Catalytic enantioselective inverse-electron demand 1,3-dipolar cycloaddition reactions of nitrones with alkenes

Article abstract of DOI:10.1021/ja983915a

A general reaction protocol for catalytic enantioselective 1,3-dipolar cycloaddition reaction of nitrones, activated by chiral Lewis acids, with electron-rich alkenes has been developed. The nitrones are activated by various chiral 2,2'-dihydroxy-1,1'-binaphthyl (BINOL)-AlMe complexes, and it has been found that 3,3'-diaryl-BINOL-AlMe complexes catalyze a highly regio- , diastereo-, and enantioselective 1,3-dipolar cycloaddition reaction of aromatic nitrones with vinyl ethers, giving the exo-diastereomer of the isoxazolidines with de's up to > 90% and up to 97% ee. The reaction has been investigated under various conditions with different nitrones and vinyl ethers (and alkenes), and a general synthetic procedure is presented. The mechanism for the reaction is discussed on the basis of a linear stereochemical effect of the catalyst, the diastereoselectivity, and absolute stereochemistry of the isoxazolidines formed, and theoretical calculations of the 3,3'-diphenyl-BINOL-AlMe-nitrone intermediate.

Full text of DOI:10.1021/ja983915a

J. Am. Chem. Soc. 1999, 121, 3845-3853
3845
Catalytic Enantioselective Inverse-Electron Demand 1,3-Dipolar
Cycloaddition Reactions of Nitrones with Alkenes
Klaus B. Simonsen, Pau Bayo´n, Rita G. Hazell, Kurt V. Gothelf, and
Karl Anker Jørgensen*
Contribution from the Center for Metal Catalyzed Reactions, Department of Chemistry, Aarhus UniVersity,
DK-8000 Aarhus C, Denmark
ReceiVed NoVember 12, 1998
Abstract: A general reaction protocol for catalytic enantioselective 1,3-dipolar cycloaddition reaction of nitrones,
activated by chiral Lewis acids, with electron-rich alkenes has been developed. The nitrones are activated by
various chiral 2,2′-dihydroxy-1,1′-binaphthyl (BINOL)-AlMe complexes, and it has been found that 3,3′-
diaryl-BINOL-AlMe complexes catalyze a highly regio-, diastereo-, and enantioselective 1,3-dipolar cycload-
dition reaction of aromatic nitrones with vinyl ethers, giving the exo-diastereomer of the isoxazolidines with
de’s up to >90% and up to 97% ee. The reaction has been investigated under various conditions with different
nitrones and vinyl ethers (and alkenes), and a general synthetic procedure is presented. The mechanism for the
reaction is discussed on the basis of a linear stereochemical effect of the catalyst, the diastereoselectivity, and
absolute stereochemistry of the isoxazolidines formed, and theoretical calculations of the 3,3′-diphenyl-BINOL-
AlMe-nitrone intermediate.
Introduction
The Lewis acid-catalyzed 1,3-dipolar cycloaddition reaction
using this approach has mainly involved acryloyloxazolidinones
as the alkene fragment reacting with various types of nitrones
(eq 2).4 By the use of this catalytic enantioselective approach,
control of both regio-, diastereo-, and enantioselectivity has been
achieved to a high degree.
The 1,3-dipolar cycloaddition reaction is a very important
reaction for the construction of five-membered heterocycles and
has been used in numerous syntheses using 1,3-dipoles such as
nitrones, nitrile oxides, azomethine ylides, and nitronates.1 The
1,3-dipolar cycloaddition reaction between nitrone 1 and alkene
2 gives the isoxazolidine 3 (eq 1), which is of importance for
the preparation of natural products such as â-amino alcohols
and alkaloids.1-3
(1)
(2)
The development of the 1,3-dipolar cycloaddition reaction
of nitrones with alkenes has recently taken a new direction, as
one of the major challenges now is to prepare optically active
isoxazolidines.3 The typical 1,3-dipolar cycloaddition reaction
of nitrones with alkenes involves a dominant interaction of the
HOMOnitrone and the LUMOalkene, resulting in eight possible
stereoisomers. These stereoisomers are two regioisomers, each
as two diastereomers, and each of these two diastereomers as
two enantiomers.3a To prepare enantiopure isoxazolidines,
several types of chiral auxiliaries have been used, including
chiral nitrones and chiral alkenes.3a However, this has several
disadvantages, such as waste of optically active material and
elevated temperatures. This can be circumvented using chiral
Lewis acids for the activation of the alkene, resulting in a
lowering of the energy of the LUMOalkene and enhancement of
the regio-, diastereo-, and enantioselectivity of the reaction.3a
In 1982 DeShong et al. reported the first 1,3-dipolar cycload-
dition reaction of electron-rich alkenes, such as vinyl acetate
and vinyl ethers, with nitrones.5 This work was extended to a
stereoselective approach when chiral nitrones were used.6
Application of chiral nitrones was investigated by Overton and
colleagues for the preparation of optically pure â-lysine.7 The
cycloaddition of chiral vinyl ethers, derived from chiral alcohols,
with nitrones has also been investigated.8
The inverse-electron demand 1,3-dipolar cycloaddition reac-
tion of nitrones with alkenes requires a dominant interaction of
the LUMOnitrone with the HOMOalkene.
3a Such a reaction requires
an activation of the nitrone, and only very few examples are
known in which the nitrone is activated by a Lewis acid followed
by the reaction with an electron-rich alkene.9 To our knowledge
no general catalytic enantioselective inverse-electron demand
1,3-dipolar cycloaddition reactions of nitrones with alkenes
showing high enantioselectivity have been reported. Seerden
et al. have used vinyl ethers and ketene dialkyl acetals for the
reaction with nitrones using oxazaborolidones as the catalysts,
leading to control of the diastereoselectivity and up to 74% ee,
but the product was formed in low yield.9a Herein we report
(1) For general reviews see: (a) Tufariello, J. J. 1,3-Dipolar Cycload-
dition Chemistry; Padwa, A., Ed; Wiley: New York, 1984. (b) Torssell,
K. B. G. Nitrile Oxides, Nitrones and Nitronates in Organic Synthesis;
VCH: New York, 1988.
(2) Tufariello, J. J. Acc. Chem. Res. 1979, 12, 396.
(3) (a) Gothelf, K. V.; Jørgensen, K. A. Chem. ReV. 1998, 98, 863. (b)
Frederickson, M. Tetrahedron 1997, 53, 403.
10.1021/ja983915a CCC: $18.00 © 1999 American Chemical Society
Published on Web 04/08/1999

3846 J. Am. Chem. Soc., Vol. 121, No. 16, 1999
Simonsen et al.
Table 1. Diastero- and Enantioselectivity of the 1,3-Dipolar
the first example of a catalytic 1,3-dipolar cycloaddition reaction
with inverse-electron demand of aromatic nitrones with vinyl
ethers catalyzed by chiral aluminum complexes giving isoxa-
zolidines in high yield, with diastereo- and enantioselectivity.
Cycloaddition Reaction of Nitrone 4a with tert-Butyl Vinyl Ether
5a Catalyzed by Different BINOL-AlMe Complexes 8a-ga
reaction
ee (exo-6a)c
entry catalyst time conversionb endo-6a:exo-6ab
(%)
1
2
3
4
5
6
7
8
9
18 hd
AlMe3 4 h
>95
>90
>90
>95
>95
>90
>90
>95
80
<5:>95
17:73
73:27
<5:>95
<5:>95
21:79
12:88
10:90
14:86
0
0
Results and Discussion
8a
8b
8c
8d
8e
8f
3 h
45 min
2 h
18 h
2.5 h
3 h
<5
89
87
79
78
80
65e
To activate the nitrone by a Lewis acid for the 1,3-dipolar
cycloaddition reaction with an electron-rich alkene, one has to
consider the interaction of the nitrone with the Lewis acid. To
activate simple nitrones such as benzylidenephenylamine N-
oxide 4a, the nitrone oxygen atom has to coordinate to the Lewis
acid, and thus chiral Lewis acids that allow a monodentate
coordination should be considered as the catalysts of choice.3a
By screening several chiral Lewis acid complexes we have found
that the combination of a chiral (R)-2,2′-dihydroxy-1,1′-binaph-
thyl (BINOL) derivative as the ligand with aluminum complexes
leads to catalysts that are very useful for the inverse-electron
demand 1,3-dipolar cycloaddition reaction of nitrones with
alkenes.
8g
48 h
a Reaction conditions: solvent: CH2Cl2; scale: 4 (0.1 mmol), 5a
(0.5 mmol), and 20 mol % 8a-g. For further details see Experimental
Section. b Determined by 1H NMR spectroscopy of the crude product.
c Ee of the exo isomer was determined by HPLC using a Daicel
Chiralcel OD column. d Performed in neat tert-butyl vinyl ether (1 mL)
at 50 °C. e Other enantiomer.
