Palladium-catalyzed kinetic and dynamic kinetic asymmetric transformation of 5-acyloxy-2-(5H)-furanone. Enantioselective synthesis of (- )-aflatoxin B lactone [8]

Full text of DOI:10.1021/ja9844229

J. Am. Chem. Soc. 1999, 121, 3543-3544
3543
ligand 86 in methylene chloride afforded 9a, however, in only
47% ee (see eq 2). Lowering the catalyst loading to 0.5% Pd2-
Palladium-Catalyzed Kinetic and Dynamic Kinetic
Asymmetric Transformation of
5-Acyloxy-2-(5H)-furanone. Enantioselective
Synthesis of (-)-Aflatoxin B Lactone
Barry M. Trost* and F. Dean Toste
Department of Chemistry
Stanford UniVersity
Stanford, California 94305-5080
ReceiVed December 28, 1998
The ability to achieve dynamic kinetic asymmetric transforma-
tions is an efficient method for asymmetric synthesis since 100%
of the racemic starting material is converted to a single enantio-
meric product.1 In considering this question within the context
of asymmetric metal catalyzed allylic alkylations, we were
attracted to γ-acyloxybutenolides as substrates. First, the utility
of γ-alkoxybutenolides as synthons for asymmetric synthesis of
natural products has spurred the development of methods for their
asymmetric preparation.2,3 Second, they offer an interesting
structural feature that may facilitate the process as illustrated in
eq 1. While ionization of racemic 1 initially generates the two
dba3 and 1.5% 8 resulted in an increase in the enantioselectivity
of 9a to 58%. We conjectured that the moderate enantioselectivity
was not due to poor recognition in the ionization step but rather
to partial equilibration of the kinetically formed π-allyl intermedi-
ate as conjectured in eq 1. Therefore, we added a base to increase
the concentration of deprotonated phenol to increase the rate of
nucleophilic attack relative to the equilibration of the kinetically
formed intermediate 2 or 3. Indeed, the addition of 15% cesium
carbonate afforded the adduct 9a in 90% yield (based on 7a since
2.2 equiv of 6 was employed) in 87-90% ee. The enantiomeric
excess of recovered 6 ranged from 16 to 44%.
Similarly, reaction of 2.2 equiv of 6 with 1 equiv of 3,5-
dimethoxyphenol (7b) or 2-iodophenol (7c) catalyzed by 1% Pd2-
dba3 and 3% ligand 8 afforded 9b and 9c in 88% ee (82% yield
based on 7b) and 86% ee (88% yield based on 7c), respectively.
Reaction with the more complicated phenol 7d7 with 6 under
similar conditions afforded adduct 9d in 70% yield and 87% ee.
Triethylborane initiated radical cyclization of 9d yielded the
tricyclic compound 10 (eq 3).8 Silyl group deprotection of 10
π-allylpalladium intermediates 2 and 4, their interconversion
through the expediency of a σ-complex like 3 provides a vehicle
for their interconversion. The aromaticity of the furan may serve
as a driving force to go from the η3 to the η1 complex. If
interconversion is fast relative to nucleophilic addition and if one
of these diastereomeric η3-complexes 2 or 4 (in the presence of
chiral ligands) undergoes reaction faster than the other, then an
effective dynamic kinetic asymmetric transformation would result.
In this paper, we explore the behavior of such butenolides in the
presence of chiral enantiomerically pure palladium complexes,
which demonstrates the feasibility of both kinetic and dynamic
kinetic asymmetric transformations and culminates in an efficient
strategy for the asymmetric synthesis of the aflatoxin family.4
We initiated our studies by examining the chiral recognition
in the ionization step. Reaction of 65 with 0.5 equiv of 4-meth-
oxyphenol 7a in the presence of 5% Pd2dba3 and 15% chiral
affords the Buchi lactone 11, which constitutes an asymmetric
formal synthesis of the aflatoxins.9,10 The absolute stereochemistry
of 11 was established by comparison to the optical rotation
reported by Marino.11
With these results in hand, our study of the dynamic kinetic
asymmetric transformation commenced (eq 2). When the reaction
of 6 with 7a, under the conditions described above, except that
equimolar quantities of phenol and butenolide were used, was
carried out to conversions of the butenolide greater than 45%,
the enantioselectivity dropped significantly (48% ee at 80%
conversion, 24% ee at 100% conversion). If the rate of nucleo-
philic attack was significantly greater than that for interconversion
of 2 and 4, speeding up the latter by, for example, addition of
chloride ion to increase the rate of interconversion of 2 and 4
(6) Trost, B. M.; Van Vranken, D. L.; Bingel, C. J. Am. Chem. Soc. 1992,
14, 9327.
