E- or Z-selective Knoevenagel condensation of acetoacetic derivatives: Effect of acylated substituent, that is, TEMPO and amines, as an auxiliary, and new accesses to trisubstituted E- and Z-2-alkenals and furans

Article abstract of DOI:10.1021/jo051952w

Knoevenagel condensation of O-acetoacetylTEMPOs (2,2,6,6- tetramethylpiperidine-1-oxyl) with aldehydes substituted with an electron-withdrawing group such as aromatic and heteroaromatic ones leads preferentially to E-adducts, while acylacetoamides includi

Full text of DOI:10.1021/jo051952w

E- or Z-Selective Knoevenagel Condensation of Acetoacetic
Derivatives: Effect of Acylated Substituent, that is, TEMPO and
Amines, as an Auxiliary, and New Accesses to Trisubstituted E- and
Z-2-Alkenals and Furans
Tsutomu Inokuchi*,† and Hiroyuki Kawafuchi‡
Department of Bioscience and Biotechnology, Faculty of Engineering, Okayama UniVersity,
Tsushima-naka, Okayama, 700-8530, Japan, and Department of Chemical and Biochemical Engineerings,
Toyama National College of Technology, Hongo-machi, Toyama, 939-8630, Japan
inokuchi@cc.okayama-u.ac.jp
ReceiVed September 17, 2005
Knoevenagel condensation of O-acetoacetylTEMPOs (2,2,6,6-tetramethylpiperidine-1-oxyl) with aldehydes
substituted with an electron-withdrawing group such as aromatic and heteroaromatic ones leads
preferentially to E-adducts, while acylacetoamides including Weinreb amides produce Z-adducts,
exclusively. These E- and Z-adducts are selectively converted to the corresponding (2E)- and (2Z)-2-
hyroxyalkyl-2-alkenals, respectively, by stepwise reductions of the acyl group with DIBALH and then
the carboxylic functions after protection of the hydroxy group. Transformation of the Knoevenagel products
by taking advantage of the E-geometry to trisubstituted furans is also developed.
Introduction
bioactive natural product syntheses.5 Thus far, to the best of
our knowledge, there is only one example where benzoyl acetate
was employed as the active methylene for the condensation with
benzaldehyde, leading to the trans relationship between the
phenyl group from benzaldehyde and the PhCO group in the
resulting adduct.6
In our continuing studies on exploring 2,2,6,6-tetramethylpi-
peridine-1-oxyl (TEMPO) substitution as a reaction-controlling
element,7 we examined Knoevenagel condensation of O-
acetoacetylTEMPO 1 with aldehydes 2 and discussed the effect
Knoevenagel condensation is highly useful for the synthesis
of functionalized enones and enoates, since addition reaction
of the methylene components, activated with two electron-
withdrawing groups, to aldehydes can be performed with
secondary amines.1 Although adducts with two different electron-
withdrawing groups are regarded to be highly versatile and
interconvertible trisubstituted olefins, stereocontrol of the reac-
tion has mainly relied on the steric effects of one of the
withdrawing groups over the other one, and hence the scope is
limited to a combination of a bulky/less bulky pattern like CO2R
versus CN,2 SO2Me versus CO2R,3 and PO(OR)2 versus CO2R.4
Furthermore, a stereoselective approach has been scarcely
achieved with respect to condensation of acetoacetic derivatives
with aldehydes,1a,2c despite such being more important in
(2) (a) Schwarz, M. J. Chem. Soc., Chem. Commun. 1969, 212-213.
(b) Zabicky, J. J. Chem. Soc. 1961, 683-687. (c) Tietze, L. F.; Beifuss, U.
Liebigs Ann. Chem. 1988, 321-329. Recently, Ti-enolate of acetoacetate
was shown to be useful for E-selectivity in some extent: (c) Hayanshi, M.;
Nakamura, N.; Yamashita, K. Tetrahedron 2004, 60, 6777-6783.
(3) Happer, D. A. R.; Steenson, B. E. Synthesis 1980, 806-807.
(4) Lehnert, W. Tetrahedron 1974, 30, 301-305.
(5) Z-isomers from acetoacetate derivatives were used in natural product
syntheses: (a) Hong, R.; Chen, Y.; Deng, L. Angew. Chem., Int. Ed. 2005,
44, 3478-3481. (b) Dineen, T. A.; Roush, W. R. Org. Lett. 2004, 6, 2043-
2046. (c) Munakata, R.; Katakai, H.; Ueki, T.; Kurosawa, J.; Takao, K.;
Tadano, K. J. Am. Chem. Soc. 2003, 125, 14722-14723.
(6) Tanikaga, R.; Konya, N.; Hamamura, K.; Kaji, A. Bull. Chem. Soc.
Jpn. 1988, 61, 3211-3216.
* Author to whom correspondence should be addressed.
† Okayama University.
‡ Toyama National College of Technology.
(1) Reviews: (a) Jones, G. Org. React. 1967, 15, 204. (b) Tietze, L. F.;
Beifuss, U. ComprehensiVe Organic Synthesis; Pergamon: New York, 1991;
Vol. 2, pp 341-394.
10.1021/jo051952w CCC: $33.50 © 2006 American Chemical Society
Published on Web 01/11/2006
J. Org. Chem. 2006, 71, 947-953
947

Inokuchi and Kawafuchi
TABLE 1. Effect of Acylated Substituents L and TEMPO on the
E/Z-Selectivity of Knoevenagel Condensationsa
SCHEME 1. Knoevenagel Condensations of O-Acetoacetyl
TEMPO
SCHEME 2. Preparation of O-AcetoacetylTEMPO and Its
Homologues
of 2,2,6,6-tetramethylpiperidine-1-oxyl (TEMPO) as an auxiliary
on the geometry of the resulting adducts 3 (Scheme 1). In the
meantime, other acyl substituents such as Me(MeO)N were also
employed for comparison. Subsequently, a new access to 2-(1-
hydroxyalkyl)-2-enals 31 of defined geometry, useful analogues
of the Morita-Baylis-Hillman reaction products,8 was devel-
oped by stepwise reductions of the acyl groups and then the
carboxylic functions of 3 and its congener. A novel method for
the synthesis of trisubstituted furans by taking advantage of
E-geometry of the Knoevenagel adducts is also explored.
entry
dicarbonyl
RCHO
product
yield (%)b
E/Z ratioc
1
2
3
4
5
6
7f
8
9
11
13
15
1
6
8
19
2a
2a
2a
2a
2a
2a
2b
2a
10a
12a
14a
16a
3a
17a
18b
20a
69
94
76
69
76
82
75
95
1:1.5
1:2.2
1:>50d
1:>50d
38:1e
20:1
3:1
>50:1d
a Carried out by using acetoacetic derivatives (1.0∼2.5 mmol) and 2
(1.5∼2.0 equiv) in EtOH or CH2Cl2 at room temperature with piperidine.
