Synthesis of (-)-amphidinolide K fragment C9-C22

Article abstract of DOI:10.1021/ol051200o

(Chemical Equation Presented) The key fragment (2a or 2b) in a total synthesis of the cytotoxic macrolide (-)-amphidinolide K (1) has been achieved from synthons C9-C14 (3) and C15-C22 (4), which have both been prepared from glutamic acid in good overall yields.

Full text of DOI:10.1021/ol051200o

ORGANIC
LETTERS
2005
Vol. 7, No. 19
4083-4086
Synthesis of (
Fragment C9
−
)-Amphidinolide K
−C22
Thanos Andreou, Anna M. Costa, Laia Esteban, Llu
Gemma Mas, and Jaume Vilarrasa*
1ısa Gonza`lez,
Departament de Qu´ımica Orga`nica, Facultat de Qu´ımica, AV. Diagonal 647,
UniVersitat de Barcelona, 08028 Barcelona, Catalonia, Spain
jVilarrasa@ub.edu
Received May 22, 2005 (Revised Manuscript Received August 9, 2005)
ABSTRACT
The key fragment (2a or 2b) in a total synthesis of the cytotoxic macrolide (
−)-amphidinolide K (1) has been achieved from synthons C9−C14
(3) and C15 C22 (4), which have both been prepared from glutamic acid in good overall yields.
−
Amphidinolide K is a macrolide isolated by Kobayashi et
al. from a laboratory-cultured dinoflagellate Amphidinium
sp.; it shows strong cytotoxic activity against murine
lymphoma L1210 and epidermoic carcinoma KB cells1 and
as such is of interest to those involved in anticancer drug
design. The structure of amphidinolide K (established from
0.3 mg of sample!) was originally reported with some
uncertainties regarding the relative stereochemistry at C2,
C4, and C18. Several years ago, we embarked on the
syntheses of different stereoisomeric fragments of this gross
structure, with the aim of finding analogues (or libraries of
analogues) of similar or higher cytotoxicity. During this
period, Williams and Meyer reported on a synthesis of the
C7-C22 segment2a and later accomplished a synthesis of
(+)-amphidinolide K and several stereoisomers,2b which
allowed the absolute configuration of the natural product to
be deduced. (-)-Amphidinolide K has the structure 1.
In previous work, we developed a route to a potentially
useful C1-C5 building block.3 We report here on our
approach to synthons 2a and 2b (cis-2,5-disubstituted) from
two key fragments: the C9-C14 + C27 segment (hence-
forward C9-C14, or 3) and the C15-C22 intermediate 4.
The synthesis was designed with the underlying ideas of
versatility and simplicity. As shown in Scheme 1, it is based
on a stereocontrolled aldol-like reaction between 3 and
aldehyde 4, as well as on a simple SN2-like cyclization
Scheme 1
(1) Ishibashi, M.; Sato, M.; Kobayashi, J. J. Org. Chem. 1993, 58, 6928.
(2) (a) Williams, D. R.; Meyer, K. G. Org. Lett. 1999, 1, 1303. (b)
Williams, D. R.; Meyer, K. G. J. Am. Chem. Soc. 2001, 123, 765.
(3) Mas, G.; Gonza`lez, L.; Vilarrasa, J. Tetrahedron Lett. 2003, 44, 8805.
10.1021/ol051200o CCC: $30.25
© 2005 American Chemical Society
Published on Web 08/19/2005

reaction (hydroxyl at C12, leaving group at C15) to create
the five-membered ring. The five carbon atoms highlighted
in 3 and 4 may arise from 5 and 6, respectively, which in
turn may come from the corresponding enantiomers of
glutamic acid or their synthetically equivalent 5-hydroxy-
methyltetrahydrofuran-2-ones (7). Both enantiomers of 7 are
commercially available, at the same price.
excellent results. However, the Swern reaction turned out to
be more practical on a larger scale (lower price of the
reagents, shorter reaction times, purification by chromatog-
raphy unnecessary). A Horner-Wadsworth-Emmons reac-
tion (EtOCOCH2PO(OEt)2, BuLi, DME, rt, 6 h)8 thereafter
gave the C9-C14 fragment 11 (cf. 3). The overall yield from
7 to 11 was 70%.
