Diastereoselective cycloadditions of chiral homoallylic alcohols with benzonitrile oxide

Article abstract of DOI:10.1016/S0040-4039(03)00100-X

The first diastereoselective cycloaddition of a nitrile oxide with a homoallylic alcohol is discussed. The reaction of benzonitrile oxide with the magnesium alkoxides of chiral homoallylic alcohols has been shown to proceed with good diastereoselectivity, favoring the syn isomer of the resulting 2-isoxazoline.

Full text of DOI:10.1016/S0040-4039(03)00100-X

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 44 (2003) 1811–1813
Diastereoselective cycloadditions of chiral homoallylic alcohols
with benzonitrile oxide
Martin G. Kociolek* and Chayanant Hongfa
Penn State Erie, The Behrend College, School of Science, Erie, PA 16563, USA
Received 27 November 2002; revised 9 January 2003; accepted 10 January 2003
Abstract—The first diastereoselective cycloaddition of a nitrile oxide with a homoallylic alcohol is discussed. The reaction of
benzonitrile oxide with the magnesium alkoxides of chiral homoallylic alcohols has been shown to proceed with good
diastereoselectivity, favoring the syn isomer of the resulting 2-isoxazoline. © 2003 Elsevier Science Ltd. All rights reserved.
Isoxazolines have long been recognized as versatile
synthetic intermediates for the construction of a variety
of 1,3-difunctionalized compounds, such as b-hydroxy
hols. It was thought that diastereoselectivity could be
achieved in these reactions by applying this Mg-
directed approach. Herein, we would like to report the
results of the reactions of benzonitrile oxide with a
number of magnesium alkoxides derived from chiral
homoallylic alcohols.
1
2
carbonyls and g-amino alcohols. Their use as latent
3
aldol adducts is also well established. Isoxazolines are
readily prepared from the reaction of nitrile oxides with
olefins and while these cycloadditions show a high
degree of regioselectivity, controlling the stereoselectiv-
The homoallylic alcohols were reacted under two sets
of conditions. Method A was the standard cycloaddi-
tion conditions (benzoyl oximinoyl chloride, triethy-
lamine, CH Cl ) in the absence of a metal. These
4
ity proves to be a challenge. A small number of
examples of diastereoselective cycloadditions of nitrile
5
oxides have been reported. One approach to stereose-
2
2
lective reactions of this type involves the use of metal
chelates to coordinate both the olefin and the nitrile
oxide, thus dictating the stereochemistry. The choice of
metals can sometimes be difficult, as the appropriate
metal must coordinate strongly enough to dictate the
stereochemistry, but not so strongly that the cycloaddi-
tion reaction is inhibited. Kanemasa and coworkers
first showed the effectiveness of magnesium in these
reactions, having reported a number of diastereoselec-
tive cycloadditions of nitrile oxides with the magnesium
conditions serve as the control, as little if any
diastereoselectivity would be expected. Method B was
adapted from Kanemasa’s work and is illustrated in
Scheme 1.
The homoallylic alcohols were initially reacted at −78°C
with ethyl magnesium bromide (1 M solution in THF),
giving the corresponding magnesium alkoxides. This
solution was then treated with a solution of benzoyl
oximinoyl chloride in methylene chloride. The nitrile
6
alkoxides of chiral allylic alcohols. In addition, they
reported the analogous reaction with a limited number
of homoallylic alcohols, including two examples pos-
6d
sessing a stereogenic center. Carreira and coworkers
expanded on this methodology by applying it to7
aliphatic nitrile oxides in the synthesis of polyketides.
In the course of our research, we found it necessary to
synthesize a series of 2-isoxazolines derived from the
cycloaddition of nitrile oxides with homoallylic alco-
*
0
Corresponding author. Tel.: 814-898-6411; fax: 814-898-6213;
e-mail: mgk5@psu.edu
Scheme 1. Mg-mediated cycloadditions.
040-4039/03/$ - see front matter © 2003 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(03)00100-X