AcOH was added when the reaction was sluggish. b Yields are based on
isolated products. c Determined on the basis of 1H NMR. d One isomer was
exclusively formed. e 6:1 when carried out in the presence of AcOH.
f Conducted in MeCN at reflux for 12 h.
Results and Discussion
The starting O-acetoacetyl-(2,2,6,6-tetramethylpiperidine)-1-
oxyl (O-acetoacetylTEMPO, 1) was prepared either by addition
of the O-trimethylsilylTEMPO, generated by silylation of the
TEMPO anion,7a to diketene9 in 81% yield from TEMPO• or
by Claisen condensation of O-acetylTEMPO (4) with isopro-
penyl acetate with LDA as a base in 49% yield (Scheme 2).10
Other O-alkanoylacetylTEMPOs, that is, 6 and 8, were prepared
by aldol reaction of 4 with the respective aldehydes followed
by oxidation of the resulting aldol adducts 5 and 7 with the
Swern reagent or the TEMPO•/RuCl2(PPh3)3/O2 system.11
As shown in Table 1, we first examined the effect of the
acylated substituents, L, that is, t-BuO, t-BuS, R2N, and Me-
(MeO)N, and TEMPO, on the E/Z selectivity of Knoevenagel
condensation with 2-furaldehyde (2a). Thus, the piperidine-
catalyzed condensation of tert-butyl acetoacetate 9 with 2a
produced the corresponding adducts 10a as a 1:1.5 E/Z mixture
(entry 1).12 Similar Z-preference was found with thioester 11,
where the adduct 12a was obtained with a 1:2.2 E/Z mixture
(entry 2). Furthermore, to our surprise, the condensation of
N-acetoacetylmorpholine 13 and the Weinreb amide 1513 with
2a afforded the corresponding (Z)-14a and (Z)-16a, respectively,
with exclusive selectivity (entries 3 and 4). On the other hand,
gratifyingly, the condensation of O-acetoacetylTEMPO 1 with
2a resulted in a reversal of the E/Z ratio to 38:1, giving the
adducts (E)-3a in 76% yield (entry 5).
Since the preferential formation of the E-isomer with O-
acetoacetylTEMPO 1 was witnessed, we next examined other
O-alkanoylacetylTEMPOs, 6 and 8, to probe the effect of
bulkiness of the alkanoyl R1 on the E/Z selectivity. The
condensation of 2-methylpropanoyl derivative 6 with 2a pro-
duced adducts 17a with an E/Z ratio of 20:1 (entry 6). Thus,
no significant steric effect causing decrease of E-selectivity was
observed with the spatially spread isopropyl group on the
condensation. Furthermore, a similar E-preference was found
(7) (a) Inokuchi, T.; Kawafuchi, H. Tetrahedron 2004, 60, 11969-11975.
(b) Inokuchi, T.; Kawafuchi, H.; Nokami, J. Chem. Commun. 2005, 537-
539.
(8) (a) Basavaiah, D.; Rao, A. J.; Satyanarayana, T. Chem. ReV. 2003,
103, 811-891. (b) Langer, P. Angew. Chem., Int. Ed. 2000, 39, 3049-
3052. (c) Ciganek, E. Org. React. 1997, 51, 201-350. (d) Basavaiah, D.;
Rao, P. D.; Hyma, R. S. Tetrahedron 1996, 52, 8001-8062.
(9) Lawesson, S.-O.; Sandberg, R. Organic Syntheses; Wiley & Sons:
New York, 1973; Collect. 5, pp 155-157.
(10) Davis, B. R.; Garratt, P. J. ComprehensiVe Organic Synthesis;
Pergamon: New York, 1991; Vol. 2, pp 795-863.
(11) Inokuchi, T.; Nakagawa, K.; Torii, S. Tetrahedron Lett. 1995, 36,
3223-3226.
(12) (a) Inokuchi, T. J. Org. Chem. 2005, 70, 1497-1500. (b) Inokuchi,
T.; Okano, M.; Miyamoto, T. J. Org. Chem. 2001, 66, 8059-8063. (c)
Inokuchi, T.; Okano, M.; Miyamoto, T.; Madon, H. B.; Takagi, M. Synlett
2000, 1549-1552.
(13) Inglesi, M.; Nicola, M.; Magnetti, S. Farmaco 1990, 45, 1327-40.
948 J. Org. Chem., Vol. 71, No. 3, 2006

KnoeVenagel Condensation of Acetoacetic DeriVatiVes
3
3
TABLE 2. Coupling Constants of J(COOT, H) or J(COL, H) of
TABLE 3. Effect of Substituent R of Aldehydes 2 on the E/
Z-Selectivity of Knoevenagel Condensations of 1, 6, and 8a
Selected Knoevenagel Condensation Products
compounds
E-isomer
Z-isomer
a
3a
3b
3c
3e
3f
17a
18b
10a
12a
14a
16a
20a
7.8
7.9
12.5
a
7.5
a
11.4
a
12.4
a
7.7
8.1
a
7.8
11.7
11.0
11.5
a
a
a
10.4
a
11.2b
a Not characterized because of scant availability of compounds. b trans-
Relationship between COOT and vinylic proton.
even in the condensation of benzoyl derivative 8 with benzal-
dehyde (2b), giving the adducts 18b in an E/Z ratio of 3:1 (entry
7), which is markedly contrasted with the exclusive formation
of the Z-isomer (E/Z ) 1:99), reported by Kaji et al., in the
condensation of ethyl benzoyl acetate with 2b.6 The condensa-
tion of phenylsulfonylacetylTEMPO 19 with 2a proceeded
smoothly but afforded the usual E-isomer 17a (trans relationship
between PhSO2 and 2-furyl groups), exclusively (entry 8).3
The E- and Z-geometries of adducts were unambiguously
deduced on the basis of line separations of the carbonyl carbons
of the COOT or COL in the 13C NMR because of coupling
with the vinylic 1H nucleus.14 As shown in Table 2, the
Z-isomers (trans relationship between COOT or COL and
vinylic H) show a larger line separation with 11.2∼12.5 Hz for
the COOT or COL carbonyl carbons, while the E isomers (cis
relationship) show reduced values of 7.5∼8.1 Hz in the line
separation.15
a Carried out by using the substrates (0.5∼1.0 mmol) and 2 (1.5∼2.0
equiv) in EtOH at 0∼10 °C with piperidine until the substrates were
consumed. AcOH was added when the reaction was sluggish. b Yields are
1
based on isolated products. c Determined on the basis of H NMR. d E-3e
isomerized to Z-3e on standing or in CDCl3. e One isomer was exclusively
produced.
