Scope and diastereoselectivity of the interrupted Feist-Benary reaction.

Article abstract of DOI:10.1021/ol0169978

[reaction: see text] The base-promoted condensation of beta-dicarbonyl compounds with alpha-haloketones, the Feist-Benary reaction, conveniently produces highly substituted dihydrofurans. We show here that this reaction is quite general with respect to the nature of the beta-dicarbonyl compound, proceeding with beta-ketoesters, beta-oxopropionates, beta-diketones, and beta-dialdehydes. We also show that the diastereoselectivity of this reaction depends on the acidity of the nucleophile.

Full text of DOI:10.1021/ol0169978

ORGANIC
LETTERS
2002
Vol. 4, No. 2
205-208
Scope and Diastereoselectivity of the
“Interrupted” Feist−Be´nary Reaction
Michael A. Calter* and Cheng Zhu
Department of Chemistry, UniVersity of Rochester, Rochester, New York 14627-0216
calter@chem.rochester.edu
Received November 3, 2001
ABSTRACT
The base-promoted condensation of â-dicarbonyl compounds with r-haloketones, the Feist−Be´nary reaction, conveniently produces highly
substituted dihydrofurans. We show here that this reaction is quite general with respect to the nature of the â-dicarbonyl compound, proceeding
with â-ketoesters, â-oxopropionates, â-diketones, and â-dialdehydes. We also show that the diastereoselectivity of this reaction depends on
the acidity of the nucleophile.
In a search for convenient methods of preparation of highly
oxidized intermediates for the synthesis of the zaragozic
acids, we became aware of the Feist-Be´nary reaction.1 This
reaction generally involved the base-promoted condensation
of â-dicarbonyl compounds with R-haloketones to produce
furans (Scheme 1). However, running the reaction under
scope of this “interrupted” reaction,2 and one group proposed
a relative stereochemistry for the dihydrofuran.2b However,
the diastereoselectivity of dihydrofuran formation had never
been rigorously determined or explained, nor had the
generality of this reaction with regard to dicarbonyl reactant
been explored. Therefore, we undertook a study to address
both of these issues.
We initially employed R-bromooxaloacetates as electro-
philes in combination with oxaloacetate nucleophiles (Scheme
2). We employed several conditions for this coupling. As
reported in the literature, the commercially available sodium
salt of diethyl oxaloacetate reacted with 0.5 equiv of bromine
in ether to afford tetraesters 1a,b.2a Similar treatment of the
sodium salt of di-tert-butyl oxaloacetate,3 formed in situ,
afforded the corresponding dihydrofurans 2a,b in moderate
yield and diastereoselectivity. We achieved similar results
using di-tert-butyl oxaloacetate in the presence of 1 equiv
of triethylamine and 0.5 equiv of bromine.
Scheme 1
(1) (a) Feist, F. Chem. Ber. 1902, 35, 1537-1544. (b) Be´nary, E. Chem.
Ber. 1911, 44, 489-492.
(2) (a) Dunlop, A. P.; Hurd, C. D. J. Org. Chem. 1950, 15, 1160-1164.
(b) Cantlon, I. J.; Cocker, W.; McMurry, T. B. H. Tetrahedron 1961, 15,
46-52.
(3) Heidelberger, C.; Hurlbert, R. B. J. Am. Chem. Soc. 1950, 72, 4704-
4706.
milder conditions allowed the isolation of a dihydrofuran
intermediate. Several groups studied the mechanism and
10.1021/ol0169978 CCC: $22.00 © 2002 American Chemical Society
Published on Web 12/28/2001

of the enolate with the chiral R-bromoketone (Scheme 3).
Such selectivity in additions to R-haloketones is well
documented4 and is rationalized by the electrostatic model
of Nguyen.5 After addition, rapid proton transfer steps
followed by cyclization led to the observed major diastere-
omer. This proposed mechanism predicted that the diaste-
reoselectivity could potentially erode if the rate of cyclization
was lowered sufficiently relative to the rate of retroaldol
reaction. This situation could arise if the pKa of the
nucleophile precursor was sufficiently low.
Scheme 2
To explore the scope of the reaction, we attempted a cross-
coupling reaction between a bromooxaloacetate and an
oxaloacetate bearing different ester substituents (Scheme 4).
Scheme 4
The reaction of diethyl oxaloacetate with bromo-di-tert-butyl
oxaloacetate (3b)6 produced a mixture of dihydrofurans with
all possible combinations of ester groups. This mixture
probably resulted from bromine transfer from 3b to the
enolate of the dicarbonyl compound, generating a mixture
of all possible bromooxaloacetates and oxaloacetates. Such
a potential side reaction is likely to occur whenever the
acidity of the dicarbonyl compound and the dehalogenated
electrophile are similar.
Cantalon assumed that the major tetraethyl ester stereo-
isomer, 1a, possessed the cis stereochemistry on the basis
of thermodynamic arguments.2b We were able to obtain single
crystals of 1a suitable for X-ray crystallographic analysis
and confirmed that this isomer did in fact possess the cis
1
stereochemistry. By analogy and comparison of H NMR
spectra, we assigned the stereochemistry of 2a as also being
cis. We performed several control experiments to establish
that the ratio of stereoisomers resulted from kinetic control.
For example, treatment of either 2a or 2b with sodium
methoxide, triethylamine, or DBU resulted in no intercon-
version.
We also attempted a cross-coupling reaction between a
bromooxaloacetate and a simple â-ketoester (Scheme 5). This
We propose that the kinetic selectivity for the cis-
dihydrofuran resulted from a diastereoselective aldol reaction
Scheme 5
Scheme 3
reaction did not produce any dihydrofuran products, again
illustrating the difficulty of carrying out this coupling reaction
(4) Cornforth, J. W.; Conforth, R. H.; Mathew, K. K. J. Chem. Soc. 1959,
2199-2204. For an example of an aldol addtion to a chiral R-chloroketone
that gives similar selectivity, see: Yasuda, M.; Oh-hata, T.; Shibata, I.;
Baba, A.; Matsuda, H.; Sonoda, N. Bull. Chem. Soc. Jpn. 1995, 68, 1180-
1186.
(5) Nguyen, T. A. Top. Curr. Chem. 1980, 146-162.
(6) General procedure for the preparation of bromo-di-alkyl oxalo-
acetates: Conover, L. H.; Tarbell, D. S. J. Am. Chem. Soc. 1950, 72, 5221-
5225.
206
Org. Lett., Vol. 4, No. 2, 2002