The results from the reaction of nitrone 4a with 5a in the
presence of AlMe3 as the catalyst stimulated the investigation
of several chiral aluminum complexes as the asymmetric
catalysts. Chiral aluminum catalysts, prepared by reacting AlMe3
or AlCl3 with chiral diols or bis(sulfonamides), have been used
successfully in Diels-Alder10 and hetero-Diels-Alder reac-
tions.11 The use of BINOL (7a) and a 3,3′-substituted BINOL
such as 3,3′-diphenyl BINOL 7b and 3,3′-bis(triphenylsilyl)-
BINOL 7g as ligands in combination with aluminum complexes
The 1,3-dipolar cycloaddition reaction of nitrone 4a with tert-
butyl vinyl ether 5a has been investigated as a model reaction
(eq 3) using chiral catalysts 8a-g obtained from the reaction
of AlMe3 with various (R)-BINOL derivatives 7a-f and (S)-
3,3′-bis(triphenylsilyl)BINOL 7g (eq 4).
(4) (a) Murahashi, S.; Imada, Y.; Kohno, M.; Kawakami, T. Synlett 1993,
395. (b) Gothelf, K. V.; Jørgensen, K. A. J. Org. Chem. 1994, 59, 5687.
(c) Gothelf, K. V.; Thomsen, I.; Jørgensen, K. A. J. Am. Chem. Soc. 1996,
118, 59. (d) Gothelf, K. V.; Hazell, R. G.; Jørgensen, K. A. J. Org. Chem.
1996, 61, 346. (e) Jensen, K. B.; Gothelf, K. V.; Hazell, R. G.; Jørgensen,
K. A. J. Org. Chem. 1997, 62, 2471. (f) Gothelf, K. V.; Jørgensen, K. A.
Acta Chem. Scand. 1996, 50, 652. (g) Jensen, K. B.; Gothelf, K. V.;
Jørgensen, K. A. HelV. Chim. Acta 1997, 80, 2039. (h) Gothelf, K. V.;
Hazell, R. G.; Jørgensen, K. A. J. Org. Chem. 1998, 63, 5483. (i) Kobayashi,
S.; Akiyama, R.; Kawamura, M.; Ishitani, H. Chem. Lett. 1997, 1039. (j)
Kobayashi, S.; Kawamura, M. J. Am. Chem. Soc. 1998, 120, 5840. (k) Hori,
K.; Kodama, H.; Ohta, T.; Furukawa, I. Tetrahedron Lett. 1996, 37, 5947.
(l) Hori, K.; Ito, J.; Ohta, T.; Furukawa, I. Tetrahedron 1998, 54, 12737.
For other types of Lewis acid-catalyzed 1,3-dipolar cycloaddition reactions
see: (m) Kanemasa, S.; Uemura, T.; Wada, E. Tetrahedron Lett. 1992, 33,
7889. (n) Kanemasa, S.; Tsuruoka, T.; Wada, E. Tetrahedron Lett. 1993,
34, 87. (o) Kanemasa, S.; Tsuruoka, T.; Yamamoto, H. Tetrahedron Lett.
1995, 36, 5019. (p) Kanemasa, S.; Tsuruoka, T. Chem. Lett. 1995, 49. (q)
Katagari, N.; Okada, M.; Morishita, Y.; Kaneko, C. J. Chem. Soc., Chem.
Commun. 1996, 2137. (r) Tokunaga, Y.; Ihara, M.; Fukumoto, K.
Tetrahedron Lett. 1996, 37, 6157. (s) Minakata, S.; Ezoe, T.; Nakamura,
K.; Ryu, I.; Komatsu, M. Tetrahedron Lett. 1998, 39, 5205. (t) Gilbertson,
S. R.; Dawson, D. P.; Lopez, O. D.; Marshall, K. L. J. Am. Chem. Soc.
1995, 117, 4431. (u) Gilbertson, S. R.; Lopez, O. D. J. Am. Chem. Soc.
1997, 119, 3399. (v) Ukaji, Y.; Taniguchi, K.; Sada, K.; Inomata, K. Chem.
Lett. 1997, 547.
(3)
(5) DeShong, P.; Dicken, C. M.; Staib, R. R.; Freyer, A. J.; Weinreb, S.
M. J. Org. Chem. 1982, 47, 4397.
(6) DeShong, P.; Dicken, C. M.; Leginus, J. M.; Whittle, R. R. J. Am.
Chem. Soc. 1984, 106, 5598.
(7) (a) Keirs, D.; Moffat, D.; Overton, K. J. Chem. Soc., Chem. Commun.
1988, 654. (b) Keirs, D.; Moffat, D.; Overton, K.; Tomanek, R. J. Chem.
Soc., Perkin Trans. 1 1991, 1041.
(8) Carruthers, W.; Coggins, P.; Weston, J. B. J. Chem. Soc., Chem.
Commun. 1991, 117.
(9) (a) Seerden, J.-P. G.; Scholte op Reimer, A. W. A.; Scheeren, H. W.
Tetrahedron Lett. 1994, 35, 4419. (b) Seerden, J.-P. G.; Kuypers, M. M.
M.; Scheeren, H. W. Tetrahedron: Asymmetry 1995, 6, 1441. (c) Seerden,
J.-P. G.; Boeren, M. M. M.; Scheeren, H. W. Tetrahedron 1997, 53, 11843.
(10) See, e.g., (a) Chapuis, C.; Jurczak, J. HelV. Chim. Acta 1987, 70,
436. (b) Corey, E. J.; Imwinkelried, R.; Pikul, S.; Xiang, Y. B. J. Am. Chem.
Soc. 1989, 111, 5493. (c) Corey, E. J.; Imai, N.; Pikul, S. Tetrahedron
Lett. 1991, 32, 7517.
The reaction of nitrone 4a with tert-butyl vinyl ether 5a gives
exclusively the 5-substituted isoxazolidine 6a, both with and
without catalyst (eq 3). This regioselectivity is normal for the
1,3-dipolar cycloaddition reaction of electron-rich alkenes, due
to the attack of the nitrone oxygen atom to the oxygen-
substituted alkene carbon atom.3a,5 In the absence of a catalyst
the reaction had to proceed at 50 °C in neat 5a for 18 h to give
complete conversion and gave the isoxazolidine exo-6a with
high exo selectivity (Table 1, entry 1). However, in the presence
of AlMe3 as the catalyst (20 mol %) the reaction was complete
within 4 h at ambient temperature, although with lower
diastereoselectivity (entry 2).
(11) (a) Maruoka, K.; Itoh, T.; Shirasaka, T.; Yamamoto, H. J. Am. Chem.
Soc. 1988, 110, 310.

1,3-Dipolar Cycloaddition Reactions of Nitrones with Alkenes
J. Am. Chem. Soc., Vol. 121, No. 16, 1999 3847
has resulted in reactions with very high enantioselectivities.10-12
We have recently reported a simple and efficient route to the
3,3′-diaryl BINOLs (R)-7b-f by a Suzuki cross-coupling
reaction, which makes this class of chiral ligands easily available
in three steps from BINOL using standard reactions.13
Table 2. Influence of Solvent, Temperature, and Catalyst Amount
on Diastereo- and Enantioselectivity of the 1,3-Dipolar
Cycloaddition Reaction between Nitrone 4a and tert-Butyl Vinyl
Ether 5a Catalyzed by BINOL-AlMe Complex 8ba
The aluminum complexes 8a-g of (R)-BINOL, (7a) 3,3′-
diaryl substituted (R)-BINOLs 7b-f, and (S)-3,3′-bis(triphenyl-
silyl) BINOL 7g14 were prepared in situ in dry CH2Cl2 by
dropwise addition of a 2 M solution of AlMe3 in heptane with
evolution of CH4 (eq 4). The mixture was agitated for 1 h and
a solution of nitrone 4a was added, followed by addition of
freshly distilled tert-butyl vinyl ether 5a (5 equiv). These
reactions were carried out on a 0.1-mmol scale with 20 mol %
catalyst, and the results are presented in Table 1. The diaste-
reoselectivity of the reaction was determined by 1H NMR
spectroscopy of the crude product using the following diagnostic
signals:5 exo-6a gives a doublet of doublets at δ ) 4.42 ppm
with J ) 7.0 and 9.0 Hz for the proton on C-3 and two doublets
of doublets of doublets at δ ) 2.34 ppm (J ) 3.3, 6.8, and
12.9 Hz) and 2.97 ppm (J ) 6.0, 9.3, and 13.2 Hz) for the two
protons on C-4, whereas endo-6a gives a doublet of doublets
at δ ) 4.81 ppm (J ) 6.6 and 9.3 Hz) and two doublets of
doublets of doublets at δ ) 2.51 ppm (J ) 5.1, 9.3, and 12.2
Hz) and 2.63 ppm (J ) 1.5, 6.9, and 12.1 Hz). The diastereomers
were separated by column chromatography and the enantiomeric
excess (ee) of the major product (exo diastereomer) was
determined by HPLC using a Daicel Chiralcel OD column. It
is important to note that the BINOL ligands can be recoVered
from the chromatography and can be reused without any
decrease in reactiVity and stereoselectiVity, which makes this
catalytic approach Very attractiVe.
