Investigation of methods for seven-membered ring synthesis: A practical synthesis of 4-oxo-5,6,7,8-tetrahydro-4h-cyclohepta[b]furan-3-carboxylic acid

Article abstract of DOI:10.1021/op0102159

Several synthetic routes to 4-oxo-5,6,7,8-tetrahydro-4H-cyclohepta[b]furan-3-carboxylic acid (1) are described, and the scale-up issues with each route are discussed. Seven-membered ring formation is a key issue with these syntheses, and several strategies are presented, including preparation from cycloheptane-1,3-dione, ring-expansion routes, Dieckmann cyclization, acetylene-furan [4 + 2] cycloaddition, and Friedel-Crafts cyclization. Two of the routes were scaled in the pilot plant to provide kilogram quantities of the title compound. The first scale-up route is outlined in Scheme 2 and utilizes a ring-expansion strategy to prepare cycloheptane-1,3-dione from cyclopentanone, via a [2 + 2] cycloaddition between dichloroketene and the silyl enol ether of cyclopentanone. The diketone is converted to the title compound by condensation with ethyl bromopyruvate and base, followed by acid hydrolysis. This route was efficient on laboratory scale but encountered problems upon scale-up due to a competing fragmentation pathway in the Zn/AcOH-mediated retro-aldol of cyclobutanone 11. The second, more successful scale-up route is described in Scheme 15, and involves Friedel-Crafts acylation of 3-carboethoxyfuran selectively at the 5-position. Reduction, lactonization, and hydrogenolysis provide acid 43, which is cyclized via a second Friedel-Crafts reaction to form the seven-membered ketone.

Full text of DOI:10.1021/op0102159

Organic Process Research & Development 2001, 5, 498−507
Investigation of Methods for Seven-Membered Ring Synthesis: A Practical
Synthesis of 4-Oxo-5,6,7,8-tetrahydro-4H-cyclohepta[b]furan-3-carboxylic Acid
John A. Ragan,*,1 Jerry A. Murry,2 Michael J. Castaldi, Alyson K. Conrad, Brian P. Jones, Bryan Li,
Teresa W. Makowski, Ruth McDermott, Barb J. Sitter, Timothy D. White, and Gregory R. Young
Chemical Research and DeVelopment, Pfizer Global Research & DeVelopment, Eastern Point Road,
Groton, CT 06340, U.S.A.
Abstract:
Several synthetic routes to 4-oxo-5,6,7,8-tetrahydro-4H-cyclo-
hepta[b]furan-3-carboxylic acid (1) are described, and the scale-
up issues with each route are discussed. Seven-membered ring
formation is a key issue with these syntheses, and several
Figure 1.
strategies are presented, including preparation from cyclohep-
theses of this diketone have been reported,3 but none were
deemed suitable for large-scale preparations. For example,
the synthesis shown in Scheme 13f involves a stoichiometric
oxymercuration of alkene 4, and issues with organomercurial
intermediates (including worker safety, tank contamination,
and potential contamination of clinical supplies) discouraged
use of this reaction beyond laboratory scale. Consequently,
while this route worked well for preparing 10-100 g
quantities of 1 to support in vitro studies of the drug
candidate derived from this intermediate, a new route was
needed for preparation of the kilogram quantities of cGMP
(current Good Manufacturing Practice) material that would
be needed for further progression of the drug candidate.
The conversion of cyclic 1,3-diketones to 3-carboxyfurans
is well-precedented for six-membered rings (e.g., cyclohex-
ane-1,3-dione4 and 5,5-dimethylcyclohexane-1,3-dione),5 and
this methodology worked well for the seven-membered
analogue (7 to 8 to 1, Scheme 1). Thus, our initial approach
was to identify a practical, scaleable synthesis of cyclohep-
tane-1,3-dione. This led to the development of a three-step
synthesis of cycloheptane-1,3-dione from cyclopentanone6
and thus an overall five-step conversion of cyclopentanone
to 1 (Scheme 2). This route was used to provide the first
multikilogram, cGMP bulk lots of 1.
tane-1,3-dione, ring-expansion routes, Dieckmann cyclization,
acetylene-furan [4 + 2] cycloaddition, and Friedel-Crafts
cyclization. Two of the routes were scaled in the pilot plant to
provide kilogram quantities of the title compound. The first
scale-up route is outlined in Scheme 2 and utilizes a ring-
expansion strategy to prepare cycloheptane-1,3-dione from
cyclopentanone, via a [2 + 2] cycloaddition between dichloro-
ketene and the silyl enol ether of cyclopentanone. The diketone
is converted to the title compound by condensation with ethyl
bromopyruvate and base, followed by acid hydrolysis. This
route was efficient on laboratory scale but encountered prob-
lems upon scale-up due to a competing fragmentation pathway
in the Zn/AcOH-mediated retro-aldol of cyclobutanone 11. The
second, more successful scale-up route is described in Scheme
15, and involves Friedel-Crafts acylation of 3-carboethoxyfuran
selectively at the 5-position. Reduction, lactonization, and
hydrogenolysis provide acid 43, which is cyclized via a second
Friedel-Crafts reaction to form the seven-membered ketone.
Introduction
4-Oxo-5,6,7,8-tetrahydro-4H-cyclohepta[b]furan-3-car-
boxylic acid (1) is an intermediate in the synthesis of a
clinical drug candidate, and we thus required access to
multikilogram quantities of this bicyclic keto-furan (Figure
1).
A central issue for all approaches investigated was the
strategy for formation of the seven-membered ring (e.g.
several approaches struggled with the kinetic preference for
formation of five-membered rings). Strategies described
herein include furan annulation onto a preexisting seven-
membered ring, ring-expansion strategies (both [5 + 2] and
[6 + 1]), intramolecular acetylene-furan [4 + 2] cyclo-
addition, and Friedel-Crafts cyclization onto a furan. The
advantages and limitations of each strategy will be discussed,
and the two routes which were scaled in the pilot plant will
be presented.
The [2 + 2] route proceeds well on laboratory scale:
starting with 19 g of cyclopentanone, 15 g of 1 was obtained
(33% yield over five steps, an average of 80% per step).
(1) To whom correspondence should be addressed. Pfizer Global Research &
Development, P.O. Box 8156-101, Groton, CT, 06335, USA. Telephone:
(860)441-6334.Fax: (860)441-5445.E-mail: john_a_ragan@groton.pfizer.com.
(2) Current address: Chemical Process Development, Merck & Co., Rahway,
NJ.
(3) (a) Eistert, B.; Haupter, F.; Schank, K. Liebigs Ann. Chem. 1963, 665, 55-
67. (b) Maclean, I.; Sneeden, R. P. A. Tetrahedron 1965, 21, 31-34. (c)
Hutmacher, H.-M.; Kruger, H.; Musso, H. Chem. Ber. 1977, 110, 3118-
3125. (d) Suzuki, M.; Watanabe, A.; Noyori, R. J. Am. Chem. Soc. 1980,
102, 2095-2096. (e) Nishiguchi, I.; Hirashima, T. Chem. Lett. 1981, 551-
554. (f) Bhushan, V.; Chandrasekaran, S. Synth. Commun. 1984, 14, 339-
345.
