Synthesis of biotinylated episilvestrol: Highly selective targeting of the translation factors eIF4AI/II

Article abstract of DOI:10.1021/ol400401d

Silvestrol (1) and episilvestrol (2) are protein synthesis inhibitors, and the former has shown efficacy in multiple mouse models of cancer; however, the selectivity of these potent cytotoxic natural products has not been described. Herein, it is demonstr

Full text of DOI:10.1021/ol400401d

ORGANIC
LETTERS
2013
Vol. 15, No. 6
1406–1409
Synthesis of Biotinylated Episilvestrol:
Highly Selective Targeting of the
Translation Factors eIF4AI/II
Jennifer M. Chambers,† Lisa M. Lindqvist,*,‡ Andrew Webb,‡ David C. S. Huang,‡
G. Paul Savage,§ and Mark A. Rizzacasa*,†
School of Chemistry, The Bio21 Institute, The University of Melbourne, Melbourne,
Victoria 3010, Australia, Walter and Eliza Hall Institute of Medical Research, Parkville,
Victoria 3052, Australia, Department of Medical Biology, The University of Melbourne,
Parkville, Victoria 3010, Australia, and CSIRO Molecular and Health Technologies,
Bayview Avenue, Victoria 3168, Australia
masr@unimelb.edu.au; lindqvist@wehi.edu.au
Received February 10, 2013
ABSTRACT
Silvestrol (1) and episilvestrol (2) are protein synthesis inhibitors, and the former has shown efficacy in multiple mouse models of cancer;
however, the selectivity of these potent cytotoxic natural products has not been described. Herein, it is demonstrated that eukaryotic initiation
factors eIF4AI/II were the only proteins detected to bind silvestrol (1) and biotinylated episilvestrol (9) by affinity purification. Our study
demonstrates the remarkable selectivity of these promising chemotherapeutics.
The approval of omacetaxine mepesuccinate (homo-
harringtonine) by the U.S. Food and Drug Administration
(FDA) for the treatment of chronic myeloid leukemia
has validated targeting the protein synthesis (translation)
machinery for the treatment of cancer. While omacetaxine
mepesuccinate inhibits translation elongation, the more
recently discovered translation initiation inhibitor silves-
trol (1) prevents ribosome recruitment.1 Both 1 and
5000-episilvestrol or episilvestrol (2) have demonstrated
potent cytotoxic activity against many human cancer cell
lines, including lung, prostate, and breast cancer with IC50
values ranging from 1 to 7 nM.2ꢀ4 Moreover, 1 has shown
remarkable activity against B-cells isolated from patients
with chronic lymphocytic leukemia as well as in vivo
activity against acute lymphoblastic leukemia in a mouse
xenograft.5 The natural product 1 can be metabolized into
silvestric acid, which was inactive in a leukemia cell line.6
(4) Pan, L.; Kardono, L. B. S.; Riswan, S.; Chai, H.; de Blanco, E. J.;
Pannell, C. M.; Soejarto, D. D.; McCloud, T. G.; Newman, D. J.;
Kinghorn, A. D. J. Nat. Prod. 2010, 73, 1873.
(5) Lucas, D. M.; Edwards, R. B.; Lozanski, G.; West, D. A.; Shin,
J. D.; Vargo, M. A.; Davis, M. E.; Rozewski, D. M.; Johnson, A. J.; Su,
B. N.; Goettl, V. M.; Heerema, N. A.; Lin, T. S.; Lehman, A.; Zhang,
X. L.; Jarjoura, D.; Newman, D. J.; Byrd, J. C.; Kinghorn, A. D.;
Grever, M. R. Blood 2009, 113, 4656.
† The Bio21 Institute, The University of Melbourne.
‡ Walter and Eliza Hall Institute of Medical Research and Department of
Medical Biology, The University of Melbourne.
§ CSIRO Molecular and Health Technologies.
(1) Bordeleau, M. E.; Robert, F.; Gerard, B.; Lindqvist, L.; Chin,
S. M. H.; Wendel, H. G.; Brem, B.; Greger, H.; Lowe, S. W.; Porco, J. A.,
Jr.; Pelletier, J. J. Clin. Invest. 2008, 118, 1.
