1,6-asymmetric induction in boron-mediated aldol reactions: application to a practical total synthesis of (+)-discodermolide.

Article abstract of DOI:10.1021/ol0270780

By relying solely on substrate-based stereocontrol, a practical total synthesis of the microtubule-stabilizing anticancer agent (+)-discodermolide has been realized. This exploits a novel aldol bond construction with 1,6-stereoinduction from the boron eno

Full text of DOI:10.1021/ol0270780

ORGANIC
LETTERS
2003
Vol. 5, No. 1
35-38
1,6-Asymmetric Induction in
Boron-Mediated Aldol Reactions:
Application to a Practical Total
Synthesis of (+)-Discodermolide
Ian Paterson,* Oscar Delgado, Gordon J. Florence, Isabelle Lyothier,
Jeremy P. Scott, and Natascha Sereinig
UniVersity Chemical Laboratory, Lensfield Road, Cambridge CB2 1EW, U.K.
ip100@cus.cam.ac.uk
Received October 11, 2002
ABSTRACT
By relying solely on substrate-based stereocontrol, a practical total synthesis of the microtubule-stabilizing anticancer agent (+)-discodermolide
has been realized. This exploits a novel aldol bond construction with 1,6-stereoinduction from the boron enolate of (Z)-enone 3 in addition to
aldehyde 2. The 1,3-diol 7 is employed as a common building block for the C −C , C −C , and C −C24 subunits.
1
5
9
16
17
(+)-Discodermolide (1) is a unique cytotoxic polyketide
isolated from the Caribbean deep-sea sponge Discodermia
dissoluta.1,2 Biological screening demonstrated cell cycle
arrest at the G2/M phase in a variety of human and murine
cancer cell lines.3 Discodermolide is now recognized as a
member of a group of antimitotic agents (including Taxol,
epothilones, eleutherobin, and laulimalide) known to act by
microtubule stabilization,4 which makes it a highly promising
candidate as a new chemotherapeutic agent for the treatment
of Taxol-resistant breast, ovarian, and colon cancer and other
multi-drug-resistant cancers.
syntheses,5 including one by our group,6 have been com-
pleted. Despite these efforts, there continues to be a pressing
demand for developing a more practical and scaleable route
to enable clinical development. With our existing total
synthesis in hand,6 we sought further refinements by inter
alia reducing the total number of steps, eliminating the use
of any chiral reagents or auxiliaries, and exploring a novel
endgame strategy. By exploiting remote 1,6-asymmetric
induction in boron-mediated aldol reactions of (Z)-enones,
we now report a highly convergent and practical total
At present, total synthesis provides the only viable means
of accessing useful quantities of discodermolide, and several
(5) (a) Nerenberg, J. B.; Hung, D. T.; Schreiber, S. L. J. Am. Chem.
Soc. 1996, 118, 11054. (b) Nerenberg, J. B.; Hung, D. T.; Somers, P. K.;
Schreiber, S. L. J. Am. Chem. Soc. 1993, 115, 12621. (c) Smith, A. B., III;
Beauchamp, T. J.; LaMarche, M. J.; Kaufman, M. D.; Qiu, Y. P.; Arimoto,
H.; Jones, D. R.; Kobayashi, K. J. Am. Chem. Soc. 2000, 122, 8654. (d)
Smith, A. B., III; Kaufman, M. D.; Beauchamp, T. J.; LaMarche, M. J.;
Arimoto, H. Org. Lett. 1999, 1, 1823. (e) Smith, A. B., III; Qiu, Y. P.;
Jones, D. R.; Kobayashi, K. J. Am. Chem. Soc. 1995, 117, 12011. (f) Harried,
S. S.; Yang, G.; Strawn, M. A.; Myles, D. C. J. Org. Chem. 1997, 62,
6098. (g) Marshall, J. A.; Johns, B. A. J. Org. Chem. 1998, 63, 7885. (h)
Halstead, D. P. Ph.D. Thesis, Harvard University, Cambridge, 1998.
(6) (a) Paterson, I.; Florence, G. J.; Gerlach, K.; Scott, J. P.; Sereinig,
N. J. Am. Chem. Soc. 2001, 123, 9535. (b) Paterson, I.; Florence, G. J.;
Gerlach, K.; Scott, J. P. Angew. Chem., Int. Ed. 2000, 39, 377. (c) Paterson,
I.; Florence, G. J. Tetrahedron Lett. 2000, 41, 6935.
(1) Gunasekera, S. P.; Gunasekera, M.; Longley, R. E.; Schulte, G. K.
J. Org. Chem. 1990, 55, 4912. Additions and corrections: J. Org. Chem.
1991, 56, 1346.
(2) Gunasekera, S. P.; Pomponi, S. A.; Longley, R. E. Patent No.
US5840750: US, Nov 24, 1998.
(3) ter Haar, E.; Kowalski, R. J.; Hamel, E.; Lin, C. M.; Longley, R. E.;
Gunasekera, S. P.; Rosenkranz, H. S.; Day, B. W. Biochemistry 1996, 35,
243.
(4) For recent reviews, see: (a) Altmann, K. H. Curr. Opin. Chem. Biol.
2001, 5, 424. (b) He, L. F.; Orr, G. A.; Horwitz, S. B. Drug DiscoVery
Today 2001, 6, 1153. (c) Stachel, S. J.; Biswas, K.; Danishefsky, S. J. Curr.
Pharm. Design 2001, 7, 1277.
10.1021/ol0270780 CCC: $25.00 © 2003 American Chemical Society
Published on Web 12/13/2002

