Total synthesis and structure assignment of the anthrone C-glycoside cassialoin

Article abstract of DOI:10.1002/anie.200704625

Configuration confirmed: The first total synthesis and structure assignment of the anthrone C-glycoside cassialoin (1) have been achieved by exploiting an isoxazole-containing stereogenic α-ketol and a subsequent intramolecular redox reaction (see retrosynthetic analysis; MOM = methoxymethyl). (Chemical Equation Presented).

Full text of DOI:10.1002/anie.200704625

Communications
DOI: 10.1002/anie.200704625
Natural Product Synthesis
Total Synthesis and Structure Assignment of the Anthrone C-Glycoside
Cassialoin**
Yasuhito Koyama, Ryo Yamaguchi, and Keisuke Suzuki*
Anthrone C-glycosides constitute an intriguing class of
natural products, in which sugars are connected to an
Scheme 1 shows the retrosynthetic analysis based on the
dissection of 1 into anthraquinone I and carbohydrate II. We
ꢀ
anthrone skeleton at its C10 position through a C C
bond. Cassialoin (1), a representative member of this
class, was isolated from plant extracts traditionally used
in Ayurvedic and Asian folk medicine, namely a
heartwood of Cassia garrettiana craib[1] or the roots
of Rheum emodi wall.[2] Such unique sugar–anthrone
hybridized structures present challenges for stereocon-
trolled synthesis, as well as for structure elucidation.
Although the sugar stereochemistry of 1 was assigned as
1
the b-C-glucoside based on H NMR spectroscopy, the
C10 configuration is not easy to assign.[2,3] While the
structure of the related aloin A (2) was established by
single-crystal X-ray analysis, the structure of 1 remained
speculative.
Scheme 1. Retrosynthetic analysis of 1.
planned to introduce the sugar as a nucleophile via glycal
anion III,[4] which posed two potential difficulties: 1) the
discrimination of two carbonyl groups at the C9 and C10
positions and 2) the facial selectivity at the C10 position. The
latter issue seemed nontrivial because acceptor I has a
pseudomirror symmetry, only differing in the C3 and C6
positions. Even if a chiral nucleophile were employed,
rigorous control of the facial selectivity seemed unlikely.
To address this stereochemical challenge, we sought to
exploit our recent approach to densely functionalized poly-
cyclic structures.[5,6] More precisely, we planned to use chiral,
nonracemic ketol IV as a selectively protected, stereogenic
surrogate of anthraquinone I (Scheme 2), with the hope that a
nucleophile would attack ketol IV from the a face.
Herein we describe the stereocontrolled first total syn-
thesis of 1, thereby establishing the stereostructure of
cassialoin as that of 1. The analysis further revealed that the
natural product contains the C10 epimer 1’ as a minor
constituent.
This key intermediate IV would be prepared by the
stereoselective benzoin cyclization of ketoaldehyde V.
Although compound V could be obtained by cyclocondensa-
tion of nitrile oxide VI and diketone VII, the dilemma was
that the diketone VII has Cs symmetry, so the cycloconden-
sation with nitrile oxide VI would inevitably produce a
racemate.
[*] Dr. Y. Koyama,[+] R. Yamaguchi, Prof. Dr. K. Suzuki
Department of Chemistry
Tokyo Institute of Technology, SORST-JST Agency
2-12-1 O-okayama, Meguro-ku, Tokyo 152-8551 (Japan)
Fax: (+81)3-5734-2788
E-mail: ksuzuki@chem.titech.ac.jp
[+] Present address:
Department of Organic and Polymeric Materials
Tokyo Institute of Technology
2-12-1 O-okayama, Meguro-ku, Tokyo 152-8551 (Japan)
[**] We are grateful to Dr. Hidehiro Uekusa for X-ray analyses. This work
was partially supported by the Global COE program (Tokyo Institute
of Technology) and a Grant-in-Aid for Scientific Research (JSPS). A
Fellowship to Y.K. from the JSPS for Young Japanese Scientists is
gratefully acknowledged. We thank Dr. Liselotte Krenn (Vienna
University) for providing an authentic sample of cassialoin and the
1H NMR spectrum.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Scheme 2. Retrosynthetic analysis of IV.
1084
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 1084 –1087

