thus preventing sequestering of Ca2+ from the cytosol,
allowing a massive secondary influx of extracellular Ca2+
leading to apoptosis.7
Scheme 2. Selenide Introduction and Enone Installation
There has been a recent discovery that a prodrug, compris-
ing Tg linked through C8 to a peptide specifically cleaved
by a prostate-specific antigen (a serine protease), may allow
site-specific targeting of Tg to prostate cancer cells.8 The
notion of thapsigargin as a real and viable prostate cancer
treatment has therefore greatly elevated the need for its
chemical synthesis.
The synthetic approach established during our syntheses
of the related natural products,2c,d trilobolide 2, nortrilobolide
3, and thapsivillosin F 4, recognizes that the bulk of the
structural variation in the family exists in the five-ring
cyclopentene unit. Thus, our approach to the synthesis of
the thapsigargins centers upon a late-stage divergent strategy
to this motif, via a common intermediate (Scheme 1). It was
Scheme 1. Late-Stage Retrosynthetic Analysis of
Thapsigargin
several reasons. First, model studies had suggested that a
bulky group was necessary to direct enolization to the C4
position rather than C2; this would be pivotal to double bond
installation. Second, it was deemed important to have a
chelating group to deliver a hydride source to the ketone at
C3 from the exo face of the five-membered ring, circum-
venting steric encumbrance at C2. The increased lability of
SEM over a MOM ether was also pertinent due to our
inability to remove MOM from the C2-O position in various
intermediates during previous studies.
Enolization at C4 initially involved treatment of 9 with
NaHMDS and attempted trapping with TMSCl. The resulting
silyl enol ether, however, was highly unstable, so a protocol
employing in situ reaction of the lithium enolate proved
preferable. Treatment of 9 with 5 equiv of LiHMDS under
strict temperature control allowed conversion to the desired
lithium enolate without enolization of the C2 position.
Reaction at -90 to -78 °C was optimal for approach of the
electrophile (phenyl selenyl chloride) from the less-hindered
(exo) face of the 5-7 fused bicycle. This resulted in a pleasing
80:20 mixture of selenide epimers 11 and 12.
envisaged that this approach could allow access to all 17
natural compounds in the series as well as to late-stage
analogues for structure-activity relationship analysis.2c
Several problems remained, however, before synthesis of
thapsigargin 1 could be accomplished. First, introduction of
the internal double bond at C4-C5 had only previously been
achieved by an unusual catalytic selenium-mediated reaction
which led to concomitant deoxygenation at C2.2c,d Second,
establishing the C3 stereochemistry via reduction proved
challenging with the presence of bulky functionality at C2.
Finally, a highly selective esterification-deprotection routine
would have to be implemented to reach the final natural
product.
The synthesis began with the installation of the SEM ether
at C2-O of our previously reported intermediate 82c,d in 93%
yield (Scheme 2). This protecting group was selected for
(7) (a) Thastrup, O.; Cullen, P. J.; Drøbak, B. K.; Hanley, M.; Dawson,
A. P. Proc. Natl. Acad. Sci. U.S.A. 1990, 87, 2466. (b) Futami, T.;
Miyagashi, M.; Taira, K. J. Biol. Chem. 2005, 280, 826.
(8) (a) Jakobsen, C. M.; Denmeade, S. R.; Isaacs, J. T.; Gady, A.; Olsen,
C. E.; Christensen, S. B. J. Med. Chem. 2001, 44, 4696. (b) Christensen, S.
B.; Andersen, A.; Kromann, H.; Trieman, M.; Tombal, B.; Denmeade, S.;
Isaacs, J. T. Bioorg. Med. Chem. 1999, 7, 1273. (c) Denmeade, S. R.;
Jakobsen, S.; Janssen, S.; Khan, S. R.; Garrett, E. S.; Lilja, H.; Christensen,
S. B.; Isaacs, J. T. J. Natl. Cancer Inst. 2003, 95, 990.
A selenide oxidation-elimination process was then imple-
mented to regioselectively install the internal double bond.
Statistically, however, removal of one of the three methyl
protons by the intermediate selenoxide (leading to the
664
Org. Lett., Vol. 9, No. 4, 2007