(4) Kneen, G.; Maddocks, P. J. Synth. Commun. 1986, 16, 1635-1640.
(5) Nagarajan, K.; Talwalker, P. K.; Goud, A. N.; Shah, R. K.; Shenoy, S. J.;
Desai, N. D. Indian J. Chem., Sect. B 1988, 27B, 1113-1123.
(6) Ragan, J. A.; Makowski, T. W.; am Ende, D. J.; Clifford, P. J.; Young, G.
R.; Conrad, A. K.; Eisenbeis, S. A. Org. Process Res. DeV. 1998, 2, 379-
381.
Discussion
The first synthesis of 1 is shown in Scheme 1 and
proceeds through cycloheptane-1,3-dione (7). Several syn-
498
•
Vol. 5, No. 5, 2001 / Organic Process Research & Development
10.1021/op0102159 CCC: $20.00 © 2001 American Chemical Society and The Royal Society of Chemistry
Published on Web 07/25/2001

Scheme 1
Scheme 2
chloro reduction. If both chlorines are reduced while the silyl
ether is still intact (e.g., 12), then hydrolysis and retro-aldol
are highly selective for the seven-membered diketone (path
A). However, if the silyl ether is cleaved while one (or both)
chlorines are still intact (e.g., 14), then hydrolysis and retro-
aldol cleavage favors formation of cyclopentanone 15 (path
B),7 presumably due to the increased acidity of the dichloro-
methyl ketone. We tested this theory by subjecting dichloro-
cyclobutanone 11 to aqueous acid in the absence of zinc; as
expected, clean formation of dichloroacetylcyclopentanone
(15) was observed by GC/MS and 1H NMR.7 Subjection of
this material to Zn/AcOH reduction generated acetylcyclo-
pentanone as the sole product. Since heat-transfer processes
can differ significantly based on reaction volume,8 we felt
that this strong temperature dependence for byproduct
formation, coupled with the lack of recrystallization methods
for product purification, did not signify a robust process.
Given the problems with the [2 + 2] route, particularly
the lack of crystalline intermediates and the instability of
cycloheptane-1,3-dione itself,9 a route which bypassed this
diketone altogether was an attractive goal. Several strategies
were investigated, as outlined below.
Dieckmann Cyclization. Scheme 4 shows the retrosyn-
thetic analysis for this approach, in which base-mediated
cyclization of tri-ester 21 was envisioned to provide seven-
membered keto-ester 18. Hydrolysis and decarboxylation
would then provide 1.
Preparation of known10 tri-ester 21 was achieved via a
furan annulation similar to that used for conversion of
cycloheptane-1,3-dione to 1 (Scheme 5). Thus, known keto
ester 2011 was treated with ethyl bromopyruvate and KOH
The sequence was operated without purification of any
intermediates, all of which are oils. 1 was formed as a brown
solid, which was purified by trituration with methyl tert-
butyl ether to provide the product as a tan, free-flowing solid.
Upon scale-up, however, the overall yield to 1 dropped
significantly: starting from 28.5 kg of cyclopentanone, 3.35
kg of 1 was obtained, representing an overall yield of just
5%. Moreover, isolation of solid 1 required a tedious series
of silica gel filtrations, which did not bode well for further
scale-up. The major contaminant in the crude product was
6-oxoheptanoic acid (17), which was traced back to acetyl-
cyclopentanone (16) formed in the Zn/AcOH reduction
(Scheme 3). Subjection of acetylcyclopentanone to the final
aqueous acid conditions cleanly formed 6-oxoheptanoic acid
via a retro-Claisen process. These byproducts had been
observed at significantly lower levels on laboratory scale (ca.
10%, vs 40-45% upon scale-up). The lack of crystalline
intermediates forced us to carry these byproducts through
the sequence, which significantly complicated the purification
of 1, thus lowering the overall yield.
Reexamination of the Zn/AcOH reduction of dichloro-
cyclobutanone 11 on laboratory scale demonstrated a strong
temperature dependence for the amount of acetylcyclopen-
tanone formed. When kept at or below 0 °C, 5-10% of this
byproduct was observed. However, when the reaction was
run at 40-50 °C, significantly higher levels (40-50%) of
the byproduct were formed. Our reasoning for this, as
outlined in Scheme 3, is the temperature dependence of two
competing processes: hydrolysis of the silyl ether versus
(7) Krepski, L. R.; Hassner, A. J. Org. Chem. 1978, 43, 3173-3179.
(8) Due to the different scale dependencies of surface area (m2) and reaction
volume (m3), and since the heat transfer capacity of a given reaction vessel
is roughly proportional to its surface area.
(9) Cyclohexane-1,3-dione and cycloheptane-1,3-dione possess remarkably
different physical properties: the former is a stable, white, crystalline solid,
whereas the seven-membered ring analogue is a colorless oil which
decomposes if stored for several weeks at room temperature. Their 1H NMR
properties in CDCl3 are also quite different: the six-membered diketone is
completely enolized, whereas the seven-membered diketone is in the diketo
form. Transannular interactions or some other form of ring strain is
presumably responsible for this difference in enolization propensity, and
may explain the greater reactivity (and thus instability) of the seven-
membered ring diketone. House has reported similar observations with
2-phenylcycloheptane-1,3-dione: House, H. O.; Wasson, R. L. J. Am. Chem.
Soc. 1956, 78, 4394-4400.
(10) Archer, S.; Pratt, M. G. J. Am. Chem. Soc. 1944, 66, 1656-1659.
(11) Thoma, H.; Spiteller, G. Liebigs Ann. Chem. 1983, 7, 1237-1248.
Vol. 5, No. 5, 2001 / Organic Process Research & Development
•
499

Scheme 3
Scheme 4
Scheme 6
Scheme 5
carbonyl band at 1745 cm-1, suggested that cyclopentanone
22 had instead been formed (this presumably occurs via
deprotonation of the methylene adjacent to the furan and
cyclization onto the ester carbonyl five atoms removed).
This is reminiscent of the reaction shown in Scheme 6,13
in which base treatment of diethyl-3-oxo-octane-1,8-dioate
generates cyclopentanone 24 instead of cycloheptanone 23.
This provides a dramatic example of the Curtin-Hammett
principle, in which seven-membered ring formation from a
more stable â-keto-ester anion fails to compete with five-
membered ring cyclization from a thermodynamically less
acidic position.
Ring Expansion. The analogous six-membered keto-furan
(25) is readily available from cyclohexane-1,3-dione,4 which,
unlike cycloheptane-1,3-dione, is an inexpensive, com-
mercially available, crystalline solid. We had investigated
several ring expansion approaches to cycloheptane-1,3-dione,
but none were superior to the [2 + 2] method described
previously (Scheme 2). However, a ring expansion subse-
quent to the furan annulation (Scheme 7) was viewed as an
attractive alternative for several reasons (e.g. keto-furan 25
is a crystalline solid, and it was hoped that intermediates
derived from it would also be crystalline).
in MeOH and then hydrolyzed with 1 N H2SO4 to provide
tri-acid 19.12 Fischer esterification (HCl-MeOH) then
provided tri-ester 21. Treatment of this compound with
KOtBu in THF led to formation of a new compound
corresponding to loss of MeOH by GC/MS, and displaying
1H and 13C NMR data consistent with the desired keto-ester
18. However, saponification and attempted decarboxylation
did not form 1. The resulting diacid was surprisingly reluctant
to decarboxylate, throwing suspicion on our initial structural
assignment. Further characterization, particularly the IR
The silyl enol ether 26 could be prepared from ketone 25
by treatment with TMSCl, NaI, and Et3N in CH3CN (Scheme
8). This particular reagent and solvent combination14 was
critical for realizing high regioselectivity: for example, use
(13) Allan, A. W.; Sneeden, R. P. A. Tetrahedron 1962, 18, 821-826.