SCHEME 3. Mechanism of Knoevenagel Reaction
We examined the effect of the kind of aldehydes on the E/Z
selectivity by employing oxa-2-cyclohexene-2-carbaldehyde
(2c),16 phenylpropargyl aldehyde (2d), trans-cinnamaldehyde
(2e), and 3-mehyl-2-butenal (2f) (Table 3). Thus, the substituent
R of enals and ynal was chosen as being of an electronically
different nature: inductively electron-withdrawing,17 more or
less electron-withdrawing, and electron-donating.
As shown in entry 1, the condensation of 1 with 2c afforded
the E-isomer 3c, predominantly. This E-preference executed with
2c reappeared in the condensation with the derivatives of
2-methylpropanoyl, 6, and benzoyl, 8, giving the corresponding
adducts 17c and 18c in 87% and 56% yields, respectively, with
an E/Z ratio of 25∼20:1(entries 2 and 3). On the other hand,
the runs of 1 with the ynal 2d and enal 2e resulted in formation
of adducts 3d and 3e with a reversed E/Z ratio of ca. 1:3-1:2
(entries 4 and 5). Furthermore, the condensation of 1 with 2f
produced the Z-adduct 3f, exclusively (entry 6).
compound 22, produced by addition of 1 to the iminium salt
21, plays an important role in the stereochemical outcome of
the reaction.6 The intermolecular or intramolecular acid-base
reaction of 22 may form the betaine 23, which would undergo
elimination of the secondary amine via the conformation A,
leading finally to thermodynamically favorable (E)- or (Z)-
isomer of adducts 3. Since the E/Z ratio of the Knoevenagel
reaction of 1 was affected by the kind of substitution R of
aldehydes 2 (Table 3), we discussed these stereochemical results
from the electronic and steric points of view as follows.
Accordingly, we postulated that these results might be
ascribable to the harmony of electron-withdrawing or -donating
nature of the substituents R of aldehydes and their steric
bulkiness. It is conceived that Z-geometry is more favorable
than its E-geometry, when R is an electron-donating nature like
the 2-methyl-2-propenyl group (Table 3, entry 6), irrespective
of steric hindrance between R and O-acylTEMPO group
(structure (Z)-3, eq 1).19 It is due to stereoelectronic reason of
As shown in Scheme 3, the amine-catalyzed Knoevenagel
reaction includes many reversible steps, in which the amine
(14) Kingsbury, C. A.; Draney, D.; Sopchik, A. Rissler, W.; Durham,
D. J. Org. Chem. 1976, 41, 3863-3868.
(15) E- and Z-geometry of 3 was further confirmed by NOE experiments
(Supporting Information).
(16) Paquette, L. A.; Schulze, M. M.; Bolin, D. G. J. Org. Chem. 1994,
59, 2043-2051.
(17) According to resonance structure ii, the oxa-2-cyclohexenyl sub-
stituent of 2c is considered to be of electron-withdrawing nature.
(18) (a) Kirby, A. J. Stereoelectronic Effect; Oxford University Press:
Oxford, U.K., 1996; Chapter 4. (b) Carey, F. A.; Sundberg, R. J. AdVanced
Organic Chemistry, 3rd ed.; Plenum Press: New York, 1993; Part A,
Chapter 3.
(19) Because of such conjugative interaction in Z-isomers as B, the 13
C
NMR absorption due to an acyl CO usually appears at higher field than
that of E-isomers, for example, (Z)-3b, 202.4 ppm versus (E)-3b, 195.4
ppm.
J. Org. Chem, Vol. 71, No. 3, 2006 949

Inokuchi and Kawafuchi
TABLE 4. Knoevenagel Condensations of O-AcetoaceylTEMPO 1
with Various Aromatic Aldehydes 2 to 3
FIGURE 1. Interpretation for preference of E- and Z-adducts
dependent on the kind of aldehyde substituents R.
R and the acyl group, that is hyper- or π-conjugative donor-
acceptor interaction, as indicated in the resonance structure B
(Figure 1, eq 1). On the other hand, when R is the electron-
withdrawing one, that is, oxa-2-cyclohexenyl (Table 3, entries
1-3), the interaction between R and the acyl group will be
weakened and the reaction may be crucially influenced by steric
hindrance between the auxiliary TEMPO and R. Therefore, the
condensation of 1 with such aldehydes leads to preferential
formation of E-isomers (eq 2), where stabilization by a hyper-
or π-conjugative interaction between R and the COOT group
is also accounted for in some extent.20
Alternatively, the marked Z-preference by the reaction of the
morpholine amide 13 and Weinreb amide 15, compared with
the alkyl ester 9 and O-acylTEMPOs 1, can be explained by
the decreased electron-withdrawing ability of the amide groups,
in which the stabilization by hyper- or π-conjugative interaction
between R and the CONR′R′′ groups on the structure of adducts
14 and 16 is canceled out.20
The conformation of Knoevenagel adduct (E)-3a was deter-
mined by X-ray crystallographic analysis (Supporting Informa-
tion). The COOT group and the 2-furyl group in trans
relationship with each other spread over the same plane of the
conjugated π-system. The carboxyl oxygen and the piperidine
nitrogen are located within the same plane, which looks like
assisting hyperconjugative interaction between the 2-furyl and
the COOT groups. The conformation of the acetyl group takes
orientation orthogonal to the conjugated π-system. Furthermore,
the calculation of the conformation of O-acylTEMPO by PM3
showed that cis conformation between CdO and O-N bonds
on the C-O bond was a favorable one.21
To assess the generality of the present method to construct
the E-adducts by Knoevenagel condensation of O-acetoacetyl-
TEMPO 1, various aromatic and heteroaromatic aldehydes 2
were examined by using piperidine as a base. As shown in Table
4, entries 1-4, 6-8, 10, and 11, a similar trend of E-preference
as in the case of 2-furaldehyde (2a) was observed. On the other
hand, low E-preference was encountered in the case of 2-chlo-
robenzaldehyde (2j) and pyrrole-2-carboxaldehyde (2n), in
which E- and Z-adducts were produced in almost equal amount
(entries 5 and 9). However, E-preference of pyrrole 2n, E/Z )
1.1:1, was greatly improved to the E/Z ratio of 14:1 by use of
a Carried out by using 1 (0.5∼2.0 mmol) and 2 (1.3∼1.5 equiv) in EtOH
(three drops) (A), in CH2Cl2 (few drops)-AcOH (two drops) (B), or in DMF
(0.5 mL)-AcOH (two drops) (C) at room temperature with piperidine base
(two drops). b Yields are based on isolated products and the value in the
parentheses is the conversion (%) of 1. c Determined on the basis of 1H
NMR. d Z-Adduct isomerized to E-isomer during recrystallization.
