Enantioselective synthesis of the complex rocaglate (-)-silvestrol

Article abstract of DOI:10.1002/anie.200702707

The total synthesis of the natural product (-)-silvestrol (1) has been accomplished and features enantioselective [3+2] photocycloaddition of a substituted 3-hydroxyflavone and methyl cinnamate promoted by a chiral Bronsted acid. Initial biological studies indicate a 5-10-fold greater activity of silvestrol as an inhibitor of protein synthesis in HeLa cells than its 1″″ diastereomer.

Full text of DOI:10.1002/anie.200702707

Angewandte
Chemie
DOI: 10.1002/anie.200702707
Natural Product Synthesis
Enantioselective Synthesis of the Complex Rocaglate (ꢀ)-Silvestrol**
Baudouin Gerard, Regina Cencic, Jerry Pelletier, and John A. Porco, Jr.*
Extracts from dried roots and stems of several species of the
plant genus Aglaia are the source of the rocaglamides, a
unique group of natural products featuring a cyclopenta[b]te-
trahydrobenzofuran skeleton.[1] The complex rocaglate sil-
vestrol (1) and its epimer 2 (Figure 1) were recently isolated
the subject of synthetic studies, an enantioselective synthesis
of silvestrol has not been reported to date. We have chosen
the promising anticancer lead compound silvestrol as a
synthetic target to further develop applications of our
enantioselective photogeneration/cycloaddition of oxido-
pyryliums derived from 3-hydroxyflavones[4] and to
develop chemistry pertaining to the synthesis and
attachment of the structurally unique dioxanyloxy seg-
ment.
Our retrosynthetic analysis of silvestrol (1) is
illustrated in Figure 2. Compound 1 may be derived
ꢀ
from C O bond formation between 1,4-dioxan-2-ol 4
and hydroxyphenyl rocaglate derivative 5. Compound 5
may be prepared using photocycloaddition methodol-
ogy developed in our laboratory to access the methyl
Figure 1. (ꢀ)-Silvestrol and related derivatives.
from the plant Aglaia foveolata by Kinghorn and co-
workers.[2,3] The structural assignment of 1 was based
on NMR spectroscopy and X-ray diffraction studies of
the bis-p-bromobenzoate derivative 3. Compound 1
was shown to exhibit very potent cytotoxic activity
(e.g. ED50 = 1.2 nm against human lung cancer cells)
which is comparable to the activity of the anticancer
agent Paclitaxel (Taxol). Mechanism-of-action studies
indicate that cytotoxicity induced by silvestrol in
human prostate cancer (LNCaP) cells is associated
with a block in the cell cycle at the G2/M checkpoint in
Figure 2. Retrosynthetic analysis of (ꢀ)-silvestrol (1).
a manner that is independent of p53.[3] In contrast to
other rocaglate derivatives, silvestrol possesses a
unique dioxanyloxy group at C6 of the cyclopenta[b]benzo-
furan core. Although the rocaglamides[1c,d] and the corre-
sponding dioxanyloxy side chain[1e,f] of silvestrol have been
rocaglate core by employing protected 3-hydroxyflavone 6 as
starting material.[4] The unusual dioxanyloxy fragment 4 may
be obtained via 1,4-dioxan-2-one precursor 7, which may be
derived from 1,2-dibenzylthreitol derivative 8. Intermediate 8
is readily obtained from the commercially available reagent
d-dimethyl tartrate.[5]
[*] B. Gerard, Prof. Dr. J. A. Porco, Jr.
Department of Chemistry
The synthesis of the protected 3-hydroxyflavone 6 began
with Friedel–Crafts acylation of phloroglucinol (9) using
benzyloxyacetyl chloride in the presence of AlCl3 following a
modified procedure (Scheme 1).[6] After MOM protection
and methylation, the derived aryl ketone 10 was subjected to
selective MOM deprotection in the presence of a catalytic
amount of iodine in methanol,[7] followed by acylation with 4-
methoxybenzoyl chloride, to afford phenyl ester 11. Baker–
Venkataraman rearrangement of 11 under basic conditions
(LiHMDS, THF) yielded diketone 12.[8] Treatment of 12 with
sodium acetate in acetic acid effected smooth cyclization/
dehydration[9] with concomitant removal of the MOM ether.