with a â-dicarbonyl nucleophile having an acidity too close
to that of the debrominated electrophile. Therefore, we turned
to nucleophiles that were more acidic than â-ketoesters.
3-Oxopropionates functioned as suitable nucleophiles for
the reaction with bromooxaloacetates (Scheme 6). Condensa-
and 8a,b. However, in contrast to the reactions involving
all previous nucleophiles, these condensations afforded the
trans stereoisomer with low selectivity. The stereochemistry
of 7b was confirmed by conversion into enone 9 and X-ray
crystallographic analysis of this compound.
The turnover in selectivity for the dialdehyde reaction
probably resulted from the relatively low pKa of the nucleo-
phile precursor. Making the nucleophile precursor more
acidic had two consequences: the retro-aldol reaction became
more favorable, and the cyclization became less favorable.
The combination of these factors most likely led to cycliza-
tion rather than addition being the product-determining step
for the dialdehyde reaction. Apparently, cyclization favored
the trans isomer to a small degree. To reverse the selectivity,
we resorted to use of the preformed sodium enolate under
anhydrous conditions (Scheme 8).11 These conditions resulted
Scheme 6
Scheme 8
tion of ester aldehyde 47 with 3b yielded dihydrofuran 5.
The stereochemistry of 5 was confirmed by X-ray crystal-
lographic analysis of 6, produced by dihydroxylation of 5
with RuO4,8 followed by lactol oxidation with N-iodosuc-
cinimide (NIS) and tetrabutylammonium iodide (TBAI).9
â-Dialdehydes also served as nucleophiles for the Feist-
Be´nary reaction (Scheme 7). Formation of an aqueous
in a modest selectivity for the cis-isomer but were also
accompanied by a moderate amount of bromine transfer and
homocoupling to yield tetraester 1a.
Scheme 7
Finally, we assayed cyclic â-diketones as the nucleophilic
component in the reaction with bromooxaloacetates (Scheme
9). The condensation of 1,3-cyclohexadione with 3a yielded
Scheme 9
dihydrofuran 10. This particular â-diketone nucleophile
yielded the trans dihydrofuran as the major product. The
stereochemistry of the major product was confirmed by X-ray
crystallographic analysis. The acidity of the diketone was
not sufficient to explain the trans selectivity. The 3-oxopro-
pionate likely had a similar pKa yet yielded the cis product
in high selectivity. As cyclohexadione was the sole cyclic
enolate precursor tested, it is possible that trans selectivity
solution of malondialdehyde,10 followed by condensation
with bromooxaloacetates 3a,b, yielded dihydrofurans 7a,b
(7) Sato, M.; Yoneda, N.; Katagari, N.; Watanabe, H.; Kaneko, C.
Synthesis 1986, 672-674.
(8) Shing, T. K. M.; Tai, V. W.-F.; Tam, E. K. W. Angew. Chem., Int.
Ed. Engl. 1994, 33, 2312-2313.
(9) Hanessian, S.; Wong, D. H.-c.; Therien, M. Synthesis 1981, 394-
395.
(10) Bond, A. M.; Deprez, P. P.; Jones, R. D.; Wallace, G. G.; Briggs,
M. H. Anal. Chem. 1980, 52, 2211-2213.
(11) Go´mez-Sa´nchez, A.; Hermos´ın, I.; Lassaletta, J.-M.; Maya, I.
Tetrahedron 1993, 49, 1237-1250.
Org. Lett., Vol. 4, No. 2, 2002
207

resulted from the restriction of the intermediate enolate to
the E-geometry.
Chemistry for support of this work. We also thank Ms.
Xiahong Zhang for preliminary experiments toward the cross-
coupling reactions, and Dr. Carla Slebodinck (Virginia Tech)
and Dr. Rene Lachicotte and Prof. William Jones (University
of Rochester) for performing the X-ray crystal structure
analyses.
In conclusion, we have proven that the “interrupted”
Feist-Be´nary reaction generally yields the cis isomer with
moderately acidic nucleophiles and the trans isomer with
more highly acid ones. We have proposed an explanation
for this selectivity, based on addition or cyclization being
the product-determining step. We have shown that a variety
dicarbonyl compounds with appropriate acidities function as
nucleophiles in this reaction. We are currently applying this
reaction to the synthesis of the zaragozic acid core and
developing an asymmetric version of the reaction.
Supporting Information Available: Full experimental
procedures and characterization data for 1a, 2a, 5, 6, 7a,b,
8a,b, 9 and 10 and representations of the X-ray crystal
structures of 1a, 6, 9, and 10.
Acknowledgment. We thank the Eastman Kodak Cor-
poration and the University of Rochester Department of
OL0169978
208
Org. Lett., Vol. 4, No. 2, 2002