N-tosyl analogue 2o (entry 10). Furthermore, some of inferior
E-preferences in Table 4 can be recovered by adopting a
contiguous chromatographic separation and Z/E isomerization
of the separated Z-adducts, as described below.
Alkyl alkylideneacetoacetates are prone to Z/E isomerization
by heating or upon treatment with a Lewis acid such as Yb-
(OTf)3.1b,10 Accordingly, we submitted individually the O-
acylTEMPO derivatives (E)- and (Z)-3b to heating at 100 °C
for 12 h in toluene, giving a 6.7:1 mixture of E- and Z-isomers
3b, respectively (Table 5, entries 1 and 2). The same treatment
of (E)-3a as above for 7 h led to a 6:1 mixture of (E)- and
(Z)-3a (entry 3), while that of (Z)-3a for 14 h resulted in a 1:2
mixture of E/Z-isomers (entry 4). It is likely that isomerization
of the (Z)-isomer of 3a to its (E)-isomer is sluggish compared
with that of (Z)-3b. On the other hand, the treatment of the
separated (Z)-isomer 3j, one of low E-preference products, with
Yb(OTf)3 in toluene at room temperature for 34 h afforded a
1:1.8 mixture of (E)- and (Z)-3k (entry 5). Thus, this procedure
can induce Z/E isomerization of a low E-preference product,
which enabled us to regain the (E)-isomers from the (Z)-isomers
and to improve totally the ratio of (E)-isomers. Similarly, the
treatment of the separated indole derivative (Z)-3n with Me3Al
in CH2Cl2 led to a 1:1.7 mixture of (E)-and (Z)-3n (entry 6).
In contrast to the E-preference condensations of ace-
toacetylTEMPO 1, as shown in Table 6, Knoevenagel conden-
sations of N-methoxy-N-methylacetoacetamide (Weinreb amide
(20) O-acylTEMPO seems to be more electron-withdrawing than Weinreb
amide, CON(OMe)Me, on the basis of IR absorption, υCdO of RCOOT )
∼1760 cm-1 (RCOOT); that of CON(OMe) ) ∼1665 cm-1
.
(21) The calculation was probed by using PM3 semiempirical energy
minimizations (MOPAC2002 in CAChe Worksystem Pro 5.04).
950 J. Org. Chem., Vol. 71, No. 3, 2006

KnoeVenagel Condensation of Acetoacetic DeriVatiVes
TABLE 5. Z/E-Isomerization by Heating or Treatment with Lewis
Acid
TABLE 7. Reduction of Knoevenagel Adducts with
Hydride-Transferring Reagents
entry
compds
R
conditions/time
E/Z ratio
1
2
3
4
5
6
(E)-3b
(Z)-3b
(E)-3a
(Z)-3a
(Z)-3k
(Z)-3n
C6H5
C6H5
2-furyl
2-furyl
2-ClC6H4
N-Ts-3-indolyl
100 °C/12 h
100 °C/12 h
100 °C/7 h
100 °C/14 h
Yb(OTf)3/RT/34 h
Me3Al/0 °C/2 h
6.7/1
6.7/1
6.0/1
1/2
1/1.8
1/1.7
product, yield, %a
entry
substrate
reagent (equiv)
1,2-red
1,4-red
SM
TABLE 6. Knoevenagel Condensations of
1
2
3
4
5
(E)-3a
(E)-3a
(E)-17a
(E)-3c
(Z)-16a
LAH (2.5)
(E)-24a, 24
(E)-24a, 83
(E)-26a, 72
(E)-24c, 78
(Z)-28a, 85
25a, 75
25a, 7
27a, 13
25c, 7
29a
N-Methoxy-N-methylacetoacetamide (Weinreb Amide 15) with
DIBALH (2.0)
DIBALH (2.0)
DIBALH (2.0)
DIBALH (1.5)
9
14
5
Various Aromatic Aldehydes to (Z)-16a
a Isolated yields are based on starting substrates.
SCHEME 4. Preparation of E- and Z-2-Alkenals by
Reduction
a Carried out by using 15 (1.0∼2.0 mmol) and 2 (1.3∼1.5 equiv) in
CH2Cl2 (few drops)-AcOH (two drops) at 28∼30 °C for 2 days. b Yields
are based on isolated products. c DMF (0.2 mL) was added.