The use of the simple (R)-BINOL-AlMe catalyst 8a resulted
in almost complete conversion after 3 h and in a diastereose-
lectivity of 73:27 in favor of endo-6a, but with <5% ee (Table
1, entry 3). The introduction of aryl or silyl substituents in the
3,3′-position of the ligand changed the diastereoselectivity
dramatically compared with the unsubstituted BINOL-AlMe
catalyst 8a. The reaction now proceeds with high exo selectivity,
as observed using the catalysts 8b-g (entries 4-9). Further-
more, these complexes induce high enantioselectivity with ee’s
in the range 65-89%. The most effective catalyst is the (R)-
BINOL-AlMe complex 8b. In this case the reaction was
completed within 45 min, giving exo-6a as the sole product
(de >90%) with 89% ee (entry 4). The same diastereoselectivity
was observed for the reaction catalyzed by 8c, but with a slight
decrease in enantioselectivity (entry 5). When two methyl
substituents were attached to the phenyl group in the 2,6-
position, as is the case for catalyst 8d (instead of the 3,5- position
as in catalyst 8c), a decrease of both reactivity and selectivity
was observed. Furthermore, the reaction has to proceed for 18
h to obtain high conversion, but at the expense of both exo
selectivity and asymmetric induction, as only 79% ee was
obtained (entry 6). Steric interactions of the methyl groups could
explain the decrease in rate with 8d. Similar chiral inductions
were observed for the catalysts 8e and 8f (entries 7, 8). These
reactions were also relatively fast and complete conversions
were achieved after 2.5-3 h with good exo selectivities. The
molar mol
scale
entry (mmol) 8b solvent
conversionb/ isolated
ee
exoc
%
reaction
time
yield,
%
endo:exob (%)
1
2
3
4
5
6
7
8
9
0.1
0.1
1.0
0.2
0.2
0.1
0.1
0.2
0.2
0.2
20
10
10
5
CH2Cl2 >95/45 min
CH2Cl2 >95/45 min
CH2Cl2 >95/60 min
CH2Cl2 50/48 h
61
70
84
32
30
71
68
67
57
74
<5:>95 89
<5:>95 87
<5:>95 89
21:79 80
23:77 79
11:89 89
2.5 CH2Cl2 48/48 h
10
10
20
20
20
d
CH2Cl2 >95/3 h
e
CH2Cl2 >90/18 h
22:78 92
THF
>95/3 h
<5:>95 89
<5:>95 92
<5:>95 93
toluene >95/2 h
toluene/ >95/45 min
hexane
10
a Reaction conditions: the molar scale is defined from 4 (1 equiv),
5a (5 equiv). For further details see Experimental Section. b Determined
by 1H NMR spectroscopy of the crude product. c Ee of the exo isomer
was determined by HPLC using a Daicel Chiralcel OD column.
d Reaction temperature 0 °C. e Reaction temperature -25 °C.
slowest reaction was observed when catalyst 8g was applied
(entry 9). Prolonged reaction time was required to get acceptable
conversion, and still only 80% conversion resulted after 48 h.
The endo:exo ratio of 14:86 is comparable to those obtained
using the other ligands, but a substantial decrease in the
enantioselectivity is observed, as only 65% ee is found. The
very bulky triphenylsilyl substituents in catalyst 8g probably
retard the coordination of the nitrone to the catalyst, resulting
in less facial discrimination.
Using a chiral (R)-BINOL-AlCl complex, prepared by react-
ing (R)-BINOL 7b with Me2AlCl as the Lewis acid instead of
AlMe3, as catalyst for the reaction between nitrone 4a and tert-
butyl vinyl ether 5a resulted in an alteration of the selectivity.
In this case the endo:exo ratio was 60:40, and, furthermore, the
chiral induction decreased to 52% ee for exo-6a and to 27% ee
for endo-6a. The reason for the change in selectivity might be
due to dimerization of the catalyst, changing the reaction sphere
of the Lewis acid site.
To optimize the enantioselectivity of the catalytic 1,3-dipolar
cycloaddition reaction of nitrone 4a with tert-butyl vinyl ether
5a using 8b as the catalyst (eq 3), the influence of catalyst
amount, molar scale, temperature, and solvent was examined.
These results are presented in Table 2.
It appears from Table 2, entries 1 and 2, that no appreciable
change is observed when going from 20 mol % catalyst to 10%,
as both reactions are complete after 45 min and proceed with
similar selectivities. However, when the amount of catalyst 8b
is reduced to 5 mol % and 2.5 mol %, the reaction is slow and
only 50% of the nitrone is consumed after 2 days. Also, a slight
decrease in diastereo- and enantioselectivity is observed (entries
4, 5). The reaction can be scaled up to 1.0 mmol by using 10
mol % catalyst without any decrease in reaction time and
selectivity, and the isolated yield of exo-6a was 84% and the
ee 89% (entry 3). When the reaction was carried out at low
temperature (-78 °C), no reaction occurred within 3 days.
However, at 0 °C a conversion of 95% was accomplished after
3 h with the same high asymmetric induction, but with a slight
decrease in the exo selectivity (entry 6). At -25 °C the reaction
was complete after 18 h with a very high ee of 92% of exo-6a,
(12) (a) Kelly, T. R.; Whiting, A.; Chandrakumar, N. S. J. Am. Chem.
Soc. 1986, 108, 3510. (b) Maruoka, K.; Hoshino, Y.; Shirasaka, T.;
Yamamoto, H. Tetrahedron Lett. 1988, 29, 3967. (c) Maruoka, K.; Banno,
H.; Yamamoto, H. J. Am. Chem. Soc. 1990, 112, 7791. (d) Maruoka, K.;
Banno, H.; Yamamoto, H. Tetrahedron: Asymmetry 1991, 2, 647.
(13) Simonsen, K. B.; Gothelf, K. V.; Jørgensen, K. A. J. Org. Chem.
1998, 63, 7536.
(14) Maruoka, K.; Itoh, T.; Araki, Y.; Shirasaka, T.; Yamamoto, H. Bull.
Chem. Soc. Jpn. 1988, 61, 2975.

3848 J. Am. Chem. Soc., Vol. 121, No. 16, 1999
Simonsen et al.
To get more information about the nature of the intermediate
complex involved in this reaction, the reaction was also tested
for nonlinear stereochemical effects of the catalyst.15 A graphic
representation of the ee of exo-6a, obtained by reaction of
nitrone 4a with tert-butyl vinyl ether 5a using the (R)-BINOL-
AlMe complex 8b as the catalyst, as a function of the optical
purity of 8b is depicted in Figure 1.
A linear relation between the ee of the catalyst and ee of the
exo-6a is observed and it is therefore reasonable to believe that
the catalytic intermediate consists of a discrete complex in which
one (R)-BINOL-ALMe is involved (see the mechanistic discus-
sion).
To illustrate a more general application of the BINOL-AlMe
8b-catalyzed 1,3-dipolar cycloaddition reaction of nitrones with
electron-rich alkenes, the reaction was investigated using various
substrates. First, the variation of the substituents of the nitrone
was investigated. The presence of an aromatic substituent on
the nitrone nitrogen atom seems to be essential in the present
reaction, as no reaction occurred when benzylidenebenzylamine
N-oxide and benzylidene-N-propylamine N-oxide were used
under the same reaction conditions. The reaction of various
aromatic nitrones 4a-d with tert-butyl vinyl ether 5a, ethyl
vinyl ether 5b and benzyl vinyl ether 5c catalyzed by the
BINOL-AlMe complex 8b (eq 5) has been studied and the
results are given in Table 3.
Figure 1. Linear relation between the optical purity of isoxazolidine
exo-6a formed in the 1,3-dipolar cycloaddition reaction of nitrone 4a
and tert-butyl vinyl ether 5a and the ee of the employed catalyst 8b.
Table 3. 1,3-Dipolar Cycloaddition Reactions of Aromatic
Nitrones 4a-d with tert-Butyl Vinyl Ether 5a, Ethyl Vinyl Ether
5b, and Benzyl Vinyl Ether 5c in the Presence of 10 mol % of
BINOL-AlMe 8b Catalysta
yield/
reaction
time
nitrone 4,
alkene 5,
ee
entry
R1
R2
product
endo:exob exoc
1
2
3
4
5
6
7
Ph-4a
p-tolyl-4b
p-ClPh-4c
p-MeOPh-4d t-Bu-5a
Ph-4a
t-Bu-5a
t-Bu-5a
t-Bu-5a
6a
6b
6c
6d
6e
6f
6g
6g
6h
6i
84/1 h
71/12 h
74/2 h
65/18 h
79/1 h
70/3 h
78/2 h
66/6 h
<5:>95
16:84
17:83
89
81
78
77
91
90
94
97
88
85
(5)
10:90
Et-5b
Et-5b
Et-5b
Et-5b
<5:>95
<5:>95
<5:>95
<5:>95
p-tolyl-4b
p-ClPh-4c
8d p-ClPh-4c
9
10
p-MeOPh-4d Et-5b
p-ClPh-4d Bn-5c
50/18 h <5:>95
68/18 h <5:>95
a The reactions were performed on a 0.5-2.0 mmol scale in CH2Cl2
at room temperature. For further details see Experimental Section. b The
endo:exo ratio was determined by 1H NMR spectroscopy of the crude
product. c Ee of the exo isomer was determined by HPLC using a Daicel
Chiralcel OD column. d Toluene as the solvent.
although at the expense of exo selectivity (entry 7). The effect
of the solvent was also investigated (entries 8-10). When the
1,3-dipolar cycloaddition reaction of nitrone 4a with 5a was
carried out in THF the reaction had to proceed for 3 h before
4a was consumed, but the same diastereo- and enantioselectivity
was observed (entry 8). In toluene the reaction was finished
within 2 h and exo-6a was obtained with 92% ee and very high
exo selectivity, but with a lower isolated yield (entry 9). To
perform the reaction in a solvent with even lower polarity than
toluene, the reaction was first carried out in hexane, but
unfortunately the reaction did not proceed at all because of the
insolubility of the nitrone in this solvent. When the 1,3-dipolar
cycloaddition reaction of 4a with 5a in the presence of catalyst
8b was carried out in a mixture of toluene and hexane, the same
high diastereo- and enantioselectivity was found but with a
significant increase in isolated yield (entry 10).
