Asymmetric total synthesis and formal total synthesis of the antitumor sesquiterpenoid (+)-eremantholide A

Article abstract of DOI:10.1021/ol0700862

Figure presented A new asymmetric total synthesis of (+)-eremantholide A is reported in which a Hoveyda-Grubbs ring-closing metathesis (RCM) reaction is used to assemble the nine-membered oxonin ring, and an enolate alkylation between the 3(2H)-furanone 2

Full text of DOI:10.1021/ol0700862

ORGANIC
LETTERS
2007
Vol. 9, No. 7
1267-1270
Asymmetric Total Synthesis and Formal
Total Synthesis of the Antitumor
Sesquiterpenoid (+)-Eremantholide A
Yi Li and Karl J. Hale*
The UCL Center for Chemical Genomics, The Christopher Ingold Laboratories,
The Chemistry Department, UniVersity College London, 20 Gordon Street,
London WC1H 0AJ, UK
k.j.hale@ucl.ac.uk
Received January 11, 2007
ABSTRACT
A new asymmetric total synthesis of (
+
)-eremantholide A is reported in which a Hoveyda−Grubbs ring-closing metathesis (RCM) reaction is
used to assemble the nine-membered oxonin ring, and an enolate alkylation between the 3(2H)-furanone 2 and O-triflate 3 is exploited for
C(9) C(10) bond construction. An Evans asymmetric aldol reaction and a Sharpless asymmetric epoxidation served to stereoselectively install
−
the C(6), C(7), and C(8) stereocenters of the target structure.
(+)-Eremantholide A is a structurally novel tumor-inhibiting
germacranolide obtained from Eremanthus elaeagnus, a rare
woody composite found in Brazil.1 Ever since its initial
discovery in 1975 by Le Quesne and Brennan, (+)-
eremantholide A has elicited significant synthetic interest2
on account of its pronounced antitumor effects against human
KB nasopharyngeal carcinoma1,3 and its architecturally
intriguing structure which provides a significant challenge
for asymmetric total synthesis.
Notwithstanding (+)-eremantholide A first being isolated
more than 30 years ago, the signal transduction pathways
and genes that it controls (within human cancer cells) remain
undefined,4 and its potential for inhibiting the growth of
xenografted solid human tumors in mice has yet to be
demonstrated. In connection with our own proposed studies
in this area and to facilitate future analogue development,
we recently opted to devise a new asymmetric total synthesis
of (+)-eremantholide A and, herein, now report success in
this venture with a route that exploits high-dilution ring-
closing metathesis (RCM) to close the highly strained nine-
membered oxonin ring present within the target.
(1) (+)-Eremantholide A isolation, structure determination, and initial
antitumor evaluation: (a) Raffauf, R. F.; Huang, P.-K. C.; Le Quesne, P.
W.; Levery, S. B.; Brennan, T. F. J. Am. Chem. Soc. 1975, 97, 6884. (b)
Le Quesne, P. W.; Levery, S. B.; Menachery, M. D.; Brennan, T. F.; Raffauf,
R. F. J. Chem. Soc., Perkin Trans. 1 1978, 1572.
(2) For past total syntheses of (+)-eremantholide A, see: (a) Boeckman,
R. K., Jr.; Yoon, S. K.; Heckendorn, D. K. J. Am. Chem. Soc. 1991,
113, 9682. (b) Takao, K.; Ochiai, H.; Yoshida, K.; Hashizuka,
T.; Koshimura, H.; Tadano, K.; Ogawa, S. J. Org. Chem. 1995, 60,
8179.
Our retrosynthetic planning for (+)-eremantholide A
(Scheme 1) disassembled the molecule across its C(4)-C(5)
(4) Although 15-acetoxy-eremantholide B inhibits NF-κB at 5 µM in
Jurkat T cells, (+)-eremantholide A does not inhibit this target at the
concentrations needed to inhibit cancer cell growth. See: Ru¨ngeler, P.;
Castro, V.; Mora, G.; Go¨ren, N.; Vichnewski, W.; Pahl, H. L.; Merfort, I.;
Schmidt, T. J. Bioorg. Med. Chem. 1999, 7, 2343.
(3) Hladon, B.; Twardowski, T. Pol. J. Pharmacol. Pharm. 1979, 31,
35.
10.1021/ol0700862 CCC: $37.00
© 2007 American Chemical Society
Published on Web 02/28/2007

