An Enantioselective Construction of the ABC System of Taxol

Full text of DOI:10.1021/ja973559g

J. Am. Chem. Soc. 1998, 120, 1337-1338
1337
Scheme 2
An Enantioselective Construction of the ABC System
of Taxol
Gilbert Stork,* Kei Manabe,† and Longbin Liu‡
Department of Chemistry, Columbia UniVersity
New York, New York 10027
ReceiVed October 13, 1997
a (i) tBuLi, ether, 78 °C, 1 h; (ii) 2-ThiophenylCu(CN)Li, 10 min;
(iii) TMSCl, THF, -78 to 0 °C, 1 h; 81%. b (i) THF, -78 °C, 45 min;
(ii) -78 to 0 °C, 50 min; >90%. c Toluene, -78 to -55 °C, 1.5 h.
d Ethanol, overnight; 54% from 6. e 4-Methoxybenzyl trichloroacetimidate,
catalyst TfOH, ether, 0 °C, 10 min; 36% (unoptimized) + bisbenzylated.
f HOAc, H2O, THF (6/2/3), 70 °C, 2.5 h; 88%. g SiO2, CH2Cl2, 1 h; 92%.
The successful use of protected cyanohydrin cyclization (cf.
A) to form the A-B system 1,1 which includes the A/B
2
framework of taxol (2), suggested the possibility of extending
the scheme to the construction of a substance such as 3 in which
the 4-pentenal array would be expected to undergo facile aldol
closure 3 with the formation of an A-B-C system in which only
oxidation states would have to be adjusted to complete the
construction of taxol.
As we had hoped, reaction of 5 with the lithium species derived
from stannyl lactone 4 proceeded in high yield to give the C2
secondary alcohol 9, which now had to be inverted to the required
epimer. The assumption that this would best be achieved by
intramolecular reduction of the C2 ketone via a hydride tethered
to the primary alcohol shown in 10 proved correct: lithium
aluminum hydride reduction of 10 gave mainly the correct alcohol
epimer at C2, while simultaneously reducing the lactone, to
produce 11. The benzylidene derivative 12 thus became available
in 49% overall yield from 10. Further transformation of 12 to
the bicyclic A-B ketone 13 confirmed the feasibility of extending
the construction we had used for the simple A-B system 1 to
Scheme 1
Scheme 3a
We assumed (incorrectly, as it developed) that either of the
secondary methyl epimers in ring B would lead to the enolate
required for the eventual aldol closure of ring C, and we initially
selected aldehyde 5 as an acceptor for stannyl lactone 4.
a TPAP, NMO, 4A molecular sieves, CH3CN, 21 h; 70%. b CH2Cl2,
H2O, 11 h, 89%. c THF, 0 °C, 2 h; 75%. d Imidazole, CH2Cl2, room
temperature, 12 h; 92%. e CSA, PhH, 1 h; 71%. f DBU, t-BuOH, PhH,
65 °C, 12 h; NaBH4, EtOH, 0 °C, 1.5 h; Ac2O, DMAP, CH2Cl2, 10 min;
THF, HOAc, H2O, 45 °C, 10 h; 64% overall. g Cyanohydrin closure: cf.
ref 1. h TBAF.
The enantiospecific synthesis of 5 is outlined below. As
starting material we chose the known butenolide 6,4 available as
the required enantiomer from (R)-glutamic acid. Conjugate
addition of the protected 4-pentenol side chain,5 followed by
methylation, set up the required absolute stereochemistry of
lactone 7, which was then easily transformed into aldehyde 5.
the more complex system shown in 14, which bears substituents
formally capable of elaboration to the A-B-C system of taxol.
It soon became clear, however, that an unforeseen difficulty lay
within structure 14 because it resisted all our attempts to cyclize
it.
A clue to the source of the difficulty came from deuterium
exchange experiments (overnight refluxing of alcohol 13 with
methoxide in deuteriomethanol). This showed no exchange of
the hydrogen on the carbon bearing the secondary methyl group.
Molecular modeling of systems such as 13 clearly showed that,
† University of Tokyo, Faculty of Pharmaceutical Sciences, Tokyo 113,
Japan.
‡ Amgen, Thousand Oaks, California.
(1) Stork, G.; Doi, T.; Liu, L. Tetrahedron Lett. 1997, 38, 7471.
(2) Taxol is the registered trademark for the substance known also as
paclitaxel.
(3) (a) Chaudhary, A. G.; Rimoldi, J. M.; Kingston, G. I. J. Org. Chem.
1993, 58, 3798. (b) Miller, R. W.; Powell, R. G.; Smith, C. R., Jr.; Arnold,
R.; Clardy, J. J. Org. Chem. 1981, 46, 1469.
(4) (a) Org. Synth. 1985, 63, 121. (b) Tetrahedron 1990, 46, 4503.
(5) (a) Lawler, D. M.; Simpkins, N. S. Tetrahedron Lett. 1988, 29, 1207.
(b) Urban, E.; Knuhl, G.; Helmchen, G. Tetrahedron 1995, 51, 3031.
S0002-7863(97)03559-2 CCC: $15.00 © 1998 American Chemical Society
Published on Web 01/30/1998