15) with various aromatic aldehydes 2 lead to (Z)-adducts 16,
exclusively. In this case, when prolonged, a very small amount
of deacetylated byproduct was found, presumably because of
retro-Claisen condensation of 22 with amine base followed by
elimination of amine.22
SCHEME 5. Conversion of Knoevenagel Adducts to Furans
Having succeeded in stereoselective Knoevenagel condensa-
tion of acetoacetic derivatives, we next examined reduction of
the adducts with hydride-transferring reagents to obtain the
corresponding 2-(1-hydroxyalkyl)-(2E)- or -(2Z)-carboxylate
derivatives, which are γ-substituted but hardly accessible
homologues by the Morita-Baylis-Hillman procedure.8 As
shown in Table 7, the LAH-reduction of (E)-3a at -78 °C gave
25a as a main product (75% yield) as a result of 1,4-addition
of hydride as well as a small amount of the carbinol (E)-24a
(24% yield) because of the 1,2-addition (entry 1). On the other
hand, the reduction of (E)-3a with DIBALH (2 equiv) at -78
°C for 15 min afforded the desired (E)-24a (83% yield) as a
major product and a small amount of 25a (7% yield, entry 2)
as a byproduct. Similarly, (E)-17a and (E)-3c were reduced with
DIBALH to give the corresponding 1,2-reduction products (E)-
26a and (E)-24c in 72 and 78% yields, respectively (entries 3
and 4). Furthermore, the reduction of Weinreb amide (Z)-16a
with DIBALH (1.5 equiv) at -78 °C gave the corresponding
(Z)-28a in 85% yield (entry 5).
According to propensity of the O-acylTEMPOs and Weinreb
amides to leave selectively the aldehydes upon reduction,7,23
we examined transformation of the above 2-hydroxyalkyl deriv-
atives (E)-24a and (Z)-28a to the corresponding E- and Z-enals
31a (Scheme 4). Thus, reduction of the THP ether (E)-30a, ac-
cessible by protection of (E)-24a with dihydropyran (78% yield),
with DIBALH (3 equiv) at -78 °C for 1 h afforded the desired
(22) Formation of a small amount of conjugated amides iii as a byproduct
was accompanied. In contrast to efficient condensations with aromatic
aldehydes, the reaction of trans-cinnamaldehyde (2e) with 15 under the
same conditions resulted in a low yield of the desired Z-adduct.
(23) Nahm, S.; Weinreb, S. M. Tetrahedron Lett. 1981, 22, 3815-3818.
J. Org. Chem, Vol. 71, No. 3, 2006 951

Inokuchi and Kawafuchi
at 0 °C. After being stirred at room temperature for 24 h, the
reaction was quenched with aqueous NH4Cl and was worked up in
the usual manner, and the products were purified by column
chromatography (SiO2, hexane-AcOEt, increasing the gradient from
40:1 to 5:1 V/V) to give 2.51 g (81%) of 1 (Rf ) 0.34, hexane-
AcOEt 5:1): bp 98-99 °C/0.06 Torr; IR (neat) 1766, 1724, 1662,
1635, 1365, 1313, 1211, 1170, 1132, 1047, 935, 904, 873, 806
aldehyde (E)-31a in 86% yield along with the corresponding
carbinol (7%) because of overreduction. The obtained E-enal
31a was contaminated with ca. 5% of Z-enal because of olefin
isomerization possibly during the reduction with DIBALH. Ole-
fin isomerization also proceeded, though slowly, on standing
in CDCl3. On the other hand, the reduction of the THP-protected
Weinreb amide (Z)-32a, obtained in 84% yield by protection
of the hydroxy group of (Z)-28a, with DIBALH at 0 °C resulted
in an intractable mixture comprised of aldehydes (ca. 1:1 mixture
of (E)/(Z)-31a), the corresponding carbinol because of overreduc-
tion, and the unidentified. However, the reduction of (Z)-32a
with lithium aluminum hydride (LAH) at 0 °C afforded cleanly
the desired Z-enal 31a in 72% yield. This conversion was ac-
companied by formation of a ca. 4:1 Z/E mixture of 3-(2-furyl)-
2-hydroxy-2-propenal (8%) because of hydrolysis of THP ether
presumably during acidic workup with aqueous tartaric acid.
We attempted transformation of (E)-24c to spiro[4.5] acetal
33 by acid-catalyzed cyclization of the hydroxy group to enol
olefin by virtue of cis configuration of the olefinic linker.
However, it turned out that the treatment of (E)-18c with PTS
in CH2Cl2 led to trisubstituted furan 34a in 93% yield as a result
of cyclization to 27 followed by aromatization.24,25 Similarly,
(E)-26c, obtained by the DIBALH reduction of the correspond-
ing precursor (E)-17c as above (85% yield), was transformed
to 34b in 94% yield.
1
cm-1; H NMR (300 MHz) δ 1.08, 1.13 (major form) and 1.05,
1.15 (minor form) (s, 12H), 1.33-1.70 (m, 6H), 1.98 (minor form)
and 2.31 (major form) (s, 3H), 3.50 (s, 2H), 5.02 and 12.19 (brs);
13C NMR (75.5 MHz) δ 16.6, 20.2 (2C), 30.2, 31.7 (2C), 38.9 (2C),
48.4, 60.0 (2C), 87.2, 166.7, 200.2. HRMS (ESI) calcd for C13H24-
NO3 (MH+) 242.3385, found 242.1750 (MH+).
Preparation of 8. A mixture of the carbinol 7 (R1 ) Ph, 307
mg, 1.0 mmol), TEMPO• (312 mg, 2 mmol), and RuCl2(PPh3)3
(96 mg, 0.1 mmol) in toluene (10 mL) was stirred at 60-62 °C for
3 days under O2.11 The mixture was passed through a short column
packed with SiO2, and the eluate was concentrated and purified by
column chromatography (SiO2, hexane-AcOEt, increasing the
gradient from 10:1 to 5:1 V/V) to give 170 mg (56%) of 8 (Rf )
0.46, hexane-AcOEt 3:1) as solids (ca. 13:1 keto/enol mixture):
mp 103-104 °C; IR (KBr) 1770, 1760, 1679, 1595, 1581, 1450,
1405, 1382, 1332, 1257, 1213, 1116, 1045, 941, 806, 748, 690,
1
632, 594 cm-1; H NMR (300 MHz) δ 1.01, 1.05 (major form)
and 1.11, 1.20 (minor form) (s, 12H), 1.33-1.73 (m, 6H), 4.01 (s,
2H), 7.46-7.52 (m, 2H), 7.57-7.63 (m, 1H), 7.97-8.00 (m, 2H);
13C NMR (75.5 MHz) δ 16.40 + 16.47, 19.90 + 20.05, 31.31 +
31.43, 38.6 (2C), 44.3, 59.8 (2C), 84.6, 125.6, 128.06, 128.10,
128.3, 130.9, 133.3, 135.7, 166.5, 191.7.