(14) Wenkert, E.; Arrhenius, T. S.; Bookser, B.; Guo, M.; Mancini, P. J. Org.
Chem. 1990, 55, 1185-1193.
(12) Hofmann, K. J. Am. Chem. Soc. 1944, 66, 51-53.
500
•
Vol. 5, No. 5, 2001 / Organic Process Research & Development

Scheme 7
Scheme 10
Scheme 11
Scheme 8
A conceptually similar approach was to cyclopropanate
enol ether 26 (Scheme 10); ring expansion under oxidative
conditions (e.g., FeCl3)18 would be expected to provide the
ring-expanded â-chlorocycloheptanone, which would elimi-
nate to form the seven-membered enone. Hydrogenation
would then provide the target ketone. Unfortunately, all
attempts to cyclopropanate enol ether 26 (e.g., CH2I2/
Zn(Cu), Simmons-Smith, etc.) gave a mixture of products,
in which starting material and mono- and bis-cyclopropa-
nation products were all observed by GC/MS. Oxidative
cleavage with FeCl3 then resulted in complex mixtures, in
which the desired enone or â-chloroketone were at best minor
components. Apparently the furan ring is too reactive towards
the reagents investigated to allow for regioselective cyclo-
propanation of the TMS enol ether olefin.
Scheme 9
[4 + 2] Cycloaddition. A fundamentally different strategy
for formation of the seven-membered ring is an intra-
molecular furan-acetylene [4 + 2] cycloaddition, outlined
retrosynthetically in Scheme 11. Such furan-acetylene [4 +
2] cycloadditions are well precedented in an intermolecular
sense,19 and Keay has reported the MeAlCl2-catalyzed
intramolecular cycloaddition of a silyl propynone-furan.20
Iwasawa has studied a similar intramolecular furan-acetylene
cycloaddition with a Co2(CO)8-complexed acetylene.21 In
addition, Jacobi has utilized intramolecular acetylene-oxazole
[4 + 2] cycloadditions in syntheses of norsecurinine22 and
stemoamide.23 In Jacobi’s work, the initial cycloadduct
extrudes CH3CN to form the furan product; application of
this strategy to our system appeared tenuous, however, as
Jacobi’s target is a 2-methoxyfuran (which is ultimately
converted to a 2-furanone). Our hope was that extrusion of
acetylene from the initial cycloadduct would lead to the target
of LDA/TMSCl provided a 1:1 mixture of R- and γ-depro-
tonation products (26 and 27).
Stork has described the one carbon ring expansion of a
silyl enol ether by treatment with dichlorocarbene (generated
by thermolysis of trichloroacetic acid) and hydrolysis to form
the ring-expanded R-chloro enone.15 We attempted a similar
ring expansion with TMS enol ether 26 but largely without
success. The best conditions found were those of Stork:
sodium trichloroacetate in refluxing tetrachloroethylene-
DME (4:1) followed by hydrolysis with 1 N HCl in THF
delivered a mixture of products, from which were isolated
R-chloro enone 29 (6%) and tropinone 30 (13%) (Scheme
9). Numerous alternative carbene-forming conditions were
less successful (e.g. CHBr3/KOtBu,16 Cl3CCO2Et/NaOEt17).
(17) (a) Parham, W. E.; Schweizer, E. E. J. Org. Chem. 1959, 24, 1733-1735.
(b) Duggan, A. J.; Hall, S. S. J. Org. Chem. 1975, 40, 2234-2237.
(18) Ito, Y.; Fujii, S.; Saegusa, T. J. Org. Chem. 1976, 41, 2073-2074.
(19) Hoffmann, K. J. Am. Chem. Soc. 1944, 66, 51-53.
(20) Roger, C.; Keay, B. A. Can. J. Chem. 1992, 70, 2929-2947.
(21) Iwasawa, N.; Sakurada, F.; Iwamoto, M. Org. Lett. 2000, 2, 871-873.
(22) Jacobi, P. A.; Blum, C. A.; DeSimone, R. W.; Udodong, U. E. S. J. Am.
Chem. Soc. 1991, 113, 5384-5392.
(15) Stork, G.; Macdonald, T. L. J. Am. Chem. Soc. 1975, 97, 1264-1265.
(16) Last, L. A.; Fretz, E. R.; Coates, R. M. J. Org. Chem. 1982, 47, 3211-
3219.
(23) (a) Jacobi, P. A.; Lee, K. J. Am. Chem. Soc. 1997, 119, 3409-3410. (b)
Jacobi, P. A.; Lee, K. J. Am. Chem. Soc. 2000, 122, 4295-4303.
Vol. 5, No. 5, 2001 / Organic Process Research & Development
•
501

Scheme 12
Scheme 13
Scheme 14
furan, although it was not clear if this would occur at
temperatures obtainable in our pilot plant (i.e., <130 °C).
Preparation of the requisite cyclization substrate is shown
in Scheme 12. Known alcohol 3424 was oxidized with
TPAP,25 and aldehyde 35 was treated with the lithium anion
of ethyl propynoate. Oxidation with MnO2 generated the
corresponding ketone by GC/MS, and a second, isomeric
product; warming the reaction mixture effected complete
conversion to the second product. We initially suspected this
product to be keto-phenol 37, but further characterization
indicated that tricyclic diene 32 had been isolated (compari-
to the 5-position of a 3-carboxyfuran. The benzylic nature
of this position facilitates subsequent reduction of the ketone.
Two variants of this basic strategy were developed; the
first route is shown in Scheme 14. Starting with the ethyl
ester of 3-carboxyfuran,28 acylation with monomethylglut-
arate was effected with trifluoroacetic anhydride (TFAA) and
phosphoric acid in acetonitrile to provide keto-furan 39 in
64% yield. Reduction of the ketone to the corresponding
methylene can be effected by treatment with Et3SiH and BF3‚
Et2O (82% yield on 4.5-g scale). Cyclization of diacid 40 is
then effected by treatment with TFAA and SnCl4, but this
reaction was capricious and frequently failed to go to
completion.
The inefficiency of the final Friedel-Crafts cyclization
in Scheme 14 was attributed to the presence of two
nonequivalent carboxylic acids; activation of the carboxylate
adjacent to the furan might lead to intermolecular oligomer-
ization processes. We reasoned that if only the desired
position was available for activation, a more efficient
cyclization could be realized. This strategy required a method
for differentiation of the two carboxylic acids, which was
achieved as outlined in Scheme 15.