Scheme 1
synthesis of (+)-discodermolide that depends solely on
substrate-based stereocontrol and makes use of a common
building block to access the three main subunits.
which is applied here to addition to formaldehyde. Thus (E)-
enolization of 9 with c-Hex2BCl/Et3N and addition of an
ethereal solution of formaldehyde8 (-78 °C) gave aldol
adduct 10 (96%, 95:5 dr) with the required 1,3-anti-
configured methyl groups.
Reduction of ketone 10 with NaBH(OAc)39 then gave the
1,3-diol 7 (90:10 dr), which could be conveniently isolated
in stereochemically pure form in 62% yield from 9 by
crystallization (mp 71-72 °C). The relative configuration
of 7 was confirmed by single-crystal X-ray analysis. Starting
from ester (S)-8, this efficient five-step sequence can be
performed routinely on a multigram scale without any
chromatography.
Our synthetic plan is outlined in Scheme 1. An aldol
disconnection across C5-C6 reveals aldehyde 2 and (Z)-
enone 3 having a γ-stereocenter at C10. We speculated that
such a remote center might induce Si-face attack (in the anti-
Felkin sense) on the aldehyde component to set up the
required (5S)-configuration in discodermolide. Further dis-
connection of 3 leads back to alcohol 4,6 an advanced
intermediate in our previous synthesis assembled by the aldol
coupling of ester 5 and aldehyde 6. Recognizing the repeating
anti,syn-stereotriad present in 2, 5, and 6, we sought to access
these three key subunits from 1,3-diol 7 as a common
building block.
The synthesis of the C1-C5 aldehyde 2 (Scheme 3) from
In contrast to the chiral reagent-based approaches adopted
by other groups,5 we chose to use substrate control to prepare
stereotriad 7 from Roche ester (S)-8 (Scheme 2). This
Scheme 3
Scheme 2
the common building block 7 proceeded in 76% overall yield.
It began with a selective TEMPO/BAIB oxidation10 of the
primary hydroxyl to give â-hydroxy aldehyde 11, which was
(7) (a) Paterson, I.; Norcross, R. D.; Ward, R. A.; Romea, P.; Lister, M.
A. J. Am. Chem. Soc. 1994, 116, 11287. (b) Paterson, I.; Goodman, J. M.;
Isaka, M. Tetrahedron Lett. 1989, 30, 7121. (c) Paterson, I.; Arnott, E. A.
Tetrahedron Lett. 1998, 39, 7185.
(8) Rodriguez, L.; Lu, N.; Yang, N. L. Synlett 1990, 227.
(9) Alternative reducing systems explored gave lower or overturned
selectivities. Evans, D. A.; Chapman. K. T.; Carreira, E. M. J. Am. Chem.
Soc. 1988, 110, 3560.
(10) De Mico, A.; Margarita, R.; Parlanti, L.; Vescovi, A.; Piancatelli,
G. J. Org. Chem. 1997, 62, 6974.
exploits the highly diastereoselective aldol chemistry of the
(E)-enol dicyclohexylborinate derivative of ketone (S)-9,6a,7
36
Org. Lett., Vol. 5, No. 1, 2003