Conclusion
General Procedure for Knoevenagel Condensaton of O-
AcetoacetylTEMPO (1), Giving 3, and Its Homologues Using
Piperidine. To a cooled (0-4 °C) mixture of O-acetoacetylTEMPO
(1, 243 mg, 1.0 mmol), 2-furaldehyde (2a, 192 mg, 2.0 mmol),
and EtOH (4 drops, 20 mg) was added piperidine (two drops, 10
mg). The mixture was allowed to warm gradually to room
temperature and was stirred for 2 days. The reaction was quenched
with cold aqueous NH4Cl, and the products were extracted with
AcOEt. The extracts were washed with brine, dried (MgSO4), and
concentrated. The crude products were purified by column chro-
matography (SiO2, hexane-AcOEt, increasing the gradient from 10:1
to 3:1 V/V) to give 243 mg (76%) of (E)-3a (Rf ) 0.34, hexane-
AcOEt 5:1) and trace of (Z)-3a (Rf ) 0.13). (E)-3a: mp 131-132
°C; IR (KBr) 3141, 3100, 2976, 1739, 1625, 1551, 1473, 1381,
In summary, the first E-selective Knoevenagel condensation
of acetoacetic derivatives has been developed by using TEMPO
for the acylated substituent. Alternatively, Z-selective Knoev-
enagel condensation was achieved by use of amide analogues
including the Weinreb amide. The E-selectivity, successful for
the kind of aldehydes bearing electron-withdrawing substituents,
can virtually be explained in terms of steric hindrance caused
by an acylated substituent TEMPO toward a substituent R at
the vinylidene carbon of the adducts. Furthermore, effect of the
kind of aldehydes as well as the acylated substituent of nucleo-
philes on the E-selectivity was discussed from the steric and
electronic viewpoints. The E- and Z-adducts are shown to be a
useful precursor of E- and Z-2-hydroxyalkyl-2-propenals owing
to propensity of O-acylTEMPO and the Weinreb amide to leave
the formyl group by reduction with either DIBALH or LAH.
Utility of E-geomety of the products was also shown by another
application to the facile synthesis of trisubstituted furans.26
1365, 1233, 1159, 1071, 1043, 1014, 957, 905, 873, 778, 753 cm-1
;
1H NMR (300 MHz) δ 1.11, 1.14 (s, 12H), 1.38-1.79 (m, 6H),
2.54 (s, 3H), 6.49 (d,d, J ) 3.3, 1.6 Hz, 1H), 6.74 (d, J ) 3.3 Hz,
1H), 7.40 (s, 1H, CH), 7.53 (m, 1H); 13C NMR (75.5 MHz) δ 16.7,
20.6 (2C), 31.0, 31.7, 38.8 (2C), 60.3 (2C), 112.5, 117.6, 126.0,
128.6, 146.0, 149.1, 164.7 (3J(COO, H) ) 7.8 Hz), 201.2. HRMS
(ESI) calcd for C18H25NO4 (MH+) 320.1862, found 320.1838
(MH+).
Experimental Section
Preparation of 1 by Reaction of the TEMPO Anion and
Diketene. In a 50-mL one-necked flask were placed C10H8 (160
mg, 1.25 mmol), TEMPO• (2.0 g, 12.8 mmol), and THF (20 mL).
To this solution was added Na metal (354 mg, 15.4 mmol) at -60
°C, and the mixture was stirred at room temperature until Na
dissolved and the blue-black color of Na+[C10H8]-• persisted. To
the mixture was added Me3SiCl (2.0 mL, 15.8 mmol) at room
temperature. The resulting mixture was transferred dropwise to a
solution of diketene (4.9 mL, 64.0 mmol) in THF (15 mL), cooled
Knoevenagel Condensation of N-methoxy-N-methylacetoac-
etamide (15) with 2, Giving 16, with Piperidine and AcOH. To
a mixture of N-methoxy-N-methylacetoacetamide (15, 360 mg, 2.48
mmol) and 2a (309 mg, 3.2 mmol) in CH2Cl2 (0.2 mL) were added
consecutively AcOH (two drops) and piperidine (three drops) at
0-4 °C. The mixture was stirred at 28-30 °C for 2 days and was
quenched with aqueous NH4Cl. Products were extracted with AcOEt
and were worked up in the usual manner. The crude products were
purified by column chromatography (SiO2, hexane-AcOEt, increas-
ing the gradient from 10:1 to 1:3 V/ V) to give 382 mg (69%) of
(Z)-16a (Rf ) 0.21, hexane-AcOEt 1:1): mp 101-103 °C (partially
decompose); IR (KBr) 3155, 3119, 3049, 2975, 1665, 1625, 1547,
1482, 1429, 1393, 1371, 1335, 1283, 1256, 1209, 1185, 1154, 1081,
1022, 993, 951, 932, 915, 883, 827, 761, 703 cm-1; 1H NMR (300
MHz) (a ca. 5.5:1 rotatory mixture on amide C-N bond) δ 2.34
(major) and 2.38 (minor) (s, 3H), 3.12 (minor) and 3.36 (major)
(s, 3H), 3.47 (major) and 3.91 (minor) (s, 3H), 6.59 (major) and
6.52 (minor) (m, 1H), 6.78 (major) and 6.85 (minor) (d, J ) 3.6
(24) Aparicio, F. J. L.; Herrera, F. J. L.; Ballesteros, J. S. Carbohydr.
Res. 1979, 69, 55-70.
(25) (a) Imagawa, H.; Kurisaki, T.; Nishizawa, M. Org. Lett. 2004, 6,
3679-3681. (b) Brown, R. C. D. Angew. Chem., Int. Ed. 2005, 44, 850-
852 and references therein.
(26) Syntheses of natural products bearing a trisubstituted furan-3-
carboxyl structure: (a) Toro, A.; Deslongchamps, P. J. Org. Chem. 2003,
68, 6847-6852. (b) Marshall, J. A.; Devender, E. A. V. J. Org. Chem.
2001, 66, 8037-8041. (c) Paterson, I.; Gardner, M.; Banks, B. J.
Tetrahedron 1989, 45, 5283-5292.