The 1,3-dipolar cycloaddition reaction of the various nitrones
4a-d proceed well with the different aryl substituents on the
nitrone carbon atom (eq 5). In general, high diastereo- and
enantioselectivities are obtained as presented in Table 3. The
addition of tert-butyl vinyl ether 5a to the nitrones 4b or 4c in
the presence of 10 mol % catalyst 8b proceeds like the reaction
involving 4a, although with a slight decrease in diastereo- and
enantioselectivity (entries 2, 3). Full conversion of 4b and 4c
was obtained after 12 h and 2 h, respectively, and the
isoxazolidines exo-6b and exo-6c were isolated in 71-74%
yield. The same enantioselectivity was also found for nitrone
(15) See, e.g., (a) Girard, C.; Kagan, H. B. Angew. Chem. 1998, 110,
3088. (b) Avalos, M.; Babiano, R.; Cintas, P.; Jime´nez, J. L.; Palacios, J.
C. Tetrahedron: Asymmetry 1997, 8, 2997.

1,3-Dipolar Cycloaddition Reactions of Nitrones with Alkenes
J. Am. Chem. Soc., Vol. 121, No. 16, 1999 3849
The reaction has also been conducted with the electron-rich
alkenes dihydrofuran 5d, vinyl acetate, and the nonactivated
alkene, styrene (5e) in the presence of 10 mol % catalyst 8b
(eq 6). However, only moderate results were obtained. The
cyclic vinyl ether 5d reacted quite slowly with nitrone 4a, and
the reaction was complete after 12 h but with low diastereose-
lectivity and very low ee (<5%). Prolonged reaction time was
also necessary for good conversion in the case of styrene 5e,
and although high diastereoselectivity was obtained, the isoxa-
zolidine exo-6k was isolated as a racemic compound. When
the reaction was carried out with vinyl acetate, no reaction was
observed after 3 days.
Figure 2. Crystal structure of exo-6g. Thermal ellipsoids are shown
at 50% probability levels; selected hydrogen atoms are drawn as small
circles of arbitrary radius.
4d, but the reaction required an increased reaction time of 18 h
to achieve an acceptable yield (entry 4). When the same nitrones
were reacted with ethyl vinyl ether 5b in the presence of 10
mol % catalyst 8b, the diastereo- and the enantioselectivities
were very high and the isoxazolidines 6e-h were isolated with
endo:exo ratios of <5:>95 and with asymmetric inductions from
88% to 97% (entries 5-9). The reaction time needed for full
conversion followed the same pattern as above and the reactivity
of the nitrones toward vinyl ethers decreased in the following
order: 4a, 4c > 4b > 4d, correlating with the electronic
properties of the phenyl substituents. The enantioselectivity can
be enhanced to 97% ee in the case of compound 6g by using
toluene as solvent, but at the expense of a slightly lower yield
(66%) (entry 8). To incorporate a protected hydroxy group in
the 5 position of the isoxazolidine ring, the 1,3-dipolar cycload-
dition reaction was investigated with a substrate having a
protected enol functionality. Nitrone 4c was reacted with benzyl
vinyl ether 5c in the presence of 10 mol % catalyst 8b and the
isoxazolidine exo-6i was isolated in 68% yield after 18 h with
high exo selectivity and 85% ee (eq 5) (Table 3, entry 10).
To get suitable crystals for X-ray crystallographic analysis
and thereby to determine the absolute configuration of the
isoxazolidines exo-6a-i, attempts to recrystallize exo-6g were
performed. Slow cooling of a hot solution of exo-6g (ee ) 94%)
in hexane resulted in clear needles, which were suitable for
crystallographic examination. Surprisingly, these crystals were
racemic as determined by X-ray analysis. HPLC and optical
rotation also confirmed the crystals of exo-6g to be racemic.
The filtrate was therefore concentrated in vacuo to give exo-6g
as the pure enantiomer (ee > 99.5%) as a clear pale yellow oil.
The tendency for the racemic complexes to form crystals was
also observed for the isoxazolidines 6e-f and 6h; thus, a simple
procedure for obtaining the isoxazolidines as enantiomerically
pure compounds is to filter the crystalline racemate from a
hexane solution, followed by concentration of the filtrate.
The X-ray structure of exo-6g was determined and the crystals
of 6g were found to contain four molecules (two of each
enantiomer) in the unit cell of space group P21c.16 The crystal
structure, presented in Figure 2, confirms the cis relation
between the phenyl and ethoxy substituents in the 3 and 5
Absolute Assignment of the Stereochemistry. To assign the
absolute stereochemistry of the two new chiral centers in the
isoxazolidine ring (carbons 3 and 5) created in this 1,3-dipolar
cycloaddition reaction, the isoxazolidines were converted into
molecules containing a known configuration of another chiral
carbon atom. First, we tried to transform the isoxazolidines 6
into â-amino aldehydes upon reduction, followed by hydrolysis
of the generated hemiacetal, and then react the aldehyde with
(S)-N-amino-2-methoxymethylpiperidine (SAMP) to obtain the
corresponding hydrazone. Unfortunately, several attempts for
the hydrogenolysis of the nitrogen-oxygen bond in the isox-
azolidines by various catalysts such as Pd/C and Pd(OH)2 proved
very difficult, and no reproducible condition was found.
Difficulties in reducing the N-O bond in isoxazolidines bearing
an alkoxy substituent at the carbon atom next to the ring oxygen
atom was also observed by DeShong et al.6
This problem was circumvented by the preparation of
isoxazolidine exo-6l with an ester functionality in the para
position of the C-3 phenyl substituent as depicted in Scheme 1.
Nitrone 4e was prepared by standard condensation between
phenylhydroxylamine and 4-carbomethoxybenzaldehyde18 and
reacted with ethyl vinyl ether 5b in the presence of 10 mol %
8b as the catalyst. Complete conversion was achieved after 2
h, and the isoxazolidine exo-6l was isolated in very high yield,
with good diastereo- and enantioselectivity (Scheme 1). The
methyl ester was hydrolyzed by using LiOH in THF/H2O to
give dextrorotatory carboxylic acid 9 in almost quantitative yield
after aqueous workup. The acid 9 formed diastereomeric salts
with (-)-ephedrine, (-)-(S)-1-phenylethylamine, (+)-(R)-1-
phenylethylamine, (-)-(S)-1-(1-naphthyl)ethylamine, (+)-(R)-
1-(1-naphthyl)ethylamine, and brucine upon mixing equimolar
amounts of the acid 9 and amine in Et2O. Of these complexes,
the salt 10 formed with (-)-(S)-1-phenylethylamine and 9 gave
the best crystals for X-ray analysis. Crystals suitable for X-ray
crystallography were prepared by slow evaporation of a solution
of 10 in MeOH. The X-ray structure of 10 is presented in Figure
3.
1
position of the isoxazolidine ring, in agreement with the H
NMR experiments. The structural data for the isoxazolidine ring
is very similar to that of other characterized isoxazolidine17 rings
and will not be discussed further.
(16) The authors have deposited atomic coordinates for the structure at
the Cambridge Crystallographic Data Centre. The coordinates can be
obtained, on request, from the Director, CCDC, 12 Union Road, Cambridge
CB2 1EZ, UK.
(17) See, e.g., ref 4d and references therein for other characterized
isoxazolidines.
(18) Yimina, C.; Tsujimoto, T.; Suda, K.; Yamuchi, M. Bull. Chem. Soc.
Jpn. 1986, 59, 2165.

3850 J. Am. Chem. Soc., Vol. 121, No. 16, 1999
Simonsen et al.
Scheme 1
Figure 4. Intermediates 11a and 12a show the optimized structures
of the nitrone 4a coordinated to the (R)-BINOL-AlMe catalyst 8b. In
11a the re face of the nitrone is available for approach of the alkene,
whereas in 12a the si face is accessible. Color code: carbon atoms of
the nitrone in intermediates 11a and 12a are shown in black to
distinguish these atoms from carbon atoms of the BINOL and Me
fragments of the (R)-BINOL-AlMe catalyst, which are grey; aluminum,
green; oxygen, red; nitrogen, blue, hydrogen, white. In 11b and 12b
an overlay of a ChemDraw representation of the approach of the alkene
(tert-butyl vinyl ether) to the two different faces, re and si, respectively,
of the coordinated nitrone and a Chem3D representation of the nitrone-
catalyst intermediate are outlined. Color code: carbon, grey; aluminum,
green; oxygen, red; nitrogen, blue, hydrogen, white.
reaction is one in which only a monomeric (R)-BINOL-AlMe
catalyst is involved.