Angewandte
Chemie
As a potential solution to this issue, we postulated that the
chiral, nonracemic diketoester VIII would serve as a useful
precursor to V, because we knew that the corresponding
racemic compound VIII (R = Et) can undergo regioselective
cyclocondensation with benzonitrile oxides to give isoxazole-
annulated product IX (Scheme 3).[5e] The remaining issue was
the availability of diketoester VIII in an enantiomerically
enriched form, for which we decided to rely on the readily
prepared, enantiomerically pure surrogate 4 reported by
Myers et al.[7]
Scheme 5 highlights the stereoselective union of the
lithiated species of glycal 9[11] and a-ketol 8 to give cis-diol
10 as a single diastereomer in 75% yield.[12] The X-ray
analysis verified the C10 configuration of 10 to be S
(Figure 1).[13]
Scheme 3. Regioselective formation of IX from diketone VIII. MS =
Molecular sieves.
The synthesis started with the cyclocondensation of stable
nitrile oxide 3[5] and 1,3-diketone 4, which proceeded
smoothly in iPrOH in the presence of powdered molecular
sieves (4 Š) to give isoxazole 5 as the major product in 72%
yield (Scheme 4).[8]
Scheme 5. Reagents and conditions: a) 9 (3.6 equiv), tBuLi (3.0 equiv),
THF, 08C, 2 h; then 8, ꢀ788C, 30 min, 75%; b) ClCH2I (2.5 equiv),
NaH (2.5 equiv), DMF, room temperature, 4 h, 84%; c) KOtBu
(10 equiv), THF, ꢀ788C!room temperature, 35 min, 76%. DMF =
N,N-dimethylformamide.
Scheme 4. Reagents and conditions: a) 4 (1.2 equiv), MS (4 Š;
1 gmmolꢀ1), iPrOH, 508C, 1 d, 72%; b) 4-DMAP (1.0 equiv), H2O,
mesitylene, 1508C, 3.5 h; c) 1m aq. H2SO4, THF, room temperature,
3.5 h, 66% (2 steps); d) 7 (0.1 equiv), DBU (0.1 equiv), tBuOH, 408C,
2 h, 99%. MOM = methoxymethyl; 4-DMAP = 4-dimethylaminopyr-
idine; THF = tetrahydrfuran; DBU = 1,8-diazabicyclo[5.4.0]undec-7-
ene.
Figure 1. ORTEP diagram of 10. Thermal ellipsoids at 50% probability.
N blue, O red, Si yellow.
By conversion of 10 into methylene acetal 11,[14] the stage
was set for the aromatization of the A ring. Through initial
unfruitful attempts based on organoselenium chemistry,[15] we
were pleased to learn that the conversion could be achieved in
a single step by exposure of isoxazole 11 to a strong base
(KOtBu), thereby giving anthrone 12 in 76% yield. The
process formally involves the cleavage of the isoxazole ring
and the aromatization of the A ring.
Scheme 6 shows the rationale for this intramolecular
redoxprocess. Deprotonation at the C2 position expels the
angular oxygen atom in X to give intermediate XI, which then
undergoes another deprotonation at the C3 position to cleave
To remove the menthyl ester moiety in 5, several classical
decarboxylation protocols were examined without success,
with each giving a complexmixture of unidentified products.
After considerable experimentation, the goal was achieved by
thermolysis of 5 in the presence of 4-DMAP.[9] Subsequent
hydrolysis of the 1,3-dioxane acetal afforded ketoaldehyde 6,
ready for the benzoin-forming reaction. Treatment of 6 with
thiazolium salt 7 (10 mol%) and DBU allowed a smooth
benzoin cyclization to give ketol 8 as a single diastereomer in
99% yield.[10]
ꢀ
the N O bond and to afford dianion XII. Upon quenching,
tautomerization gives imine 12 with the A ring aromatized.[16]
Angew. Chem. Int. Ed. 2008, 47, 1084 –1087
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1085