The synthesis of the C15-C22 fragment was first at-
tempted according to the strategy shown in Scheme 1. Methyl
(S)-4,5-epoxypentanoate (6) was prepared from (S)-79 by
tosylation and ring opening with sodium methoxide. The
organocuprate prepared from (E)-1-bromopropene, Li,10 and
CuCN was active on an epoxide model but gave a complex
mixture of products when reacted with 6 and starting material
was also recovered. The desired product (isolated as its
lactone 12) was obtained in poor yields (ca. 30%). Coupling
of the organocuprate prepared from (E)-1-propenyl-1-lithium
and typically CuI or CuCN with the O-tosyl derivative 13
(commercially available), to reach 12 in an alternative way,
was also unsuccessful. Thus, we decided to replace epoxy-
ester 6 with the corresponding epoxyalcohol,11 protected as
a silyl ether.
Thus, stereoisomers of 2a or 2b would be synthesized
using the same protocols. Employing either butyrolactone
as a starting material, varying the chiral auxiliary used in
the aldol reaction, and carrying out the cyclic ether formation
either way (leaving group at C12 and hydroxyl at C15 or
vice versa) would allow one to achieve alternate configura-
tions at C12, C15, and/or C18. In other words, the outlined
strategy is in principle really versatile.
Conversion of (R)-7 into an appropriate derivative of 5
was first investigated (Scheme 2). Protection of the hydroxy
Scheme 2
Reaction of 13 with an excess of DIBALH, protection of
the primary alcohol as its TBS ether (14), and treatment with
NaH in THF afforded 15 as shown in Scheme 3.12 Addition
Scheme 3
group of (R)-7 as a p-methoxybenzyl ether (8),4 opening of
the lactone with base, and silylation of the alcohol with tert-
butyldimethylsilyl chloride in two steps gave 9. Activation
of the carboxyl group in 9 in the usual way and reaction
with the lithium salt of (R)-4-benzyl-1,3-oxazolidin-2-one5
at -78 °C for 2 h in THF and 0 °C for 10 h gave the C11-
C14 fragment 10 in 95% yield, ready to be subjected to the
desired aldol reactions.6 After removal of the PMB group
from 10 with DDQ (H2O, CH2Cl2, rt, 3 h),4 oxidation of the
alcohol under Swern conditions (Me2SO/ClCOCOCl, CH2-
Cl2, -78 °C, then Et3N)7a or using the Dess-Martin
periodinane (1.5 equiv, CH2Cl2)7b gave the aldehyde with
(4) Cf. Greene, T. W.; Wuts, P. G. M. ProtectiVe Groups in Organic
Synthesis, 3rd ed.; Wiley: New York, 1999, and references therein.
(5) (a) Cage, J. R.; Evans, D. A. Org. Synth. 1989, 68, 77. (b) Evans, D.
A.; Urp´ı, F.; Sommers, T. C.; Clark, J. S.; Bilodeau, M. T. J. Am. Chem.
Soc. 1990, 112, 8215.
(6) On the other hand, opening of the lactone ring of 7 with (R,R)-
pseudoephedrine and protection/activation of the vicinal diol as a sulfite
ester, followed by reaction with 3-methylbutanal (as a model) and
tetrahydrofuran ring formation with NaH posed diverse problems, such as
a moderate yield of the aldol reaction and no cyclization in the last step.
Use of a stronger EWG-sulfate instead of sulfite-posed problems also,
i.e., poor yields of the sulfate preparation and a premature elimination
reaction.
of the organocuprate of (E)-1-propenyl-1-lithium (2 mol per
mol of CuCN) gave the desired product, 16, in 99% isolated
(8) Review: Maryanoff, B. E.; Reitz, A. B. Chem. ReV. 1989, 89, 863.