1338 J. Am. Chem. Soc., Vol. 120, No. 6, 1998
Communications to the Editor
Scheme 4
transformations via 21, the usual cyanohydrin sequence1 was used
for the closure of ring B. The resulting 22 (X, Y ) 2-methoxy-
isopropyl, cyano; then, carbonyl) was converted, after Johnson-
Claisen elongation of the pendant allyl alcohol, into 23, the sought
after secondary methyl epimer of 14. The cyclization of 23 to
the tricyclic aldol 24 (57%) now took place normally. The
assigned stereochemistry was entirely consistent with nOe
measurements.
in the lowest energy conformation, the hydrogen on the carbon
bearing the R-methyl is held in the plane of the carbonyl, so that
14 is unable to form the enolate required for aldol cyclization.
Models showed, however, that proper overlap should be obtained
with the secondary methyl epimer of 14 which should, therefore,
cyclize normally.6 A stereoselective route to that epimer is now
described.
Scheme 6a
Scheme 5a
a CSA, 60 °C, 2 h. b PhH, t-BuOH, 80 °C, 6 h. c CeCl3, 2:1 THF-
H2O, 3 h. d Pyr., DMAP, overnight. e THF, 4A sieves; 40% from 22.
f TMSCN, H+, 2-methoxypropene. g Cf. ref 1. h TBAF, HMPA, 4A
molecular sieves, overnight. i THF-(aq)HCl, room temperature, 20 min;
1 N NaOH, 10 min. j MeCH2CO2H, PhMe, 120 °C, 12 h; 63%. k (i) PhMe,
-78 °C, 40 min, (ii) Dess-Martin; 84%. l MeOH, 18-crown-6, reflux, 3
h; 55%.
a THF, -90 to -100 °C, 15 min. (b) 15, THF, -78 °C to room
temperature 1.5 h; 40% 16. b C2 â epimer, recycled by (i) Dess Martin
oxidation and (ii) NaBH4-CeCl3 reduction. c CSA, MeOH, 4 h. d DMF,
0 °C to room temperature THF-H2O. e CH2Cl2, pyridine, catalyst DMAP,
0 °C to room temperature, 12 h. f (i) DMPU, THF, -78 °C, (ii) LDA,
-78 °C, then to room temperature, 1.5 h. g (i) 110 °C, 12 h, (ii) CH2Cl2,
CH2N2; 54% and recyclables. h CH3CN-H2O, 11 h. i (iPr)2NEt, catalyst
Bu4Nl, 7 h; 78% from 19. j THF, -78 to 0 °C, 2 h; 57%.
It is worthy of note that the tricyclic system 24 differs from
the tetracyclic system of taxol only in the oxidation state.10
Acknowledgment. We thank Professor Clark Still for his much
appreciated help with modeling, especially of 14, and the National
Institutes of Health and the Kanagawa Academy of Science and
Technology for support of this work.
Addition of the usual lithiolactone from 4 to the protected
dihydroxyaldehyde 157 gave a 1:1.4 mixture of the C2 alcohol
epimers. Oxidation of the unwanted â-hydroxyl to the ketone,
and reduction with borohydride, led to the predominant formation
of the required secondary alcohol 16. Desilylation was followed
by conversion to the allylic propionate of the isopropylidene
derivative shown in 17, a structure designed to show a face
differentiation which would permit use of a Claisen rearrangement
to establish the correct streochemistry at C3. Importantly, a chair
transition state for such a rearrangement (cf. 18) should, in
addition, lead to the required secondary methyl stereochemistry
shown in 19. In the event, application of the Ireland-Claisen
conditions8 to propionate 17 did produce the required structure
and sterochemistry shown in 19, as could be verified by X-ray
Supporting Information Available: Spectral data for compounds
5, 7, 8, 9, 10, 11, 12, and 13 and experimental and spectral data for
compounds 4, 16, 17, 19, 20, 21, 22, 23, 24 and i, footnote 9 (37 pages).
See any current masthead page for ordering information and Web access
instructions.
JA973559G
(9) We thank Dr. John C. Huffman of the Indiana University Molecular
Structure Center for this determination. The γ lactone (i) had mp 85-86 °C,
[R]D +183 (0.34, CHCl3), IR 1781 cm-1
9
structure determination of the γ-lactone obtained by treatment
of 19 with methanolic acid. Hydrolysis of the isopropylidene
group of 19, protection of the released allylic alcohol, LAH
reduction, and silylation gave lactol 20. Following standard
(6) Our observations related to the failure of 14 to undergo aldol cyclization,
our rationalization of this result, and our corollary successful cyclization of
23 were first presented publicly in 1995. Related observations have been
published recently, See: (a) Wender, P. A. et al. J. Am. Chem. Soc. 1997,
119, 2757. (b) Mukaiyama, T., et al. Chem. Lett. 1996, 483.
(7) Made from dihydroxyacetone by silylation (Shao, X.; Dolder, M.;
Tamm, C. HelV. Chim. Acta 1990, 73, 483), followed by condensation with
dimethyl carbomethoxymethyl phosphonate, reduction (DIBAL), and oxidation
(PCC).
.
(10) It is encouraging that silylation of 24 (triethyl silyl triflate) followed
by SeO2 oxidation (tert-butyl peroxide, hexane, catalyst acetic acid, reflux
1.5 h) appeared to result in the expected formation (∼62%) of the desired
5R-hydroxy derivative.
(8) Inter alia: Ireland, R. E.; Wipf, P.; Armstrong, J. D. J. Org. Chem.
1991, 56, 352 and earlier papers.