Enantioselective 1,3-Dipolar Cycloaddition Reaction of Nitrones with α,β-Unsaturated Aldehydes Catalyzed by Cationic 3-Oxobutylideneaminatocobalt(III) Complexes

Article abstract of DOI:10.1246/bcsj.76.2197

The enantioselective 1,3-dipolar cycloaddition reaction of nitrones with α,β-unsaturated aldehydes was realized using 3-oxobutylideneaminatocobalt complex catalysts. Varieties of the cobalt(III) and cobalt(III) complexes were screened and the cationic cob

Full text of DOI:10.1246/bcsj.76.2197

Ó 2003 The Chemical Society of Japan
Bull. Chem. Soc. Jpn., 76, 2197–2207 (2003) 2197
Enantioselective 1,3-Dipolar Cycloaddition Reaction of Nitrones
with ꢀ,ꢁ-Unsaturated Aldehydes Catalyzed by Cationic
3-Oxobutylideneaminatocobalt(c) Complexes
Satoko Kezuka,# Natsuki Ohtsuki, Tsuyoshi Mita, Youichi Kogami, Tomoko Ashizawa,
ꢀ
Taketo Ikeno, and Tohru Yamada
Department of Chemistry, Faculty of Science and Technology, Keio University,
Hiyoshi, Kohoku-ku, Yokohama 223-8522
Received June 13, 2003; E-mail: yamada@chem.keio.ac.jp
The enantioselective 1,3-dipolar cycloaddition reaction of nitrones with ꢀ,ꢁ-unsaturated aldehydes was realized us-
ing 3-oxobutylideneaminatocobalt complex catalysts. Varieties of the cobalt(b) and cobalt(c) complexes were screened
and the cationic cobalt(c) complex with hexafluoroantimonate was found to be the most effective for the catalytic enan-
tioselective 1,3-dipolar cycloaddition reaction. In the presence of the cobalt(c) hexafluoroantimonate complex, the enan-
tioselective 1,3-dipolar cycloaddition reaction of various nitrones with ꢀ,ꢁ-unsaturated aldehydes afforded the corre-
sponding isoxazolidines in high yields and with high enantioselectivities. The absolute configuration of the optically ac-
tive products was determined by X-ray analysis. Reasonable explanations for the enantioselection in the present 1,3-di-
polar cycloaddition reaction catalyzed by the 3-oxobutylideneaminatocobalt complex were proposed.
The 1,3-dipolar cycloaddition reactions of nitrones with al-
kenes are some of the most useful and reliable strategies for
preparing isoxazolidine derivatives. Because its N–O bond
was readily cleaved to form 3-amino alcohol equivalents under
the mild reducing conditions, these cyclic compounds have
been applied to synthetic intermediates for useful compounds
such as alkaloids, ꢁ-lactams, amino acids, amino sugars, etc.1
Since the pioneering contributions to this field of the 1,3-dipo-
lar cycloaddition by Huisgen et al. in the 1960s,2 various 1,3-di-
poles have been developed. Among them, nitrones are some of
the most readily available 1,3-dipoles; they are relatively stable
and easily isolated. Thus the 1,3-dipolar cycloaddition of ni-
trones has been widely investigated. The remarkable feature
of this cycloaddition reaction is the concerted [4ꢂ þ 2ꢂ] supra-
facial process like the Diels–Alder reaction to simultaneously
construct three new chiral centers. The asymmetric approach
of the 1,3-dipolar cycloaddition for the isoxazolidine was first
tried by employing nitrones and alkenes with chiral
auxiliaries.3 These reactions provided effective methods to pre-
pare several natural products,3 but stoichiometric chirality was
required and the stereoselectivities were not always satisfactory
due to the relatively high reaction temperature. The 1,3-dipolar
cycloaddition reaction could be explained as a concerted
[4ꢂ þ 2ꢂ] process; therefore, it is expected that a Lewis acid
catalyst could accelerate the reaction and also improve the re-
gio- and diastereo-4 as well as enantioselectivities. Various chi-
ral Lewis acids have been developed for the catalytic and enan-
tioselective versions of electron-rich5 alkenes or electron-defi-
cient6 alkenes. In the former combination, the LUMO of the
nitrone was lowered by coordinating to a Lewis acid to accel-
erate the reaction. On the contrary, LUMO-lowering of the
electron-deficient alkene, such as alkenoyl derivatives, is ex-
pected by coordinating of the carbonyl to a Lewis acid; howev-
er, the competitive coordination of the nitrone should be con-
sidered over coordination of the monodentate carbonyl
compound.7 Chelation of alkenoyl oxazolidinones was prefer-
red for the cycloaddition reactions5 (Scheme 1), whereas the
study of the 1,3-dipolar cycloaddition of the monodentate
ꢀ,ꢁ-unsaturated aldehyde has been limited. Recently, this type
of asymmetric 1,3-dipolar cycloaddition has ultimately been
achieved using chiral DBFOX/Ph–Zn or Ni catalysts,8
CpRu–BIPHOP-F or Fe catalysts,9 and efficient organocata-
lysts10 (Fig. 1).
The optically active 3-oxobutylideneaminatocobalt com-
plexes (Fig. 2) were originally developed as effective catalysts
of the enantioselective borohydride reductions;11 they were re-
cently reported to catalyze the enantioselective cyclopropa-
nation.12 The enantioselectivities and diastereoselctivities in
these reactions could be tuned by the steric and/or electronic
factor of the 3-oxobutylideneaminato ligands. During the
course of our continuing study of the catalysis using these co-
balt complexes, they were found to act as chiral Lewis acid cat-
alysts for the 6ꢂ-electrocyclic concerted reaction, an enantiose-
lective hetero Diels–Alder reaction.13 The corresponding cat-
ionic cobalt(c) complexes could be employed more effectively
than the original cobalt(b) complex; their counter anions signif-
icantly influenced their Lewis acidities and catalytic abilities.14
These cationic cobalt(c) complexes also catalyzed the enantio-
selective carbonyl–ene reaction of glyoxal derivatives to afford
the corresponding homoallylic alcohols with high enantio-
selectivities.15 In this article, we will fully disclose the enantio-
selective 1,3-dipolar cycloaddition reactions of the ꢀ,ꢁ-unsat-
# Present address: Department of Chemistry, Faculty of Science
and Engineering, Aoyama Gakuin University, Fuchinobe,
Sagamihara, Kanagawa 229-8558
Published on the web November 15, 2003; DOI 10.1246/bcsj.76.2197

2198 Bull. Chem. Soc. Jpn., 76, No. 11 (2003)
Enantioselective 1,3-Dipolar Cycloaddition
Scheme 1. Equilibrium among aldehyde, nitrone, and Lewis acid catalyst.
Fig. 1. Effective catalysts for the 1,3-dipolar cycloaddition reaction with ꢀ,ꢁ-unsaturated aldehydes.
ined for the enantioselective 1,3-dipolar cycloaddition reaction
of N-benzylideneaniline N-oxide (1a) with 1-cyclopentene-1-
carbaldehyde (2a) (Table 1).16 The cobalt(b) complex 4a cat-
alyzed the reaction at ꢁ40 ꢂC to selectively provide the corre-
sponding endo-bicyclo adduct in 43% yield in 89 h, but the op-
tical yield was 32% ee. The iodo cobalt(c) 4b and the cationic
cobalt(c) complexes with various counter anions 4c–4e were
used to improve the reactivity and the selectivity. The reaction
rate was maintained using the iodo cobalt(c) complex, and the
enantiomeric excess was up to 46%. The corresponding
cobalt(c) tetrafluoroborate 4c and triflate 4d more smoothly
catalyzed the reaction to obtain the corresponding isoxazoli-
dine with good enantioselectivities (Entries 3 and 4). Screening
of the counter anions revealed that the cobalt(c) hexafluoroan-
timonate 4e could be effectively employed as a highly active
cationic cobalt(c) complex catalyst and that the resulting 4-
hydroxymethylisoxazolidine with 51% ee was obtained in 48
h (Entry 5).
Optimization of Temperature and Solvents for the Enan-
tioselective 1,3-Dipolar Cycloaddition Reaction. The tem-
perature and reaction solvents generally influence the selectiv-
ities and catalytic activities of the Lewis acid catalysis. The re-
gio-, diastereo-selectivity as well as enantioselectivity should
be regulated by the reaction solvent and temperature during
the enantioselective 1,3-dipolar cycloaddition reaction. The re-
Fig. 2. 3-Oxobutylideneaminato cobalt complex.
urated aldehyde with nitrones catalyzed by cationic cobalt(c)
complexes with optically active 3-oxobutylideneaminato
ligands.