Reintroduction of the MOM protecting group and subse-
quent hydrogenolysis of the benzyl group provided the
requisite 3-hydroxyflavone 6.
Center for Chemical Methodology and Library Development
Boston University
590 Commonwealth Avenue, Boston, MA 02215 (USA)
Fax: (+1)617-358-2847
E-mail: porco@bu.edu
R. Cencic, Prof. Dr. J. Pelletier
McGill University
Department of Biochemistry
Room 810, 3655 Drummond St., Montreal, QC, H3G 1Y6 (Canada)
[**] Financial support from the National Institutes of Health (J.A.P.,Jr.,
GM-073855) and from the NCI Canada (J.P., 017099), Bristol-Myers
Squibb, and Merck Research Laboratories is gratefully acknowl-
edged. We thank Dr. R. Murray Tait for providing natural silvestrol
and Dr. Siva Dandapani, Dr. Jean-Charles MariØ, Dr. Aaron Beeler,
and Mr. Suwei Dong (Boston University) for helpful discussions.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2007, 46, 7831 –7834
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
7831

Communications
dioxan-2-one 7 was based on a three-
step synthesis starting from 8 involving
protection of the primary alcohol, fol-
lowed by esterification with dimethoxy-
acetic acid, and final deprotection of the
primary alcohol and attempted trans-
acetalization of 18 under acid conditions
(e.g. p-TsOH, CSA, K-10 clay)
(Scheme 3). However, using this path-
way we could not isolate the desired
dioxane derivative 7. The major product
isolated was characterized as the pri-
mary ester derivative obtained from acyl
transfer during the deprotection step. A
similar approach led us to investigate a
one-pot process for functionalization of
the 1,2-diol 8 through regioselective
alkylation followed by lactonization.
We thus envisioned the use of a tin
acetal as a reactive intermediate for
Scheme 1. a) Benzyloxyacetyl chloride, AlCl3, CH2Cl2/Et2O (1:1), 508C, 24 h, 58%; b) MOMCl,
K2CO3, acetone, RT, 3 h, 60%; c) Me2SO4, K2CO3, acetone, 608C, 6 h, 93%; d) I2, MeOH, RT,
4 h, 94%; e) 4-methoxybenzoyl chloride, Et3N, DMAP, CH2Cl2, RT, 4 h, 85%; f) LiHMDS
(3 equiv), THF, ꢀ208C, 1 h, 88%; g) AcOH, AcONa (2.5 equiv), 1008C, 3 h; h) MOMCl,
acetone, K2CO3, RT, 6 h, 70% over two steps; i) H2, Pd(OH)2, EtOH/THF (1:1), RT, 45 min,
92%. Bn=benzyl, MOM=methoxymethyl, DMAP=4-dimethylaminopyridine, HMDS=hexame-
thyldisilazide.
With compound 6 in hand, we
proceeded to evaluate the asymmetric
synthesis of methyl rocaglate frag-
ment 5 by employing enantioselective
[3+2] photocycloaddition mediated
by functionalized TADDOL deriva-
tives (TADDOL= a,a,a’,a’-tetraaryl-
2,2-dimethyl-1,3-dioxolan-4,5-dime-
thanol).[4b] During our previous inves-
tigations, we found that both the
nature of the aryl substituent and
ketal side chain of the TADDOL
framework, as well as low-tempera-
ture reaction conditions, were crucial
factors for high enantioselectivity.
