Total synthesis of (-)-episilvestrol and (-)-silvestrol

Article abstract of DOI:10.1002/anie.200702700

Sugar and spice... The total synthesis of the rare but potent anticancer natural product (-)-episilvestrol and its 5? epimer (-)-silvestrol was accompushed from D-glucose, naringenin, and methyl cinnamate (see scheme). The key steps of the sequence were inspired by the possible biogenesis of these compounds.

Full text of DOI:10.1002/anie.200702700

Angewandte
Chemie
DOI: 10.1002/anie.200702700
Natural Product Synthesis
Total Synthesis of (À)-Episilvestrol and (À)-Silvestrol**
Mariana El Sous, Mui Ling Khoo, Georgina Holloway, David Owen, Peter J. Scammells, and
MarkA. Rizzacasa*
Extracts of the bark of the woody South-East Asian shrub
Aglaia leptantha, Miq. (Meliaceae) show potent cytotoxic
activity, which is attributed to two new molecules 1 and 2.[1]
Both silvestrol (1) and episilvestrol (2) show comparable
potent cytotoxic activity against several human tumor cell
lines including lung, prostate, and breast cancer with IC50
values ranging from1 to 7 n m.[1,8] Recently, it has been
demonstrated that the cytotoxicity induced by silvestrol (1) in
human prostate cancer (LNCaP) cells is associated with a
block in the cell cycle at the G2/M checkpoint.[9] Specifically,
compound 1 alters the expression of genes regulating
apoptosis and cell cycle in a way that is not dependent on
human tumor protein 53 (p53). Interestingly, compounds that
possess only the parent cyclopenta[b]benzofuran, such as 3
and 4, are significantly less active, suggesting that the
presence of the 1,4-dioxanyloxy group is critical for activity.[1]
The in vivo activity of 1 has also been demonstrated.
Administration of silvestrol (1) by intraperitoneal injection
(3 mgkgÀ1) to athymic mice with human prostate cell
xenografts reduced tumor weights by about 60% after
29 days, while the body weight of the mouse was unaffected.[1]
Unfortunately, the paucity of both 1 and 2 fromthe natural
source (0.01% isolated yield) precludes isolation as a
sufficient supply of these important compounds. Therefore,
an efficient chemical synthesis is an attractive alternative for
accessing quantities of 1 or 2 as well as analogues for further
biological evaluation. We now report the total synthesis of
(À)-episilvestrol (2) and (À)-silvestrol (1) from common
simple precursors which could address the supply problem.[10]
Our retrosynthesis of episilvestrol (2; Scheme 1) is
inspired, in part, by the proposed biosynthesis of the two
Compounds 1 and 2 are diastereoisomers (epimers at 5’’’) that
contain a common cyclopenta[b]benzofuran core,[2] found in
the related Aglaia metabolites aglafolin[3] (methyl rocaglate,
3)[4–6] and rocaglamide (4),[6,7] as well as a novel, unprece-
dented 1,4-dioxanyloxy pseudosugar substituent. In addition,
two metabolites, named silvestrol and episilvestrol were
isolated from Aglaia foveolata Pannell by Kinghorn and co-
workers and found to be identical to 1 and 2, respectively.[8]
The structure of silvestrol (1) has been determined by NMR
spectroscopy and X-ray analysis of a derivative which served
to confirmthe relative and absolute configuration of this
compound.
[*] Dr. M. El Sous, M. L. Khoo, Prof. Dr. M. A. Rizzacasa
School of Chemistry
Bio21 Institute, The University of Melbourne
Melbourne, Victoria 3010 (Australia)
Fax: (+61)3-9347-8396
E-mail: masr@unimelb.edu.au
Homepage: http://www.chemistry.unimelb.edu.au/staff/masr/
research/MARHomepage.html
Dr. D. Owen[+]
Cerylid Biosciences
Richmond, Victoria (Australia)
Dr. G. Holloway, Prof. Dr. P. J. Scammells
Department of Medicinal Chemistry
Victorian College of Pharmacy, Monash University
381 Royal Parade, Melbourne, Victoria 3052 (Australia)
[+] Current address:
Starpharma Pty Ltd., Baker Building
75 Commercial Rd, Melbourne, Victoria 3004 (Australia)
[**] Financial support for this work was provided by the Australian
Research Council (Discovery Grant DP0770408) and Cerylid
Biosciences. We thank Prof. T. Burgess and Dr. F. Walker (Ludwig
Institute for Cancer Research, Melbourne) for conducting the cancer
cell assays and Prof. A. D. Kinghorn (Ohio State University) and Dr.
M. Tait (Cerylid Biosciences) for authentic samples of 1 and 2.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Scheme 1. Retrosynthetic analysis of episilvestrol (2). TBS=tert-butyl-
dimethylsilyl.
Angew. Chem. Int. Ed. 2007, 46, 7835 –7838
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
7835