of 11 would be installed through an Evans asymmetric aldol
reaction8 between 13 and 149 which would yield 12.
Theoretically, the conversion of 12 into 11 would require
functional group interchange at O(8), O(9), and C(6),
oxidative cleavage of the C(12) alkene, reduction of the
product aldehyde, and introduction of a TBDPS protecting
group on the newly elaborated oxygen atom. Thereafter,
Wittig chain extension of a suitable C(6) aldehyde and enoate
1,2-reduction would furnish the desired allylic alcohol 11.
The synthesis of lactone 7 (Scheme 2) commenced with
a low-temperature Evans aldol reaction8 between oxazoli-
dinone 13 and aldehyde 14,9 which delivered a single syn-
aldol adduct 12 in 81% yield. Reductive removal of the chiral
auxiliary from 12 with LiBH4 in MeOH/THF furnished the
diol 15 which underwent O-desilylation with KF in aqueous
DMF to provide the triol 16 in 71% combined yield.
Although n-Bu4NF also proved effective at cleaving the
TBDPS group from 15, the problems it caused in the
purification of 16 made KF the reagent of choice for this
operation. Next, we selectively protected the terminal 1,2-
diol grouping in 16 to allow the C(6) primary alcohol to be
selectively manipulated. For this, O-cyclohexylidenation with
cyclohexanone and p-TsOH10 proved optimal; the desired
alcohol 17 was formed in 75% yield without contamination
of the crude reaction mixture with any of the 1,3-dioxane.
Following O-benzoylation of 17 with benzoyl chloride,
the double bond of 18 was ozonolytically cleaved and the
ozonides were reduced with NaBH4; alcohol 19 was isolated
in 78% overall yield. The latter was then O-silylated to obtain
20, and its O-benzoate grouping was removed with K2CO3/
MeOH (without causing O-silyl migration). The desired
alcohol 21 was now in a format appropriate for further
oxidation and Wittig homologation.
Scheme 1. Our Retrosynthetic Analysis of (+)-Eremantholide
A
A range of methods were investigated for converting
alcohol 21 into aldehyde 23, including Swern oxidation
(COCl2/Me2SO) which, rather surprisingly, caused cata-
strophic loss of the cyclohexylidene grouping. Most unusu-
ally, the catalytic TPAP/NMO oxidation11 protocol also
proved problematic, causing significant epimerization (ca.
10-15%) of the C(7) stereocenter, notwithstanding it
performing efficiently in the chemical conversion of 21 into
the aldehyde (95% yield). As a result, we evaluated alterna-
tive oxidants for accomplishing this transformation and,
after much effort, eventually discovered that catalytic
TEMPO and excess [bis(acetoxy)iodo]benzene (BAIB)12 in
CH2Cl2 could successfully produce the desired aldehyde 23
without any epimerization at C(7). Aldehyde 23 was
thereafter condensed with carbomethoxymethylenetriph-
enylphosphorane in CH2Cl2 to give the (E)-enoate 24 as
essentially a single geometrical isomer in 94% yield.
Compound 24 was readily reduced to 11 with DIBAL-H
in PhMe/CH2Cl2 at -78 °C over a 1.5 h period. We found
trisubstituted olefin and opted to construct its synthetically
challenging nine-membered oxonin ring via RCM. For C(10)
quaternary stereocenter installation, we proposed to react the
O-triflate 3 or iodide 5 with the “chiral” lithium enolate
obtained from 25 by kinetic deprotonation with lithium bis-
[(S)-R-methylbenzylamide] (or its (R)-optical antipode),6 in
the hope that such a device would permit a diastereoselective
union between the AB- and D-ring fragments. Intermediates
3 and 5 would themselves be prepared from alcohol 6, which
potentially would be fashioned from the lactone 7 by enolate
C-acylation/C-methylation and Fischer glycosidation with
MeOH.
Our preferred strategy for securing the desired butyrolac-
tone 7 would entail us chemoselectively oxidizing the diol
9 and assembling the latter from iodo-epoxide 10 by
reductive elimination. Compound 10 would be fashioned
from the allylic alcohol 11 by Sharpless asymmetric epoxi-
dation (AE)7 and iodination. The C(7) and C(8) stereocenters
(5) Chimichi, S.; Boccalini, M.; Cosimelli, B.; Dall’Acqua, F.; Viola,
G. Tetrahedron 2003, 59, 5215.
(6) (a) Whitesell, J. K.; Felman, S. W. J. Org. Chem. 1980, 45, 755. (b)
Marshall, J. A.; Lebreton, J. J. Am. Chem. Soc. 1988, 110, 2925. (c) Cain,
C. M.; Cousins, R. P. C.; Coumbarides, G.; Simpkins, N. S. Tetrahedron
1990, 46, 523.
(8) Evans,D.A.;Bartroli,J.;Shih,T.L. J.Am.Chem.Soc. 1981,103,2127.
(9) Hale, K. J.; Frigerio, M.; Manaviazar, S. Org. Lett. 2001, 3, 3791.
(10) Hale, K. J.; Hummersone, M. G.; Bhatia, G. S. Org. Lett. 2000, 2,
2198.
(11) Ley, S. V.; Norman, J.; Griffith, W. P.; Marsden, S. P. Synthesis
1994, 639.
(7) Gao, Y.; Hanson, R. M.; Klunder, J. M.; Ko, S. Y.; Masamune, H.;
Sharpless, K. B. J. Am. Chem. Soc. 1987, 109, 5765.
(12) De Mico, A.; Margarita, R.; Parlanti, L.; Vescovi, A.; Piancatelli,
G. J. Org. Chem. 1997, 62, 6974.
1268
Org. Lett., Vol. 9, No. 7, 2007