Photocycloaddition (hn > 350 nm) of
3-hydroxyflavone 6 and methyl cinna-
mate (13) in the presence of chiral
additive 14, bearing a 1-pyrenyl sub-
Scheme 2. a) hn>350 nm, CH2Cl2/toluene, ꢀ708C, 10 h, 66%; b) MeONa (2.5 equiv), MeOH,
608C, 30 min, 89%; c) Me4NBH(OAc)3, MeCN, AcOH, 57% (16), 13% (17); d) TMSBr, CH2Cl2,
3 h, ꢀ788C, 84% yield, 87% ee after recrystallization. TMS=trimethylsilyl.
stituent and cyclooctyl ketal, at ꢀ708C
using PhCH3/CH2Cl2 (2:1) as solvent
led to the formation of cycloadduct 15
as well as its ketol-shift isomer[4a] after
purification on SiO2 (Scheme 2). After an a-ketol rearrange-
ment/hydroxy-directed reduction sequence,[4a] endo-rocaglate
derivative 16 was isolated in 57% yield and 71% ee along
with the corresponding exo stereoisomer 17. Compound 16
was then subjected to MOM deprotection using TMSBr in
CH2Cl2 to afford hydroxyphenyl rocaglate derivative 5.
Fortunately, we were able to increase the enantiomeric
excess of 5 through recrystallization to afford centrosymmet-
ric racemate crystals[4b] and 5 with 87% ee (75% recovery) in
the mother liquor.
regioselective alkylation. Formation of the O-stannylene
acetal[10] derived from 8, followed by addition of freshly
prepared methyl 2-bromo-2-methoxy acetate,[11] afforded 7
(45%) and its stereoisomer 19 (33%). The overall process
8!7 represents
a
tandem alkylation/lactonization
sequence.[12] Attempted epimerization of dioxanyl derivative
19 was unsuccessful and led mostly to decomposition. Finally,
DIBAL reduction of 7 produced 1,4-dioxan-2-ol 4 as a
mixture of diastereoisomers.[13]
With the two fragments 4 and 5 in hand, we evaluated a
series of conditions for their coupling (Scheme 4).[14] Utiliza-
tion of glucosidation methods using fluoride or trichloroace-
tamidate reagents derived from 4 led to unsatisfactory yields
of coupling products and significant decomposition. After
The synthesis of the dioxanyloxy fragment 4 was initiated
with (2S,3S)-1,2-di-O-benzylidenethreitol (8), which was
readily obtained in four steps from commercially available
d-dimethyl tartrate.[5] Our first strategy to obtain 1,4-
7832
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 7831 –7834

Angewandte
Chemie
cies of silvestrol (1) and its 1’’’ diastereoisomer 22 in
an in vitro translation system utilizing a rabbit
reticulocyte lysate programmed with firefly lucifer-
ase (FLuc) mRNA.[20] Titration of silvestrol (1)
revealed an IC50 value of approximately 0.4 mm,
whereas 22 showed an IC50 value of about 2 mm
(Figure 3a), indicating that 22 is fivefold less
active than 1. The relative potency of 1 and 22
was also assessed in vivo by exposing cells to the
two compounds, followed by monitoring the incor-
poration of 35S-methionine into proteins. The
results indicate a 10-fold difference in IC50 values
between silvestrol (1) and 22, with silvestrol being
more potent for inhibition of protein synthesis in
HeLa cells (Figure 3b).
Scheme 3. a) TBSCl, imidazole, Et3N, DMF, RT, 6 h; b) dimethoxyacetic acid, DCC,
DMAP, THF, RT, 82% over two steps; c) nBu2SnO, benzene, reflux 9 h,
CH3OCHBrCO2Me, TBAI, benzene, 708C, 2 h 45% (7), 33% (18); d) DIBAL-H,
toluene, ꢀ788C, 1 h, 83%. TBS=tert-butyldimethylsilyl, DCC=dicyclohexyl carbo-
diimide, TBAI=tetra-n-butylammonium iodide, DIBAL-H=diisobutylaluminum hy-
dride.