Results and Discussion
Design of the Cationic Cobalt Complexes and Their
Counter Anions. The cationic character of the central cobalt
atom has a crucial effect on the reaction rate and on the enan-
tioselectivity, based on the previous examinations of the hetero
Diels–Alder reaction and the carbonyl–ene reaction catalyzed
by 3-oxobutylideneaminatocobalt complexes.14,15 When the
iodo cobalt(c) complex was employed for the hetero Diels–
Alder reaction, the reaction time was shortened; the cationic
cobalt(c) complex was thus proved to be a highly active
catalyst. In the carbonyl–ene reactions, the counter anions of
the cationic cobalt(c) complex significantly influenced the re-
activity and selectivity. Various 3-oxobutylideneaminato-
cobalt(b) and cobalt(c) complex catalysts were therefore exam-
Table 1. Various Cobalt Complex Catalysts
Entry
Catalyst
Time/h
Yield/%a)
Endo/Exob)
Ee (Endo)/%c)
1d)
2d)
3
4
5
Co
Co–I
Co–BF4
Co–OTf
Co–SbF6
4a
4b
4c
4d
4e
89
97
53
53
48
43
59
46
63
89
>99=1
>99=1
>99=1
>99=1
>99=1
32
46
40
47
51
a) Isolated yield after the reduction with NaBH4. b) Determined by 1H NMR analysis. c) Determined by HPLC
analysis using Daicel Chiralpak AD-H or Chiralcel OD-H. d) 0.1 mol. amt. catalyst was employed.

S. Kezuka et al.
Table 2. Temperature Effect
Bull. Chem. Soc. Jpn., 76, No. 11 (2003) 2199
Entry
Temp/ꢂC
Time/h
Yield/%a)
Endo/Exob)
Ee (Endo)/%c)
1
2
3
4
r.t.
0
6
24
24
48
91
92
94
89
87/13
94/6
97/3
3
15
27
51
ꢁ20
ꢁ40
>99=1
Reaction conditions: Cobalt(c)–SbF6 catalyst 4e 0.015 mmol (0.05 mol. amt.), aldehyde 0.45 mmol, and nitrone
0.3 mmol in CH2Cl2. a) Isolated yield after the reduction with NaBH4. b) Determined by 1H NMR analysis. c)
Determined by HPLC analysis using Daicel Chiralpak AD-H or Chiralcel OD-H.
Table 3. Various Solvents
Entry
Solvent
Time/h
Yield/%a)
Endo/Exob)
Ee (Endo)/%c)
1
2
3
4
5
6
CHCl3
PhCH3
PhCl
C2H5CN
THF
CH2Cl2
66
65
66
53
53
48
29
38
70
23
22
89
98/2
98/2
98/2
>99=1
>99=1
>99=1
33
37
31
45
46
51
a) Isolated yield after the reduction with NaBH4. b) Determined by 1H NMR analysis. c) Determined by HPLC
analysis using Daicel Chiralpak AD-H or Chiralcel OD-H.
action temperature was therefore examined and the results are
shown in Table 2. The endo-selective products were obtained
in 91% yield, but with no enantioselectivity at room tempera-
ture (Entry 1). The diastereo- and enantio-selectivities were
slightly improved at 0 C and ꢁ20 C (Entries 2 and 3). At
ꢁ40 ꢂC, the corresponding isoxazolidine was obtained in
89% yield in 48 h with excellent endo-selectivity and with
51% ee (Entry 4). At lower temperatures, the selectivities in-
creased but the yield of the isoxazolidine decreased; therefore,
the reaction temperature was fixed based on the reactivity and
selectivity.
tions indicated that dichloromethane was the most suitable sol-
vent for the present enantioselective 1,3-dipolar cycloaddition
reaction with nitrone.
Enantioselective 1,3-Dipolar Cycloaddition Reaction of
Various Nitrones and ꢀ,ꢁ-Unsaturated Aldehydes. The
highly active catalysts, cobalt(c) hexafluoroantimonate com-
plexes (Fig. 3), were successfully applied to the enantioselec-
ꢂ
ꢂ
Various solvents were next examined for the enantioselec-
tive 1,3-dipolar cycloaddition reaction of nitrones. The yields
and enantiomeric excesses of the resulting product are shown in
Table 3. In chloroform, the most suitable solvent for the enan-
tioselective carbonyl–ene reaction, the isoxazolidine was ob-
tained only in 29% yield in 66 h and with 33% ee. The product
was obtained in moderate yields but with similar selectivities in
toluene and chlorobenzene (Entries 2 and 3). The coordinating
solvents, such as propiononitrile and tetrahydrofuran (THF),
decreased the catalytic activity but the selectivity was improved
up to 45% ee (Entries 4 and 5). In dichloromethane, the reac-
tion was completed in 48 h to afford the corresponding endo-se-
lective-isoxazolidine with 51% ee (Entry 6). These examina-
Fig. 3. Various cobalt(c) hexafluoroantimonate complexes.

2200 Bull. Chem. Soc. Jpn., 76, No. 11 (2003)
Enantioselective 1,3-Dipolar Cycloaddition
tive 1,3-dipolar cycloaddition reaction of various nitrones
(Scheme 2) with 1-cyclopentene-1-carbaldehyde (2a)
(Table 4). The chiral diaryl diamine parts of the complexes
were examined for the reaction with N-benzylideneaniline N-
oxide (1a) and the corresponding isoxazolidine was obtained
with 51% ee using the complex 4e derived from the optically
active cyclohexanediamine (Entries 1–4). Moreover, the abso-
lute configuration was reversed when the complexes 6 and 7,
derived from 1,2-bis(3,5-dimethylphenyl)ethylenediamine
Scheme 2. Preparation of nitrones.
Table 4. Nitrone Cycloadditions of Various Nitrones
Entry
Nitrone
Product
Catalyst
Time/h
Yield/%a)
Endo/Exob)
Ee (Endo)/%c)
1
2
3
4
4e
5
6
48
65
41
55
89
87
68
86
>99=1
>99=1
>99=1
>99=1
ꢁ51
ꢁ8
7
1a
1b
1c
1d
1e
1f
3a
7
37
5
6
7
3b
3c
3d
3e
3f
4e
4e
4e
134
48
82
91
83
99/1
>99=1
99/1
ꢁ53
ꢁ43
ꢁ52
73
8
9
4e
6
108
96
94
76
>99=1
98/2
5
22
10
11
12
13
4e
5
6
73
96
96
48
84
89
96
98
>99=1
99/1
99/1
99/1
65
78
80
67
7
14
15d)
6
6
96
96
85
18
99/1
>99=1
85
91
1g
1h
1i
3g
3h
3i
16
17
6
6
60
96
quant
quant
>99=1
>99=1
87
83
18
1j
3j
6
83
93
>99=1
85
1
a) Isolated yield after the reduction with NaBH4. b) Determined by H NMR analysis. c) Determined by HPLC
ꢂ
analysis using Daicel Chiralpak AD-H or Chiralcel OD-H. d) Reaction temp: ꢁ78 C.

S. Kezuka et al.
Table 5. Various ꢀ,ꢁ-Unsaturated Aldehydes
Bull. Chem. Soc. Jpn., 76, No. 11 (2003) 2201
Entry
1d)
Aldehyde
R1
R2
Product Yield/%a) rs(a/b)b)
Endo/Exob) Ee(Endo)/%c)
2a
2b
–(CH2)3–
3h
8
quant
90
>99=1
89/11
>99=1
>99=1
>99=1
98/2
87
79
63
2e)
3
H
H
H
2c CH3
9
94
4f)
5d)
6d)
7d)
2d
2e
2f
H
H
H
H
CH3
C4H9
100
110
120
130
91
60
80
93
5/95
>99=1
>99=1
84/16
>99=1
82
79
77
92
1/>99
C6H13
3/97
1/99
2g
C6H5CH2
8d,g)
2h
H
140
quant
1/>99
>99=1
90
Reaction time: 24–144 h. Reaction temp: ꢁ40 ꢂC. a) Isolated yield after the reduction with NaBH4. b) Determined
by 1H NMR analysis. c) Determined by HPLC analysis using Daicel Chiralpak AD-H or Chiralcel OD-H. d) Com-
plex 6 was employed. e) Reaction temp: ꢁ78 ꢂC. f) 0.08 mol. amt. complex 6 was employed at ꢁ60 ꢂC and five
portions of nitrone was added at 24 h intervals. g) Reaction temp: ꢁ30 ꢂC.
and 1,2-bis(2,4,6-trimethylphenyl)ethylenediamine, were em-
ployed. The reactions with the nitrone derived from the
p-substituted benzaldehyde, N-(4-methoxybenzylidene)aniline
N-oxide, N-(4-methylbenzylidene)aniline N-oxide and N-(4-
chlorobenzylidene)aniline N-oxide, were examined. The prod-
ucts were obtained with moderate enantioselectivities (Entries
5–7). In contrast, the optical yield of the cyclo adduct derived
from nitrone 1e decreased to 5% (Entry 8, vs Entry 7) but, in the
case of the o-chloro derivative 1f, it increased to 65% (Entry 10,
vs Entry 7). The optically active ligands were screened again
and the complex 6 was found to be the most suitable (Entries
10–13). The cobalt-complex-catalyzed cycloaddition reaction
using nitrone 1g also occurred and the corresponding endo-ad-
duct was selectively obtained with 85% ee (Entry 14). At ꢁ78
ꢂC, the resulting isoxazolidine was obtained in low yield, but
the optical yield was improved to 91% ee (Entry 15). The ni-
trones 1h, 1i, and 1j derived from 2,3-dichlorobenzaldehyde,
2,4-dichlorobenzaldehyde, and 2,3,5-trichlorobenzaldehyde
were reacted to afford the corresponding isoxazolidines with
perfect endo-selectivities in high yields with high enantioselec-
tivities (Entries 16–18). Examination of the various types of ni-
trones revealed that the 1,3-dipolar cycloaddition reaction cat-
alyzed by the cationic cobalt complexes proceeded with excel-
lent endo-selectivity and that a high enantioselectivity was
achieved with nitrones derived from the ortho-substituted
benzaldehyde.