The coordination of the nitrone to the (R)-BINOL-AlMe
catalyst leads to the active intermediate, and the highly regio-,
diastereo-, and enantioselective catalytic properties of the
reaction indicate that the BINOL fragment provides a “nearly
sterically perfect environment”.
Figure 3. Crystal structure of the diasteromeric salt 10 (Scheme 1)
used for the assignment of the absolute stereochemistry of the chiral
centers in the isoxazolidine ring formed by reaction of nitrone 4e with
ethyl vinyl ether 5b catalyzed by (R)-BINOL-AlMe 8b. Only the major
component is shown. Thermal ellipsoids are shown at 50% probability
levels; selected hydrogen atoms are drawn as small circles of arbitrary
radius. The hydrogen bonds between the carboxyl group and the -NH3
groups of three different cations are indicated by the thin lines.
In an attempt to understand the coordination of the nitrone
to the (R)-BINOL-AlMe catalyst and the approach of the alkene
to the various possible intermediates, theoretical calculations
of different intermediates have been performed. The size of the
intermediate, the BINOL-AlMe 8b catalyst and nitrone 4a,
precludes the use of ab initio DFT calculations. However,
semiempirical calculations of the intermediates, although prob-
ably less reliable than the DFT calculations, might give
important information about the structure of the intermediate.
We have thus applied AM-1 calculations19 to the different
coordination modes of 4a with the catalyst 8b. The results,
which account for both the diastereo- and enantioselectivity in
the reaction, are presented below.
On the basis of the absolute stereochemistry of (-)-(S)-1-
phenylethylamine in the X-ray structure 10, Figure 3, the
absolute stereochemistry of the chiral centers formed in the
isoxazolidine ring of the major enantiomer by the 1,3-dipolar
cycloaddition reaction of nitrone 4e with ethyl vinyl ether
catalyzed by 8b is assigned to be 3R,5S. The absolute stereo-
chemistry of the isoxazolidine indicates that the alkene ap-
proaches the re face (on the basis of assignment relative to the
nitrone carbon atom) of the nitrone when coordinated to the
catalyst.
The geometry of various intermediates consisting of nitrone
4a coordinated in two different ways to the (R)-BINOL-AlMe
8b catalyst has been optimized. The coordination of 4a to 8a
Mechanistic Aspects of the (R)-BINOL-AlMe-catalyzed
1,3-Dipolar Cycloaddition Reaction. The observation of a
linear relation between the ee of the catalyst and the ee of exo-
6a (Figure 1) indicates that the catalytic intermediate in the
(19) The structures of intermediates 11a and 12a have been optimized
using AM1 calculations: (a) Dewar, M. J. S.; Zoebisch, E. G.; Healy, E.
F.; Stewart, J. J. P. J. Am. Chem. Soc. 1985, 107 3902. (b) Dewar, M. J. S.;
Holder, A. J. Organometallics 1990, 9, 508.

1,3-Dipolar Cycloaddition Reactions of Nitrones with Alkenes
J. Am. Chem. Soc., Vol. 121, No. 16, 1999 3851
can take place in two different ways, one which allows approach
of the alkene to the re face of the nitrone and the other approach
to the si face. The coordination of the nitrone to catalyst 8b
leads to a change of geometry of the aluminum center from
planar to tetrahedral, and the structures of these two intermedi-
ates, 11a and 12a, are shown in Figure 4. The carbon atoms of
the nitrone in intermediates 11a and 12a are shown in black to
distinguish the carbon atoms of the nitrone from the carbon
atoms of the BINOL and Me fragments of the (R)-BINOL-AlMe
catalyst.
In intermediate 11a, Figure 4, nitrone 4a is coordinated to
the BINOL-AlMe 8b catalyst in such a way that the re face of
4a is available for attack by the alkene, whereas in intermediate
12a, the si face is accessible. The total energy of 11a is
calculated to be -7571.38 eV, whereas the total energy of 12a
is 6 kcal mol-1 higher, indicating that the former is slightly
more stable than the latter. However, we are fully aware that
the level of the present calculations does not allow one to
distinguish between the two intermediates from an energetic
point of view. The major structural difference between the two
intermediates is, besides the orientation of the nitrone, the Al-O
(nitrone) bond length. In 12a it has been calculated to be 2.36
Å, which is significantly longer than the same bond in 11a (1.81
Å).
alkenes is developed. It is shown that 3,3′-diaryl-BINOL-AlMe
complexes are highly regio-, diastereo-, and enantioselective
catalysts for the 1,3-dipolar cycloaddition reaction of nitrones,
which can coordinate in a monodentate fashion to the catalyst,
with vinyl ethers. The reaction leads to exo-isoxazolidines in
good yield with endo:exo ratios <5:>95 and with ees up to
97%. A preferred coordination of the nitrone to the 3,3′-diaryl-
BINOL-AlMe catalyst leads to a shielding of the si face of the
nitrone, and a preferred exo-selective and enantioselective
approach of the vinyl ether to the re face of the nitrone accounts
for the absolute stereochemical outcome of the 1,3-dipolar
cycloaddition reaction.
Experimental Section
1
General Methods. The H and 13C NMR spectra were recorded in
CDCl3, unless otherwise stated, at 300 MHz and 75 MHz, respectively.
The chemical shifts are reported in ppm downfield to tetramethylsilane
(TMS). Optical rotations were measured on a Perkin-Elmer 241
polarimeter. Solvents were dried using standard conditions and stored
over molecular sieves (4 Å). THF and Et2O were distilled from sodium
benzophenone before use. All glass equipment was flame-dried under
vacuum before use. HPLC were performed using a Daicel Chiracel
OD column with hexane/i-PrOH 98:2 to 99.5:0.5 with UV detection at
247 nm.
Materials. The ligands 7a,20 7b-f, 13 and 7g14 were prepared as
described in the literature. Benzylidenephenylamine N-oxide 4a,21 (4-
methylbenzylidene)phenylamine N-oxide 4b,22 (4-chlorobenzylidene)-
phenylamine N-oxide 4c,23 (4-methoxybenzylidene)phenylamine N-
oxide 4d,24 and (4-carbomethoxybenzylidene)phenylamine N-oxide 4d18
were synthesized according to the literature. tert-Butyl vinyl ether, ethyl
vinyl ether, dihydrofuran, vinyl acetate, and styrene were received from
Aldrich and distilled from sodium immediately before use. Benzyl vinyl
ether was prepared as described in the literature from benzyl alcohol
using butyl vinyl ether and Hg(OAc)2 and distilled from sodium before
use.25 Millex filter units, 45-µm pore size, were received from Millipore.
General Procedure for the Stereoselective 1,3-Dipolar Cycload-
dition Reactions. The appropriate ligand 7 (0.02 mmol) was placed in
a 5-mL Schlenk flask flushed three times with N2, and CH2Cl2 (1 mL)
was added with a syringe. To this solution was added a 2 M solution
of AlMe3 in heptane (10 µL, 0.02 mmol), whereupon the solution turned
yellow under the evolution of CH4. The solution was stirred for 1 h
and a solution of nitrone 4 (0.1 mmol) in CH2Cl2 (1 mL) was added
together with the vinyl ethers 5 (0.5 mmol). After the appropriate
reaction time, the reaction was quenched with MeOH (1 mL) and
filtered through a 20-mm plug of silica. The silica was washed with
Et2O (3 mL) and the combined fractions were evaporated. The crude
material was purified by column chromatography (silica gel, petroleum
ether/Et2O 9:1) to give the single diastereomer of 6. Generally, the
endo isomers appeared with lower Rf values than the exo isomers (∆Rf
) 0.1).
The two intermediates, 11a and 12a (Figure 4), give important
information about the mechanism for the highly diastereo- and
enantioselective 1,3-dipolar cycloaddition reaction of aromatic
nitrones with vinyl ethers catalyzed by the (R)-BINOL-AlMe
complexes. For 11a, the vinyl ether, exemplified with tert-butyl
vinyl ether, can approach the re face of nitrone, whereas in 12a
the si face is available.
Structures 11b and 12b (Figure 4) show the exo approach of
the vinyl ether to the nitrone when coordinated to the catalyst,
11a and 12a, respectively [please observe that the carbon atoms
of the nitrone now have the same gray color as the carbon atoms
of the BINOL and Me fragments of the (R)-BINOL-AlMe
catalyst], as an overlay of a ChemDraw representation of tert-
butyl vinyl ether and a Chem3D representation of the nitrone-
catalyst intermediate. For intermediate 11b the vinyl ether fits
into “the pocket” formed by the arrangement of the 3-phenyl
substituent of the BINOL ligand and the coordinated nitrone.