immediately oxidized to the acid (NaClO2)11 and then
converted into the methyl ester (MeI, K2CO3, DMF; or
TMSCHN2). Protection of the secondary hydroxyl as the TBS
ether 12, PMB deprotection (DDQ) followed by Swern
oxidation, then gave the C1-C5 subunit 2.
The preparation of the C17-C24 aldehyde 6 from the
common 1,3-diol 7 required sequential protecting group
manipulations (Scheme 4). Selective TBS protection of the
alcohol was converted into iodide 18 (I2, imidazole, PPh3)
and treated with the lithium enolate of 2,6-dimethyl-4-
methoxyphenyl acetate (LiHMDS, THF) to produce aryl ester
5 (65% over three steps).16
We next sought to explore the potential influence of the
γ-stereogenic center at C10 of methyl ketone 3 (Scheme 1)
in the proposed C5-C6 aldol coupling to access discoder-
molide. Our previous studies of related reactions of γ-chiral
(Z)-enals6c led us to speculate that such a remote center would
indeed act as a stereodeterminant in boron-mediated aldol
additions. Model studies with (Z)-enone substrate 19 indi-
cated a useful level of remote 1,6-stereoinduction was
achievable (Scheme 6). Enolization of 19 (c-Hex2BCl, Et3N)
Scheme 4
Scheme 6
primary hydroxyl and treatment with anhydrous DDQ12
afforded PMP acetal 13 (82%). Regioselective acetal opening
with DIBAL,13 followed by Dess-Martin oxidation of the
resultant primary hydroxyl and installation of the (Z)-diene,
following our previously established protocol,14 gave the
known intermediate 16,6a which was routinely converted into
aldehyde 6.
The synthesis of the remaining C9-C16 subunit 5 started
out from the already prepared â-hydroxy aldehyde 11
(Scheme 5), where Still-Gennari HWE olefination15a and
Scheme 5
and addition of i-PrCHO (-78 °C) gave 81:19 dr in favor
of adduct 20.17 A similar addition to R-chiral aldehyde 21
gave 22 with comparable selectivity (82:18 dr). For chiral
aldehyde 23 (from oxidation of 14), the adduct 24 was
isolated with >95:5 dr (84%), indicative of matched double
(11) (a) Lindgren, B. O.; Nilsson, T. Acta Chem. Scand. 1973, 27, 888.
(b) Bal, B. S.; Childers, W. E. J.; Pinnick, H. W. Tetrahedron 1981, 37,
2091.
(12) Oikawa, Y.; Yoskioka, T.; Yonemitsu, O. Tetrahedron Lett. 1982,
23, 889.
(13) Takano, S.; Akiyama, M.; Sato, S.; Ogasawara, K. Chem. Lett. 1983,
1593.
(14) Paterson, I.; Schlapbach, A. Synlett 1995, 498.
(15) (a) Still, W. C.; Gennari, C. Tetrahedron Lett. 1983, 24, 4405. (b)
Wensheng, Y.; Su, M.; Jin, Z. Tetrahedron Lett. 1999, 40, 6725.
(16) To achieve an efficient alkylation with iodide 18, incorporation of
an electron-donating 4-methoxy group into the aryl acetate was required.
The subsequent aldol coupling of 5 with 6 and ester reduction steps
proceeded in comparable yield to that obtained previously for the corre-
sponding Heathcock aryl ester (ref 6a).
hydroxyl protection gave TBS ether 17 (72% from 7).
Following reduction of ester 17 with DIBAL, the resulting
Org. Lett., Vol. 5, No. 1, 2003
37