The cycloaddition reaction with various ꢀ,ꢁ-unsaturated al-
dehydes was then attempted in the presence of cobalt(c) hexa-
fluoroantimonate (Table 5). The reaction with acrylaldehyde
smoothly proceeded to afford the corresponding isoxazolidine
with 89% regioselectivity, 99% diastereoselectivity, and 79%
enantioselectivity (Entry 2). Crotonaldehyde, a ꢁ-substituted
ꢀ,ꢁ-unsaturated aldehyde, was also allowed to react with the
nitrone 1h and the resulting alcohol was 63% ee (Entry 3).
The regioselectivity was reversed in the case of employing
the ꢀ-substituted aldehydes; this reversal could be attributed
to their steric factor rather than their electronic density.17
The regio- and diastereo-selectivities of the corresponding iso-
xazolidines were high-to-excellent in all cases for the ꢀ-substi-
tuted ꢀ,ꢁ-unsaturated aldehydes. The reaction with methacryl-
aldehyde was induced by the cationic cobalt complex and the
corresponding endo-adduct was selectively obtained with 82%
ee (Entry 4). The ꢀ-alkyl derivatives, 2e and 2f, were next at-
tempted and the corresponding endo products were regioselec-
tively obtained with moderate enantioselectivities (Entries 5
and 6). For the reaction of 2-benzyl-2-propenal (2g) and its de-
rivatives 2h, only the corresponding endo cycloadducts were
obtained and their enantioselectivities were up to 92% and

2202 Bull. Chem. Soc. Jpn., 76, No. 11 (2003)
Enantioselective 1,3-Dipolar Cycloaddition
Scheme 3. Determination of the absolute configuration of cycloadducts 3a and 3g.
90%, respectively (Entries 7 and 8).
contribution of the coordination depicted in Fig. 6 should be
considered because of the less hindered blocking of the disfa-
vored coordination. Thus the corresponding cycloadduct with
the opposite chirality was obtained. It should be pointed out
here that the ortho-halo substituted nitrones such as 1f–1j could
increase the enantioselectivity. It can be explained as follows:
The disfavored coordination should be suppressed by the steric
repulsion between the ortho-halo substitution and the bulky
side chain of the ligand (Fig. 7), whereas the halogen atom
on the ortho-position of the donor molecule could coordinate
with the metal-complex to improve the enantioselectivity in
the present 1,3-dipolar cycloaddition (Fig. 5).
In Table 5, that shows the cycloaddition reaction with a va-
riety of aldehydes, the 4-formyl cycloadducts were selectively
obtained from the alkenals 2a, 2b and 2c and the 5-formyl cy-
cloadducts from the alkenals 2d–2h. The regioselectivity for
the 4-formyl cycloadducts can be attributed to the frontier orbi-
tal interaction. The secondary orbital interaction could enhance
the regioselectivity (Fig. 8). On the contrary, the steric effect
should be considered for the regioselectivity of the 5-formyl cy-
cloadducts from the alkenals 2d–2h. The absolute configura-
tion of the cycloadduct 130 derived from 2g was determined
to be (S,S) by X-ray analysis of the 2,4-dinitrophenylhydrazone
derivative 15 (Fig. 9). The ꢀ-substituted 2-alkenal could coor-
dinate the cobalt(c) complex in an s-cis conformer. Avoiding
the steric repulsion between the ꢀ-substitutent of the aldehyde
and the N-phenyl group of the nitrone, the cycloaddition reac-
tion produced the 5-formyl adduct with excellent regio-
selectivity. The enantioselection of the 5-formyl adduct was
also regulated by the steric-demanding aryl groups of the opti-
cally active diamine part and the side chain of the 3-oxobutyl-
ideneaminato ligand (Fig. 10).
Absolute Configuration of the Optically Active Isoxazoli-
dine Derivatives. The absolute configuration of the cycload-
duct 3j was determined by X-ray analysis. The cycloadduct 3j
of 85% ee was recrystallized from ethanol to afford the optical-
ly pure 3j, which was again crystallized from ethanol for prep-
aration of the single crystals. The X-ray analysis of the optical-
ly pure 3j confirmed that it was an endo-adduct and revealed
that its absolute configuration for the reaction catalyzed by
the (S,S)-cobalt complex catalyst was the (S,S,S) form as de-
picted in Fig. 4. The absolute configurations of 3a and 3g were
then determined. As shown in Scheme 3, the hydroxy group of
the cycloadduct 3g was protected with the tert-butyldimethyl-
silyl group and then it was treated with n-butyllithium and
TBAF to obtain the isoxazolindine 3a in high yield without
any loss of optical purity. The HPLC retention time and peak
area of the obtained 3a were compared with those of compound
3a from the nitrone 1a. It was confirmed that they had the same
absolute configuration (1S,4R,5S) and that the enantioselective
sense was not affected by the halogen on the ortho position of
the nitrones.
On the basis of these observations and the related discussions
on the hetero Diels–Alder and the carbonyl–ene reaction cata-
lyzed by the cationic cobalt(c) complexes, the sense of chiral
induction of the present cycloaddition reaction could be ex-
plained as follows. The coordination of the ꢀ,ꢁ-unsaturated al-
dehyde with the cobalt atom would be similar to that of alde-
hydes in the hetero Diels–Alder reaction. The axial coordina-
tion site of cobalt(c) was occupied by the oxygen atom of the
aldehyde carbonyl anti to the alkyl group, which was expected
to orient between two coordinating oxygen atoms on the planar
3-oxobutylideneaminato ligand. It is also reasonable to assume
that the conformation of the ꢀ,ꢁ-unsaturated aldehyde during
the reaction could be s-trans. The nitrone could favorably ap-
proach the formed complex of the ꢀ,ꢁ-unsaturated aldehyde
with the cobalt catalyst, as illustrated as Fig. 5, avoiding the
steric hindrance of the chiral diamine and the bulky aryl group
of the side chain of the 3-oxobutylideneaminato ligand, to af-
ford the corresponding cycloadduct with high regio-, diastereo-
and enantioselectivities. In the reactions catalyzed by the
cobalt(c) complex 4e derived from cyclohexanediamine, the
Conclusions
In summary, the cationic cobalt(c) hexafluoroantimonate
complex with an optically active 3-oxobutylideneaminato li-
gand effectively catalyzed the asymmetric 1,3-dipolar cycload-
dition reaction of various nitrones with aldehydes to afford the
corresponding isoxazolidines in high yield and with high enan-
tioselectivities.

S. Kezuka et al.
Bull. Chem. Soc. Jpn., 76, No. 11 (2003) 2203
Fig. 8. Regioselectivity for 4-formyl cycloadducts.
Fig. 4. Absolute configuration of the cycloadduct 3j.
Fig. 5. Reasonable explanation for the enantioselection.
Fig. 9. Absolute configuration of a hydrazone of cycload-
duct 130.
Fig. 6. Favored coordination of complex 4e catalyzed reaction.
Fig. 10. Enantio- and regioselectivity for 5-formyl adduct.
or C6D6 as the solvent and tetramethylsilane as the internal
standard. High-resolution mass spectra were obtained with a
Hitachi M-80B. For the thin-layer chromatography (TLC) analysis
throughout this study, Merck precoated TLC plates (silica gel 60
F254, 0.25 mm) were used. The products were purified by prepar-
ative column chromatography on silica gel (silica gel 60N). High-
performance liquid chromatography (HPLC) analyses were per-
formed using a Shimadzu LC-6A chromatograph with an optically
active column (Chiralcel OB-H, Chiralcel OD-H, and Chiralpak
AD-H columns, Daicel Ltd., Co.); the peak areas were obtained
with a Shimadzu chromatopack C-R4A or a Varian Dynamax
MacIntegrator. Optical rotations were measured with a JASCO
DIP-370 digital polarimeter. N-Benzylideneaniline N-oxide (1a),
N-(4-methoxybenzylidene)aniline N-oxide (1b), N-(4-methylben-
Fig. 7. Effect of o-substituted halogen.