(9) Compound (S)-7 is usually obtained through diazotization of L-
glutamic acid, lactonization, and reduction of the carboxyl group with borane
(or related procedures). See: (a) Lehmann, J.; Pieper, B. Tetrahedron:
Asymmetry 1992, 3, 1537. (b) Figade`re, B.; Harmange, J.-C.; Laurens, A.;
Cave´, A. Tetrahedron Lett. 1991, 32, 7539 and references therein. For an
alternative synthesis of 6 from D-mannitol, see: (c) Chattopadhyay, S.;
Mamdapur, V. R.; Chadha, M. S. Tetrahedron 1990, 46, 3667.
(7) (a) Mancuso, A. J.; Huang, S. L.; Swern, D. J. Org. Chem. 1978,
43, 2480. (b) Dess, D. B.; Martin, J. C. J. Org. Chem. 1983, 48, 4155.
4084
Org. Lett., Vol. 7, No. 19, 2005

yield. Use of lithium [cyano(E-1-propen-1-yl)(2-thienyl)-
cuprate]13 afforded 16 in 90% yield but now expended only
1 mol of (E)-1-bromopropene per mol of the commercially
available Li[Cu(CN)(2-thienyl)].
Scheme 5
Protection of the hydroxy group in 16 as its tert-
butyldiphenylsilyl ether and selective removal of the TBS
group with pyridinium tosylate (PPTS) gave 17. Oxidation
to 18 was carried out by standard procedures. The overall
yield of 18 (cf. 4, PG ) TBDPS) from 7 was 68%.
The intended aldol reaction between 18 and 10 was first
probed on a model setting using substrate 10 and 3-methyl-
butanal (Scheme 4). Via the chlorotitanium enolate derived
Scheme 4
CH2Cl2)14a and aldehyde 18 (addition at -78 °C, then 5 h at
room temperature), 50% of stereopure 22 was isolated,
whereas with the dibutylboron enolate of 10 (Bu2BOTf, Et3N,
CH2Cl2) and aldehyde 18 (addition at -78 °C, then 5 h at 0
°C) an improved product yield of 75% (with recovery of
15% and 20% of 10, respectively) was observed.15
from 10 (with TiCl4),6b only one aldol diastereomer (19) was
obtained in 72% yield.14 The configuration of the carbon
atom R to the COAux* (C14) was not important, as a double
bond would later be installed between C14 and C27.
Nevertheless, only the expected syn adduct was detected.
Reduction of 19 with LiBH4/MeOH in THF, to give the diol,
protection of the primary alcohol as its trityl ether 20, and
activation of the secondary alcohol in 20 as a methanesulfo-
nyl derivative (MsO group) took place smoothly. Treatment
of this mesylate with Bu4N+F-‚3H2O (TBAF) thereafter
induced tetrahydrofuran ring formation in excellent yield,
leading to model compound 21; its relative stereochemistry
was established by a NOESY experiment.
Application of the two enolization protocols to 11 and 18
(Scheme 5) provided 23 as a single diastereomer, in 61%
(14) (a) To avoid a partial cleavage of the acid-sensitive groups, the
reagents were added to 10 as follows: 0.1 equiv of EtiPr2N (DIPEA), then
1.2 equiv of TiCl4, and finally 1.1 equiv of DIPEA. (b) Enolization with
TiCl3(OiPr) afforded a mixture of two aldols: Mas, G. PhD Thesis,
Universitat de Barcelona, 2000. (c) When parallel aldol-like reactions were
carried out in our lab with thiazolidine-2-thione chiral auxiliaries (instead
of the oxazolidin-2-one derivatives shown in Schemes 2 and 4), the results
were disappointing: Batlle, M. Master’s Thesis, Universitat de Barcelona,
2003. For leading references on thiazolidine-2-thione auxiliaries, see: (d)
Nagao, Y.; Yamada, S.; Kumagai, T.; Ochiai, M.; Fujita, E. J. Chem. Soc.,
Chem. Commun. 1985, 1418. (e) Nagao, Y.; Hagiwara, Y.; Kumagai, T.;
Ochiai, M.; Inoue, T.; Hashimoto, K.; Fujita, E. J. Org. Chem. 1986, 51,
2391 [acetate aldol reactions of Sn(II) enolates]. (f) Delaunay, D.; Toupet,
L.; Corre, M. L. J. Org. Chem. 1995, 60, 6604 and references therein
(preparation). (g) Gonza´lez, A.; Aiguade´, J.; Urp´ı, F.; Vilarrasa, J.