(7) The preparation of phenol 7d (3 steps from phloroglucinol) is detailed
in the Supporting Information.
(8) The tributyltinhydride/AIBN method failed to produce any of the desired
tricyclic compound 10. (a) Sloan, C. P.; Cueras, J. C.; Quesnelle, C.; Snieckus,
V. Tetrahedron Lett. 1988, 20, 1685. (b) Hoffmann, H. M. R.; Scmidt, B.;
Wolff, S. Tetrahedron 1989, 45, 6126. (c) Wolff, S.; Hoffman, H. M. R. Synlett
1998, 760.
(1) For a review on dynamic kinetic transformations, see: Ward, R. S.
Tetrahedron: Asymmetry 1995, 6, 1475.
(2) (a) Faber, W. S.; Koh, J.; de Lange, B.; Feringa, B. L. Tetrahedron
1994, 50, 4775. (b) van der Deen, H.; Hof, R. P.; van Oeveren, A.; Feringa,
B. L.; Kellogg, R. M. Tetrahedron Lett. 8441. (c) van der Deen, H.; Hof, R.
P.; van Oeveren, A.; Feringa, B. L.; Kellogg, R. M. J. Am. Chem. Soc. 1996,
118, 3801. (d) Kinderman, S. S.; Feringa, B. L. Tetrahedron: Asymmetry 1998,
9, 1215.
(3) For representative examples, see: (a) Krief, A.; Lecomte, Ph.; De-
mounte, J. P.; Dumont, W. Synthesis 1990, 275. (b) van Oeveren, A.; Jansen,
J. F. G. A.; Feringa, B. L. J. Org. Chem. 1994, 59, 5999 and references therein;
c) Harcken, C.; Bru¨ckner, R. Angew. Chem., Int. Ed. Engl. 1997, 36, 2751.
(4) Schuda, P. F. Top. Curr. Chem. 1980, 91, 75.
(5) Butenolide 6 is prepared by the reaction of furfural and singlet oxygen
(Gollick, K.; Griesbeck, A. Tetrahedron 1985, 2059) followed by reaction of
the product alcohol with di-tert-butyl dicarbonate.
(9) (a) Buchi, G.; Foulkes, D. M.; Kurone, M.; Mitchell, G. F.; Schneider,
R. S. J. Am. Chem. Soc. 1967, 89, 6745. (b) Buchi, G.; Weinreb, S. M. J.
Am. Chem. Soc. 1971, 93, 746.
(10) For alternative enantioselective formal syntheses of the aflatoxins
(aflatoxin B2): see: (a) Rapoport, H.; Civitello, E. H. J. Org. Chem. 1994,
59, 9, 3775. (b) Bando, T.; Shishido, K. Synlett 1997, 665.
(11) Measured [R]D ) -137.5 (87% ee). Reported [R]D ) -128 (80%
ee). Marino, J. P. Pure Appl. Chem. 1993, 65, 667.
10.1021/ja9844229 CCC: $18.00 © 1999 American Chemical Society
Published on Web 03/25/1999

3544 J. Am. Chem. Soc., Vol. 121, No. 14, 1999
Communications to the Editor
Scheme 1. Formal Synthesis of Aflatoxin B
through coordination to palladium which would favor the η1-
complex 312 should increase the ee. Indeed, addition of 30% tetra-
n-butylammonium chloride to the reaction mixture resulted in 9a
being isolated in 80% yield and 75% ee, up from 24% ee in the
absence of chloride. Slowing the rate of nucleophilic attack by
removing the cesium carbonate afforded 9a in 84% ee (74%
yield). Compound 9c was prepared under identical conditions, in
83% yield and 87% ee, by the reaction of 7c with 6. A single
recrystallization increases the ee to 99%. Reductive Heck cy-
clization affords tricyclic lactone 12 in 64% yield without
significant deterioration of the enantiomeric excess (93% ee) (eq
4).13
A significant portion of the synthetic utility of γ-alkoxybuteno-
lides relies on transfer of chirality from the γ-alkoxy center to
the R and â carbons.3 To this end, we prepared butenolide 13 in
97% ee (87% yield), from 6 and 2-naphthol 7e under our standard
palladium-catalyzed conditions for dynamic kinetic asymmetric
transformations (eq 5). The steric bulk of 2-naphthol presumably
kinetic asymmetric transformation with butenolide 6, affording
the adduct 19 in 89% yield. Unfortunately, at this stage, we were
unable to determine its enantioselectivity. Unlike the attempted
but failed palladium-catalyzed cyclization of 9d, subjecting 19
to the standard reductive Heck cyclization conditions afforded
the tetracyclic coumarin 20 in 93% yield. At this stage, the
1
enantiomeric excess of 20 was determined to be g95% by a H
NMR chiral shift experiment (Eu(hfc)3). In accord with earlier
observations, the phenol substituents which are electron-
withdrawing, such as the coumarin ring, may be essential for
efficient cyclization. Acid-catalyzed hydrolysis of the ethyl ester
proceeds smoothly under the conditions reported by Bu¨chi.9a
Introduction of ring E is accomplished in one step by treatment
with scandium triflate and lithium perchlorate.18 Thus, Bu¨chi’s
intermediate 22 is available in six steps from two simple building
blocks 15 and 16.