(2) Meurer-Grimes, B. M. Yu, J. Vario, G. L. WO 2002002566, A1
20020110, PCT Int. Appl., 2002.
(3) Hwang, B. Y.; Su, B.-N.; Chai, H. M.; Mi, Q.; Kardono, L. B. S.;
Afriastini, J. J.; Riswan, S.; Santarsiero, B. D.; Mesecar, A. D.; Wild, R.;
Fairchild, C. R.; Vite, G. D.; Rose, W. C.; Farnsworth, N. R.; Cordell,
G. A.; Pezzuto, J. M.; Swanson, S. M.; Kinghorn, A. D. J. Org. Chem.
2004, 69, 3350. Correction: J. Org. Chem. 2004, 69, 6156.
(6) Saradhi, U. V. R.; Gupta, S. V.; Chiu, M.; Wang, J.; Ling, Y.; Liu,
Z.; Newman, D. J.; Covey, J. M.; Kinghorn, A. D.; Marcucci, G.; Lucas,
D. M.; Grever, M. R.; Phelps, M. A.; Chan, K. K. AAPS J. 2001, 13, 347.
(7) Lindqvist, L. M.; Vikstrom, I.; Chambers, J. M.; McArthur, K.;
Anderson, M. A.; Henley, K. J.; Happo, L.; Cluse, L.; Johnstone, R. W.;
Roberts, A. W.; Kile, B. T.; Croker, B. A.; Burns, C. J.; Rizzacasa,
M. A.; Strasser, A.; Huang, D. C. S. Cell Death & Disease 2012, 3, e409.
r
10.1021/ol400401d
Published on Web 03/05/2013
2013 American Chemical Society

Silvestrol can induce multiple forms of cell death, which
varies between cell types.7
accessible by our synthetic approach and has equivalent
activity to 1 (Figure 2A), so 2 was selected for func-
tionalization.
Both 1 and 2 are metabolites of Aglaia, which were
isolated independently by two groups.2,3 These com-
pounds bear a unique 1,4-dioxylanyloxy structural moiety
that is connected to a cyclopentabenzofuran ‘rocaglate’8
ring system. Two additional isomers were later identified
from another Aglaia species as the diastereoisomers
2000-episilvestrol (3) and2000,5000-diepisilvestrol (4) (Figure 1).4
Figure 1. Structures of silvestrol (1), 5000-episilvestrol (2), and
related compounds.
Figure 2. Biological activity of acid derivatives at the C2 posi-
tion. (A) In vitro translations performed in rabbit reticulocyte
lysate programmed with capped Firefly luciferase/HCV/Renilla
luciferase mRNA. Data are represented as Firefly luciferase/
Renilla luciferase (FF/Ren) relative to vehicle (DMSO) controls
(n = 3). HCV-driven Renilla served as an internal control. (B)
Rate of translation of endogenous mRNAs after 1 h treatment
of MEFs with 1 μM compound (n = 3ꢀ4). (C) Cytotoxicity of
compounds after 72 h. The number of viable cells was measured
by the relative ATP concentration in MEFs after 72 h (n = 3ꢀ4).
All graphs depict the mean and standard error of the mean
(SEM).
The paucity of these compounds from natural sources
has led to the total synthesis of 1,9ꢀ11 2,10,11 and 2000,5000-
diepisilvestrol (4),12 as well as the unnatural analogues
40-desmethoxyepisilvestrol10 and 1000,2000,5000-triepisilvestrol.12
Pelletier and co-workers have demonstrated that silvestrol
and selected rocaglate analogues inhibit translation initiation
by modulating the activity of eIF4A and removing the RNA
helicase from the eIF4F complex.1,13,14 While a related
rocaglamide 1-O-formylaglafoline which also targets eIF4A
does not inhibit Ded1p activity or inhibit mRNA splicing,1
the selectivity of silvestrol or episilvestrol has not been
systematically defined. Herein, we disclose the remarkable
and unusual selectivity silvestrol and episilvestrol for eIF4AI
and eIF4AII.