1
son of H and 13C data from two closely related [4 + 2]
cycloadducts was particularly revealing20,26). Even following
24 h treatment at 100 °C in DMSO, no evidence of
conversion to the desired furan was seen. Evaluation of
higher temperatures to effect expulsion of C2H2 (or hydro-
genation followed by C2H4 expulsion) was not pursued, as
the practicality of scaling such conditions was of concern.
Friedel-Crafts Cyclization. The final and most suc-
cessful route to 1 was an intramolecular Friedel-Crafts
cyclization to form the seven-membered ring (Scheme 13).27
A key retrosynthetic manipulation was introduction of a
ketone adjacent to the furan, which allows introduction of
the side chain via an intermolecular Friedel-Crafts addition
(24) Foehlisch, B.; Herter, R. Chem. Ber. 1984, 117, 2580-2596.
(25) Griffith, W. P.; Ley, S. V.; Whitcombe, G. P.; White, A. D. J. Chem. Soc.,
Chem. Commun. 1987, 1625-1627.
(26) Aicart, M.; Mavoungou-Gomes, L. J. Heterocycl. Chem. 1985, 22, 921-
925.
(28) Although somewhat expensive on research scale ($2-3/g from Aldrich),
3-furoic acid was available to us as a raw material from an unrelated
development candidate, and we felt that the price would likely decrease if
larger orders were placed in the future.
(27) Danishefsky recently reported a similar intramolecular furan Friedel-Crafts
cyclization in his synthesis of Frondosin B: Inoue, M.; Frontier, A. J.;
Danishefsky, S. J. Angew. Chem., Int. Ed. 2000, 39, 761-764.
502
•
Vol. 5, No. 5, 2001 / Organic Process Research & Development

Scheme 15
Reduction of ketone 41 with NaBH4 provided alcohol
formation of the seven-membered ring. Two of these routes
have been utilized to provide multikilogram quantities of
material: the ketene [2 + 2] route to cycloheptane-1,3-dione
(Scheme 2) and the intramolecular Friedel-Crafts cyclization
(Scheme 15). The latter route was particularly efficient, and
is viewed as the optimal route to this compound identified
to date.
42,29 which cyclized to the corresponding δ-lactone upon
treatment with AcOH in toluene. Hydrogenolysis of the
pseudobenzylic C-O bond then provided carboxylic acid
43, in which the desired carboxylate was now differentiated
from the ethyl ester.
As predicted, Friedel-Crafts cyclization of this substrate
was more efficient, although significant optimization was
required. Several combinations of acyl activation (trifluo-
roacetate mixed anhydride, acid chloride), Lewis acid (AlCl3,
BF3‚OEt2, SnCl4, H3PO4, CF3SO3H), and solvent (dichlo-
romethane, CH3CN, nitrobenzene, dichloroethane) were
investigated. The major side products appeared to be higher-
molecular weight oligomers, from intermolecular Friedel-
Crafts processes which competed with the desired cyclization.
Minimization of these intermolecular processes was com-
plicated by the fact that the higher oligomers were generally
not observable by our HPLC method and evidenced them-
selves more in the form of tarring, particularly during
workup. Thus, an experiment would seem to have proceeded
cleanly to the desired product based on its HPLC profile,
but upon workup and purification, the isolated yield of 1
would be low. Inverse addition or high dilution conditions
were found to help minimize the formation of these oligo-
mers.
Optimal conditions were found to be formation of the
mixed anhydride with trifluoroacetic anhydride and Et3N,
followed by slow addition to triflic acid in dichloroethane
(DCE) at reflux (83 °C).30 This addition protocol minimized
intermolecular side reactions (running the reaction under
high-dilution conditions achieved the same effect), resulting
in a 62-68% isolated yield of 1 following aqueous hydroly-
sis of the ethyl ester.
Experimental Section
All chemicals were used as purchased unless otherwise
noted. Dimethylformamide (DMF) and tetrahydrofuran (THF)
were purchased in anhydrous form from Aldrich in “Sure-
Seal” glass bottles; all other solvents were reagent grade.
Reactions were run under a positive pressure of nitrogen
unless otherwise stated. H (400 MHz) and 13C (100 MHz)
1
NMR spectra were obtained on a Varian Unity+400
spectrometer equipped with two RF channels, indirect
detection, and pulsed-field gradients (z-axis only). 1H spectra
were acquired using 45° acquisition pulses, 3.0 s recycle
delay, and 16 scans at a resolution of 0.2 Hz/point. The
acquisition window was typically 6800 Hz, and processing
used 0.1 Hz line broadening. 13C spectra were acquired using
45° acquisition pulses, 0.7 s recycle delay, and 512 scans at
a resolution of 0.8 Hz/point. Proton decoupling was applied
at 3 W of power during acquisition, and 2 Hz line broadening
was used during processing. Mass spectral data was collected
on either a Hewlett-Packard GC/MS (electron impact ioniza-
tion), or a Micromass (Fisons) Platform II mass spectrometer
(atmospheric pressure chemical ionization). Thin-layer chro-
matography (TLC) was performed on precoated sheets of
60 F254 (Merck Art. 5719). Visualization was achieved by
UV, and by staining with iodine, phosphomolybdic acid, ceric
ammonium molybdate, or p-anisaldehyde solutions and
heating. Combustion analyses were performed by Schwarz-
kopf Microanalytical Laboratory, Woodside, NY.
Conclusions
In summary, a variety of methods for the preparation of
1 have been investigated, all of which struggled with
4-Oxo-5,6,7,8-tetrahydro-4H-cyclohepta[b]furan-3-car-
boxylic Acid (1). Method A (from Cycloheptane-1,3-
dione). To a 0 °C solution of cycloheptane-1,3-dione (18.5
g, 147 mmol, prepared as described previously6 without
distillation) in isopropyl alcohol (148 mL) was added ethyl
bromopyruvate (18.5 mL, 147 mmol) and K2CO3 (20.3 g,
147 mmol). The resulting slurry was warmed to room
temperature and stirred for 16 h. The reaction was quenched
(29) Initial reductions in EtOH provided a mixture of methyl and ethyl esters,
which were converted to the bis-ethyl ester 42 to simplify characterization.
On scale-up, MeOH was used to avoid this issue.
(30) Similar cyclization conditions were utilized by workers at Sakai Research
Laboratories: Chujo, I.; Masuda, Y.; Fujino, K.; Kato, S.; Mohri, S.; Ogasa,
T.; Kasai, M. Abstracts of Papers, 217th National Meeting of the American
Chemical Society, Anaheim, CA.; American Chemical Society: Washington,
DC, 1999; ORG 77.