alcohol 25. Again, the latter was produced with high
stereocontrol, as essentially a single diastereoisomer. It was
immediately iodinated at room temperature with Ph3P, I2,
and imidazole13 in THF and MeCN. Significantly, the
inclusion of MeCN proved critical to the success of this
reaction, enabling the desired iodo-epoxide 10 to be formed
efficiently in good yield (97%). Treatment of 10 with Zn
dust in MeOH then effected the desired reductive elimination
to provide the allylic alcohol 26 in 91% yield. O-Desilylation
with n-Bu4NF subsequently provided 9.
Scheme 2. Route Developed to Lactone 7
Although the direct conversion of epoxy alcohol 25 into
26 was examined, using Yadav’s titanocene chloride (Cp2-
TiCl/THF) reduction procedure,14 sadly, we were unable to
detect any of 26 in the reaction mixture.
The time had now come for us to attempt the all-important
chemoselective oxidation of 9 to the chiral butyrolactone 7.
After abortively screening a number of normally quite selec-
tive oxidants for generating γ-butyrolactones from 1,4-diols,
including TPAP/NMO,15 we eventually found that the cata-
lytic TEMPO/BAIB oxidation system12,16 could very cleanly
deliver the desired lactone 7 in 94% yield. We were now in
an excellent position to complete the synthesis of alcohol 6
(Scheme 3). Initially, we successfully C-acylated 7 with 8
and thereafter C-methylated the resulting â-keto ester with
MeI (see Scheme 1) but observed that the subsequent Fischer
glycosidation with MeOH/(MeO)3CH/PPTS did not yield any
of the desired methyl glycoside 6. Instead, an internal acetal
was formed involving the C(9) oxygen (for the precise
structure of this undesired product, see structure 45 on pages
163-170 of our Supporting Information).
A decision was therefore made to remove the cyclohexy-
lidene grouping from 7 with 1,3-propanedithiol and BF3-
Et2O10,17 (Scheme 3), and selective O-pivaloylation was
subsequently attempted on the C(9) OH of diol 27. Following
O-silylation of the product monopivaloate with Et3SiOTf,
the desired C-acylation of 28 proceeded admirably using
LiN(SiMe3)2 as the base and 8 as the electrophile. The enolate
derived from 29 was thereafter reacted with MeI to give 30.
Exposure of 30 to PPTS/MeOH/(MeO)3CH now cleaved the
TES group from O(8) and brought about the desired Fischer
glycosidation at C(16) to give 31. The Piv group was
detached from 31 with NaOMe/MeOH, and alcohol 6 was
converted into iodide 5 (see Scheme 1 for this structure) by
successive O-mesylation and displacement with NaI/butanone
at reflux. Unfortunately, iodide 5 proved unreactive toward
all the metal enolates that we prepared from 2. We therefore
conducted all subsequent alkylation chemistry with the
O-triflate 3, analogously to Tadano et al.2b
Initially, we used a “chiral” lithium enolate generated from
2 with lithium bis[(S)-R-methylbenzylamide]6 in PhMe at
(13) Garegg, P. J.; Samuelsson, B. J. Chem. Soc., Perkin Trans. 1 1980,
2866.
(14) Yadav, J. S.; Shekharam, T.; Gadgil, V. R. Chem. Commun. 1990, 843.
(15) Shi, H.; Liu, H.; Bloch, R.; Mandville, G. Tetrahedron: Asymmetry
2002, 13, 1423.
(16) Hansen, T. M.; Florence, G. J.; Lugo-Mas, P.; Chen, J.; Abrams, J.
N.; Forsyth, C. J. Tetrahedron Lett. 2003, 44, 57.
(17) Important Note: Under these conditions, we find that 27 is always
formed alongside approximately 14-17% of a lactone migration product
that is difficult to remove at this stage; the latter impurity arises from attack
of the C(8)-OH in 27 on its own lactone carbonyl.
it necessary to keep the temperature of this reduction at -78
°C throughout, otherwise unwanted cleavage of the primary
TBDPS group began to occur, even at temperatures as low
as -40 °C. However, by adhering to the -78 °C procedure
that we have outlined (in the Supporting Information),
excellent results (86% yield) were generally obtained.
With the desired allylic alcohol 11 in hand, we next
investigated the Sharpless AE7 needed to secure 2,3-epoxy
Org. Lett., Vol. 9, No. 7, 2007
1269