Scheme 7
diastereoface selection. This outcome, in the anti-Felkin sense
However, reduction at C7 of â-hydroxy ketone 26 proved
troublesome, affording mixtures of epimeric alcohols with
a variety of reagents. Gratifyingly, reduction of the readily
derived δ-lactone (AcOH, aq THF) with K-Selectride5c
proceeded cleanly in favor of the desired (7S)-configuration
in 27 (85%, 97:3 dr). Global deprotection of 276a,b then gave
(+)-discodermolide (1) in 81% yield, which was identical
to an authentic sample in all respects.
with respect to the aldehyde, correctly introduces the C2-
C5 stereotetrad sequence for the δ-lactone of discodermolide.
Our working model for rationalizing Si-face attack on all
these aldehydes invokes the preferred chair transition struc-
ture TS-1, in which the dienolate is constrained in the lower
energy s-trans conformation, A(1,3) strain is minimized, and
other steric clashes are avoided.
With these supportive results in hand, the C5-C6 bond
formation for discodermolide itself was now addressed
(Scheme 7). The aldol coupling of the C9-C16 and C17-C24
subunits 5 and 6 and elaboration into alcohol 4 was realized
in an analogous manner16 to that described previously.6a In
our new route, selective olefination of the derived aldehyde
(TEMPO, BAIB, CH2Cl2)6a,10 under modified Still-Gennari
HWE conditions15b gave (Z)-enone 25 in 90% yield from 4
(12:1 Z/E). Carbamate installation18 then gave the required
C6-C24 subunit 3 (97%). As expected from the model
studies, enolization of methyl ketone 3 (c-Hex2BCl, Et3N,
Et2O) and addition of aldehyde 2 (1.3 equiv, -78 °C) gave
adduct 26 in 64% isolated yield with a high level of control
over the (5S)-center (g92:8 dr).19 In comparison, the reversed
aldol coupling at C6-C7 used in our original route6 depends
on stereoinduction from a chiral boron reagent to overturn
the substrate bias and at best proceeds with 84:16 dr.
Up to this point, we had successfully configured all of
the new stereocenters by relying on only substrate control.
In summary, we have completed a highly convergent and
practical total synthesis of (+)-discodermolide (5.1% over
24 linear steps, with 35 steps in total), making use of 1,3-
diol 7 as a readily prepared common building block. In
contrast to other reported syntheses of discodermolide that
start out from the Roche ester,5a-g,6 the present route relies
solely on substrate control to configure all of the remaining
stereocenters. The exploitation of remote 1,6-asymmetric
induction in the boron-mediated aldol reactions of γ-chiral
(Z)-enones, as used in 3 f 26, is also notable. This new
route should be applicable to the preparation of multigram
quantities, enabling further biological and clinical studies of
discodermolide, and provide access to further novel ana-
logues for SAR studies.
Acknowledgment. We thank the EPSRC (GR/L41646;
G.J.F. and J.P.S.), EC (Marie Curie Postdoctoral Fellowship
to N.S.; Network HPRN-CT-2000-00018), Cambridge Eu-
ropean Trust, Gobierno de Canarias (O.D.), and Novartis
Pharma AG for support.
(17) The configuration of the aldol adducts 20 and 22 was established
by 1H NMR analysis of their Mosher esters: Kusumi, T.; Hamada, T.;
Ishitsuka, M. O.; Ohtani, I.; Kakisawa, H. J. Org. Chem. 1992, 57, 1033.
The configuration of adduct 24 was determined by 1H NMR analysis of its
derived PMP acetal (see the Supporting Information).
Supporting Information Available: Spectroscopic data
for all new compounds. This information is available free
of charge via the Internet at http://pubs.acs.org.
(18) Kocovsky, P. Tetrahedron Lett. 1986, 27, 5521.
(19) To date, we have performed this key aldol coupling on a gram scale.
In contrast, by using the lithium enolate from 3, the addition proceeds under
Felkin-Anh induction from the aldehyde 2 and leads to exclusively the
undesired (5R) configuration.
OL0270780
38
Org. Lett., Vol. 5, No. 1, 2003