As it is not practical to obtain silvestrol (1) and episil-
vestrol (2) from the natural source, the total synthesis of
these compounds was required for these studies (Figure 1).
Synthetic 1 and 2 were obtained as previously reported,10,11
using recent modifications to the synthesis of the enantio-
merically pure cyclopentabenzofuran core.12 2 is more
In order to determine the protein target(s) of 1 and 2, we
sought an active analogue of 2 with a suitable biotin tag.
Previous work has demonstrated that while modification
at the 40 position retained activity,11 the addition of larger
substituents at this location is not tolerated. In addition,
analogues of the dioxane fragment often compromised the
bioactivity;15 therefore, the C2 ester was the next target for
modification. We were originally hesitant to attempt func-
tionalization of this site since it had been previously
reported that the conversion of silvestrol to silvestric acid
in plasma renders the derivative inactive.6 Surprisingly, we
found that silvestric acid, obtained by base hydrolysis3 of
synthetic 1,11 had a similar ability to inhibit cap-dependent
translation as the parent compound in vitro (Figure 2A).
Episilvestric acid (5), prepared by hydrolysis of synthetic
2 (Scheme 1), was also equally potent. We therefore
(8) Proksch, P.; Edrada, R.; Ebel, R.; Bohnenstengel, F. I.; Nugroho,
B. W. Curr. Org. Chem. 2001, 5, 923.
(9) Gerard, B.; Cencic, R.; Pelletier, J.; Porco, J. A., Jr. Angew.
Chem., Int. Ed. 2007, 46, 7831.
(10) El Sous, M.; Khoo, M. L.; Holloway, G.; Owen, D.; Scammells,
P. J.; Rizzacasa, M. A. Angew. Chem., Int. Ed. 2007, 46, 7835.
(11) Adams, T. E.; El Sous, M.; Hawkins, B. C.; Hirner, S.; Holloway,
G.; Khoo, M. L.; Owen, D. J.; Savage, G. P.; Scammells, P. J.; Rizzacasa.,
M. A. J. Am. Chem. Soc. 2009, 131, 1607.
(12) Chambers, J. M.; Huang, D. C. S.; Lindqvist, L. M.; Savage,
G. P.; White, J. M.; Rizzacasa, M. A. J. Nat. Prod. 2012, 75, 1500.
ꢀ
(13) Cencic, R.; Carrier, M.; Galicia-Vazquez, G.; Bordeleau, M. -E.;
(15) Liu, T.; Nair, S. J.; Lescarbeau, A.; Belani, J.; Peluso, S.; Conley,
J.; Tillotson, B.; O’Hearn, P.; Smith, S.; Slocum, K.; West, K.; Helble, J.;
Douglas, M.; Bahadoor, A.; Ali, J.; McGovern, K.; Fritz, C.; Palombella,
V. J.; Wylie, A.; Castro, A. C.; Tremblay, M. R. J. Med. Chem. 2012, 55,
8859.
Sukarieh, R.; Bourdeau, A.; Brem, B.; Teodoro, J. G.; Greger, H.;
Tremblay, M. L.; Porco, J. A., Jr.; Pelletier, J. PLoS ONE 2009, 4, e5223.
(14) Rodrigo, C. M; Cencic, R.; Roche, S. P.; Pelletier, J.; Porco,
J. A., Jr. J. Med. Chem. 2012, 55, 558.
Org. Lett., Vol. 15, No. 6, 2013
1407

speculated that the reduction of activity observed pre-
viously was due to impairment of cellular uptake, not an
inability to bind the target(s). Since hydrolysis may reduce
the activity in vivo, it will be interesting to determine if
compounds which cannot be metabolized into the acid will
have increased potency in animal models.
as a Cu(I) ligand19 gave biotinylated episilvestrol 9 in a
good yield.
Scheme 2. Synthesis of Biotinylated Episilvestrol (9)
To examine this hypothesis, we compared the ability of 1
and 2, along with silvestric acid3 and 5, to inhibit transla-
tion in mouse embryonic fibroblasts (MEFs), and indeed
both acid derivatives were less potent (Figure 2B). In
addition, both silvestric acid and episilvestric acid had a
100-fold increase in IC50 compared to their parent com-
pounds when cytotoxicity was compared (Figure 2C).