Vol. 5, No. 5, 2001 / Organic Process Research & Development
•
503

by addition of 185 mL of water, and extracted with two 200-
mL portions of dichloromethane. Concentration provided a
pale yellow oil (GC/MS and 1H NMR analysis indicated this
to be hydroxy-ester 8), which was treated with 185 mL of 1
N H2SO4 and warmed in a 95 °C oil bath for 16 h. After
cooling to room temperature, the mixture was extracted with
two 300-mL portions of dichloromethane. The organic
extracts were washed with brine and concentrated to provide
24.8 g of a brown solid, which was purified by trituration
with 124 mL of hot methyl tert-butyl ether (with warming
on a steam bath). The resulting slurry was allowed to
granulate overnight at ambient temperature, and the solids
were collected by vacuum filtration to provide 14.9 g of a
tan, free-flowing solid (52% yield, 33% overall from
cyclopentanone): 1H NMR (CDCl3) δ 8.07 (s, 1H), 3.14-
3.11 (m, 2H), 2.90-2.87 (m, 2H), 2.05-1.94 (m, 4H). 13C
NMR (CDCl3) δ 202.5, 168.7, 162.3, 150.6, 120.2, 119.9,
43.1, 29.0, 23.5, 21.7. MS (CI): m/z 195 (M + 1, 100). Anal.
Calcd for C10H10O4: C, 61.85; H, 5.19. Found: C, 61.66;
H, 5.17.
4-Oxo-5,6,7,8-tetrahydro-4H-cyclohepta[b]furan-3-car-
boxylic Acid (1). Method B (Friedel-Crafts Cyclization
of 43). A 250-mL round-bottom flask was charged with 43
(10.0 g, 41.7 mmol) and 100 mL of dichloroethane. To this
solution was added trifluoroacetic anhydride (17.7 mL, 125
mmol), and the resulting solution was stirred at room
temperature for 25 min. A separate, 500-mL round-bottom
flask was charged with trifluoromethanesulfonic acid (3.7
mL, 42 mmol) and 100 mL of dichloroethane and was
warmed to reflux under nitrogen. The solution of mixed
anhydride from the first flask was then added to the triflic
acid solution via addition funnel over 40 min, and the
resulting solution was stirred at reflux for an additional 45
min. HPLC analysis at this point shows complete conversion
to the ethyl ester of 1. The solution was then treated with
100 mL of 1 N HCl, a distillation head was added, and the
bulk of the dichloroethane (ca. 200 mL) was removed by
distillation (this required ca. 2 h). HPLC analysis showed
complete conversion to 1. After cooling to room temperature,
the solution was diluted with brine (100 mL) and extracted
with two 100-mL portions of dichloromethane. The com-
bined organic extracts were filtered through a 12-g pad of
silica gel, rinsing with an additional 100 mL of dichloro-
methane. Concentration on the rotary evaporator provided
8.50 g of a brown solid, which was granulated overnight
with 40 mL of methyl tert-butyl ether. Filtration provided
4.86 g of 1 as a tan solid (60% yield). Concentration of the
filtrate and granulation in isopropyl ether provided an
additional 0.68 g (8%), for a total of 5.54 g (68%). HPLC
conditions: Symmetry C-8 reverse phase column, elution
with 30/70 acetonitrile/0.2% aqueous phosphoric acid, moni-
tor at 210 nm. Retention times were as follows: 43 (7.3 min),
ethyl ester of 1 (4.8 min), 1 (2.2 min).
and the solution stirred for 1 h at ambient temperature. A
separate 100-L nitrogen-purged reactor was charged with
triflic acid (650 mL, 8.4 mol) and 30 L of dichloroethane
and warmed to reflux (jacket temperature was 85 °C). The
gas outlet from this reactor was passed through a 5% NaOH
scrubber. The mixed anhydride in the first reactor was then
added to the triflic acid/DCE solution over a period of 45
min. After the addition, 10 L of trifluoroacetic anhydride/
DCE was distilled atmospherically from the reactor, until
the distillate temperature reached 83 °C. The reaction mixture
was cooled and analyzed for cyclization completion by
HPLC, which indicated complete conversion to the ethyl ester
of the desired product. HCl (30 L, 1 N) was added, and the
dichloroethane was removed by distillation at atmospheric
pressure. HPLC analysis indicated complete ester hydrolysis
to the desired product. The reaction mixture was cooled,
diluted with 30 L of brine, and extracted with three portions
of CH2Cl2 (30, 15, and 15 L, respectively). The combined
organic extracts were filtered through 4.0 kg of silica gel,
rinsing with an additional 16 L CH2Cl2. The filtrate was
concentrated at atmospheric pressure to a volume of ca. 15
L and gradually displaced with 12 L of methyl tert-butyl
ether, to a final volume of 10 L. After cooling to room
temperature, the product was isolated via filtration to provide
a free-flowing, tan solid (1.5 kg, 7.7 mol, 62% yield).
2-(4-Carboxybutyl)-furan-3,4-dicarboxylic Acid (19).
3-Oxooctanedioic acid methyl ester9 (1.7 g, 7.8 mmol) was
dissolved in 10 mL of MeOH and treated with ethyl
bromopyruvate (0.98 mL, 7.8 mmol) and KOH pellets (0.50
g, 7.8 mmol), initially in an ice bath (0 °C), with gradually
warming to room-temperature overnight. After 16 h, TLC
indicated complete consumption of starting material. The
reaction mixture was poured onto ice and extracted with CH2-
Cl2 (4 × 25 mL). The organic extracts were washed with
brine, dried over MgSO4, filtered, and concentrated to
provide 2.13 g of a reddish brown oil. This material was
treated with 16 mL of 1 N H2SO4 and warmed to reflux.
After 16 h, TLC indicated complete conversion to a more
polar material. The reaction mixture was cooled and extracted
with CH2Cl2 (5 × 25 mL), followed by 10:1 CH2Cl2-MeOH
(5 × 25 mL). The organic extracts were combined, dried
over MgSO4, filtered, and concentrated to provide the desired
triacid as an orange-white solid (1.23 g, 62% yield). This
material was suitable for the next reaction or could be further
purified by recrystallization from CH2Cl2 to provide a white,
free-flowing solid, albeit in modest recovery (25%). 1H NMR
(CD3OD) δ 8.22 (s, 1H), 3.11-3.06 (m, 2H), 2.34-2.28 (m,
2H), 1.76-1.67 (m, 2H), 1.65-1.57 (m, 2H). 13C NMR
(CD3OD) δ 177.2, 168.3, 168.2, 167.2, 150.8, 118.7, 112.4,
34.4, 28.5, 28.2, 25.4.
2-(4-Methoxycarbonylbutyl)-furan-3,4-dicarboxylic Acid
Dimethyl Ester (21). 2-(4-Carboxybutyl)-3,4-dicarboxyfuran
(19) (0.77 g, 3.0 mmol) was dissolved in 15 mL of MeOH
and treated with 5 drops of concentrated H2SO4 and then
warmed to reflux for 3 h. After cooling to room-temperature
overnight, the solution was diluted with 50 mL of 1:1
isopropyl ether-ethyl acetate, washed with aqueous NaHCO3
(2 portions) and brine, dried over MgSO4, filtered, and
4-Oxo-5,6,7,8-tetrahydro-4H-cyclohepta[b]furan-3-car-
boxylic Acid (1). Method C (Scale-Up of Method B). A
clean, dry, nitrogen-purged 50-L reactor was charged with
43 (3.00 kg, 12.5 mol) and 30 L of dichloroethane.