The exo approach of the vinyl ether to the re face of the nitrone
as outlined in 11b is in good accordance with the experimental
results. If the vinyl ether approaches, in an endo selective
fashion, the re face of the nitrone when coordinated to the (R)-
BINOL-AlMe catalyst (11a), the ether substituent (ethyl, tert-
butyl, or benzyl) will suffer an unfavorable steric repulsion with
the 3-phenyl substituent of the BINOL fragment, which reduces
the possibility for the reaction path. In 12b the exo approach
of the vinyl ether to the si face of the nitrone coordinated to
the (R)-BINOL-AlMe catalyst is outlined. For this reaction path
it appears that a more severe steric repulsion between the tert-
butyl substituent of the vinyl ether and the 3-phenyl substituent
of the BINOL ligand takes place, thus eliminating this reaction
path compared with the former.
The reactions with 10 mol % catalyst were performed in the same
manner, whereas the reactions using 2.5 mol % and 5 mol % catalyst
were performed on a 0.2-mmol scale using the same amount of solvent.
The reactions on 1.0-mmol scale are performed in the same manner
by scaling everything up 10 times. Racemic mixtures of the isoxazo-
lidines exo-6a-l for HPLC analysis were prepared by stirring a mixture
of the nitrone 4 (0.1 mmol) and the vinyl ether 5 (1.0 mmol) in CH2-
Cl2 overnight in the presence of AlMe3 (25 mol %).
On the basis of the calculated structure of the nitrone
coordinated to the (R)-BINOL-AlMe catalyst, we are thus able
to account for both how the nitrone coordinates to the catalyst
and how the alkene approaches the activated nitrone, giving
the high regio-, diastereo-, and enantioselectivity obtained in
this new 1,3-dipolar cycloaddition reaction.
(20) Cai, D.; Hughes, D. L.; Verhoeven, T. R.; Reider, P. J. Tetrahedron
Lett. 1995, 36, 7991.
(21) (a) Kamm, O. Organic Synthesis; Wiley: New York, 1941; Collect.
Vol. 1, p 445. (b) Wheeler, O. H.; Gore, P. H. J. Am. Chem. Soc. 1956, 78,
3363.
(22) Beckett, A. H.; Coutts, R. T.; Ogunbona, F. A. Tetrahedron 1973,
29, 4189.
Conclusion
(23) Boyland, E.; Nery, R. J. Chem. Soc. 1963, 3141.
(24) Nunno, L.; Scilimati, A. Tetrahedron 1991, 47, 4121.
(25) Corey, E. J.; Bass, J. D.; LeMahieu, R.; Mitra, R. B. J. Am. Chem.
Soc. 1964, 86, 5570.
A new highly selective catalytic inverse-electron demand 1,3-
dipolar cycloaddition reaction of nitrones with electron-rich

3852 J. Am. Chem. Soc., Vol. 121, No. 16, 1999
Simonsen et al.
(+)-(3R,5S)-5-tert-Butoxy-2,3-diphenylisoxazolidine (exo-6a) was
synthesized according to the general procedure on a 1.0-mmol scale
with catalyst 8b (10 mol %): yield 84%; ee ) 89%; semicrystalline
catalyst 8b (10 mol %): yield 79%; ee ) 91%; clear oil; [R]D ) +217.3
1
(c 1.0, CHCl3). H NMR9c 13C NMR δ 15.2, 46.8, 63.6, 69.2, 100.6,
116.6, 122.5, 127.4, 127.6, 128.5, 128.8, 141.4, 150.3. HPLC (Daicel
Chiralcel OD, hexane/i-PrOH ) 99:1, flow rate ) 1.0 mL/min) tr )
11.2 min (minor), tr ) 25.3 min (major). HRMS (EI) calcd for C17H19-
NO2 (M+) 269.1415, found 269.1402.
1
solid; [R]D ) +160.9 (c 1.0, CHCl3). H NMR δ 1.39 (s, 9H), 2.34
(ddd, J ) 3.3, 6.8, and 12.9 Hz, 1H), 2.97 (ddd, J ) 6.0, 9.3, and 13.2
Hz, 1H), 4.42 (dd, J ) 7.0 and 9.0 Hz, 1H), 5.65 (dd, J ) 3.3 and 6.1
Hz, 1H), 6.93 (m, 2H), 7.15-7.38 (m, 6H), 7.56 (m, 2H). 13C NMR δ
28.9, 47.0, 68.9, 75.1, 96.4, 115.7, 121.9, 127.3, 127.4, 128.5, 128.7,
142.0, 150.9. HPLC (Daicel Chiralcel OD, hexane/i-PrOH ) 99:1, flow
rate ) 0.7 mL/min) tr ) 12.2 min (minor), tr )23.6 min (major). HRMS
(EI) calcd for C19H23NO2 (M+) 297.1723, found 297.1732.
(+)-(3R,5S)-5-Ethoxy-3-(4-methylphenyl)-2-phenylisoxazoli-
dine (exo-6f) was synthesized according to the general procedure on a
2-mmol scale with catalyst 8b (10 mol %): yield 70%; ee ) 90%;
1
[R]D ) +171.2 (c 1.0, CHCl3). H NMR δ 1.34 (t, J ) 7.2 Hz, 3H),
2.37 (s, 3H), 2.40 (ddd, J ) 2.4, 7.2, and 13.2 Hz, 1H), 3.04 (ddd, J
) 6.0, 9.3, and 13.2 Hz, 1H), 3.65 (dq, J ) 7.2 and 9.8 Hz, 1H), 4.04
(dq, J ) 7.2 and 9.3 Hz, 1H), 4.30 (dd, J ) 6.6 and 9.3 Hz, 1H), 5.38
(dd, J ) 2.4 and 6.0 Hz, 1H), 6.92-7.02 (m, 3H), 7.14-7.25 (m, 4H),
7.44 (d, J ) 8.1 Hz, 2h). 13C NMR δ 15.2, 21.4, 47.0, 63.6, 69.1,
100.6, 116.7, 122.5, 127.4, 128.5, 129.5, 137.2, 138.2, 150.3. HPLC
(Daicel Chiralcel OD, hexane/i-PrOH ) 99:1, flow rate ) 1.0 mL/
min) tr ) 5.3 min (minor), tr ) 12.2 min (major). HRMS (EI) calcd
for C18H21NO2 (M+) 283.1572, found 283.1752.
(+)-(3R,5R)-5-tert-Butoxy-2,3-diphenylisoxazolidine (endo-6a) was
synthesized according to the general procedure on a 0.5-mmol scale
using BINOL 7b and AlMe2Cl (10 mol %) as the catalyst: yield 40%;
1
ee ) 26.6%; semicrystalline solid; [R]D ) +4.4 (c 1.0, CHCl3). H
NMR δ 1.26 (s, 9H), 2.50 (ddd, J ) 5.1, 9.3, and 12.2 Hz, 1H), 2.63
(ddd, J ) 1.5, 6.9, and 12.1 Hz, 1H), 4.81 (dd, J ) 6.6 and 9.3 Hz,
1H), 5.65 (br. d, J ) 4.2 Hz, 1H), 6.88 (t, J ) 7.0 Hz, 1H), 6.98 (d, J
) 7.8 Hz, 2H), 7.14-7.40 (m, 5H), 7.49 (d, J ) 7.2 Hz, 2H). 13C
NMR δ 28.8, 47.5, 67.5, 74.9, 97.1, 115.5, 121.1, 126.5, 127.4, 128.2,
128.8, 141.6, 152.4. HPLC (Daicel Chiralcel OD, hexane/i-PrOH )
99.5:0.5, flow rate ) 0.7 mL/min) tr ) 31.0 min (minor), tr ) 33.5
min (major). HRMS (EI) calcd for C19H23NO2 (M+) 297.1723, found
297.1738.
(+)-(3R,5S)-3-(4-Chlorophenyl)-5-ethoxy-2-phenylisoxazolidine
(exo-6 g) was synthesized according to the general procedure on a
1-mmol scale with catalyst 8b (10 mol %): yield 78%; ee ) 94%;
1
[R]D ) +210.5 (c 1.0, CHCl3). H NMR δ 1.31 (t, J ) 7.2 Hz, 3H),
2.34 (ddd, J ) 2.1, 6.6, and 13.2 Hz, 1H), 3.02 (ddd, J ) 6.0, 9.9, and
13.2 Hz, 1H), 3.64 (dq, J ) 7.2 and 9.6 Hz, 1H), 4.01 (dq, J ) 7.2 and
9.9 Hz, 1H), 4.33 (dd, J ) 6.6 and 9.6 Hz, 1H), 5.39 (dd, J ) 2.1 and
6.0 Hz, 1H), 6.90-7.02 (m, 3H), 7.18-7.25 (m, 2H), 7.35 (d, J ) 8.1
Hz, 2h), 7.49 (d, J ) 8.1 Hz, 2H). 13C NMR δ 15.2, 46.5, 63.6, 68.4,
100.6, 116.4, 122.6, 128.6, 128.8, 128.9, 133.3, 140.2, 150.1. HPLC
(Daicel Chiralcel OD, hexane/i-PrOH ) 99.5:0.5, flow rate ) 0.7 mL/
min) tr ) 13.1 min (minor), tr ) 14.1 min (major). HRMS (EI) calcd
for C17H18ClNO2 (M+) 303.1026, found 303.1042.