Tetrahedron Lett. 1996, 37, 8949 (acetate aldol reactions of Ti enolates).
Also see: (h) Crimmins, M. T.; King, B. W.; Tabet, E. A. J. Am. Chem.
Soc. 1997, 119, 7883 (Ti enolates of oxazolidine-2-thione derivatives). (i)
Crimmins, M. T.; King, B. W.; Tabet, E. A.; Chaudhary, K. J. Org. Chem.
2001, 66, 894 and references therein. For a review, see: (j) Evans, D. A.;
Shaw, J. T. Actualite´ Chim. 2003, 35.
The reactivity of various enolates derived from 10 and
the more advanced intermediate 11 were then compared
(Scheme 5). Using the titanium enolate of 10 (TiCl4, EtiPr2N,
(10) Whitesides, G. M.; Casey, C. P.; Krieger, J. K. J. Am. Chem. Soc.
1971, 93, 1379.
(11) Conversion of 6 to the corresponding epoxyalcohol with titanium
tetraisopropoxide and polymethylhydrosiloxane (Reding, M. T.; Buchwald,
S. L. J. Org. Chem. 1995, 60, 7884) was not useful as the 4,5-epoxy-1-
pentanol isomerized to 5-hydroxymethyloxolane (THF derivative) under
the reaction conditions, as might be expected.
(12) For related procedures, see: (a) Nagumo, S.; Furukawa, T.; Ono,
M.; Akita, H. Tetrahedron Lett. 1997, 38, 2849. (b) Harrowven, D. C.;
Dennison, S. T.; Hayward, J. S. Tetrahedron Lett. 1994, 35, 7467.
(13) Lipshutz, B. H.; Koerner, M.; Parker, D. A. Tetrahedron Lett. 1987,
28, 945.
(15) For the cleavage of the final boron complexes, in all cases the
reaction mixtures were quenched at low temperature with MeOH and
buffered water (pH 7). The resulting solutions or suspensions were stirred
for ca. 12 h (no hydrogen peroxide or bases were added).
Org. Lett., Vol. 7, No. 19, 2005
4085

yield for TiCl4/EtiPr2N conditions, and 85% yield for Bu2-
BOTf/Et3N conditions (15% of 11 was recovered in both
cases, so that the conversion was ca. 75% and ca. 100%,
respectively).14a,15 After several trials, reduction of the
complex multifunctionnal substrate 23 to its diol could be
accomplished with LiBH4 (220 mol %) in the presence of
MeOH (250 mol %),16 in such a way that only 5-7%
reduction of the conjugate double bond took place, while
the C20-C21 double bond was left intact. After protection
of the primary alcohol with a trityl group (24a), the secondary
OH of 24a was converted to a MsO group and selective
cleavage of the TBS vs TBDPS ether with TBAF directly
gave 2a, as the result of a one-pot cyclization.
A significant NOE between H12 and H15 indicated a cis-
arrangement. Additional correlation data (COSY and NOE-
SY) for H13, H14, H15, and H27 confirmed the relative
configuration of all stereocenters.
Alternatively, the primary hydroxy group at C27 could
be quantitatively converted into a selenoaryl ether (with
2-nitrophenylselenyl cyanide and Bu3P)17 or, more quickly
and without requiring any chromatographic separation, into
its 2-pyridylselenyl ether (24b) with di-2-pyridyl diselenide
and trimethylphosphine (PySeSePy/Me3P).18
Compound 2b was characterized by NMR, as well as by
its strong [M + H]+ and [M + Na]+ peaks in the FABMS
(in sharp contrast to other analogues of type 2, the corre-
sponding peaks of which could hardly be detected). Inde-
pendent experiments with other selenopyridyl ethers and with
2b (to be reported elsewhere) showed that their oxidation
with hydrogen peroxide or other peroxo derivatives afforded
the desired double bond (to be introduced between C14 and
C27 in a later stage of the total synthesis of 1).