Trifluoroacetic anhydride (5.3 L, 38 mol) was then added,
504
•
Vol. 5, No. 5, 2001 / Organic Process Research & Development

concentrated to provide 0.83 g of a clear, orange oil. Column
chromatography (20 mm column, eluting with 7:1 to 5:1
hexanes-ethyl acetate) provided the desired product as a
clear, yellow oil (0.36 g, 1.20 mmol, 40% yield). Spectral
data were consistent with that previously reported.10
Rf ) 0.23 was 4-oxo-4H-cyclohepta[b]furan-3-carboxylic
acid methyl ester (30): 183 mg clear, orange oil (0.90 mmol,
13% yield). H NMR (CDCl3) δ 7.94 (s, 1H), 7.36 (d, J )
11 Hz, 1H), 7.10 (dd, J ) 12, 8 Hz, 1H), 7.02 (d, J ) 12
Hz, 1H), 6.83 (dd, J ) 11, 8 Hz, 1H), 3.86 (s, 3H). MS
(CI): m/z 205 (M + 1, 100).
1
2-(2-Oxocyclopentyl)-furan-3,4-dicarboxylic Acid Di-
methyl Ester (22). 2-(4-Methoxycarbonylbutyl)-furan-3,4-
dicarboxylic acid dimethyl ester (21) (155 mg, 0.52 mmol)
was dissolved in 5 mL of THF and treated with KOtBu (117
mg, 1.04 mmol). After 20 h at ambient temperature, the
solution was quenched with 5 mL of 1 N HCl, and extracted
with two 10-mL portions of EtOAc. The combined organics
were washed with brine, dried over MgSO4, filtered, and
concentrated to provide 154 mg of a dark red oil. The major
nonbaseline spot was isolated by flash chromatography
(gradient elution from 4:1 to 2:1 hexanes-EtOAc), to provide
the title compound as a clear, pale yellow oil (23 mg, 0.086
5-(2-Furanyl)-pentanal (35). To a solution of 5-furan-
2-yl-pentan-1-ol (537 mg, 3.48 mmol) in CH2Cl2 (25 mL)
were added 4 Å molecular sieves (1.75 g), N-methylmor-
pholine-N-oxide (612 mg, 5.22 mmol) and tetrapropylam-
monium perruthanate (122 mg, 0.35 mmol). After 60 min
no starting material was observed by TLC. The reaction
mixture was concentrated in vacuo to a thick slurry and
filtered through a pad of Celite. The filter cake was washed
with ethyl acetate (2 × 10 mL). The filtrate was concentrated
in vacuo to give the desired product as a light brown oil:
1
IR (film) 1723 cm-1; H NMR (CDCl3) δ 9.75 (t, J ) 1.7
1
mmol, 17%). IR (thin film): ν (cm-1) 1745. H NMR
Hz, 1 H), 7.29-7.28 (m, 1 H), 6.27-6.26 (m, 1 H), 5.99-
5.96 (m, 1 H), 2.66-2.60 (m, 2 H), 2.47-2.43 (m, 2 H),
1.70-1.62 (m, 4 H); 13C NMR (CDCl3) δ 202.7, 155.8,
141.1, 110.3, 105.2, 43.8, 27.9, 27.7, 21.7.
(CDCl3) δ 7.76 (s, 1H), 3.97-3.92 (m, 1H), 3.81 (s, 6H),
2.46-2.16 (m, 5H), 1.95-1.87 (m, 1H). 13C NMR (CDCl3,
results of a DEPT experiment are in parentheses, where C
is quarternary, CH is methine, CH2 is methylene, and CH3
is methyl) δ 213.4 (C), 163.3 (C), 162.3 (C), 158.5 (C), 146.3
(CH), 118.9 (C), 114.7 (C), 52.0 (CH3), 51.9 (CH3), 48.7
(CH), 37.7 (CH2), 29.2 (CH2), 21.0 (CH2). MS (CI): m/z
267 (M + 1, 25), 235 (M - OCH3, 100).
4-Trimethylsilyloxy-6,7-dihydro-benzofuran-3-carbox-
ylic Acid Methyl Ester (26). Keto ester 254 (283 mg, 1.46
mmol) was dissolved in 2 mL of CH3CN and treated with
Et3N (0.25 mL, 1.8 mmol), NaI (273 mg, 1.8 mmol), and
TMSCl (0.23 mL, 1.8 mmol). After 48 h, the reaction mixture
was diluted with 1:1 hexanes-ethyl acetate, washed with
aqueous NaHCO3 and brine, dried over K2CO3, filtered, and
concentrated to provide the desired enol ether as a clear, pale
yellow oil (320 mg, 1.20 mmol, 82%). 1H NMR (CDCl3) δ
7.73 (s, 1H), 4.78 (t, J ) 5 Hz, 1H), 3.76 (s, 3H), 2.65 (t, J
) 9 Hz, 2H), 2.42-2.36 (m, 2H), 0.19 (s, 9H). MS (EI):
m/z 266 (M + 1, 30), 251 (M-CH3, 100).
Dichlorocarbene Ring Expansion of TMS Enol Ether
26. TMS enol ether 26 (1.81 g, 6.81 mmol) and Cl3CO2Na
(2.53 g, 13.6 mmol) were combined in tetrachloroethylene
(16 mL) and dimethoxyethane (4 mL), and the resulting
slurry was warmed to reflux. After 4 h, the mixture was
cooled to room temperature and concentrated to provide a
viscous, dark brown oil. This material was dissolved in THF
(20 mL) and 1 N HCl (20 mL), and stirred at room
temperature. After 3 h, the solution was diluted with
isopropyl ether (50 mL), washed with brine (filtration through
Celite was required at this point to break an emulsion), dried
over MgSO4, filtered, and concentrated to provide 1.69 g of
a dark brown oil. Flash chromatography provided two
products.
Rf ) 0.40 was 5-chloro-4-oxo-7,8-dihydro-4H-cyclohepta-
[b]furan-3-carboxylic acid methyl ester (29): 93 mg clear,
pale yellow oil (0.39 mmol, 6% yield). 1H NMR (CDCl3) δ
7.61 (s, 1H), 7.03 (t, J ) 7 Hz, 1H), 3.80 (s, 3H), 3.01-
2.98 (m, 2H), 2.65-2.60 (m, 2H). MS (CI): m/z 241 (M +
1, 50).