Communications
unstable oxirane, which was reduced with BH3·THF[21] to give
alcohol 15 (J1’,2’ = 10.6 Hz, J2’,3’ = 8.5 Hz) as a single product
in 66% yield.
The MOM group in 15 was selectively removed with PPTS
in the presence of molecular sieves (4 Š)[22] and the product
was treated with TBAF to remove the silyl protecting groups.
In a pleasant surprise, the remaining nosyl group was also
detached during concentration of the resulting mixture in
vacuo. The TBAF-derived byproducts were removed by the
procedure described by Kaburagi and Kishi;[23] reversed-
phase column chromatography (MeOH/H2O (65:35) with
0.1% trifluoroacetic acid) gave cassialoin (1). The spectro-
scopic data (1H and 13C NMR spectra, mass spectra, IR data,
TLC mobilities in several different solvent systems) of this
product were consistent with those reported for the natural
product.
Scheme 6. Intramolecular redox reaction. SB = strong base.
The next stage was the hydrolysis of imine 12 to the
corresponding ketone (Scheme 7), which was not achievable
under conventional acidic conditions. After unfruitful trials,
Having the homogeneous synthetic material with a
1
stereochemical identity, we noticed that H NMR spectrum
of the naturally obtained sample contained additional minor
peaks, which suggested the presence of an impurity or minor
component. Assuming that it could well be the C10 epimer,
we prepared a comparison sample by starting with 16, the
racemate of 8 (Scheme 8). The same series of transformations
Scheme 7. Reagents and conditions: a) NsCl (5.0 equiv), K2CO3
(10 equiv), DMF, 08C, 2 h; b) SiO2, CH2Cl2, room temperature, 14 h,
86% (2 steps); c) dimethyldioxirane, acetone/CH2Cl2 (1:2), 08C, 5 h;
d) BH3·THF (3.5 equiv), THF, 08C, 7 h, 66% (2 steps); e) PPTS
(1.0 equiv), tBuOH, reflux, 1 h, 81%; f) TBAF (5.0 equiv), THF, room
temperature, 10 h, 53%. Ns = 2-nitrobenzenesulfonyl; PPTS =
pyridinium p-toluenesulfonate; TBAF = tetrabutylammonium fluoride.
Scheme 8. Synthesis of C10 epimer 1’.
the issue was solved by protecting the phenol with an
as those described above gave an easily separable mixture of
epimers 15 (b-OH) and 17 (a-OH); and deprotection of the
latter gave the C10 epimer 1’.[24] The 1H NMR spectrum of 1’
coincided with the minor peaks present in the 1H NMR
spectrum of the natural sample.
In summary, this synthesis established the stereochemistry
of 1 and also revealed that the epimer 1’ is present as a minor
component in the natural material. The synthesis features the
first application of isoxazole-containing stereogenic a-ketol 8
as an anthraquinone equivalent and may be extended to the
syntheses and structure assignments of other anthrone
C-glycosides.
electron-withdrawing
2-nitrobenzenesulfonyl
(nosyl)
group,[17] which rendered the imine easily hydrolyzable.
Upon treatment of nosylate 12’ with silica gel, the corre-
sponding ketone 13 was obtained in high yield. The strong
electron-withdrawal by the nosylate group served also to
increase the stability of the tertiary alcohol under acidic
conditions.[18]
The remaining task was the hydration of glycal 13 to
complete the b-glucoside structure. While attempts by hydro-
boration or hydrosilylation uniformly failed, due presumably
to the decomposition path from metallacycle 14,[19] the
projected hydration was realized by epoxidation followed
by reductive opening of the oxirane ring. The epoxidation
with dimethyldioxirane[11,20] proceeded smoothly to give an
Received: October 8, 2007
Published online: December 28, 2007
1086 www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 1084 –1087

