrate
the stereoselectivity of the glycosylation of this sugar donor, be-
cause of its lack of a neighboring participatory group (a similar
reaction used a 2,3-branched trisaccharide donor which afforded
an ꢁ configuration7). To overcome this problem, a short ap-
proach via a stepwise glycosylation strategy has been adopted
by Li and Yu recently.5 In this strategy, a 3-O-substituted kryp-
togenin (2, or its isomer 3, Figure 2) was required for the instal-
lation of the glucose moiety at 26-OH. In order to extend the ap-
plication of Yu’s strategy in the construction of furostanol sap-
onins, further study of this intermediate has been conducted in
the present work.
4a, R1 = TPS 4b, R1 = TBS
O
OPiv
O
4c, R1 = R
O
O
PivO
O
O
OH
BzO
R =
O
BzO
BzO
4
BzO
BzO
BzO
O
Figure 2. Molecular structures of the derivatives of 3-O-substi-
tuted kryptogenin and its isomer, and the intermediates in the
synthesis of methyl protodioscin.
most reported cases, and we observed that either 2 or 3 would
readily convert to a complex mixture containing the two isomers
at least. Secondly, glycosylation employing Lewis acid was also
problematic,5,11 which afforded 26-O-glycoside by either 2 or a
mixture with 3. This was evidence of rapid tautomerization,
which proceeded more quickly than an ordinary glycosylation.
As such, tautomerization would be responsible for similar yield
in the utilization of either pure 2 or the mixture according to our
experiments.
For potent synthetic use, acetic acid is also worthy of
discussion among the acids. Instead of employing a combination
of HOAc and Ac2O, HOAc was enough to open the hemiketal
via the aforementioned tautomerization, which afforded the
ring-opened form as the major product. The acetyl anhydride,
in fact, acted merely as an acylating agent of 26-OH, and provid-
ed no other structural change, including the possible acylation at
16-OH.12 Based on the above phenomena, the Zn/KI/HOAc/
Ac2O reduction carried out in the previous study4 was replaced
with Zn/KI/HOAc to reduce the epoxide 4a (Figure 2). As
anticipated, 2a13 was the major product (eq 1), and was obtained
in a 42% isolated yield due to interconversion during silica
gel separation as discussed above. Through this method, the
necessary expensive metal salts (i.e., Ag2CO3 and AgNO3)
used to transform the 26-iodide in the formal procedure,5 were
no longer needed, providing a shorter sequence of steps to
accomplish the strategy without any operation of protective
groups at 26-OH.
Compar able to the hydride shift in a steroidal 1,5-ketol,8
a tautomerization9–11 generally occurrs in compound 2 which
bears a 16-keto group in addition to the identical side chain of
the 1,5-ketol. It was found that, besides silica gel,9,10 other acids
could also catalyze the equilibrium, which directly impacts the
subsequent chemistry of the molecule. Above all, intercon-
version during chromatographic separation was common in
OMe
26
O
O
OH
1
O
OH
O
OH
HO
HO
O
O
HO
3
O
O
HO
HO
O
HO
1
HO
HO
Zn/KI/HOAc
4a ꢁꢁꢁꢁꢁꢁꢁꢁ! 2a
HO
ð1Þ
r.t.
Figure 1. Molecular structure of methyl protodioscin.
Copyright Ó 2008 The Chemical Society of Japan

Chemistry Letters Vol.37, No.7 (2008)
781
caused an unexpected deprotection of the TBS group, while
the former did not, affording 2b directly from 4b13 (Figure 2).
Considering the widespread application of silyl ethers and other
acid-sensitive protective groups in synthesis, the application
of the latter reduction might be problematic in some cases.
Even the TPS-protected equivalent of 4a, can not endure such
conditions shown by our results. Therefore, the Zn/KI/HOAc
reduction which was derived from Conforth’s method16 will
become another valuable method in the synthesis of furostanol
saponins despite a lower yield. The final product prepared by
this method was identical to the product in Ref 4 according to
Oxone,r.t.,100%
Zn/KI/HOAc
r.t.
5
4c
2c
acetone,CH2Cl2,H2O
TMSOTf,CH2Cl2,0°C,
30% in 2 steps
7
1
OBz
O
O
CCl3
BzO
BzO
OBz
NH
6
Scheme 1. Formal synthesis of 1.
Again, 1 was chosen as the target to validate the synthetic
feasibility of the above-described protocol (Scheme 1). The
ester of dioscin 5,14 as the starting material, was oxidized by
in-situ generated DMDO to afford the epoxide 4c as a 1:1 mix-
ture of diastereomers in quantitative yield. Then the key step,
Zn/KI/HOAc reduction, was accomplished in isolated yield of
30%. However, a higher ratio of Zn and KI was required, due
to the steric hindrance of the sugar chain at 3-OH. In order to
complete the reaction, both extended reaction times and addi-
tional equivalents Zn/KI were permissible. Higher temperatures
(above 50 ꢂC) were avoided, because of the prominent acylation
at 26-OH caused by HOAc. The unpurified 2c was glycosylated
directly with 2,3,4,6-tetra-O-benzoyl-D-glucopyranosyl trichlo-
roacetimidate (6) to afford the desired ester in 30% yield. This
ester corresponded directly to that of cholestanol saponin 7 in the
previous study.4 This sequence realized the brief conversion
from spirostan ester to the required cholestan ester in overall
three steps and with acceptable yields.13
1
a comparison of H NMR spectra.