8-Furan-2-yl-4-hydroxy-oct-2-ynoic Acid Ethyl Ester
(36). Lithium bis(trimethylsilyl)amide (4.13 mL of a 1.0 M
solution in THF, 4.13 mmol) was dissolved in THF (20 mL)
and cooled to -78 °C. Ethyl propynoate (405 mg, 4.13
mmol) dissolved in THF (10 mL) was added slowly via
addition funnel, and the resulting solution was stirred for 10
minutes. 5-(2-Furanyl)-pentanal (35) was dissolved in THF
(10 mL) and added slowly to the reaction mixture via additon
funnel. After 60 min no starting material was observed by
TLC. HCl (20 mL, 1 N) was added, and the reaction was
warmed to room temperature. The phases were separated,
and the aqueous portion was washed with EtOAc (2 × 25
mL). The combined organics were dried (Na2SO4), filtered,
and concentrated in vacuo to produce the desired product as
a brown oil (537 mg, 2.15 mmol, 52% yield): IR (film) 3407,
1
2236, 1710 cm-1; H NMR (CDCl3) δ 7.29 (dd, J ) 1.7,
0.9 Hz, 1 H), 6.27 (dd, J ) 2.9, 1.6 Hz, 1 H), 5.98 (dd, J )
2.9, 0.8 Hz, 1 H), 4.48 (dd, J ) 6.7, 6.7 Hz, 1 H), 4.24 (q,
J ) 7.1 Hz, 2 H), 2.64 (t, J ) 7.5 Hz, 2 H), 2.10 (br s, 1 H),
1.82-1.76 (m, 2 H), 1.71-1.64 (m, 2 H), 1.56-1.47 (m, 2
H), 1.31 (t, J ) 7.1 Hz, 3 H); 13C NMR (CDCl3) δ 156.1,
153.6, 141,0, 110.3, 105.1, 87.8, 76.9, 62.5, 62.2, 36.7, 28.0,
27.8, 24.7, 14.2. Anal. Calcd for C14H18O4: C, 67.18; H,
7.25. Found: C, 66.85; H, 7.23.
6-Oxo-12-oxatricyclo[7.2.1.01,7]dodeca-7,10-diene-8-
carboxylic Acid Ethyl Ester (32). 8-Furan-2-yl-4-hydroxy-
oct-2-ynoic acid ethyl ester (36) (50 mg, 0.20 mmol) was
dissolved in CH2Cl2 (10 mL), and MnO2 (695 mg, 7.99
mmol) was added. After the mixture stirred for 15 min, no
starting material was observed by TLC. The reaction was
filtered through a pad of Celite, the cake was washed with
CH2Cl2 (2 × 5 mL), and the filtrate was concentrated in
vacuo. NMR in DMSO-d6 showed a 1:1 mixture of two
products. After heating for 1 h at 100 °C, NMR showed
conversion to a single product, assigned as [4 + 2]
cycloadduct 32. IR (film) 1709, 1626 cm-1; 1H NMR
(CDCl3) δ 7.29 (dd, J ) 5.2, 1.9 Hz, 1 H), 7.02 (d, J ) 5.2
Vol. 5, No. 5, 2001 / Organic Process Research & Development
•
505

Hz, 1 H), 5.61 (d, J ) 1.9 Hz, 1 H), 4.20 (q, J ) 7.1 Hz, 2
H), 2.67-2.61 (m, 1 H), 2.54-2.48 (m, 1 H), 2.31-2.24
(m, 1 H), 2.22-2.15 (m, 1 H), 2.01-1.82 (m, 3 H), 1.67-
1.62 (m, 1 H); 13C NMR (CDCl3) δ 202.0, 165.7, 162.8,
148.8, 146.2, 143.6, 95.4, 83.6, 61.5, 43.4, 30.3, 25.8, 24.4,
14.2; MS (CI), m/z 249.2 (M+).
desired product as an off-white solid (4.47 g, 21.1 mmol,
1
82% yield): IR (film) 1692 cm-1; H NMR (DMSO-d6) δ
8.10 (s, 1 H), 6.33 (s, 1 H), 2.58 (t, J ) 7.1 Hz, 2 H), 2.20
(t, J ) 7.1 Hz, 2 H), 1.57-1.47 (m, 4 H); 13C NMR (DMSO-
d6) δ 175.1, 164.7, 157.8, 147.4, 120.8, 105.8, 33.9, 27.4,
27.3, 24.5; MS (CI), m/z 211.2 (M-).
Furan-3-carboxylic Acid Ethyl Ester (38).31 To a clean
and dry glass-lined reactor was charged 87 L of 2B ethanol
and 3-carboxyfuran (8.7 kg, 78 mol). Concentrated H2SO4
(4.3 L, ca. 77 mol) was added, and the resulting reaction
mixture was warmed to reflux for 16 h, at which point HPLC
analysis indicated complete conversion. The reaction was
cooled to below 75 °C and concentrated under reduced
pressure to approximately half the original volume (ca. 50
L). After cooling to room temperature, 87 L of CH2Cl2 was
added, followed by slow addition of a 10% aqueous NaHCO3
solution (125 L). After stirring for 30 min, the organic layer
was separated and concentrated under reduced pressure to
give the desired product as a brown oil (6.8 kg, 49 mmol,
63% yield), which was used without further purification in
the next step. 1H NMR data for this compound was consistent
with that reported previously.32
5-(4-Methoxycarbonylbutyryl)-furan-3-carboxylic Acid
Ethyl Ester (41). A clean, dry, 200-L reactor attached to a
5% NaOH off-gas scrubber was purged with nitrogen and
then charged with trifluoroacetic anhydride (22.5 L, 160
mol), monomethylglutarate (7.5 L, 60 mol), and furan-3-
carboxylic acid ethyl ester (38) (5.6 kg, 40 mol). The
resulting solution was warmed to a gentle reflux while
phosphoric acid (85% w/w in H2O, 500 mL, 7.3 mol) was
added dropwise. (Caution: exothermic!) The solution was
refluxed for an additional 2 h post-addition, at which point
HPLC analysis of an aliquot indicated complete conversion.
The reaction mixture was cooled to 15 °C, and water (56 L,
10 volumes) was slowly added at a rate such that the internal
temperature remained <35 °C. Ethyl acetate (57 L) was then
added, and the resulting mixture stirred for 30 min. The
layers were separated (addition of brine facilitated separation
of the layers), and the organic extract was carefully treated
with 10% aqueous NaHCO3 solution (sufficient to reach pH
8). The layers were separated (with brine additions as
needed), and the aqueous layer was extracted with an
additional 11 L of ethyl acetate. The combined organic layers
were concentrated under reduced pressure to provide the
product as a brown oil, which solidified upon standing (10.7
kg, 40 mol, 100% yield). This material was used in the next
reaction without further purification: IR (film) 1723, 1684
cm-1; 1H NMR (CDCl3) δ 8.09 (s, 1 H), 7.45 (s, 1 H), 4.30
(q, J ) 7.1 Hz, 2 H), 3.65 (s, 3 H), 2.89 (t, J ) 7.3 Hz, 2
H), 2.40 (t, J ) 7.3 Hz, 2 H), 2.02 (dd, J ) 7.3, 7.3 Hz, 2
H), 1.33 (t, J ) 7.1 Hz, 3 H); 13C NMR (CDCl3) δ 188.7,
173.8, 162.2, 153.1, 150.3, 121.7, 116.4, 61.3, 51.9, 37.5,
33.1, 19.1, 14.5; MS (CI), m/z 269 (M + H). HRMS (FAB)
m/e calcd for C13H16O6 (M + H) 269.1025, found 269.1023.
5-(4-Ethoxycarbonyl-1-hydroxybutyl)-furan-3-carbox-
ylic Acid Ethyl Ester (42). A clean, dry, glass-lined 200-L
reactor was charged with 58 L of 2B ethanol, 58 L of THF,
660 mL of methanol (16.3 mol), and 5-(4-methoxycarbon-
ylbutyryl)-furan-3-carboxylic acid ethyl ester (41) (10.7 kg,
40 mol). The reaction mixture was cooled to 0-5 °C, and
NaBH4 pellets were added in three portions (total of 616 g,
16.3 mol) such that the internal temperature remained below
5 °C. The reaction mixture was stirred for 6 h at 0-5 °C, at
which point HPLC analysis indicated complete reduction to
the secondary alcohol. 1H NMR and HPLC of a concentrated
aliquot were consistent with the desired alcohol. The toluene
solution was carried directly into the next reaction lacton-
ization.