In conclusion, we have accomplished the enan-
tioselective synthesis of the rocaglate natural prod-
uct and antitumor agent (ꢀ)-silvestrol. The key
strategy involves an enantiose-
lective dipolar cycloaddition of
oxidopyrylium ylides derived
from excited-state intramolecu-
lar proton transfer (ESIPT) of 3-
hydroxyflavones using specifi-
cally functionalized TADDOL
derivatives as chiral Brønsted
acids. The unusual 1,4-dioxanyl
unit was generated from readily
available starting materials using
a
tandem alkylation/lactoniza-
tion sequence. Initial biological
studies indicate that silvestrol
has approximately a 5–10-fold
greater activity as an inhibitor
of protein synthesis in vivo and in
vitro in HeLa cells than its 1’’’
diastereomer, illustrating the
influence of the stereochemistry
of the dioxanyl moiety on bio-
logical activity.[21] Further studies
toward the synthesis of related
rocaglamide derivatives and bio-
Scheme 4. a) DIAD, F-PPh3, toluene, 4-Š MS, RT, 42% (20), 20% (21); b) H2, Pd(OH)2, EtOH, 87%.
DIAD=diisopropylazodicarboxylate, F-PPh3 =diphenyl-[4-(1H,1H,2H,2H-perfluorodecyl)phenyl]phos-
phine.
considerable experimentation, we identified the Mitsunobu
reaction[1f,15] as a suitable transformation for fragment
coupling of 4 and 5. However, in this particular instance,
separation of the triphenylphosphine oxide by-product from
the desired products was found to be problematic. Fortu-
nately, use of a fluorous-tagged triphenylphosphine reagent
(F-PPh3)[16] enabled facile product purification after filtration
through fluorous silica gel. The desired coupling product was
isolated as a separable mixture of diastereoisomers 20 (42%)
and 21 (20%). Finally, hydrogenation of 20 and 21 using
Pearlmanꢀs catalyst afforded silvestrol (1) and its 1’’’ stereo-
isomer 22. Data for synthetic 1 were confirmed to be identical
logical evaluation of silvestrol and related molecules will be
reported in due course.
Received: June 20, 2007
Published online: September 5, 2007
Keywords: antitumor agents · cycloaddition · natural products ·
.
photochemistry · total synthesis
[1] For isolation and characterization, see: a) P. Proksch, R. Edrada,
R. Ebel, F. I. Bohnenstengel, B. W. Nugroho, Curr. Org. Chem.
2001, 5, 923 – 938; b) S. Kim, A. A. Salim, S. M. Swanson, A. D.
Kinghorn, Anti-Cancer Agents Med. Chem. 2006, 6, 319 – 345.
For synthetic studies, see: c) G. A. Kraus, J. O. Sy, J. Org. Chem.
1989, 54, 77 – 83; d) B. M. Trost, P. D. Greenspan, B. V. Yang,
M. G. Saulnier, J. Am. Chem. Soc. 1990, 112, 9022 – 9024; e) M.
El Sous, M. A. Rizzacasa, Tetrahedron Lett. 2005, 46, 293 – 295;
1
with those reported for natural (ꢀ)-silvestrol[2] including H
and 13C NMR, mass, and IR spectra, [a]D values, HPLC, and
TLC Rf values.[17]
Cyclopenta[b]benzofurans are known protein synthesis
inhibitors.[18,19] Accordingly, we compared the relative poten-
Angew. Chem. Int. Ed. 2007, 46, 7831 –7834
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7833

Communications
[3] Q. Mi, S. Kim, B. Y. Hwang, B.-N. Su, H. Chai, Z. H. Arbieva,
D. A. Kinghorn, S. M. Swanson, Anticancer Res. 2006, 26, 3349 –
3356.
[4] a) B. Gerard, G. Jones II, J. A. Porco, Jr., J. Am. Chem. Soc.
2004, 126, 13620 – 13621; b) B. Gerard, S. Sangji, D. OꢀLeary,
J. A. Porco, Jr., J. Am. Chem. Soc. 2006, 128, 7754 – 7756.
[5] S. Takano, A. Kurotaki, Y. Sekiguchi, S. Satoh, M. Hirama, K.
Ogasawara, Synthesis 1986, 811 – 817.
[6] L. Crombie, R. C. F. Jones, C. J. Palmer, J. Chem. Soc. Perkin.
Trans. 1. 1987, 317 – 331.