Experimental
General. The infrared (IR) spectra were recorded using a JAS-
CO Model FT/IR-410 infrared spectrometer on KBr pellets or liq-
uid film on NaCl. The 1H NMR spectra and 13C NMR spectra were
measured using a JEOL Model GX-400 spectrometer with CDCl3

2204 Bull. Chem. Soc. Jpn., 76, No. 11 (2003)
Enantioselective 1,3-Dipolar Cycloaddition
zylidene)aniline N-oxide (1c), N-(4-chlorobenzylidene)aniline N-
oxide (1d), N-(3-chlorobenzylidene)aniline N-oxide (1e), N-(2-
chlorobenzylidene)aniline N-oxide (1f), N-(2-bromobenzyli-
dene)aniline N-oxide (1g), N-(2,3-dichlorobenzylidene)aniline N-
oxide (1h), N-(2,4-dichlorobenzylidene)aniline N-oxide (1i), and
N-(2,3,5-trichlorobenzylidene)aniline N-oxide (1j) were prepared
using previously reported methods.18 Acrylaldehyde (2b), croton-
aldehyde (2c), and methacrylaldehyde (2d) were purchased from
Tokyo Kasei Kogyo (TCI) Co., Ltd. 1-Cyclopentene-1-carbalde-
hyde (2a),19 2-butyl-2-propenal (2e),20 2-hexyl-2-propenal (2f),20
2-benzyl-2-propenal (2g),20 and 2-(4-methylbenzyl)-2-propenal
(2h)20 were prepared by previously published methods.
CHCl3). HPLC (Chiralcel OD-H, 2-propanol 5.0% in hexane,
Flow 1.0 mL/min, 11.9 min (major), 29.2 min (minor), 53% ee,
(R,R)-4e).
ꢀ
ꢀ
ꢀ
(1R ,4R ,5R )-5-Hydroxymethyl-4-(4-methylphenyl)-3-
phenyl-2-oxa-3-azabicyclo[3.3.0]octane (3c): 1H NMR (CDCl3)
ꢃ 1.05–1.12 (1H, m), 1.41–1.70 (4H, m), 1.79 (1H, s), 2.00–2.06
(1H, m), 2.34 (3H, s), 3.49–3.55 (2H, m), 4.54–4.55 (1H, m),
4.83 (1H, s), 6.88 (1H, t, J ¼ 7:3 Hz), 6.98 (2H, d, J ¼ 7:8 Hz),
7.13–7.21 (4H, m), 7.32 (2H, d, J ¼ 7:8 Hz); 13C NMR (CDCl3)
ꢃ 21.2, 24.8, 31.2, 31.5, 66.7, 67.4, 75.3, 86.7, 114.5, 121.1,
127.1, 128.7, 128.8, 136.5, 136.6, 151.1; IR (KBr) 3414, 3024,
2953, 1596, 1514, 1488, 1249, 1041, 758, 694 cmꢁ1
. HRMS:
Preparation of Optically Active 3-Oxobutylideneaminato-
cobalt Complexes. Complexes 4a–4e, 5, 6, and 7 were prepared
by a reported method.14b
Calcd for C20H23NO3: Mþ = 309.1729. Found: m=z 309.1720.
27
½ꢀꢃD ꢁ76:0ꢂ (c 0.467, CHCl3). HPLC (Chiralpak AD-H, 2-pro-
panol 2.0% in hexane, Flow 1.0 mL/min, 33.8 min (major), 44.7
min (minor), 43% ee (R,R)-4e).
General Procedure for Enantioselective 1,3-Dipolar Cyclo-
addition Reaction of ꢀ,ꢁ-Unsaturated Aldehyde. Under nitro-
gen, to the N,N0-bis[3-oxo-2-(2,4,6-trimethylbenzoyl)butylidene]-
(1R,2R)-cyclohexane-1,2-diaminatocobalt(c) hexafluoroantimo-
nate (12.9 mg, 0.015 mmol, 0.05 molar amount) was added 1-cy-
clopentene-1-carbaldehyde (43.8 mg, 0.46 mmol) in dichloro-
methane (1.0 mL). A solution of the N-benzylideneaniline N-oxide
(59.3 mg, 0.30 mmol) in dichloromethane (1.0 mL) was then added
at ꢁ40 ꢂC. The mixture was stirred for 48 h at ꢁ40 ꢂC, followed
by treatment with a solution of sodium borohydride (21.0 mg, 0.56
mmol) in ethanol at ꢁ40 ꢂC for 1 h; it was finally left to room
temperature. After standard work-up and chromatography on sili-
ꢀ
ꢀ
ꢀ
(1R ,4R ,5R )-4-(4-Chlorophenyl)-5-hydroxymethyl-3-
phenyl-2-oxa-3-azabicyclo[3.3.0]octane (3d): 1H NMR (CDCl3)
ꢃ 1.09–1.17 (1H, m), 1.36–1.42 (1H, m), 1.50–1.70 (4H, m), 2.02–
2.07 (1H, m), 3.54 (2H, s), 4.55–4.56 (1H, m), 4.91 (1H, s), 6.91
(1H, t, J ¼ 7:3 Hz), 6.97 (2H, d, J ¼ 7:8 Hz), 7.21–7.28 (2H,
m), 7.32 (2H, d, J ¼ 8:3 Hz), 7.41 (2H, d, J ¼ 8:3 Hz); 13C NMR
(CDCl3) ꢃ 24.8, 31.0, 31.8, 66.9, 67.6, 75.0, 86.9, 114.3, 121.3,
128.2, 128.7, 128.8, 132.6, 138.4, 150.8; IR (KBr) 3427, 2958,
2873, 1597, 1489, 1091, 1041, 1014, 756, 695 cmꢁ1
.
HRMS:
Calcd for C19H20ClNO2: Mþ
= 329.1183.
Found: m=z
329.1199. ½ꢀꢃD ꢁ82:3ꢂ (c 0.300, CHCl3). HPLC (Chiralpak
AD-H, 2-propanol 1.5% in hexane, Flow 1.0 mL/min, 38.6 min
(major), 58.1 min (minor), 52% ee, (R,R)-4e).
28
ꢀ
ꢀ
ꢀ
ca gel (Hexane:AcOEt = 4:1) afforded (1R ,4S ,5R )-5-hydroxy-
methyl-3,4-diphenyl-2-oxa-3-azabicyclo[3.3.0]octane (79.5 mg,
diastereomer ratio >99=1) in 89% yield.
ꢀ
ꢀ
ꢀ
(1R ,4R ,5R )-4-(3-Chlorophenyl)-5-hydroxymethyl-3-
phenyl-2-oxa-3-azabicyclo[3.3.0]octane (3e): 1H NMR (CDCl3)
ꢃ 1.10–1.18 (1H, m), 1.39–1.43 (1H, m), 1.44–1.73 (5H, m), 2.03–
2.08 (1H, m), 3.53 (2H, s), 4.55–4.56 (1H, m), 4.91 (1H, s), 6.91
(1H, t, J ¼ 7:3 Hz), 6.98 (2H, d, J ¼ 7:8 Hz), 7.21–7.28 (4H,
m), 7.35 (1H, d, J ¼ 7:3 Hz), 7.49 (1H, s); 13C NMR (CDCl3) ꢃ
24.8, 30.8, 31.8, 67.1, 67.5, 75.2, 87.0, 114.1, 121.3, 125.5,
127.2, 127.3, 128.8, 129.4, 134.1, 142.1, 150.8; IR (NaCl) 3389,
3065, 2962, 1596, 1487, 1202, 1043, 757, 694, 517 cmꢁ1. HRMS:
Calcd for C19H20ClNO2: Mþ = 329.1183. Found: m=z 329.1185.
ꢀ
ꢀ
ꢀ
(1R ,4S ,5R )-5-Hydroxymethyl-3,4-diphenyl-2-oxa-3-aza-
bicyclo[3.3.0]octane (3a, Chart 1): 1H NMR (CDCl3) ꢃ 1.08–
1.14 (1H, m), 1.39–1.90 (5H, m), 2.03–2.06 (1H, m), 3.55 (2H,
s), 4.56–4.57 (1H, m), 4.89 (1H, s), 6.87–6.90 (1H, m), 6.99
(2H, dd, J ¼ 1:0, 8.8 Hz), 7.18–7.28 (3H, m), 7.34 (2H, t, J ¼
7:6 Hz), 7.45 (2H, d, J ¼ 7:3 Hz); NOE: H4–Ha 5.0%, Ha–H4
4.1%, H1–H4 0.9%, H4–H1 0.4%, H1–Ha 4.9%, Ha–H1 3.3%;
13C NMR (CDCl3) ꢃ 24.8, 31.1, 31.6, 66.9, 67.5, 75.5, 86.8,
114.4, 121.1, 127.0, 127.3, 128.1, 128.7, 139.8, 151.1; IR (NaCl)
3427, 3087, 3061, 3027, 2957, 2873, 1710, 1598, 1488, 1250,
1075, 1042, 923, 752, 700 cmꢁ1. HRMS: Calcd for C19H21NO2:
26
½ꢀꢃD ꢁ34:8ꢂ (c 0.756, CHCl3). HPLC (Chiralcel OD-H, 2-pro-
panol 2.0% in hexane, Flow 1.0 mL/min, 12.8 min (major), 32.1
min (minor), 22% ee, (S,S)-6).
Mþ = 295.1572. Found: m=z 295.1566. ½ꢀꢃD ꢁ76:3ꢂ (c
25
ꢀ
ꢀ
ꢀ
0.832, CHCl3). HPLC (Chiralpak AD-H, 2-propanol 2.0% in hex-
ane, Flow 1.0 mL/min, 29.5 min (major), 40.5 min (minor), 51%
ee, (R,R)-4e).