Angewandte
Chemie
[13] See the Supporting Information.
[14] J. S. Brimacombe, A. B. Foster, B. D. Jones, J. J. Willard, Chem.
Commun. 1967, 2404.
Keywords: configuration determination · glycosides ·
heterocycles · natural products · total synthesis
.
[15] For reviews on organoselenium chemistry, see: Organoselenium
Chemistry: Modern Development in Organic Synthesis (Topics in
Current Chemistry, Vol. 208) (Ed.: T. Wirth), Springer, Heidel-
berg, 2000.
[16] Formation of the methylene acetal was essential, as the cis-diol
10 remained unchanged under such reaction conditions.
[17] T. Fukuyama, C.-K. Jow, M. Cheung, Tetrahedron Lett. 1995, 36,
6373.
[18] Conversely, if electron-donating groups (para-methoxyphenyl-
methyl, MOM, or methylthiomethyl) were used for the phenol
protection, the product was unstable on silica gel and gave a
complexmixture of unidentified byproducts.
[19] The anthraquinone lacking the sugar moiety was identified as a
side product in the complexmixture.
[20] a) R. L. Halcomb, S. J. Danishefsky, J. Am. Chem. Soc. 1989, 111,
6661; b) J. D. Rainier, S. P. Allwein, J. Org. Chem. 1998, 63, 5310;
c) M. Inoue, S. Yamashita, A. Tatami, K. Miyazaki, M. Hirama,
J. Org. Chem. 2004, 69, 2797.
[21] The hydride attack proceeded only from the a face, to give 15.
For a related report, see; H. C. Brown, N. M. Yoon, J. Am.
Chem. Soc. 1968, 90, 2686.
[1] K. Hata, K. Baba, M. Kozawa, Chem. Pharm. Bull. 1978, 26,
3792.
[2] L. Krenn, R. Pradhan, A. Presser, G. Reznicek, B. Kopp, Chem.
Pharm. Bull. 2004, 52, 391.
[3] a) P. Manitto, D. Monti, G. Speranza, J. Chem. Soc. Perkin Trans.
1 1990, 1297; b) P. Manitto, D. Monti, G. Speranza, J. Chem. Soc.
Perkin Trans. 1 1993, 1577.
[4] a) K. A. Parker, C. A. Coburn, J. Am. Chem. Soc. 1991, 113,
8516; b) for a recent review on metalated heterocycles, see: R.
Chinchilla, C. Nµjera, M. Yus, Chem. Rev. 2004, 104, 2667.
[5] a) J. W. Bode, Y. Hachisu, T. Matsuura, K. Suzuki, Tetrahedron
Lett. 2003, 44, 3555; b) J. W. Bode, K. Suzuki, Tetrahedron Lett.
2003, 44, 3559; c) Y. Hachisu, J. W. Bode, K. Suzuki, J. Am.
Chem. Soc. 2003, 125, 8432; d) Y. Hachisu, J. W. Bode, K. Suzuki,
Adv. Synth. Catal. 2004, 346, 1097; e) T. Matsuura, J. W. Bode, Y.
Hachisu, K. Suzuki, Synlett 2003, 1746; f) J. W. Bode, Y. Hachisu,
T. Matsuura, K. Suzuki, Org. Lett. 2003, 5, 391; g) J. W. Bode, H.
Uekusa, K. Suzuki, Org. Lett. 2003, 5, 395.
[6] For an enantioselective route to a-ketols, see: H. Takikawa, Y.
Hachisu, J. W. Bode, K. Suzuki, Angew. Chem. 2006, 118, 3572;
Angew. Chem. Int. Ed. 2006, 45, 3492.
[7] A. G. Myers, N. J. Tom, M. E. Fraley, S. B. Cohen, D. J. Madar, J.
Am. Chem. Soc. 1997, 119, 6072.
[22] H. Monti, G. Lꢀandri, M. Klos-Ringquet, C. Corriol, Synth.
Commun. 1983, 13, 1021.
[23] The TBAF-derived materials were removed by treatment with
Dowex50WX8-400 and CaCO 3 in MeOH (room temperature,
1 d); see: Y. Kaburagi, Y. Kishi, Org. Lett. 2007, 9, 723.
[24] With normal silica gel chromatography, the behaviors of the
epimers 1 and 1’ were very similar; the separation was only
realized by HPLC with a chiral stationary phase, which revealed
that the natural sample was composed of a 4:1 mixture of 1 and
1’.
[8] The regioisomer was obtained in 17% yield. Structural assign-
ments of 5 and the regioisomer were based on the HMBC
correlations, after hydrolysis of the cyclic acetal.
[9] D. F. Taber, R. E. Ruckle, Jr., Tetrahedron Lett. 1985, 26, 3059.
[10] The stereostructure of 8 was assigned by the NOE observed
between the tertiary alcohol and the C3 methine proton.
[11] S. W. Roberts, J. D. Rainier, Org. Lett. 2005, 7, 1141.
1
[12] The H NMR spectra of 10 is highly dependent on the concen-
tration and temperature, due most likely to the change in the
aggregation state.
Angew. Chem. Int. Ed. 2008, 47, 1084 –1087
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1087