952 J. Org. Chem., Vol. 71, No. 3, 2006

KnoeVenagel Condensation of Acetoacetic DeriVatiVes
Hz, 1H), 7.31 (s, 1H), 7.53 (major) and 7.57 (minor) (d, J ) 1.6
Hz); 13C NMR (75.5 MHz) (major isomer) δ 26.4, 32.5, 61.4, 112.7,
117.7, 125.3, 131.7, 146.1, 149.3, 168.7 (3J(COO, H) ) 10.4 Hz),
194.2. HRMS (ESI) calcd for C11H14NO4 (MH+) 224.0923, found
224.0928 (MH+).
give 25 mg (86%) of (E)-31a (Rf ) 0. 52, hexane-AcOEt 2:1) and
2 mg (7%) of the corresponding carbinol (Rf ) 0.32) because of
overreduction. (E)-31a: IR (neat) 1679, 1625, 1469, 1371, 1199,
1076, 1022, 983, 939, 815 cm-1; 1H NMR (300 MHz) δ 1.45 and
1.51 (d, J ) 6.9 Hz, 3H), 1.50-1.90 (m 6H), 3.30-3.65 (m 2H),
4.41 and 4.79 (m, 1H), 5.27 and 5.40 (q, J ) 6.9 Hz, 1H), 6.55
(m, 1H), 7.01 and 7.11 (s, 1H), 7.12 and 7.15 (d, J ) 3.57 Hz,
1H), 7.61 (m, 1H), 9.579 and 9.583 (s, 1H); 13C NMR (75.5 MHz)
δ 18.6 and 19.2, 19.3 and 19.5, 25.2 and 25.3, 30.57 and 30.69,
62.1 and 62.5, 65.5 and 68.4, 96.4 and 98.1, 112.73 and 112.89,
118.43 and 118.91, 131.69 and 133.69, 137.18 and 145.73, 150.15
and 150.47, 192.83 and 192.99. HRMS (ESI) calcd for C14H18O4
(MH+) 250.1205, found 250.1184 (MH+).
General Procedure for Reduction of Knoevenagel Adducts
3 to O-2-(1-Hydroxyalkyl)-2-alkenoylTEMPOs (E)-24. To a
cooled (-78 °C) solution of (E)-3a (R ) 2-furyl, 80 mg, 0.25
mmol) in toluene (5 mL) was added dropwise DIBALH (1.0 N in
toluene, 0.5 mL, 2.0 equiv) over 1 min. The mixture was stirred at
the same temperature for 15 min, and excess DIBALH was decom-
posed with AcOEt and the reaction was quenched with aqueous
NaHCO3. The mixture was diluted with AcOEt and a small amount
of benzene, and the supernatant pipetted off was dried (MgSO4),
and concentrated. The crude products were analyzed by GC (150
°C) and were purified by column chromatography (SiO2, hexane-
AcOEt, by increasing the gradient from 10:1 to 2:1) to give 6 mg
(7%) of the 1,4-reduction product 25a (Rf ) 0.46, hexane-AcOEt
3:1), 7 mg (9%) of the starting (E)-3a (Rf ) 0.27), and 67 mg (83%)
of the 1,2-reduction product (E)-24a (Rf ) 0.21). (E)-24a: mp 112-
113 °C; IR (KBr) 3469, 3102, 1716, 1631, 1552, 1251, 1201, 1122,
1093, 767 cm-1; 1H NMR (300 MHz) δ 1.086, 1.09, 1.21, 1.22 (s,
12H), 1.40-1.46 (m, 1H), 1.50-1.78 (m, 5H), 1.53 (d, J ) 6.6
Hz, 3H), 3.94 (br d, J ) 10.4 Hz, 1H), 5.39 (m, 1H), 6.50 (d,d, J
) 3.6, 1.7 Hz, 1H), 6.65 (d, J ) 3.6 Hz, 1H), 7.24 (s, 1H), 7.56 (d,
J ) 1.7 Hz, 1H); 13C NMR (75.5 MHz) δ 16.9, 20.93 and 20.96,
23.1, 31.87 and 31.91, 39.1 (2C), 60.33 and 60.41, 65.9, 112.2,
116.7, 123.4, 130.3, 145.0, 150.3, 167.7 (3J(COO, H) ) 7.78 Hz).
HRMS (ESI) calcd for C18H28NO4 (MH+) 322.2018, found
322.2023 (MH+). 25a: IR (neat) 1770, 1722, 1508, 1380, 1365,
Similarly, (Z)-31a was obtained from (Z)-28a as follows. The
protection of (Z)-28a with THP as above afforded (Z)-32a (84%
yield), and reduction of the resulting THP ether (Z)-32a with LAH
(7 equiv) at 0 °C for 10 min followed by usual workup including
treatment with aqueous tartaric acid and purification on SiO2
(hexane-AcOEt by increasing the gradient from 10:1 to 1:1 V/V)
gave (Z)-31a (Rf ) 0.57, hexane-AcOEt 2:1) in 72% yield and a
ca. 4:1 Z/E mixture of 3-(2-furyl)-2-hydroxyethyl-2-propenal (Rf
) 0.19) in 8% yield. (Z)-31a: IR (neat) 1660, 1625, 1473, 1365,
1301, 1201, 1120, 1076, 1022, 985, 929, 815 cm-1; 1H NMR (300
MHz) δ 1.29 and 1.36 (d, J ) 6.9 Hz, 3H), 1.50-1.90 (m 6H),
3.40-3.53 and 3.72-3.95 (m 2H), 4.49 and 4.75 (m, 1H), 4.87
and 4.95 (q, J ) 6.9 Hz, 1H), 6.51 (m, 1H), 6.69 (m, 1H), 7.19
and 7.35 (s, 1H), 7.57 and 7.59 (d, J ) 1.9 Hz, 1H), 10.61 and
10.64 (s, 1H); 13C NMR (75.5 MHz) δ 19.8 and 19.9, 21.2 and
22.8, 25.3 and 25.4, 30.94 and 30.97, 62.1 and 62.8 and 62.9, 67.8
and 68.2, 97.1 and 97.2, 112.2 and 112.3, 117.1 and 117.3, 128.2
and 128.5, 138.9, 145.9 and 146.2, 151.4, 192.6 and 192.9. HRMS
(ESI) calcd for C14H18O4 (MH+) 250.1205, found 250.1194 (MH+).