In conclusion, a detailed chemical property study of 2 was
developed. Tautomerization under a variety of acidic conditions
was described, which provided the opportunity for a direct con-
version from the epoxide to 2 through a Zn/KI/HOAc reduction
step. This methodology was applied to the synthesis of 1 with an
acceptable yield (30% from 5 to 7). Further, due to the mildness
of Zn/KI/HOAc reduction, the strategy allows for the potential
applications of Lewis acid-sensitive protecting groups (such as
the silyl ethers) in the construction of the sugar chain at the
3-OH position of the steroidal aglycone, which provides a
possibility to the synthesis of diverse furostanol saponins.
The authors acknowledge National Natural Science Founda-
tion of China (Nos. 20472054 and 30772641) for financial
support of this research.
References and Notes
1
The tautomerization could be activated immediately by
acids, such as TMSOTf (Table 1, Entries 1 and 2) or BF3,
Lewis acids which were generally used in glycosylation, even
at ꢁ78 ꢂC; and acetic acid (Table 1, Entries 3 and 4). Certain
spectroscopic characteristics of the tautomerization were ob-
served. When CDCl3 was used as the solvent for NMR detection
of each isomer, a result of tautomerization was received in both
cases based on the concurrence of respective characteristic 13C
signal (ꢂ 218.0 and 214.5 for the C16 and C22 of 2 respectively,
whereas ꢂ 116.0 and 110.9 for the counterparts of 3). To avoid
this phenomenon, it was necessary to choose pyridine-d5 as
the solvent for its inhibition of the tautomerization.15
K. Hostettmann, A. Marston, Saponins, Cambridge University
Press, Cambridge, 1995.
2
3
K. Hu, X. Yao, Cancer Invest. 2003, 21, 389.
G. Wang, H. Chen, M. Huang, N. Wang, J. Zhang, Y. Zhang, G.
Bai, W. Fong, M. Yang, X. Yao, Cancer Lett. 2006, 241, 102.
M. S. Cheng, Q. L. Wang, Q. Tian, H. Y. Song, Y. X. Liu, Q. Li,
X. Xu, H. D. Miao, X. S. Yao, Z. Yang, J. Org. Chem. 2003, 68,
3658.
4
5
6
M. Li, B. Yu, Org. Lett. 2006, 8, 2679.
T. Ikeda, H. Miyashita, K. Tetsuya, T. Kajimoto, T. Nohara,
Tetrahedron Lett. 2001, 42, 2353. The anomeric 1H NMR shifts
of the chacotriosyl at 3-OH of our product was identical to that
of dioscin, but not the ꢁ isomer.
It is important to compare the Zn/KI/HOAc epoxide
reduction with the TMSCl/NaI/CH3CN epoxide reduction.
The latter afforded products in 81–90% yields, but readily
7
8
9
S. Hou, P. Xu, L. Zhou, D. Yu, P. Lei, Bioorg. Med. Chem. Lett.
2006, 16, 2454.
A. F. Kluge, M. L. Maddox, L. G. Partridge, J. Org. Chem.
1985, 50, 2359.
Table 1. Tautomerization by treatment with acida
´
´
´
R. F. Barreira, A. G. Gonzalez, J. A. S. Rocıo, E. S. Lopez,
acid
2a
3a
Phytochemistry 1970, 9, 1641.
10 K. Nakano, Y. Kashiwada, T. Nohara, T. Tomimatsu, H.
Tsukatani, T. Kawasaki, Yakugaku Zasshi. 1982, 102, 1031.
11 B. Yu, J. Liao, J. Zhang, Y. Hui, Tetrahedron Lett. 2001, 42, 77.
12 i.e., a mixture of 3a and acetyl anhydride was stirred at room
temperature for several hours, but no product was afforded by
TLC detection.
13 See Supporting Information for experimental details. The
material is available electronically on the CSJ-Journal web site,
http://www.csj.jp/journals/chem-lett/index.html.
14 Z. Yang, E. L.-M. Wong, T. Y.-T. Shum, C.-M. Che, Y. Hui,
Org. Lett. 2005, 7, 669.
Entry
Compds
Conditions
2a/3a Ratiob
1
2
3
4
5
2a
3a
2a
3a
2a
TMSOTfc
TMSOTfc
HOAc
HOAc
Dowex (Hþ)d
3.42:1
2.45:1
10.08:1
8.76:1
1.82:1
aThe tautomerization was also catalyzed by certain bases,
such as the deprotection of acetyl at 26-OH by K2CO3 (2a/
3a, 2.02:1)11 and the removal of HOAc by NaHCO3 (2a/
b
3a, 2.53:1) in the Zn/KI/HOAc reduction. The ratio were
determined by HPLC (ODS, MeOH, and UV detector). cOne
drop of 5% (v/v) TMSOTf (in CH2Cl2) was added to a 1-mL
CH2Cl2 solution. In a CH2Cl2 solution with pH 2.
15 See the Supporting Information of reference 5.
16 J. W. Cornforth, R. H. Cornforth, K. K. Mathew, J. Chem. Soc.
1959, 2539.
d
Published on the web (Advance View) June 21, 2008; doi:10.1246/cl.2008.780