(+)-(3R,5S)-5-tert-Butoxy-3-(4-methylphenyl)-2-phenylisoxazoli-
dine (exo-6b) was synthesized according to the general procedure on
a 1.0-mmol scale with catalyst 8b (10 mol %): yield 71%; ee ) 81%;
clear oil; [R]D ) +145.4 (c 1.0, CHCl3). 1H NMR δ 1.42 (s, 9H), 2.36
(ddd, J ) 3.3, 7.2, and 13.2 Hz, 1H), 2.40 (s, 3H), 2.99 (ddd, J ) 6.0,
9.3, and 12.9 Hz, 1H), 4.40 (dd, J ) 7.2 and 8.7 Hz, 1H), 5.67 (dd, J
) 3.3 and 6.0 Hz, 1H), 6.90-7.05 (m, 3H), 7.16-7.25 (m, 4H), 7.49
(d, J ) 7.8 Hz, 2H). 13C NMR δ 21.2, 29.0, 47.4, 68.8, 75.0, 96.2,
115.8, 121.8, 127.2, 128.5, 129.4, 137.0, 138.9, 151.0. HPLC (Daicel
Chiralcel OD, hexane/i-PrOH ) 99:1, flow rate ) 1.0 mL/min) tr )
7.0 min (minor), tr ) 11.6 min (major). HRMS (EI) calcd for C20H25-
NO2 (M+) 311.1885, found 311.1899.
(+)-(3R,5S)-5-Ethoxy-3-(4-methoxyphenyl)-2-phenylisoxazoli-
dine (exo-6h) was synthesized according to the general procedure on
a 0.5-mmol scale with catalyst 8b (10 mol %): yield 50%; ee ) 88%;
1
[R]D ) +145.7 (c 1.0, CHCl3). H NMR δ 1.32 (t, J ) 7.2 Hz, 3H),
2.37 (ddd, J ) 2.1, 7.2, and 13.2 Hz, 1H), 3.01 (ddd, J ) 6.0, 9.3, and
13.2 Hz, 1H), 3.63 (dq, J ) 7.2 and 9.9 Hz, 1H), 3.82 (s, 3H), 4.01
(dq, J ) 7.2 and 9.3 Hz, 1H), 4.25 (dd, J ) 7.2 and 9.3 Hz, 1H), 5.37
(dd, J ) 2.1 and 6.0 Hz, 1H), 6.91 (d, J ) 9.0 Hz, 2H), 6.97 (d, J )
7.8 Hz, 2H), 7.16-7.22 (m, 3H), 7.44 (d, J ) 9.0 Hz, 2H). 13C NMR
δ 15.2, 46.9, 55.3, 63.6, 69.0, 100.6, 114.1, 116.8, 122.6, 124.3, 128.5,
128.7, 150.3, 160.1. HPLC (Daicel Chiralcel OD, hexane/i-PrOH )
99:1, flow rate ) 1.0 mL/min) tr ) 7.4 min (minor), tr ) 13.5 min
(major). HRMS (EI) calcd for C18H21NO3 (M+) 299.1521, found
299.1536.
(+)-(3R,5S)-5-Benzyloxy-3-(4-chlorophenyl)-2-phenylisoxazoli-
dine (exo-6i) was synthesized according to the general procedure on a
0.2-mmol scale with catalyst 8b (10 mol %): yield 68%; ee ) 85%.
1H NMR δ 2.41 (ddd, J ) 1.5, 6.0, and 13.2 Hz, 1H), 3.03 (ddd, J )
5.8, 9.4, and 13.2 Hz, 1H), 4.33 (dd, J ) 6.0 and 9.9 Hz, 1H), 4.68 (d,
J ) 12 Hz, 1H), 4.98 (d, J ) 12 Hz, 1H), 5.44 (dd, J ) 1.5 and 6.0
Hz, 1H), 6.90-7.02 (m, 3H), 7.18-7.45 (m, 9H), 7.48 (d, J ) 9.0 Hz,
2H). HPLC (Daicel Chiralcel OD, hexane/i-PrOH ) 99:1, flow rate )
0.7 mL/min) tr ) 12.9 min (minor), tr ) 15.3 min (major). HRMS (EI)
calcd for C22H20ClNO2 (M+) 365.1182, found 365.1211.
(+)-(3R,5S)-5-tert-Butoxy-3-(4-chlorophenyl)-2-phenylisoxazoli-
dine (exo-6c) was synthesized according to the general procedure on
a 1.0-mmol scale with catalyst 8b (10 mol %): yield 74%; ee ) 78%;
clear oil; [R]D ) +145.8 (c 1.0, CHCl3). 1H NMR δ 1.37 (s, 9H), 2.27
(ddd, J ) 3.0, 6.1, and 13.2 Hz, 1H), 2.95 (ddd, J ) 6.0, 9.3, and 12.6
Hz, 1H), 4.40 (dd, J ) 6.0 and 9.3 Hz, 1H), 5.66 (dd, J ) 3.3 and 6.0
Hz, 1H), 6.90-7.00 (m, 3H), 7.16-7.25 (m, 2H), 7.34 (d, J ) 8.6 Hz,
2H), 7.52 (d, J ) 8.5 Hz, 2H). 13C NMR δ 28.9, 46.8, 68.1, 75.1, 96.3,
115.5, 122.0, 128.6, 128.7, 128.8, 133.1, 140.7, 150.6. HPLC (Daicel
Chiralcel OD, hexane/i-PrOH ) 99:1, flow rate ) 1.0 mL/min) tr )
5.7 min (minor), tr ) 7.0 min (major). HRMS (EI) calcd for C19H22-
ClNO2 (M+) 331.1339, found 331.1325.
1
endo-6c: Yield 10%; ee ) 0%; clear oil; H NMR δ 1.25 (s, 9H),
2.44 (ddd, J ) 5.1, 9.3, and 12.5 Hz, 1H), 2.61 (ddd, J ) 1.8, 7.2, and
12.2 Hz, 1H), 4.78 (dd, J ) 6.6 and 9.3 Hz, 1H), 5.62 (br. d, J ) 4.5
Hz, 1H), 6.85-6.98 (m, 3H), 7.14-7.22 (m, 2H), 7.32 (d, J ) 8.4 Hz,
2H), 7.42 (d, J ) 8.1 Hz, 2H). 13C NMR δ 28.8, 47.4, 66.9, 75.0, 97.0,
115.5, 121.4, 127.9, 128.3, 128.9, 133.1, 140.1, 152.1. HPLC (Daicel
Chiralcel OD, hexane/i-PrOH ) 99:1, flow rate ) 1.0 mL/min) tr )
11.7 min, tr ) 13.8 min (major). HRMS (EI) calcd for C19H22ClNO2
(M+) 331.1339, found 331.1321.
(rac)-2,3,3a,4,5,6a-Hexahydro-2,3-diphenylfuro[3,2-d]isoxazoli-
dine (exo-6j) was synthesized according to the general procedure on a
0.5-mmol scale with catalyst 8b (10 mol %): yield 51%; ee ) <5%;
clear oil; 1H NMR.26 13C NMR δ 28.3, 53.5, 69.3, 71.0, 105.3, 119.2,
123.7, 127.4, 127.7, 128.4, 128.8, 137.8, 148.8. HPLC (Daicel Chiralcel
OD, hexane/i-PrOH ) 98:2, flow rate ) 1.0 mL/min) tr ) 12.5 min
(minor), tr ) 18.6 min (major). HRMS (EI) calcd for C17H17NO2 (M+)
267.1259, found 267.1281.
(+)-(3R,5S)-5-tert-Butoxy-3-(4-methoxyphenyl)-2-phenylisoxazo-
lidine (exo-6d) was synthesized according to the general procedure
on a 1.0-mmol scale with catalyst 8b (10 mol %): yield 65%; ee )
1
77%; clear oil; [R]D ) +141.6 (c 1.0, CHCl3). H NMR δ 1.38 (s,
9H), 2.32 (ddd, J ) 3.3, 7.2, and 12.6 Hz, 1H), 2.95 (ddd, J ) 6.0,
9.0, and 13.0 Hz, 1H), 3.81 (s, 3H), 4.35 (dd, J ) 7.2 and 8.7 Hz, 1H),
5.64 (dd, J ) 3.3 and 6.0 Hz, 1H), 6.88-7.0 (m, 5H), 7.19 (t, J ) 7.9
Hz, 2H), 7.46 (t, J ) 8.4 Hz, 2H). 13C NMR δ 29.0, 47.3, 55.3, 68.6,
75.0, 96.2, 114.1, 115.9, 121.8, 128.5, 133.8, 150.9, 159.0. HPLC
(Daicel Chiralcel OD, hexane/i-PrOH ) 99:1, flow rate ) 1.0 mL/
min) tr ) 10.1 min (minor), tr ) 13.3 min (major). HRMS (EI) calcd
for C20H25NO3 (M+) 327.1834, found 327.1851.
(endo-6j): Yield 12%; ee ) 0%; clear oil; 1H NMR.26 HPLC (Daicel
Chiralcel OD, hexane/i-PrOH ) 98:2, flow rate ) 1.0 mL/min) tr )
10.1 min (minor), tr ) 13.3 min (major). HRMS (EI) calcd for C17H17-
NO2 (M+) 267.1259, found 267.1278.