Preparation of Furan (34). To a solution of the carbinol (E)-
26a (15 mg, 0.04 mmol) in CH2Cl2 (2 mL) was added p-TsOH
(catalytic). The mixture was stirred at room temperature for 5 h
and was worked up in the usual manner, and the crude product
was purified by column chromatography (SiO2, hexane-AcOEt,
increasing the gradient from 7:1 to 1:1 V/V) to give 14 mg (93%)
of 34b (Rf ) 0. 51, hexane-AcOEt 1:1) as colorless oil: IR (neat)
3455, 3120, 1735 1727, 1618, 1583, 1456, 1365, 1265, 1232, 1182,
1
1245, 1182, 1132, 1010, 954, 734 cm-1; H NMR (300 MHz) δ
0.95, 1.01, 1.11, 1.12 (s, 12H), 1.35-1.46 (m, 1H), 1.47-1.73 (m,
5H), 2.30 (s, 3H), 3.26 (d, J ) 7.7 Hz, 2H), 4.00 (t, J ) 7.7 Hz,
1H), 6.07 (m, 1H), 6.25 (m, 1H), 7.29 (m, 1H); 13C NMR (75.5
MHz) δ 16.8, 20.5 (2C), 26.9, 30.0, 31.7 (2C), 39.08 and 39.18,
57.0, 60.39 and 60.48, 106.9, 110.4, 141.5, 151.6, 168.0, 202.1.
Similarly, (E)-26c was obtained in 85% yield from (E)-17c. (E)-
26c: mp 125-126 °C; IR (KBr) 3398, 1745, 1693, 1635, 1596,
1463, 1384, 1365, 1245, 1236, 1201, 1120, 1045, 1022, 960, 941,
1
767 cm-1; H NMR (300 MHz) δ 0.80 and 1.04 (d, J ) 6.9 Hz,
1
6H), 1.04 (s, 6H), 1.16 and 1.18 (s, 6H), 1.35-1.75 (m, 6H), 1.80-
1.95 (m, 3H), 2.20 (m, 2H), 3.82 (t, J ) 11.3 Hz, 1H), 4.07 (m,
2H), 4.83 (m, 1H), 5.24 (m, 1H), 6.79 (s, 1H); 13C NMR (75.5
MHz) δ 16.9, 19.55 and 19.63, 20.86 and 20.88, 21.49 and 21.53,
31.80 and 31.83, 39.1 (2C), 60.15 and 60.34, 65.9, 74.5, 113.2,
130.3, 134.1, 150.9, 168.4.
1132, 1052, 1041, 989, 952, 914, 873, 767, 690 cm-1; H NMR
(300 MHz) δ 1.09 and 1.21 (s, 12H), 1.40-1.80 (m, 11H), 2.57 (s,
3H), 2.61 (t, J ) 7.1 Hz, 2H), 3.68 (t, J ) 6.3 Hz, 2H), 6.27 (s,
1H); 13C NMR (75.5 MHz) δ 14.0, 17.0, 21.0 (2C), 24.1, 27.4,
31.9, 32.1 (2C), 39.0 (2C), 60.0 (2C), 62.5, 105.7, 112.4, 153.9,
157.4, 164.2. HRMS (ESI) calcd for C19H32NO4 (MH+) 338.2331,
found 338.2347 (MH+).
Similarly, (Z)-28a was obtained in 85% yield from (Z)-16a as a
ca. 1.7:1 mixture of rotatory isomers of amide bond. (Z)-28a: IR
(neat) 3411, 1629, 1556, 1486, 1388, 1151, 1110, 1018, 979, 885,
Acknowledgment. We appreciate supports by Okayama
University (T. I.), Wesco Scientific Promotion Foundation (T.
I.), and a Grant-in-Aid for Scientific Research from Japan
Society for the Promotion of Science (H.K. No. 16550104) for
this work. We are grateful to the SC-NMR laboratory of
Okayama University for high-field NMR experiments and to
the Venture Business Laboratory for X-ray crystallographic
analysis. We thank Assistant Professor Toshinobu Korenaga of
this University for help in the measurement of X-ray. We are
indebted to Professor Junzo Nokami, Okayama University of
Science, for HRMS analyses and to Koei Chemical Co., Ltd.
for a research sample of TEMPO•.
1
746, 595 cm-1; H NMR (300 MHz) δ 1.40 (d, J ) 6.3 Hz, 3H),
2.52 (br), 3.11 and 3.30 (brs, 3H), 3.47 and 3.84 (brs, 3H), 4.57
(m, 1H), 6.35 (br m, 2H), 6.43 (brs, 1H), 7.36 (brs, 1H); 13C NMR
(75.5 MHz) δ 21.8 and 22.2, 32.0 and 32.2, 60.96 and 61.2, 68.8,
110.3, 111.3 and 111.5, 114.6, 137.3, 142.6 and 143.0, 150.0 and
150.4, 166.1 and 170.4.
General Procedure for THP-Protection of (E)-24a and (Z)-
28a Followed by DIBALH or LAH Reduction to (E)- and (Z)-
31a. To a mixture of (E)-24a (37 mg, 0.12 mmol) and DHP (170
mg, 2.0 mmol) in CH2Cl2 (3 mL) was added PPTS (catalytic). The
mixture was stirred at room temperature for 43 h and was worked
up in the usual manner including washing with aqueous NaHCO3
followed by purification on column chromatography (SiO2, hexane-
AcOEt by increasing the gradient from 10:1 to 3:1 V/V) to give
48 mg (78%) of (E)-30a. To a solution of the THP-protected (E)-
30a (0.11 mmol) in toluene (5 mL) was added dropwise DIBALH
(0.33 mL, 3 equiv) at -78 °C. After being stirred at the same
temperature for 1 h, the mixture was worked up as above and the
crude products were purified by column chromatography (SiO2,
hexane-AcOEt by increasing the gradient from 10:1 to 1:1 V/V) to
Supporting Information Available: Spectral data including IR,
1H NMR, and 13C NMR spectra of 1, 6, 8, (E)- and (Z)-3a-p, (E)-
and (Z)-12a, (Z)-14a, (Z)-16a,b,h,k,l,m,o, (E)-17a,c, (E)-18b,c, (E)-
20a, (E)-24a,c, 25a,c, (E)-26a,c, (E)-27a, (E)-28a, (E)-30a, (E)-
and (Z)-31a, (Z)-32a, and 34a,b, and the crystallographic data of
(E)-3a are provided. This material is available free of charge via
the Internet at http://pubs.ac.org.
JO051952W
J. Org. Chem, Vol. 71, No. 3, 2006 953