Therefore, while the acid metabolites could still efficiently
inhibit translation in extracts, they had reduced activity
when tested on intact cells. These results suggest that
modification of silvestrol at C2 to prevent acid formation
could increase cellular uptake and improve pharmacoki-
netics. With this new information, further C2 analogues
were synthesized and their biological activity examined.
Scheme 1. Synthesis of Methyl 6 and Propargyl Amide 7
Derivatives of Episilvestrol (2)
Figure 3. Biotinylated episilvestrol (9) inhibits protein synthesis.
(A) In vitro translations in rabbit reticulocyte lysate. Data
are represented as FF/Ren relative to vehicle (DMSO) controls
(n = 4). (B) Viability of MEFs after 72 h treatment with
indicated compounds (n = 3). All graphs depict the mean and
standard error of the mean (SEM).
Two secondary amide analogues of 2 were prepared by
coupling 5 with the corresponding amine (Scheme 1). The
coupling of 5 with methylamine or propargyl amine using
DCC and DMAP afforded the methyl amide 6 and the
propargyl amide 7 in reasonable yields over two steps.
Incorporation of a terminal alkyne in amide 7 enabled
biotinylation of 2 via a Huisgen 1,3-dipolar cycloaddition
reaction (Scheme 2).16 The biotin linker 817 was synthe-
sized following a modified procedure.18 Cycloaddition in
the presence of tris(benzyltriazolylmethyl)amine (TBTA)
Amides 6 and 7 had similar activity as 2 for the inhibition
of cap-dependent translation in vitro (Figure 3A) and
also inhibited proliferation and/or the survival of fibroblasts
(Figure 3B). Notably, while all have modifications of the C2
methyl ester, 6 and 7 were more potent than 5 (Figure 2C
and 3B). Though the in vitro activity of biotinylated
episilvestrol (9) was reduced (Figure 3A), we considered it
sufficient to continue with a streptavidin pulldown assay.
(16) Rostovtsev, V. V.; Green, L. G.; Fokin, V. V.; Sharpless, K. B.
Angew. Chem., Int. Ed. 2002, 41, 2596.
(17) Stewart, S. G.; Braun, C. J.; Polomska, M. E.; Karimi, M.;
Abraham, L. J.; Stubbs, K. A. Org. Biomol. Chem. 2010, 8, 4059.
(18) See Supporting Information for details.
(19) Donnelly, P. S.; Zanatta, S. D.; Zammit, S. C.; White, J. M.;
Williams, S. J. Chem. Commun. 2008, 21, 2459.
1408
Org. Lett., Vol. 15, No. 6, 2013

Using 9 and free biotin as a negative control, we
performed pulldown experiments to determine the direct
target(s) of 2. Pulldownsamples werewashedinlysis buffer
and eluted with successive treatments of 50 μM 1, 25 mM
glycine pH 2.5 at rt, and finally in sample buffer containing
lithium dodecyl sulfate (LDS) at 95 °C. Elutions with
silvestrol, low pH, or detergent yielded eIF4AI/II from
samples containing 9, but not free biotin (Figure 4A).
Antibodies used in Western blots were specific for their
respectiveeIF4A isoform, asdeterminedwithrecombinant
protein (Supporting Information Figure 1). In contrast
to eIF4AI/II, eIF4AIII (DDX48);another DEAD-box
helicase with 65% identity to eIF4AI;was not eluted.
DDX3 and DHX29, which are both DEAD/H-box heli-
cases reported to function in protein synthesis,20ꢀ22 were
also not detected in any elutions (Figure 4A). As expected,
the other two components of the eIF4F complex (eIF4G
and eIF4E) were no longer bound to eIF4A in the presence
of 9.13 Therefore, biotinylated episilvestrol and silvestrol
are specific for eIF4AI and II with respect to the DHX
helicases known to be involved in translation.