(rac)-2,3,5-Triphenylisoxazolidine (exo-6k) was synthesized ac-
cording to the general procedure on a 0.5-mmol scale with catalyst 8b
(+)-(3R,5S)-5-Ethoxy-2,3-diphenylisoxazolidine (exo-6e) was syn-
thesized according to the general procedure on a 0.5-mmol scale with
(26) Paul, R.; Tchelitcheff, S. Bull. Soc. Chim. Fr. 1967, 4179.

1,3-Dipolar Cycloaddition Reactions of Nitrones with Alkenes
J. Am. Chem. Soc., Vol. 121, No. 16, 1999 3853
1
(10 mol %): yield 57%; ee ) 0%; semicrystalline solid; H NMR.27
graphite-monochromated MoKR radiation collecting a hemisphere of
data out to θ ) 27.7°. The structure was solved by direct methods
(SIR9729) and refined by full matrix least squares on Fsq.30 Of 3306
unique reflections, the 2643 with I > 3σ(I) were used in the refinement
of 263 parameters, giving R ) 0.033, Rw ) 0.032.
Complex 10. This family of compounds seem to be very difficult
to crystallize in the optically active form and gave, with different
cations, either no crystals or tiny crystals with unit cell volumes of
5000-8000 Å3, and with far too few significant reflections. The
phenethylammonium salt, however, gave usable, though tiny and poorly
diffracting, crystals.
13C NMR δ 48.7, 71.6, 80.6, 114.0, 121.4, 126.3, 126.9, 127.4, 128.4,
128.6, 128.9, 129.0, 137.9, 142.9, 152.5. HPLC (Daicel Chiralcel OD,
hexane/i-PrOH ) 99:1, flow rate ) 1.0 mL/min) tr ) 12.5 min, tr )
14.6 min. HRMS (EI) calcd for C21H19NO (M+) 301.1460, found
301.1444.
(+)-(3R,5S)-3-(4-Carbomethoxyphenyl)-5-ethoxy-2-phenylisoxa-
zolidine (exo-6l) was synthesized according to the general procedure
on a 2.0-mmol scale with catalyst 8b (10 mol %). The reaction was
finished after 2 h: yield 90%; ee ) 91%; [R]D ) +173.8 (c 1.0, CHCl3).
1H NMR δ 1.28 (t, J ) 7.2 Hz, 3H), 2.36 (ddd, J ) 2.1, 6.0, and 13.2
Hz, 1H), 3.03 (ddd, J ) 5.7, 9.6, and 13.2 Hz, 1H), 3.62 (dq, J ) 6.9
and 9.9 Hz, 1H), 3.92 (s, 3H), 3.98 (dq, J ) 6.9 and 9.6 Hz, 1H), 4.41
(dd, J ) 6.0 and 9.9 Hz, 1H), 5.39 (dd, J ) 1.8 and 6.0 Hz, 1H),
6.85-6.95 (m, 3H), 7.18 (t, J ) 8.1 Hz, 2H), 7.61 (d, J ) 8.1 Hz,
2H), 8.03 (d, J ) 8.1 Hz, 2H). 13C NMR δ 15.2, 46.3, 52.1, 63.6, 68.6,
100.6, 116.3, 122.6, 127.3, 128.6, 129.4, 130.1, 146.9, 150.1, 166.9.
HPLC (Daicel Chiralcel OD, hexane/i-PrOH ) 99:1, flow rate ) 1.0
mL/min) tr ) 14.3 min (major), tr ) 18.9 min (minor). HRMS (EI)
calcd for C19H21NO4 (M+) 327.1470, found 327.1492.
(+)-(3R,5S)-3-(4-Carboxyphenyl)-5-ethoxy-2-phenylisoxazoli-
dine (exo-9). The methyl ester exo-6l (0.63 g, 2 mmol) was dissolved
in 50 mL of THF/H2O (3:1), and LiOH (0.165 g, 4 mmol) was added.
The reaction mixture was stirred overnight, diluted with H2O (50 mL),
and the THF was removed in vacuo. The aqueous phase was extracted
with EtOAc (2 × 40 mL) to remove unreacted starting material, and
neutralized with 1 M HCl, whereupon a white solid starts to precipitate.
The mixture was extracted with EtOAc (3 × 50 mL), and the organic
phase was dried (Na2SO4) and evaporated to give the free acid exo-9
as a pale yellow glass; yield 0.59 g (94%); [R]D ) +193.5 (c 1.0,
CHCl3). 1H NMR δ 1.32 (t, J ) 7.2 Hz, 3H), 2.38 (ddd, J ) 1.5, 6.0,
and 12.9 Hz, 1H), 3.05 (ddd, J ) 6.3, 9.9, and 12.9 Hz, 1H), 3.62 (dq,
J ) 7.2 and 9.9 Hz, 1H), 3.99 (dq, J ) 7.2 and 9.9 Hz, 1H), 4.45 (dd,
J ) 6.0 and 9.9 Hz, 1H), 5.41 (dd, J ) 1.8 and 6.0 Hz, 1H), 6.90-
7.00 (m, 3H), 7.18 (t, J ) 8.1 Hz, 2H), 7.65 (d, J ) 8.1 Hz, 2H), 8.10
(d, J ) 9.0 Hz, 2H), 10.5 (s br., 1H). 13C NMR δ 15.2, 46.2, 63.6,
68.6, 100.7, 116.3, 122.7, 127.5, 128.6, 128.7, 130.8, 148.0, 150.0,
172.0. HRMS (EI) calcd for C18H19NO4 (M+) 313.1313, found
313.1328.
X-ray Analysis of 10, C18H18NO4-, C8H12N+, is orthorhombic,
P212121, with a ) 6.1488(4), b ) 17.506(1), c ) 22.638(2) Å, V )
2436.8 Å3, Z ) 4. Same procedure as for exo-6g gave 31 422 reflections
measured, 5396 unique, 2042 significant [I > 3σ(I)]. The structure was
solved without difficulty, but anisotropic displacement parameters for
the central part of the molecule showed disorder. Each atom was split
into two isotropic atoms, and geometric considerations showed
unambiguously a superposition of two anions of identical conformation
except for rotation about the single bonds outside the rings. To stabilize
refinement and save parameters, constraints were introduced according
to Pawley:31 all phenyl rings identical with mm2 symmetry; the 5-ring
with its attached oxygen identical for the two disordered anions; thermal
motion described by the TLS model for three sections: phenyl group-
(s) attached to N2, phenyl group(s) at C3, 5-rings with ethoxy group;
one common occupation parameter for disordered part; hydrogen atoms
+
in calculated positions, methyl and -NH3 refined with a parameter
for rotation from the staggered position; carboxyl group and phenyl-
ethylammonium ion ordered, refined anisotropically. Parameters (232)
were refined, final R ) 0.059, Rw ) 0.070. Atomic coordinates, bond
lengths and angles, and thermal parameters for exo-6g and 10 have
been deposited at the Cambridge Crystallographic Data Center (CCDC).
Acknowledgment. This work was made possible by a grant
from The Danish National Research Foundation and The
Carlsberg Foundation.
Supporting Information Available: Copies of NMR spec-
tra, MS spectra and X-ray data (PDF). This material is available
free of charge via the Internet at http://pubs.acs.org.
Formation of the Diastereomeric Salt 10. To a solution of 9 (0.022
mg, 0.07 mmol) in Et2O (2 mL) in a 10-mL round-bottomed flask was
added a solution of (-)-(S)-1-phenylethylamine (10 µL, 0.08 mmol)
in Et2O (1 mL); the flask was capped and left overnight. After 1/2 h
the formation of a white crystalline compound starts. The resulting white
crystals were filtered, washed with ether, and dried in vacuo: yield:
29 mg (95%).
X-ray Analysis of exo-6 g and Complex 10. Exo-6g, C17H18NO2-
Cl, is monoclinic, P21/c, with a ) 10.1400(5), b ) 16.8233(8), c )
9.6098(5) Å, â ) 114.163(1), V ) 1495.7(1), Z ) 4. Reflections
(14 675) were measured on a SMART diffractometer28 at 120 K using
JA983915A
(28) Siemens SMART and SAINT Area Detector Control and Integration
Software. Siemens Analytical X-ray Instruments Inc., Madison, WI, 1995.
(29) Cascarano, G.; Altomare, A.; Giacovazzo, C.; Guagliardi, A.;
Moliterni, A. G. G.; Siliqi, D.; Burla, M. C.; Polidori, G.; Camalli, M. Acta
Crystallogr. 1996, A52, C-79.
(30) Busing, W. R.; Martin, K. O.; Levy, H. A. ORFLS Report ORNL-
TM-305, Oak Ridge National Laboratory: Oak Ridge, TN, 1962.
(31) (a) Pawley, G. S. AdV. Struct. Res. Diff. Methods 1971, 4, 1. (b)
Hazell, A. KRYSTAL, an integrated system of crystallographic programs;
Aarhus University, Denmark, 1995.
(27) Huisgen, R.; Grashey, R.; Hauck, H.; Seidl, H. Chem. Ber. 1968,
101, 4548.