In summary, completion of the total synthesis of 1 seems
really feasible soon, as a route to the main fragment (C9-
C22, 2a/2b) has now been disclosed and key steps have been
optimized. The generation of titanium enolates without
cleavage of PMB and/or TBS ether functionality, the
excellent yields of the boron-mediated aldol reactions after
appropriate workup, and the use of PySeSePy/PMe3 are all
worthy of noting. In the future different stereoisomers of
the C9-C22 building block will be combined with the C1-
C5 stereoisomers already reported,3 to obtain a set of
diastereomers and analogues of (-)-amphidinolide K, with-
out major deviations from the original strategy.
Acknowledgment. Financial support, a studentship to
G.M. (1997-2000), and a postdoctoral fellowship to L.G.
(1997-1999) from the Ministerio de Educacio´n y Cultura
(Madrid) are acknowledged. A.M.C. thanks the Ministerio
de Ciencia y Tecnolog´ıa for a Ramo´n y Cajal fellowship.
The Research Directorate-General of the European Com-
mission (Brussels, HPRN-CT-2000-018 Network) is also
acknowledged. The Generalitat de Catalunya (Barcelona) has
partially contributed with a grant (2001SGR 051) and an
IGSoC studentship to T.A. since 2003. Thanks are due to
Prof. D. R. Williams, Indiana University, for sending us the
manuscript of his (+)-amphidinolide K total synthesis before
publication.
Activation of the C15 hydroxy group of 24b as its Ms
derivative, followed by removal of its TBS group, as above,
gave directly the tetrahydrofuran derivative 2b in 95% yield.
The deprotection and cyclization steps could be easily
followed by TLC, as the three spots were clearly distin-
guished.
(16) (a) Penning, T. D.; Djuric, S. W.; Haack, R. A.; Kalish, V. J.;
Miyashiro, J. M.; Rowell, B. W.; Yu, S. S. Synth. Commun. 1990, 307. (b)
Evans, D. A.; Cage, J. R.; Leighton, J. L. J. Am. Chem. Soc. 1992, 114,
9434. (c) LiBEt3H (“Superhydride”, 2.2 equiv) in THF at -78 °C cleaved
exclusively the chiral auxiliary of 11; however, in the case of 23, several
products were produced.
(17) (a) Sharpless, K. B.; Young, M. W. J. Org. Chem. 1975, 40, 947.
(b) Reich, H. J.; Wollowitz, S.; Trend, J. E.; Chow, F.; Wendelborn, D. F.
J. Org. Chem. 1978, 43, 1697. (c) Sayama, S.; Onami, T. Tetrahedron Lett.
2000, 41, 5557.
(18) (a) Mart´ın, M.; Mart´ınez, G.; Urp´ı, F.; Vilarrasa, J. Tetrahedron
Lett. 2004, 45, 5559. (b) Esteban, J.; Costa, A. M.; Urp´ı, F.; Vilarrasa, J.
Tetrahedron Lett. 2004, 45, 5563. For the use of the PySe group in syn
selenoxide eliminations, see: (c) Toshimitsu, A.; Owada, H.; Terao, K.;
Uemura, S.; Okana, M. J. Org. Chem. 1984, 49, 3796. (d) Toshimitsu, A.;
Hayashi, G.; Terao, K.; Uemura, S. J. Chem. Soc., Perkin Trans. 1 1988,
2113 and references therein.
Supporting Information Available: Experimental pro-
cedures and characterization data for the main compounds.
This material is available free of charge via the Internet at
http://pubs.acs.org.
OL051200O
4086
Org. Lett., Vol. 7, No. 19, 2005