(1R ,4R ,5R )-4-(2-Chlorophenyl)-5-hydroxymethyl-3-
phenyl-2-oxa-3-azabicyclo[3.3.0]octane (3f): 1H NMR (CDCl3)
ꢃ 1.14–1.19 (1H, m), 1.25–1.61 (5H, m), 1.99–2.04 (1H, m), 3.64
(1H, d, J ¼ 10:7 Hz), 3.71 (1H, d, J ¼ 10:7 Hz), 4.72–4.73 (1H,
m), 5.23 (1H, s), 6.90 (1H, t, J ¼ 7:3 Hz), 7.00 (2H, d, J ¼ 7:8
Hz), 7.21–7.30 (4H, m), 7.38 (1H, dd, J ¼ 1:5, 7.8 Hz), 7.81
(1H, dd, J ¼ 1:5, 7.8 Hz); 13C NMR (CDCl3) ꢃ 25.1, 30.8, 30.9,
67.0, 67.2, 73.0, 88.0, 113.5, 121.0, 126.8, 128.4, 128.7, 129.0,
129.6, 132.8, 137.7, 150.6; IR (KBr) 3400, 2958, 2874, 1596,
ꢀ
ꢀ
ꢀ
(1R ,4R ,5R )-5-Hydroxymethyl-4-(4-methoxyphenyl)-3-
phenyl-2-oxa-3-azabicyclo[3.3.0]octane (3b): 1H NMR (CDCl3)
ꢃ 1.03–1.08 (1H, m), 1.35–1.64 (5H, m), 1.96–2.01 (1H, m), 3.49
(2H, s), 3.74 (3H, s), 4.48–4.49 (1H, m), 4.77 (1H, s), 6.80–6.84
(3H, m), 6.91–6.93 (2H, m), 7.14–7.16 (2H, m), 7.30 (2H, d, J ¼
8:8 Hz); 13C NMR (CDCl3) ꢃ 24.8, 31.3, 31.6, 55.2, 66.7, 67.6,
75.1, 86.8, 113.5, 114.6, 121.1, 128.4, 128.7, 131.8, 151.1,
158.4; IR (KBr) 3414, 2956, 1598, 1512, 1488, 1249, 1174,
1487, 1470, 1056, 1035, 751, 706 cmꢁ1
.
HRMS: Calcd for
C19H20ClNO2: Mþ = 329.1183. Found: m=z 329.1170. ½ꢀꢃD
ꢁ62:7ꢂ (c 0.733, CHCl3). HPLC (Chiralcel OD-H, 2-propanol
1.0% in hexane, Flow 1.0 mL/min, 18.1 min (major), 26.2 min
(minor), 80% ee, (S,S)-6).
25
1032, 753, 695 cmꢁ1
.
HRMS: Calcd for C20H23NO3: Mþ
=
325.1678. Found: m=z 325.1687. ½ꢀꢃD ꢁ87:6ꢂ (c 0.189,
27
ꢀ
ꢀ
ꢀ
(1R ,4R ,5R )-4-(2-Bromophenyl)-5-hydroxymethyl-3-
phenyl-2-oxa-3-azabicyclo[3.3.0]octane (3g): 1H NMR (CDCl3)
ꢃ 1.15–1.62 (6H, m), 2.00–2.05 (1H, m), 3.68 (1H, d, J ¼ 10:8
Hz), 3.76 (1H, d, J ¼ 10:8 Hz), 4.72–4.74 (1H, m), 5.21 (1H, s),
6.90 (1H, t, J ¼ 7:3 Hz), 6.99 (2H, d, J ¼ 7:8 Hz), 7.13–7.17
(1H, m), 7.23–7.34 (3H, m), 7.56 (1H, dd, J ¼ 1:2, 8.1 Hz), 7.81
Chart 1.

S. Kezuka et al.
Bull. Chem. Soc. Jpn., 76, No. 11 (2003) 2205
(1H, dd, J ¼ 1:7, 8.1 Hz); 13C NMR (CDCl3) ꢃ 25.1, 30.8, 30.9,
67.1, 67.3, 75.0, 88.1, 113.5, 121.0, 123.4, 127.4, 128.7, 129.0,
130.0, 131.9, 139.3, 150.5; IR (KBr) 3418, 2956, 2874, 1595,
1487, 1236, 1040, 1024, 750, 695 cmꢁ1
.
HRMS: Calcd for
C19H20BrNO2: Mþ = 373.0677. Found: m=z 373.0674. ½ꢀꢃD
ꢁ36:8ꢂ (c 0.937, CHCl3). HPLC (Chiralcel OD-H, 2-propanol
3.0% in hexane, Flow 1.0 mL/min, 8.9 min (major), 11.0 min (mi-
nor), 85% ee, (S,S)-6).
27
Chart 2.
ꢀ
ꢀ
ꢀ
(1R ,4R ,5R )-4-(2,3-Dichlorophenyl)-5-hydroxymethyl-3-
phenyl-2-oxa-3-azabicyclo[3.3.0]octane (3h): 1H NMR (CDCl3)
ꢃ 1.08–1.10 (1H, m), 1.26–1.42 (2H, m), 1.47–1.59 (2H, m), 1.86
(1H, s), 1.98–2.02 (1H, m), 3.60 (1H, d, J ¼ 10:7 Hz), 3.67 (1H, d,
J ¼ 10:7 Hz), 4.71–4.72 (1H, m), 5.24 (1H, s), 6.90 (1H, t, J ¼ 7:3
Hz), 6.97 (2H, d, J ¼ 8:3 Hz), 7.19–7.27 (3H, m), 7.40 (1H, d, J ¼
6:8 Hz), 7.74 (1H, d, J ¼ 7:3 Hz); 13C NMR (CDCl3) ꢃ 25.1, 30.6,
30.9, 66.7, 67.2, 73.5, 88.1, 113.4, 121.0, 127.2, 127.8, 128.9,
129.0, 230.9, 132.1, 140.3, 150.3; IR (NaCl) 3425, 2959, 2875,
had the correct absolute configurations.
Procedure for Enantioselective 1,3-Dipolar Cycloaddition
Reaction of N-2,3-Dichlorobenzylideneaniline N-oxide and
Methacrylaldehyde. Under nitrogen, to the N,N0-bis[3-oxo-2-
(2,4,6-trimethylbenzoyl)butylidene]-(1S,2S)-1,2-bis(3,5-dimethyl-
phenyl)ethylenediaminatocobalt(c) hexafluoroantimonate (14.8
mg, 0.015 mmol, 0.08 molar amount) was added methacrylalde-
hyde (115.6 mg, 1.5 mmol) in dichloromethane (1.0 mL). Three
portions of N-2,3-dichlorobenzylideneaniline N-oxide (16.0 mg,
0.06 mmol) in dichloromethane (0.5 mL) was added at 24 h inter-
vals to the reaction mixture at ꢁ60 ꢂC. The mixture was stirred for
1596, 1488, 1420, 1238, 1039, 755, 696 cmꢁ1
. HRMS: Calcd
for C19H19Cl2NO2: Mþ = 363.0793. Found: m=z 363.0799.
26
ꢂ
½ꢀꢃD ꢁ37:5ꢂ (c 1.029, CHCl3). HPLC (Chiralcel OD-H, 2-pro-
90 h at ꢁ60 C, followed by treatment with a solution of sodium
ꢂ
panol 1.0% in hexane, Flow 1.0 mL/min, 22.7 min (major), 35.5
min (minor), 87% ee, (S,S)-6).
borohydride (85.2 mg, 2.25 mmol) in ethanol at ꢁ60 C for 1 h,
and then left to room temperature. Standard work-up and chroma-
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
(1R ,4R ,5R )-4-(2,4-Dichlorophenyl)-5-hydroxymethyl-3-
phenyl-2-oxa-3-azabicyclo[3.3.0]octane (3i): 1H NMR (CDCl3)
ꢃ 1.11–1.16 (1H, m), 1.26–1.36 (2H, m), 1.45–1.59 (2H, m), 1.67
(1H, s), 1.98–2.02 (1H, m), 3.59 (1H, d, J ¼ 10:7 Hz), 3.65 (1H, d,
J ¼ 10:7 Hz), 4.69–4.70 (1H, m), 5.17 (1H, s), 6.90 (1H, t, J ¼ 7:3
Hz), 6.97 (2H, d, J ¼ 7:8 Hz), 7.23–7.26 (3H, m), 7.39 (1H, d, J ¼
1:5 Hz), 7.75 (1H, d, J ¼ 8:3 Hz); 13C NMR (CDCl3) ꢃ 25.1, 30.7,
30.9, 66.8, 67.2, 72.6, 88.1, 113.4, 121.1, 127.1, 128.4, 129.0,
130.6, 133.29, 133.33, 136.5, 150.3; IR (NaCl) 3434, 2959,
2875, 1596, 1488, 1469, 1039, 868, 752, 695 cmꢁ1. HRMS: Calcd
for C19H19Cl2NO2: Mþ = 363.0793. Found: m=z 363.0796.
tography on silica gel (Hexane:AcOEt = 5:1) afforded (3R ,5R )-
3-(2,3-dichlorophenyl)-5-hydroxymethyl-5-methyl-2-phenylisox-
azolidine (55.3 mg, regioisomer ratio 95/5, diastereomer ratio
>99=1) in 91% yield.