Communications
key fragments: the cyclopentabenzofuran core[2,11] and the
1,4-dioxanyloxy fragment.[12] Episilvestrol (2) could be pro-
duced by deprotection of precursor 5, which in turn could be
secured by a modified Mitsunobu-type[13] coupling between
the hemiacetal 6 and core phenol 7. Compound 6 could arise
fromthe known d-glucose derivative 8 by following a route
based on a biosynthetic proposal suggested by us.[12] The
cyclopentabenzofuran core 7 could be secured from commer-
cially available (Æ )-4’,5,7-trihydroxyflavanone (naringenin,
9) and methyl cinnamate by adapting the elegant oxidopyr-
illium[3 +2] cycloaddition approach to these systems as
reported by Porco and co-workers.[5,6]
Scheme 3. Model Mitsunobu coupling. DIAD=diisopropylazodicarbox-
ylate.
The synthesis of the dioxane fragment 6 (Scheme 2) is
based on our approach to a model system.[12] Glycosylation of
commercially available bromide 8 with p-methoxybenzyl
sieves was critical to the success of this reaction, and the
adducts 18 and 19 were obtained in 54% yield favoring the
desired axial isomer 18. With the success of this coupling in a
model system, we next prepared the aromatic core 7 required
for the production of 2.
The synthesis of the required cyclopentabenzofuran core
7 was achieved as detailed in Scheme 4. Selective O6
benzylation of (Æ )-naringenin (9) using Na2CO3 as base
afforded ether 20, which was oxidized to the flavone 21 upon
exposure to iodine in pyridine. Methylation provided the
ether 22, which could be converted into the cycloaddition
precursor 23 through C3 deprotonation with LDA[19] and
Scheme 2. Synthesis of 1,4-dioxane fragment 6. PMB=p-methoxyben-
zyl, Bn=benzyl, CSA=10-camphorsulfonic acid, DiBAL-H=diisobutyl-
aluminum hydride, OTf=trifluoromethanesulfonate, DDQ=dichlorodi-
cyanobenzoquinone.
alcohol mediated by silver carbonate followed by methanol-
ysis of the acetate groups and benzylidene formation gave
acetal 10[14] in a reasonable overall yield. Exposure of 10 to
[15]
BH3·THF in the presence of Cu(OTf)2
effected highly
À
selective C O6 benzylidene cleavage to afford the O4-
protected sugar 11. Periodate cleavage[16] followed by imme-
diate selective reduction of the resultant aldehyde-lactol
afforded lactol as an approximately 1:3 mixture of anomers.
Selective protection of the primary alcohol as the TBS
ether and subsequent methylation of the lactol mixture using
[17]
butyllithiumand methyl triflate
gave the desired axial
product 15 in high yield, along with a small amount of
equatorial ketal 14. The observed selectivity can be explained
by rapid alkylation of the intermediate axial alkoxide at faster
rate than the corresponding equatorial alkoxide. A protect-
ing-group exchange was then carried out to swap the benzyl
ether for a TBS group to provide 16. DDQ-induced removal
of the PMB group gave the lactols 6 as a roughly 1:1 mixture.
After some experimentation, we found that the Misunobu
reaction as utilized by Roush and Lin for the synthesis of O-
aryl glycosides[18] could be applied to couple the model phenol
17 and lactols 6 (Scheme 3). The addition of 4-Š molecular
Scheme 4. Synthesis of the cyclopentabenzofuran core (Æ)-7.
Bn=benzyl, LDA=lithium diisopropylamide, pyr.=pyridine, PMP=p-
methoxyphenyl.
7836
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 7835 –7838