Even though the most prominent band visible after Nu-
PAGE separation and silver staining was consistent with the
size of eIF4AI/II (Figure 4B), we extended the study by
analyzing the entire silvestrol and glycine elutions by mass
spectrometry to determine if less prominent proteins would
be detected. Strikingly, only peptides corresponding to eIF4A
were found in samples containing 9 but not the negative
control. Because of the high amino acid identity between
eIF4AI and eIF4AII, 50% of the peptides were not able to be
distinguished as coming from either eIF4AI or eIF4AII. No
eIF4AII-specific sequences were detected by mass spec-
trometry, potentially because eIF4AII is a much less abundant
protein.23 Mass spectrometry showed that sequence coverage
of eIF4AI was 35% and 75% for the silvestrol and glycine
elutions, respectively (Supporting Information Figure 2). Of
note, Prohibitin 1 and 2, which are the reported targets of
another member of the rocaglamide family that acts up-
stream of the protein synthesis machinery,24 were also absent.
In conclusion, we have demonstrated for the first time
the direct interaction of silvestrol (1) and biotinylated
episilvestrol (9) with eIF4AI and eIF4AII. This study
was undertaken to determine unequivocally the protein
targets of the anticancer compounds 1 and episilvestrol (2).
Using an unbiased approach, we demonstrated that 1 and
2 are some of the most specific eIF4A-targeting translation
inhibitors to date, validating these compounds as powerful
Figure 4. eIF4AI/II are specifically pulled down by 9. (A) Strepta-
vidin sepharose pulldowns with 9 or free biotin control were washed
and eluted sequentially with 50 μM silvestrol, 25 mM glycine pH 2.5,
and LDS loading buffer. Samples were separated by NuPAGE,
transferred to polyvinylidene difluoride (PVDF), and probed with
antibodies raised against the indicated proteins. (B) Pulldown
samples visualized by silver stain after NuPAGE separation.
chemical tools to elucidate the role of eIF4AI/II in both cell
and cancer biology. Importantly, the exclusive specificity of
these compounds is a highly desirable trait in drug devel-
opment and one that is very rarely achieved with enzymatic
inhibitors.25 This insight furthers our belief that 1 and 2
warrant additional investigation as therapeutic candidates.
Acknowledgment. Financial support for this work was
provided by a University of Melbourne Research Support
Scheme Grant and NHMRC Program Grants (461221 and
1016701). L.M.L. is supported by an NHMRC Early Career
Fellowship and a CIHR Postdoctoral Fellowship. J.M.C.
thanks CRC for financial support. We are grateful to Prof.
D. Vaux (WEHI) for helpful discussions and reagents.
J.M.C. and L.M.L. contributed equally to this work.
(20) Soto-Rifo, R.; Rubilar, P. S.; Limousin, T.; de Breyne, S.;
Decimo, D.; Ohlmann, T. EMBO J. 2012, 31, 3745.
(21) Pisareva, V. P.; Pisarev, A. V.; Komar, A. A.; Hellen, C. U. T.;
Pestova, T. V. Cell 2008, 135, 1237.
(22) Parsyan, A.; Shahbazian, D.; Martineau, Y.; Petroulakis, E.;
Alain, T.; Larsson, O.; Mathonnet, G.; Tettweiler, G.; Hellen, C. U.;
Pestova, T. V.; Svitkin, Y. V.; Sonenberg, N. Proc. Natl. Acad. Sci.
U.S.A. 2009, 106, 22217.
(23) Nielsen, P. J.; Trachsel, H. EMBO J. 1988, 7, 2097.
(24) Polier, G.; Neumann, J.; Thuaud, F.; Ribeiro, N.; Gelhaus, C.;
Supporting Information Available. Experimental details
as well as characterization data and copies of the NMR
spectra of all new compounds. This material is available
free of charge via the Internet at http://pubs.acs.org.
€
Schmidt, H.; Giaisi, M.; Kohler, R.; Muller, W. W.; Proksch, P.; Leippe,
M.; Janssen, O.; Desaubry, L.; Krammer, P. H.; Li-Weber, M. Chem.
€
ꢀ
Biol. 2012, 19, 1093.
(25) Davies, S. P.; Reddy, H.; Caivano, M.; Cohen, P. Biochem. J.
2000, 351, 95.
The authors declare no competing financial interest.
Org. Lett., Vol. 15, No. 6, 2013
1409