ꢀ
ꢀ
(3R ,4R )-3-(2,3-Dichlorophenyl)-4-hydroxymethyl-2-phen-
ylisoxazolidine (8, Chart 2): 1H NMR (CDCl3) ꢃ 1.58 (1H, brs),
2.26–2.86 (1H, m), 3.59–3.73 (1H, m), 3.95 (1H, dd, J ¼ 4:7, 10.7
Hz), 4.09 (1H, dd, J ¼ 5:6, 8.8 Hz), 4.25 (1H, dd, J ¼ 6:8, 8.8 Hz),
6.85–7.00 (3H, m), 7.18–7.33 (3H, m), 7.42 (1H, dd, J ¼ 1:5, 7.8
Hz), 7.74 (1H, dd, J ¼ 1:5, 7.8 Hz); 13C NMR (CDCl3) ꢃ 55.9,
62.9, 68.8, 69.3, 114.0, 121.6, 126.9, 127.9, 128.9, 129.3, 130.1,
132.9, 142.1, 150.4; IR (NaCl) 3410, 2931, 2877, 1599, 1489,
25
½ꢀꢃD ꢁ76:6ꢂ (c 1.020, CHCl3). HPLC (Chiralcel OD-H, 2-pro-
panol 1.0% in hexane, Flow 1.0 mL/min, 21.0 min (major), 33.0
min (minor), 83% ee, (S,S)-6).
1450, 1421, 1043, 756, 696 cmꢁ1
.
HRMS: Calcd for
C16H15Cl2NO2: Mþ = 323.0480. Found: m=z 323.0472. ½ꢀꢃD
ꢁ35:3ꢂ (c 0.307, CHCl3). HPLC (Chiralpak AD-H, 2-propanol
3.0% in hexane, Flow 1.0 mL/min, 17.1 min (major), 22.9 min
(minor), 79% ee, (S,S)-7).
29
(1S,4S,5S)-5-Hydroxymethyl-3-phenyl-4-(2,3,5-trichloro-
phenyl)-2-oxa-3-azabicyclo[3.3.0]octane (3j): 1H NMR (CDCl3)
ꢃ 1.08–1.26 (1H, m), 1.31–1.42 (2H, m), 1.53–1.58 (3H, m), 2.02–
2.06 (1H, m), 3.62 (1H, d, J ¼ 10:7 Hz), 3.66 (1H, d, J ¼ 10:7 Hz),
4.71–4.72 (1H, m), 5.23 (1H, s), 6.93 (1H, t, J ¼ 7:3 Hz), 6.97 (2H,
d, J ¼ 8:4 Hz), 7.25–7.29 (2H, m), 7.43 (1H, d, J ¼ 2:0 Hz), 7.77
(1H, d, J ¼ 2:0 Hz); 13C NMR (CDCl3) ꢃ 25.2, 30.5, 31.0, 66.7,
67.3, 73.6, 88.2, 113.3, 121.3, 128.0, 128.9, 129.1, 129.4, 132.9,
133.1, 141.8, 150.1; IR (KBr) 3588, 3081, 2941, 2868, 1490,
1219, 1056, 863, 764, 702 cmꢁ1. Found: C, 57.08; H, 4.70; N,
3.47%. Calcd for C19H18Cl3NO2: C, 57.24; H, 4.55; N, 3.51%.
ꢀ
ꢀ
ꢀ
(3R ,4R ,5R )-3-(2,3-Dichlorophenyl)-4-hydroxymethyl-5-
methyl-2-phenylisoxazolidine (9): 1H NMR (CDCl3) ꢃ 1.43 (3H,
d, J ¼ 6:4 Hz), 1.33–1.60 (1H, brs), 2.28–2.40 (1H, m), 3.66 (1H,
dd, J ¼ 7:3, 10.7 Hz), 3.91 (1H, dd, J ¼ 3:9, 10.7 Hz), 4.20–4.33
(1H, m), 6.84–7.01 (3H, m), 7.16–7.31 (3H, m), 7.39 (1H, dd,
J ¼ 1:5, 8.1 Hz), 7.77 (1H, dd, J ¼ 1:5, 7.8 Hz); 13C NMR
(CDCl3) ꢃ 18.1, 61.6, 63.2, 69.9, 78.0, 113.5, 121.3, 127.2,
128.0, 129.0, 129.2, 129.9, 132.6, 142.8, 151.2; IR (KBr) 3288,
2885, 1595, 1489, 1419, 1072, 1030, 785, 760, 698 cmꢁ1. HRMS:
Calcd for C17H17Cl2NO2: Mþ = 337.0636. Found: m=z 337.0620.
25
Mp 185.8–187.1 ꢂC ½ꢀꢃD ꢁ61:2ꢂ (c 1.043, CHCl3). HPLC
(Chiralcel OD-H, 2-propanol 1.0% in hexane, Flow 1.0 mL/min,
23.9 min (major), 35.8 min (minor), 85% ee, (S,S)-6). The optical-
27
½ꢀꢃD þ2:6ꢂ (c 0.749, CHCl3). HPLC (Chiralpak AD-H, 2-pro-
26
ly pure 3j was obtained with recrystallization from ethanol. ½ꢀꢃD
panol 2.0% in hexane, Flow 1.0 mL/min, 23.3 min (major), 38.9
min (minor), 63% ee, (S,S)-7).
ꢁ71:962ꢂ (c 0.650, CHCl3). The absolute configuration of cyclo-
adduct 3j was determined by X-ray analysis.
ꢀ
ꢀ
(3R ,5R )-3-(2,3-Dichlorophenyl)-5-hydroxymethyl-5-
methyl-2-phenylisoxazolidine (100): 1H NMR (CDCl3) ꢃ 1.42
(3H, s), 2.03 (1H, dd, J ¼ 7:3, 12.5 Hz), 3.12 (1H, dd, J ¼ 8:8,
12.5 Hz), 3.47 (1H, d, J ¼ 11:7 Hz), 3.54 (1H, d, J ¼ 11:7 Hz),
5.10–5.14 (1H, m), 6.85 (2H, d, J ¼ 7:8 Hz), 6.91 (1H, t, J ¼
7:3 Hz), 7.18–7.26 (3H, m), 7.39 (1H, dd, J ¼ 1:5, 7.8 Hz), 7.67
(1H, dd, J ¼ 1:5, 7.8 Hz); 13C NMR (CDCl3) ꢃ 22.1, 45.7, 66.2,
66.7, 84.2, 113.8, 121.3, 125.9, 127.9, 129.0, 129.2, 130.1,
133.1, 142.0, 150.4; IR (NaCl) 3429, 2933, 1598, 1490, 1420,
X-ray Analysis of 3j: The crystals were grown from an ethanol
solution over the course of one day. A crystal specimen of approx-
imate dimension 0:68 ꢄ 0:50 ꢄ 0:40 mm3 was chosen. The X-ray
intensities were measured on a Rigaku AFC-7R diffractometer
with Mo Kꢀ radiation. The structure was solved by direct methods
using SIR92 and refined by full-matrix least-squares calculations
on F2 using SHELXL-97. The refinement of the absolute config-
urations was performed and the structure presented yielded a value
of ꢁ0:10ð7Þ for the Flack parameter, indicating that the structure
1181, 1043, 872, 753, 694 cmꢁ1
. HRMS: Calcd for C17H17-

2206 Bull. Chem. Soc. Jpn., 76, No. 11 (2003)
Enantioselective 1,3-Dipolar Cycloaddition
29
Cl2NO2: Mþ = 337.0636. Found: m=z 337.0620. ½ꢀꢃD ꢁ55:1ꢂ
(c 0.487, CHCl3). HPLC (Chiralcel OD-H, 2-propanol 3.0% in
hexane, Flow 1.0 mL/min, 11.4 min (major), 18.0 min (minor),
82% ee, (S,S)-6).
137.0, 143.5, 151.9; IR (NaCl) 3442, 3022, 2922, 1597, 1489,
1448, 1421, 1041, 754, 694 cmꢁ1
. HRMS: Calcd for C24H24-
28
Cl2NO2: (M + 1)þ = 428.1184. Found: m=z 428.1205. ½ꢀꢃD
ꢁ75:4ꢂ (c 0.991, CHCl3). HPLC (Chiralpak AD-H, EtOH 2.0%
in hexane, Flow 1.0 mL/min, 13.2 min (major), 26.0 min (minor),
90% ee, (S,S)-6).