The current study validates the strategy depicted in eq 1 as an
efficient approach for the asymmetric synthesis of butenolides.
These substitutions occur with excellent ee despite the ease with
which the products may racemize due to their propensity toward
enolization. Since the nucleophilic substitution event occurs
concomitantly with the deracemization event, this strategy is more
efficient than performing this transformation as two separate
operations.19 These successes, highlighted by the emergence of
an efficient asymmetric synthesis of the aflatoxins, stimulates the
exploration of other situations for dynamic kinetic asymmetric
transformations.
decreases the rate of nucleophilic attack relative to equilibration
and thus may be responsible for the enhanced enantioselectivity.
Conjugate addition of diethylzinc promoted by tert-butyldimeth-
ylsilyl triflate gave a 10:1 ratio of diastereomers in 51% (71%
based on recovered 13) yield.14 This reaction demonstrates that
the 2-naphthyloxy moiety is an efficient auxiliary for this type
of chirality transfer.
With these results in hand, we investigated the use of the
palladium-catalyzed dynamic kinetic asymmetric transformation
toward an enantioselective synthesis of the aflatoxins. Our
synthetic plan (see Scheme 1) was to utilize a coumarin as the
phenol nucleophile in the allylic alkylation. This would make the
synthesis highly convergent and allow for some flexibility in ring
E, an important consideration since members of this class of
compounds vary at this ring. To this end, we prepared coumarin
17 by the Pechman condensation15 of monomethyl phloroglucinol
1516 with ketoester 16,17 which was separated from the 1.4:1
regioisomeric mixture of coumarins by fractional recrystallization
(CH2Cl2-hexane) in 47% yield. Treatment of 17 with iodinemon-
ochloride resulted in regioselective iodination to produce 18 in
92% yield. The sterically demanding coumarin 18 readily
participated as the nucleophile in the palladium-catalyzed dynamic
Acknowledgment. We thank the National Science Foundation and
the General Medical Sciences Institute of the National Institutes of Health
for their generous support of our work. Mass spectra were provided by
the Mass Spectrometry Regional Center of the University of California-
San Francisco supported by the NIH Division of Research Resources.
Supporting Information Available: Experimental procedure and
characterization data for 6, 7d, 9a-d, 10-14, 17-22 (12 pages (PDF).
This material is available free of charge via the Internet at http://pubs.acs.org.
(12) Trost, B. M.; Toste, F. D., J. Am. Chem. Soc., in press. Cf. Burckhardt,
U.; Baumann, M.; Togni, A. Tetrahedron: Asymmetry 1997, 8, 155.
(13) Schmidt, B.; Hoffmann, H. M. R. Tetrahedron 1991, 47, 9357.
(14) Lutz, C.; Knochel, P. J. Org. Chem. 1997, 62, 7895.
(15) Asao, T.; Buchi, G.; Abdel-Kader, M. M.; Chang, S. B.; Wick, E. L.;
Wogan, G. N. J. Am. Chem. Soc. 1965, 87, 882.
(16) Thomsen, I.; Torssell, K. B. G. Acta Chem. Scand. 1991, 45, 539.
(17) Wilen, S. H.; Shen, D.; Licata, J. M.; Balwin, E.; Russell, C. S.
Heterocycles 1984, 22, 1747.
JA9844229
(18) Kobayshi, S. Moriwaki, M.; Hachiya, I. Tetrahedron Lett. 1996, 37,
4183.
(19) Cuiper, A. D.; Kellogg, R. M.; Feringa, B. L. Chem. Commun. 1998,
655.