[7] J. M. Keith, Tetrahedron Lett. 2004, 45, 2739 – 2742.
[8] a) A. Fougerousse, E. Gonzalez, R. Brouillard, J. Org. Chem.
2000, 65, 583 – 586; b) C. J. Bennett, S. T. Caldwell, D. B.
McPhail, P. C. Morrice, G. C. Duthie, R. C. Hartley, Bioorg.
Med. Chem. 2004, 12, 2079 – 2098.
[9] W. Baker, J. Chem. Soc. 1933, 1381 – 1389.
[10] For a review on functionalization of diols using O-stannylene
acetals and related derivatives, see: S. David, S. Hanessian,
Tetrahedron 1985, 41, 643 – 663.
[11] For radical bromination of a-alkoxyesters, see: a) A. L. J. Beck-
with, C. L. L. Chai, Tetrahedron 1993, 49, 7871 – 7882; b) R.
Clausen, M. Bols, J. Org. Chem. 2000, 65, 2797 – 2801.
[12] For a tandem alkylation/lactonization sequence using a tin
acetal, see: A. Rathjens, J. Thiem, Pol. J. Chem. 2005, 79, 211 –
222.
[13] M. Kawasaki, F. Matsuda, S. Terashima, Tetrahedron 1988, 44,
5713 – 5725.
[14] For reviews of O-glycosylation methods, see: a) K. Toshima, K.
Tatsuta, Chem. Rev. 1993, 93, 1503 – 1531; b) M. Jacobsson, J.
Malmberg, U. Ellervik, Carbohydr. Res. 2006, 341, 1266 – 1281.
[15] For a recent review concerning the Mitsunobu reaction, see: R.
Dembinski, Eur. J. Org. Chem. 2004, 2763 – 2772.
[16] For recent applications of fluorous-tagged reagents, see: a) D. P.
Curran, S. Dandapani in Handbook of Fluorous Chemistry
(Eds.: J. A. Gladysz, D. P. Curran, I. T. Horvꢁth), Wiley-VCH,
Weinheim, 2004, 436 – 437; b) S. Dandapani, D. P. Curran, J. Org.
Chem. 2004, 69, 8751 – 8757.
[17] See Supporting Information for complete experimental details.
[18] S. K. Lee, B. Cui, R. R. Mehta, A. D. Kinghorn, J. M. Pezzuto,
Chem.-Biol. Interact. 1998, 115, 215 – 228.
Figure 3. a) Dose-dependent inhibition of translation in vitro by 1 and
22 in rabbit reticulocyte lysates. Firefly luciferase activity of FF/HCV/
Ren mRNA in the presence of compound 1 or 22 was normalized to
the value obtained in the presence of vehicle (DMSO). b) Dose-
dependent inhibition of protein synthesis in vivo in HeLa cells by 1
and 22. TCA: separation of radiolabeled protein from unincorporated
35S-methionine by precipitation. The material was quantified by liquid
scintillation counting.
[19] T. Ohse, S. Ohba, T. Yamamoto, T. Koyano, K. Umezawa,J. Nat.
Prod. 1996, 59, 650 – 652.
[20] O. Novac, A. S. Guenier, J. Pelletier, Nucleic Acids Res. 2004, 32,
902 – 915.
[21] For recent examples concerning the effect of stereochemistry on
biological activity, see: a) B. A. Harrison, T. M. Gierasch, C.
Neilan, G. W. Pasternak, G. L. Verdine, J. Am. Chem. Soc. 2002,
124, 13352 – 13353; b) Y-k. Kim, M. A. Arai, T. Arai, J. O.
Lamenzo, E. F. Dean III, N. Patterson, P. A. Clemons, S.
Schreiber, J. Am. Chem. Soc. 2004, 126, 14740 – 14745.
f) D. J. Owen, M. A. Rizzacasa, M. El Sous, M. Spiniello, P. J.
Scammells, G. Holloway, WO 2006007634 PCTInt. Appl., 2006.
[2] B. Y. Hwang, B.-N. Su, H. 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 – 3358.
7834
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 7831 –7834