Angewandte
Chemie
quenching of the trimethylborate followed by oxidative
workup. Irradiation[20] of the 3-hydroxyflavone 23 (350 nm,
450-W medium-pressure Hg lamp) generated the dipole 24,
which underwent cycloaddition[5] with methyl cinammate to
give a complex mixture of the adduct 25 along with the
cyclobutane 26, presumably as a result of an alternative
[2+2] cycloaddition[21] or a-ketol rearrangement. Treatment
of this mixture with base, in analogy to that reported by Porco
and co-workers,[5] induced a-ketol rearrangement to a ther-
modynamically preferred b-ketoester, and subsequent selec-
tive anti reduction[22] afforded the desired cyclopentabenzo-
furan 27 along with the corresponding exo isomer in a ratio of
4.6:1, respectively. The stereochemistry of the major endo
version was lower (14%, 49% based on recovered 7). As
expected, the combination of racemic core 7 with optically
pure 6 afforded two major axial diastereoisomers 28 and 5 in
equal amounts which were inseparable by conventional flash
chromatography. Subsequent normal-phase HPLC (5 mm
silica, 250 mm ” 10 mm, 2 mLminÀ1 with 50% EtOAc/petro-
leumether eluent) cleanly separated the isomers
5 (Rt =
10.10 min) and 28 (Rt = 10.48 min). It was difficult at this
stage to determine which was the correct isomer required for
the production of 2, however, deprotection the slower eluting
compound with TBAF afforded diastereoisomer 29 while
deprotection of the isomer with the shorter retention time on
HPLC, namely compound 5, then afforded (À)-episilvestrol
(2).
1
isomer was assigned on the basis of its H NMR spectrum,
which displayed signals at d = 5.03 ppm(1H, d, J = 6 Hz, H1),
4.31 ppm(1H, d, J = 14 Hz, H3), and 3.90 ppm(1H, dd, J =
14, 6 Hz, H2) suggesting a 1,2-cis, 2,3-trans arrangement.
These H1–H3 signals matched those observed for episilves-
trol (2).[8] Hydrogenolysis of 27 gave the desired core 7 as a
racemate. This short sequence provided useful quantities of
(Æ )-7 ready for coupling to the dioxane fragment.
Synthetic episilvestrol (2) ([a]2D5 = À91.3, c = 0.03, CHCl3)
was identical to natural 2[1,8] (lit.[8] [a]2D0 = À94.5, c = 0.43,
CHCl3) in all respects (1H and 13C NMR, IR, and mass
spectra), while isomer 26 ([a]2D4 = À23.9, c = 0.09, CHCl3)
displayed slightly different physical data. In addition, syn-
thetic episilvestrol (2) displayed anticancer activity against
human colon cancer cells LIM 1215 (IC50 = 2.0 nm) compa-
rable to that reported for the natural product, whilst
compound 29 was less active by at least one order of
magnitude (IC50 = 56 nm). Thus, it appears that the stereo-
chemistry of the core has some influence on the cytotoxicity.
Finally, conversion of episilvestrol (2) into silvestrol (1) was
easily achieved by a selective “double Mitsunobu” reaction,
which resulted in inversion at 5’’’ (Scheme 5).[10]
The pivotal Mitsunobu coupling between 6 and (Æ )-7
proceeded with selectivity for the desired axial diastereoiso-
mers in reasonable yield based on recovered core
7
(Scheme 5). A higher selectivity for the axial isomers ( ꢀ 6:1
ax/eq ratio) was obtained when the coupling was run at 08C
and not warmed to room temperature, however, the con-
In conclusion, we have achieved a total synthesis of
episilvestrol (2; 21 steps in total; longest linear sequence of
13 steps) fromreadily available precursors through a route
inspired by the possible biogenesis of these compounds. In
addition, episilvestrol (2) was converted into silvestrol (1).
With the development of an asymmetric approach to the
cyclopentabenzofuran core,[6] the efficient synthesis of 2 and
analogues which are not available fromthe natural products
themselves is possible.
Received: June 20, 2007
Published online: September 6, 2007
Keywords: antitumor agents · cycloaddition · natural products ·
.
total synthesis
[1] B. M. Meurer-Grimes, J. Yu, G. L. Vario, WO 2002002566, A1
20020110, PCT Int. Appl., 2002.
[2] P. Proksch, R. Edrada, R. Ebel, F. I. Bohnenstengel, B. W.
Nugroho, Curr. Org. Chem. 2001, 5, 923.
[3] T.-S. Wu, M.-J. Liou, C.-S. Kuoh, C.-M. Teng, T. Nagao, K.-H.
Lee, J. Nat. Prod. 1997, 60, 606.
[4] F. N. Ko, T. S. Wu, M. J. Liou, T. F. Huang, C. M. Teng, Eur. J.
Pharmacol. 1992, 218, 129.
[5] Total synthesis: B. Gerard, G. Jones II, J. A. Porco, Jr., J. Am.
Chem. Soc. 2004, 126, 13620.
[6] Asymmetric total synthesis: B. Gerard, S. Sangji, D. J. OꢀLeary,
J. A. Porco, Jr., J. Am. Chem. Soc. 2006, 128, 7754.
[7] M. Lu King, C. C. Chiang, H. C. Ling, E. Fujita, M. Ochiai, A. T.
McPhail, J. Chem. Soc. Chem. Commun. 1982, 1150; total
synthesis: a) C2,3 epimer G. A. Kraus, J. O. Sy, J. Org. Chem.
1989, 54, 77; b) (À)-rocaglamide: B. M. Trost, P. D. Greenspan,
Scheme 5. Completion of the total synthesis of episilvestrol (2) and
silvestrol (1). TBAF=tetra-n-butylammonium fluoride, DEAD=diethyl-
azodicarboxylate, PNBA=p-nitrobenzoic acid.
Angew. Chem. Int. Ed. 2007, 46, 7835 –7838
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7837