ꢀ
ꢀ
(3R ,5R )-5-Butyl-3-(2,3-dichlorophenyl)-5-hydroxymeth-
yl-2-phenylisoxazolidine (110): 1H NMR (CDCl3) ꢃ 0.90 (3H, t,
J ¼ 7:1 Hz), 1.20–1.47 (4H, m), 1.59–1.70 (1H, m), 1.78 (1H, t,
J ¼ 6:7 Hz), 1.82–1.93 (1H, m), 2.06 (1H, dd, J ¼ 8:1, 12.7
Hz), 3.03 (1H, dd, J ¼ 8:1, 12.7 Hz), 3.47 (1H, dd, J ¼ 6:7, 12.0
Hz), 3.58 (1H, dd, J ¼ 6:7, 12.0 Hz), 5.11 (1H, t, J ¼ 8:1 Hz),
6.81–6.86 (2H, m), 6.90 (1H, t, J ¼ 7:8 Hz), 7.17–7.26 (3H, m),
7.39 (1H, dd, J ¼ 1:5, 7.8 Hz), 7.64 (1H, dd, J ¼ 1:5, 7.8 Hz);
13C NMR (CDCl3) ꢃ 14.0, 23.2, 26.5, 34.7, 44.3, 64.7, 66.0, 86.6
113.6, 121.2, 125.8, 127.9, 129.0, 129.2, 130.0, 133.1, 142.1,
150.5; IR (NaCl) 3431, 2956, 2870, 1599, 1490, 1450, 1419,
ꢀ
ꢀ
ꢀ
Transformation of 3g into 3a. Silylation of (1R ,4R ,5R )-
4-(2-Bromophenyl)-5-hydroxymethyl-3-phenyl-2-oxa-3-azabi-
ꢀ
ꢀ
ꢀ
cyclo[3.3.0]octane (3g): To a solution of (1R ,4R ,5R )-4-(2-
bromophenyl)-5-hydroxymethyl-3-phenyl-2-oxa-3-azabicyclo-
[3.3.0]octane (3g, 70.6 mg, 0.189 mmol) in DMF (2.5 mL) was
added imidazole (38.4 mg, 0.56 mmol) and tert-butyldimethylsilyl
chloride (87.2 mg, 0.58 mmol) at 0 ꢂC. The reaction mixture was
then warmed to r.t. and stirred for 12 h. Standard work-up and
chromatography on silica gel (Hexane:AcOEt = 40:1) afforded
the corresponding silyl ether 3g0 (85.1 mg) in 92% yield.
1041, 785, 752, 694 cmꢁ1
. HRMS: Calcd for C20H22Cl2NO2:
(M ꢁ 1)þ = 378.1027. Found: m=z 378.0999. ½ꢀꢃD ꢁ64:5ꢂ (c
0.684, CHCl3). HPLC (Chiralcel OD-H, 2-propanol 5.0% in hex-
ane, Flow 1.0 mL/min, 6.0 min (major), 7.6 min (minor), 79% ee,
(S,S)-6).
(1R ,4R ,5R )-4-(2-Bromophenyl)-5-(tert-butyldimethylsil-
oxymethyl)-3-phenyl-2-oxa-3-azabicyclo[3.3.0]octane
(3g0):
28
ꢀ
ꢀ
ꢀ
1H NMR (CDCl3) ꢃ ꢁ0:143 (3H, s), ꢁ0:136 (3H, s), 0.75 (9H,
s), 1.07–1.14 (1H, m), 1.25–1.56 (4H, m), 1.97–2.02 (1H, m),
3.63 (1H, d, J ¼ 9:3 Hz), 3.70 (1H, d, J ¼ 9:3 Hz), 4.71–4.72
(1H, m), 5.21 (1H, s) 6.87 (1H, t, J ¼ 7:3 Hz), 6.96 (1H, d, J ¼
7:3 Hz), 7.13 (1H, dt, J ¼ 1:5, 7.8, 13.7 Hz), 7.21–7.20 (2H, m),
7.31 (1H, t, J ¼ 7:3 Hz), 7.55 (1H, dd, J ¼ 1:5, 8.1 Hz), 7.81
(1H, dd, J ¼ 1:5, 7.8 Hz); 13C NMR (CDCl3) ꢃ 18.3, 25.2,
25.75, 25.85, 30.67, 30.70, 67.0, 67.6, 75.0, 88.5, 113.3, 120.6,
123.3, 127.3, 128.5, 128.9, 130.1, 131.9, 140.1, 151.1; IR (NaCl)
3066, 2954, 2928, 2856, 1596, 1487, 1469, 1440, 1361, 1254,
ꢀ
ꢀ
(3R ,5R )-3-(2,3-Dichlorophenyl)-5-hexyl-5-hydroxymeth-
yl-2-phenylisoxazolidine (120): 1H NMR (CDCl3) ꢃ 0.87 (3H, t,
J ¼ 6:8 Hz), 1.15–1.48 (8H, m), 1.55–1.71 (1H, m), 1.74–1.94
(1H, m), 2.05 (1H, dd, J ¼ 7:6, 12.4 Hz), 3.03 (1H, dd, J ¼ 8:5,
12.4 Hz), 3.46 (1H, dd, J ¼ 5:1, 12.1 Hz), 3.57 (1H, dd, J ¼ 5:9,
12.1 Hz), 5.11 (1H, t, J ¼ 8:0 Hz), 6.83 (2H, d, J ¼ 7:8 Hz),
6.90 (1H, t, J ¼ 7:6 Hz), 7.15–7.30 (3H, m), 7.39 (1H, dd,
J ¼ 1:5, 8.3 Hz), 7.63 (1H, dd, J ¼ 1:5, 8.1 Hz); 13C NMR
(CDCl3) ꢃ 14.1, 22.6, 24.3, 29.8, 31.7, 34.9, 44.3, 64.7, 66.0,
86.6, 113.6, 121.2, 125.8, 127.9, 129.0, 129.1, 130.0, 133.1,
142.0, 150.5; IR (NaCl) 3444, 2929, 2858, 1599, 1491, 1419,
1271, 1041, 752, 694 cmꢁ1. HRMS: Calcd for C22H27Cl2NO2:
1105, 839, 779, 749, 695 cmꢁ1
.
Debromination of 3g0: Under nitrogen atmosphere, to a solu-
tion of the silyl ether 3g0 (45 mg, 0.092 mmol) in THF (1.5 mL)
was added n-butyllithium (1.58 M (1 M = 1 mol dmꢁ3) solution
in hexane, 0.1 mL) at ꢁ78 ꢂC. The mixture was stirred for 5
min, followed by treatment with water and warmed to room
temperature. The organic extracts were concentrated to afford
the crude debrominated product. The debrominated silyl ether
was treated with tetrabutylammonium fluoride (1 mol/L in THF,
2 mL) and stirred for 30 min at room temperature. Standard
work-up and chromatography on silica gel (Hexane:AcOEt =
Mþ = 407.1419. Found: m=z 407.1406. ½ꢀꢃD ꢁ63:0ꢂ (c
27
0.552, CHCl3). HPLC (Chiralcel OD-H, 2-propanol 3.0% in hex-
ane, Flow 1.0 mL/min, 7.0 min (major), 12.7 min (minor), 77% ee,
(S,S)-6).
(3S,5S)-3-(2,3-Dichlorophenyl)-5-hydroxymethyl-5-benzyl-
2-phenylisoxazolidine (130): 1H NMR (CDCl3) ꢃ 1.84 (1H, s),
2.12 (1H, dd, J ¼ 7:6, 12.5 Hz), 2.94 (1H, d, J ¼ 13:7 Hz), 2.98
(1H, dd, J ¼ 8:8, 12.5 Hz), 3.24 (1H, d, J ¼ 13:7 Hz), 3.44 (1H,
dd, J ¼ 5:6, 12.0 Hz), 3.51 (1H, dd, J ¼ 5:6, 12.0 Hz), 5.05 (1H,
t, J ¼ 8:3 Hz), 6.84 (2H, d, J ¼ 7:8 Hz), 6.92 (1H, t, J ¼ 7:8
Hz), 7.14 (1H, t, J ¼ 7:8 Hz), 7.18–7.33 (7H, m), 7.38 (1H, dd,
J ¼ 1:5, 7.8 Hz), 7.48 (1H, dd, J ¼ 1:5, 7.8 Hz); 13C NMR
(CDCl3) ꢃ 41.0, 44.0, 64.2, 65.9, 86.1, 114.2, 121.5, 125.8,
126.6, 128.0, 128.2, 128.9, 129.2, 130.0, 130.3, 133.0, 136.3,
141.8, 150.1; IR (NaCl) 3444, 3028, 2927, 1712, 1597, 1489,
1261, 1041, 752, 528 cmꢁ1. HRMS: Calcd for C23H21Cl2NO2:
ꢀ
ꢀ
ꢀ
4:1) afforded (1R ,4S ,5R )-5-hydroxymethyl-3,4-diphenyl-2-
oxa-3-azabicyclo[3.3.0]octane (3a, 25.0 mg) in 92% yield. HPLC
(Chiralpak AD-H, 2-propanol 2.0% in hexane, Flow 1.0 mL/min,
21.3 min (major), 32.1 min (minor), 87% ee, corresponding to
(S,S)-6).
Preparation of 15: To a solution of 2,4-dinitrophenylhydra-
zine (55.2 mg, 0.28 mmol) and sulfuric acid (370 mg) in H2O
(0.3 mL)–EtOH (1 mL), was added a solution of aldehyde (95.7
mg, 0.23 mmol), precursor of cycloadduct 130, in MeOH (15
mL) under nitrogen atmosphere. After 10 min at room tempera-
ture, the precipitate was filtered to afford the product, (3S,5S)-5-
benzyl-3-(2,3-dichlorophenyl)-5-(2,4-dinitrophenylhydrazono)-2-
phenylisoxazolidine (15), in 95% yield (130.0 mg).