5-(4-Carboxybutyryl)-furan-3-carboxylic Acid (39).
3-Carboethoxyfuran (50 g, 360 mmol) was dissolved in
acetonitrile (500 mL), and monomethyl glutarate (57 g, 390
mmol) and 85% phosphoric acid (27 mL) were added.
Trifluoroacetic anhydride (300 g, 1.43 mol) was added slowly
via addition funnel over 30 min to ensure the temperature
remained below 30 °C. After the mixture stirred overnight,
no starting material remained, and the reaction was cooled
to 0 °C. The solution was neutralized (pH 8) by slow addition
of 50% aqueous NaOH. This solution was stirred overnight.
Concentrated HCl was used to adjust to pH 4, and the
resulting mixture was extracted with EtOAc (3 × 1000 mL).
The combined organics were dried (Na2SO4), filtered, and
concentrated to about 1/10 volume and diluted with hexanes
(200 mL). The resulting heterogeneous solution was heated
on a steam bath until homogeneous and was then cooled
with vigorous stirring and solids began forming. This solution
was granulated overnight and then filtered to provide the
title compound as an off-white solid (52 g, 230 mmol, 64%
1
yield). IR (film) 3422, 1682 cm-1; H NMR (DMSO-d6) δ
8.16 (s, 1 H), 7.46 (s, 1 H), 2.81 (t, J ) 7.5 Hz, 2 H), 2.24
(t, J ) 7.5 Hz, 2 H), 1.76 (dd, J ) 7.5, 7.5 Hz, 2 H); 13C
NMR (DMSO-d6) δ 189.1, 175.0, 165.2, 152.2, 150.2, 140.0,
119.1, 37.5, 33.7, 20.0; MS (CI), m/z 225.1 (M-).
5-(4-Carboxybutyl)-furan-3-carboxylic Acid (40). 5-(4-
Carboxybutyryl)-furan-3-carboxylic acid (39) (5.81 g, 25.7
mmol) was suspended in dichloromethane (116 mL) followed
by addition of BF3‚OEt2 (43.8 g, 308 mmol) and triethylsilane
(12.0 g, 103 mmol). The resulting mixture was refluxed
overnight at which time the reaction was homogeneous. The
reaction was cooled to room temperature and quenched with
H2O (110 mL). This solution was stirred vigorously as solids
began crashing out. After 3 h the mixture was filtered and
washed with H2O. The solids were air-dried to provide the
An earlier run in straight EtOH (which upon workup trans-
esterified to the bis-ethyl ester) was purified to provide 5-(4-
ethoxycarbonyl-1-hydroxybutyl)-furan-3-carboxylic acid eth-
yl ester (42) for characterization: IR (film) 3459, 1717 cm-1;
1H NMR (CDCl3) δ 7.92 (s, 1 H), 6.57 (s, 1 H), 4.66 (t, J )
6.6 Hz, 1 H), 4.28 (q, J ) 7.1 Hz, 2 H), 4.10 (q, J ) 7.1 Hz,
(31) Kubota, T.; Matsuura, T. J. Chem. Soc. 1958, 3667-3673.
(32) Divald, S.; Chun, M. C.; Jouillie, M. M. J. Org. Chem. 1976, 41, 2835-
2846.
506
•
Vol. 5, No. 5, 2001 / Organic Process Research & Development

2 H), 3.66 (s, 1 H), 2.42-2.31 (m, 2 H), 1.95-1.64 (m, 4
H), 1.32 (t, J ) 7.1 Hz, 3 H), 1.23 (t, J ) 7.1 Hz, 3 H); 13
NMR (CDCl3) δ 173.8, 163.4, 158.1, 147.2, 120.2, 106.2,
67.3, 60.8, 60.7, 34.8, 34.0, 20.9, 14.5, 14.4.
the layers were separated, and the toluene layer was extracted
with an additional 15 L of saturated aqueous NaHCO3. The
combined aqueous extracts were washed with toluene (2 ×
30 L) and then adjusted to pH 1.5 by careful addition of
concentrated HCl. The resulting solids were filtered, and the
filter cake was rinsed with 10 L of water. After drying (45
°C vacuum oven for 48 h, Karl Fischer analysis showed
<0.1% water), the desired acid was obtained as a pale yellow
solid (4.0 kg, 16.7 mol, 42% over three steps, from 41: IR
C
5-(4-Carboxybutyl)-furan-3-carboxylic Acid Ethyl Es-
ter (43). The solution of 5-(4-methoxycarbonyl-1-hydroxy-
butyl)-furan-3-carboxylic acid ethyl ester in toluene from
above (ca. 40 mol) was treated with 10 L of acetic acid, and
warmed to reflux at a rate whereby 10 L of distillate per
hour was collected. After 6 h, an additional 90 L of toluene
and 10 L of acetic acid were added, and the distillation was
continued. After an additional 60 L of distillate was collected,
lactonization was complete by HPLC analysis. Toluene (30
L) was added, and atmospheric distillation was continued
to a volume of ca. 20 L. Another 30 L of toluene was added,
and distillation continued until the distillate boiling point
reached 111 °C, indicating complete removal of acetic acid.
The toluene solution was used directly in the hydrogenolysis.
A clean, water-rinsed, glass-lined reactor was charged
with 590 g of 10% Pd/C under a nitrogen atmosphere,
followed by addition of the toluene solution of lactone from
above and an additional 165 L of toluene. The resulting
reaction mixture was subjected to 50 psi hydrogen for 40 h,
at which point HPLC analysis indicated complete conversion.
The mixture was filtered through Celite, concentrated to a
volume of ca. 40 L, and 35 L of saturated aqueous NaHCO3
was carefully added. The mixture was stirred for 60 min,
1
(film) 3487, 1709 cm-1; H NMR (CDCl3) δ 7.86 (s, 1 H),
6.33 (s, 1 H), 4.25 (q, J ) 7.1 Hz, 2 H), 2.65-2.57 (m, 2
H), 2.40-2.30 (m, 2 H), 1.71-1.60 (m, 4 H), 1.31 (t, J )
7.1 Hz, 3 H); 13C NMR (CDCl3) δ 179.7, 163.9, 157.3, 146.5,
120.1, 105.3, 60.6, 33.9, 27.7, 27.3, 24.2, 14.5. Anal. Calcd
for C12H16O5: C, 59.99; H, 6.71. Found: C, 59.45; H, 6.90.
HRMS (FAB) m/e calcd for C12H16O5 (M + H) 241.1076,
found 241.1087.
Acknowledgment
David am Ende provided thermochemical data and helpful
insights on the Zn/AcOH reduction, and Frank Urban, Bob
Dugger, and Stephane Caron provided numerous helpful
suggestions on synthetic strategies.
Received for review May 1, 2001.
OP0102159
Vol. 5, No. 5, 2001 / Organic Process Research & Development
•
507