Communications
B. Yang, V. , M. G. Saulnier, J. Am. Chem. Soc. 1990, 112, 9022;
[12] M. A. Rizzacasa, M. El Sous, Tetrahedron Lett. 2005, 46, 293.
[13] O. Mitsunobu, Synthesis 1981, 1.
c) A. E. Davey, M. J. Schaeffer, R. J. K. Taylor, J. Chem. Soc.
Perkin Trans. 1 1992, 2657; d) M. R. Dobler, I. Bruce, F.
Cederbaum, N. G. Cooke, L. J. Diorazio, R. G. Halla, E. Irving,
Tetrahedron Lett. 2001, 42, 8281.
[14] E. Kaji, F. W. Lichtenthaler, Y. Osa, K. Takahashi, E. Matsui, S.
Zen, Chem. Lett. 1992, 707; E. Kaji, F. W. Lichtenthaler, Y. Osa,
K. Takahashi, S. Zen, Bull. Chem. Soc. Jpn. 1995, 68, 2401.
[15] C.-R. Shie, Z.-H. Tzeng, S. S. Kulkarni, B.-J. Uang, C.-Y. Hsu, S.-
C. Hung, Angew. Chem. 2005, 117, 1693 – 1696; Angew. Chem.
Int. Ed. 2005, 44, 1665.
[16] T. Heidelberg, J. Thiem, J. Prakt. Chem./Chem.-Ztg. 1998, 340,
223.
[17] R. R. Schmidt, M. Reichrath, U. Moering, Carbohydr. Chem.
1984, 3, 67.
[18] W. R. Roush, X.-F. J. Lin, J. Org. Chem. 1991, 56, 5740.
[19] A. M. Costa, F. M. Dean, M. A. Jones, R. S. Varma, J. Chem. Soc.
Perkin Trans. 1 1985, 799.
[20] a) P.-T. Chou, J. Chin. Chem. Soc. 2001, 48, 651; b) A. N. Bader,
F. Ariese, C. Gooijer, J. Phys. Chem. A 2002, 106, 2844.
[21] H. C. Hailes, R. A. Raphael, J. Staunton, Tetrahedron Lett. 1993,
34, 5313.
[8] B. Y. Hwang, B.-N. Su, H. M. Chai, Q. Mi, L. B. S. Kardono, J. J.
Afriastini, S. Riswan, B. D. Santarsiero, A. D. Mesecar, R. Wild,
C. R. Fairchild, G. D. Vite, W. C. Rose, N. R. Farnsworth, G. A.
Cordell, J. M. Pezzuto, S. M. Swanson, A. D. Kinghorn, J. Org.
Chem. 2004, 69, 3350; correction: B. Y. Hwang, B.-N. Su, H. M.
Chai, Q. Mi, L. B. S. Kardono, J. J. Afriastini, S. Riswan, B. D.
Santarsiero, A. D. Mesecar, R. Wild, C. R. Fairchild, G. D. Vite,
W. C. Rose, N. R. Farnsworth, G. A. Cordell, J. M. Pezzuto, S. M.
Swanson, A. D. Kinghorn, J. Org. Chem. 2004, 69, 6156.
[9] Q. Mi, S. Kim, B. Y. Hwang, B.-N. Su, H. Chai, Z. H. Arbieva,
A. D. Kinghorn, S. M. Swanson, Anticancer Res. 2006, 26(5A),
3349.
[10] D. J. Owen, M. A. Rizzacasa, M. El Sous, M. Spiniello, P. J.
Scammells, G. Holloway, WO 2006007634, A1 2006 0126, PCT
Int. Appl., 2006.
[11] M. Bacher, O. Hofer, G. Brader, S. Vajrodaya, H. Greger,
Phytochemistry 1999, 52, 253.
[22] D. A. Evans, K. T. Chapman, E. M. Carreira, J. Am. Chem. Soc.
1988, 110, 3560.
7838
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 7835 –7838