X-ray Analysis of 15: The crystals were grown from THF–
hexane solution over the course of one day. A crystal specimen
of suitable approximate dimensions was chosen. The X-ray inten-
sities were measured on a Rigaku AFC-7R diffractometer with
Mo Kꢀ radiation. The structure was solved by direct methods us-
ing SIR92 and refined by full-matrix least-squares calculations on
F2 using SHELXL-97. The refinement of the absolute configura-
tions was performed and the structure presented yielded a value of
ꢁ0:1ð2Þ for the Flack parameter, indicating that the structure had
the correct absolute configurations.
Mþ = 413.0962. Found: m=z 413.0997. ½ꢀꢃD ꢁ84:1ꢂ (c
29
0.895, CHCl3). HPLC (Chiralpak AD-H, 2-propanol 10.0% in
hexane, Flow 1.0 mL/min, 10.2 min (major), 11.6 min (minor),
92% ee, (S,S)-6).
ꢀ
ꢀ
(3R ,5R )-3-(2,3-Dichlorophenyl)-5-hydroxymethyl-5-(4-
methylbenzyl)-2-phenylisoxazolidine (140): 1H NMR (CDCl3) ꢃ
1.89 (1H, s), 2.15 (1H, dd, J ¼ 8:3, 12.7 Hz), 2.31 (3H, s), 2.88
(1H, d, J ¼ 13:7 Hz), 2.96 (1H, dd, J ¼ 8:3, 12.7 Hz), 3.20 (1H,
d, J ¼ 13:7 Hz), 3.42 (1H, d, J ¼ 11:7 Hz), 3.50 (1H, d, J ¼
11:7 Hz), 5.04 (1H, t, J ¼ 8:3 Hz), 6.81–6.87 (2H, m), 6.91 (1H,
t, J ¼ 7:3 Hz), 7.05–7.17 (5H, m), 7.18–7.25 (2H, m), 7.37 (1H,
dd, J ¼ 1:6, 8.2 Hz), 7.50 (1H, dd, J ¼ 1:6, 8.2 Hz); 13C NMR
(CD3OD) ꢃ 21.2, 41.2, 44.8, 65.0, 67.1, 87.4, 115.6, 122.3,
127.4, 129.1, 129.5, 129.6, 130.1, 130.9, 131.6, 133.9, 134.9,

S. Kezuka et al.
Bull. Chem. Soc. Jpn., 76, No. 11 (2003) 2207
addition Reaction in Organic Synthesis,’’ ed by S. Kobayashi and
K. A. Jꢀrgensen, Wiley-VCH, Weinheim (2002), p. 274.
The authors thank Dr. Takashiro Akitsu, Keio University, for
X-ray analyses.
9
Soc., 124, 4968 (2002).
10 W. S. Jen, J. J. M. Wiener, and D. W. C. MacMillan, J. Am.
Chem. Soc., 122, 9874 (2000).
F. Viton, G. Bernardinelli, and E. P. Kundig, J. Am. Chem.
¨
References
1
a) J. J. Tufariello, ‘‘1,3-Dipolar Cycloaddition Chemistry,’’
ed by A. Padwa, Wiley, New York (1984), Vol. 2, p. 83. b) K. B. G.
Trossell, ‘‘Nitrile Oxides, Nitrones and Nitronates in Organic Syn-
thesis,’’ VCH, New York (1988). c) ‘‘Synthetic Applications of
1,3-Dipolar Cycloaddition Chemistry toward Heterocycles and
Natural Products,’’ ed by A. Padwa and P. W. Pearson, Wiley &
Sons, New York (2002).
11 a) T. Nagata, K. Yorozu, T. Yamada, and T. Mukaiyama,
Angew. Chem., Int. Ed. Engl., 34, 2145 (1995). b) K. D. Sugi, T.
Nagata, T. Yamada, and T. Mukaiyama, Chem. Lett., 1997, 493.
c) T. Yamada, Y. Ohtsuka, and T. Ikeno, Chem. Lett., 1998, 1129.
12 a) T. Yamada, T. Ikeno, H. Sekino, and M. Sato, Chem.
Lett., 1999, 719. b) T. Ikeno, M. Sato, and T. Yamada, Chem. Lett.,
1999, 1345. c) T. Ikeno, A. Nishizuka, M. Sato, and T. Yamada,
Synlett, 2001, 406. d) T. Ikeno, M. Sato, H. Sekino, A. Nishizuka,
and T. Yamada, Bull. Chem. Soc. Jpn., 74, 2139 (2001).
13 T. Yamada, S. Kezuka, T. Mita, and T. Ikeno, Heterocycles,
52, 1041 (2000).
2
3
4
R. Huisgen, Angew. Chem., 75, 604 (1963).
M. Fredrickson, Tetrahedron, 53, 403 (1997).
S. Kanemasa, T. Uemura, and E. Wada, Tetrahedron Lett.,
33, 7889 (1992).
5
a) K. V. Gothelf and K. A. Jꢀrgensen, J. Org. Chem., 59,
5687 (1994). b) K. V. Gothelf, R. G. Hazell, and K. A. Jꢀrgensen,
J. Org. Chem., 61, 346 (1996). c) S. Crosignani, G. G. Desimoni,
G. Faita, S. Fillippone, A. Mortoni, P. Righetti, and M. Zema, Tet-
rahedron Lett., 40, 7007 (1999). d) A. I. Sanchez-Blanco, K. V.
Gothelf, and K. A. Jꢀrgensen, Tetrahedron Lett., 38, 7923
(1997). e) S. Kobayashi and M. Kawamura, J. Am. Chem. Soc.,
120, 5840 (1998). f) S. Kanemasa, Y. Oderaotoshi, J. Tanaka,
and E. Wada, J. Am. Chem. Soc., 120, 12355 (1998). g) K. Hori,
H. Kodama, T. Ohta, and I. Furukawa, J. Org. Chem., 64, 5017
(1999). h) S. Iwasa, S. Tsushima, T. Shimada, and H. Nishiyama,
Tetrahedron, 58, 227 (2002).
14 a) S. Kezuka, T. Mita, N. Ohtsuki, T. Ikeno, and T.
Yamada, Chem. Lett., 2000, 824. b) S. Kezuka, T. Mita, N.
Ohtsuki, T. Ikeno, and T. Yamada, Bull. Chem. Soc. Jpn., 74,
1333 (2001).
15 a) S. Kezuka, T. Ikeno, and T. Yamada, Org. Lett., 3, 1937
(2001). b) S. Kezuka, Y. Kogami, T. Ikeno, and T. Yamada, Bull.
Chem. Soc. Jpn., 76, 49 (2003).
16 The reaction was quenched with an ethanolic solution of
NaBH4 for the transformation of the produced aldehyde into the
stable primary alcohol. It was confirmed that the endo/exo ratio
and the enantioselectivity were completely retained after this treat-
ment. Without NaBH4 treatment in the reaction of 1a and 2a cata-
lyzed by the complex 4e (Entry 5, Table 1), the resulting aldehyde
was isolated by the rapid work-up in 81% yield with 99% endo-se-
lectivity and 51% ee. In this article, the chemical yield, diastereo-
selectivity, and enantioselectivity listed in Tables 1–5 were deter-
mined as the corresponding alcohol obtained after reductive
quenching.
6
a) J.-P. G. Seerden, A. W. A. Scholte op Reimer, and H. W.
Scheeren, Tetrahedron Lett., 35, 4419 (1994). b) J.-P. G. Seerden,
M. M. M. Boeren, and H. W. Scheeren, Tetrahedron, 53, 11843
(1997). c) K. B. Jensen, R. G. Hazell, and K. A. Jꢀrgensen, J.
´
Org. Chem., 64, 2353 (1999). d) K. B. Simonsen, P. Bayon, R.
G. Hazell, K. V. Gothelf, and K. A. Jꢀrgensen, J. Am. Chem.
Soc., 121, 3845 (1999).
7
K. B. Jensen, K. V. Gothelf, R. G. Hazell, and K. A.
Jꢀrgensen, J. Org. Chem., 62, 2471 (1997).
a) M. Shirahase, Y. Oderaotoshi, and S. Kanemasa, The
17 S. Kanemasa, Nippon Kagaku Kaishi, 2000, 155.
18 a) O. Kamm, Org. Synth., Coll. Vol. 2, 445 (1941). b) O. H.
Wheeler and P. H. Gore, J. Am. Chem. Soc., 78, 3363 (1956).
19 J. B. Brown, H. B. Henbest, and E. R. H. Jones, J. Chem.
Soc., 1950, 3634.
8
78th Annual Meeting of the Chemical Society of Japan, March
2000, 4G1 08, Funabashi, Japan. b) M. Shirahase and S. Kanemasa,
The 80th Annual Meeting of the Chemical Society of Japan, March
2001, 2H6 08 and 2H6 09, Kobe, Japan. c) S. Kanemasa, ‘‘Cyclo-
20 J. L. Roberts, P. S. Borromeo, and C. D. Poulter, Tetrahe-
dron Lett., 19, 1621 (1977).