Scheme 3. Final Stages of Our Total Synthesis of
(+)-Eremantholide A
Scheme 4. Intersection with Tadano’s Alcohol 39
behavior, we conducted the union of 2 and 3 under “achiral”
enolization conditions. The best result came when KHMDS
was used as the base in PhMe; it produced a 1:1 mixture of
32/33. However, instead of separating the two individual
isomers at this stage, we generally found it more convenient
to deprotonate the mixture of 32 and 33 at C(4) and trap the
resulting dienolates with gaseous formaldehyde to obtain 34
as a mixture of diastereomers. Exposure of 34 to MsCl/Et3N,
followed by DBU, very cleanly induced elimination of the
intermediary O-mesylate esters to give 1 and 35 in 69% yield;
the two trienes were separated by chromatography. Intrigu-
ingly, triene 1 cyclized in respectable yield (54%) to 42b when
subjected to RCM with the Hoveyda-Grubbs catalyst 36,18
whereas compound 35 showed no tendency to cyclize;
instead, it preferentially cross-metathesized.
The final step of our total synthesis was the acid-induced
hydrolysis of known2b methyl glycoside 4 under the condi-
tions previously reported by Tadano et al.;2b the end result
was (+)-eremantholide A, formed in 82% yield.
In furtherance of our work, we also successfully converted
lactone 31 into Tadano’s advanced intermediate 392b for (+)-
eremantholide A by the three-step sequence shown in Scheme
4; this now shortens the route needed to access Tadano’s
AB-alcohol2b by eight steps and provides 39 by a pathway
that is 26 steps overall from a commercially available chiral
auxiliary. Our new pathway to 39 thus constitutes a formal
total synthesis of (+)-eremantholide A.
Future work in our laboratory will attempt to use our new
route to construct novel eremantholide analogues and pho-
toaffinity labeled probes to help us elucidate the mechanism-
(s) of antitumor action of (+)-eremantholide A.4
Acknowledgment. We thank the EPSRC (GR/S81582/
01) for generous financial support of this program.
Supporting Information Available: Full experimental
procedures and detailed spectral data of all key compounds
are reported. Copies of 500 MHz 1H and 125 MHz 13C NMR
spectra are provided along with IR spectra and HRMS data.
This material is available free of charge via the Internet at
http://pubs.acs.org.
low temperature (-78 to 0 °C) to effect the diastereoselective
alkylation of 3 and observed that a 2.48:1 mixture of 32/33
was formed, in which the undesired isomer 32 predominated.
Somewhat surprisingly, a similar ratio of products (2.33:1
of 32/33) was formed when lithium bis[(R)-1-R-methylben-
zylamide]6 was used for the kinetic deprotonation of 2 in
PhMe (see Supporting Information)! In light of this unusual
OL0700862
(18) (a) Kingsbury, J. S.; Harrity, J. P. A.; Bonitatebus, P. J., Jr.; Hoveyda,
A. H. J. Am. Chem. Soc. 1999, 121, 791. (b) Grubbs, R. H.; Chang, S.
Tetrahedron 1998, 54, 4413. (c) For the successful use of 36 to construct
the highly strained ring system of ingenol, see: Nickel, A.; Maruyama, T.;
Tang, H.; Murphy, P. D.; Greene, B.; Yusuff, N.; Wood, J. L. J. Am. Chem.
Soc. 2004, 126, 16300.
1270
Org. Lett., Vol. 9, No. 7, 2007