1
812
M. G. Kociolek, C. Hongfa / Tetrahedron Letters 44 (2003) 1811–1813
oxide, generated in situ, was then allowed to react by
slowly warming to room temperature and stirring
overnight. After aqueous acidic workup, the ratio of
diastereomers was determined by HPLC and the iso-
mers purified by column chromatography (silica gel;
hexanes/ethyl acetate).
Table 2. Cycloaddition reactions of b-substituted homoal-
lylic alcohols
Cmpd.
R
Method
% Yielda
Ratio (4/5)b
3
3
3
3b
3c
3
3
3
3
3
a
a
b
Me
Me
Et
A
B
A
B
A
B
A
B
A
B
80
81
72
85
72
80
75
83
65
78
54:46
85:15
66:34
92:8
65:35
90:10
56:44
84:16
54:46
91:9
Initially, several a-substituted homoallylic alcohols (1a–
c) were reacted by both method A and B, resulting in
syn/anti mixtures of the b-hydroxy isoxazolines (2a–c).
Alcohol 1a was commercially available (Aldrich) and
Et
n-Pr
n-Pr
Bn
Bn
Ph
c
d
d
e
e
1
b–c were synthesized by reaction of allylmagnesium
chloride (2M solution in THF) with the appropriate
Ph
1
carbonyl. H NMR spectra for compounds 1b–c were
consistent with those reported in the literature.
8
a
Combined yield of both isomers.
Determined by HPLC (silica gel, LC-18; MeOH/H O).
b
2
sized by the reaction of 3,4-epoxy-1-butene with the
9
appropriate organomagnesium halide in diethyl ether.
1
H NMR spectra for compounds 3b–e were consistent
10
with those reported in the literature. It is also of note,
that in all cases the diastereomers could be separated by
column chromatography. All compounds (4a–e, 5a–e)
1
were isolated pure and gave satisfactory H NMR and
HRMS.
The results are shown in Table 1. In the case of these
a-substituted alcohols, little to no diastereoselectivity
was observed under either set of conditions. This was
not entirely unexpected, as the larger substituent (R or
R%) would prefer to be in an equatorial position in what
is presumably a chair-like transition state. In this con-
formation, the substituent would have very little inter-
action with the incoming nitrile oxide. In light of the
lack of selectivity, no effort was made to establish the
relative stereochemistry of the major isomer in any of
these cases. In addition, it was observed that the yields
from method B were considerably lower than that of
method A. This is primarily due to the competing
formation of what has been tentatively identified as
bis-homoallylic ether products.
The results are shown in Table 2. It was observed that
the presence of a stereogenic center in the b-position
(
relative to the hydroxyl), adjacent to the olefin, has a
Next, a series of b-substituted homoallylic alcohols
significant influence on the diastereoselectivity. Even a
relatively small group, such as methyl influences the
stereochemical outcome of the reactions. Increasing the
size of the substituent seemed to have only minor
impact, increasing the selectivity only slightly. Interest-
ingly, the ethyl and n-propyl substituted compounds
showed modest selectivity even in the absence of a
metal (method A). In all cases, an increase in the
overall yield (both isomers) for method B was observed.
(
3a-e) were reacted under the same two sets of condi-
tions resulting in syn/anti mixtures of the b-hydroxy
isoxazolines (4a–e, syn; 5a–e, anti). Alcohol 3a was
commercially available (Aldrich) and 3b–e were synthe-
Table 1. Cycloaddition reactions of a-substituted homoal-
lylic alcohols
Cmpd.
R
R%
Method
% Yielda
Ratiob
The relative stereochemistry of the major isomer was
1
assigned by examination of the H NMR chemical
1a
1a
1b
1b
1c
1c
CH3
H
H
H
H
Ph
Ph
A
B
A
B
A
B
50
60
80
60
79
25
50:50
51:49
51:49
55:45
56:44
51:49
shifts and coupling constants of the proton attached to
C-5 of the isoxazoline ring. It has been previously
demonstrated, in analogous b-hydroxy isoxazolines,
that the chemical shift of the C-5 proton on the syn
diastereomers occurs at a slightly higher field than that
CH3
Ph
Ph
CH3
CH3
11
of the anti. In all the cases examined above, the
a
chemical shift of the C-5 proton of the major isomer
Combined yield of both isomers.
Determined by HPLC (silica gel, LC-18; MeOH/H O).
b
occurred at higher field (0.1–0.2 ppm) than that of the
2

M. G. Kociolek, C. Hongfa / Tetrahedron Letters 44 (2003) 1811–1813
1813
anti. In addition the splitting patterns observed in 4a–e
and 5a–e were analogous to those reported in the
literature. As such, we are reasonably confident that
the major isomer in our case is indeed the syn isomer.
1980, 101–120; (c) Jager, V.; Muller, V.; Paulus, E. F.
Tetrahedron Lett. 1985, 26, 2997–3000.
3. (a) Kozikowski, A. P. Acc. Chem. Res. 1984, 17, 410–416;
(b) For a general discussion, see: 1,3-Dipolar Cycloaddi-
tion Chemistry; Padwa, A., Ed.; Wiley: New York, 1984;
Vol. 1.
4. For reviews, see: (a) Padwa, A. In Comprehensive Organic
Synthesis; Trost, B. M., Fleming, I., Eds.; Pergamon
Press: Oxford, 1991; Vol. 4, pp. 1069–1109; (b) Kane-
masa, S.; Tsuge, O. Heterocycles 1990, 30, 719–736; (c)
Caramella, P.; Grunanger, P. In 1,3-Dipolar Cycloaddi-
tion Chemistry; Padwa, A., Ed.; John Wiley & Sons: New
York, 1984; Vol. 1, p. 291.
5. (a) Annunziata, R.; Cinquini, M.; Cozzi, F.; Raimondi,
L. J. Chem. Soc., Chem. Commun. 1987, 529–530; (b)
Annunziata, R.; Cinquini, M.; Cozzi, F.; Raimondi, L.
Tetrahedron 1988, 44, 4645–4652; (c) Curran, D. P.; Kim,
B. H.; Daugherty, J.; Heffner, T. A. Tetrahedron Lett.
1988, 29, 3555–3558.
6. (a) Fukuda, S.; Kamimura, A.; Kanemasa, S.; Hori, K.
Tetrahedron 2000, 56, 1637–1647; (b) Kanemasa, S.;
Okuda, K.; Yamamoto, H.; Kaga, S. Tetrahedron Lett.
1997, 38, 4095–4098; (c) Kanemasa, S.; Tsuruoka, T.;
Yamamoto, H. Tetrahedron Lett. 1995, 36, 5019–5022;
(d) Kanemasa, S.; Nishiuchi, M.; Kamimura, A.; Hori,
K. J. Am. Chem. Soc. 1994, 116, 2324–2339; (e) Kane-
masa, S.; Nishiuchi, M.; Wada, E. Tetrahedron Lett.
1992, 33, 1357–1360; (f) Kanemasa, S.; Kobayashi, S.;
Nishiuchi, M.; Yamamoto, H.; Wada, E. Tetrahedron
Lett. 1991, 32, 6367–6370.
1
1
The diastereoselectivity of these reactions has also been
investigated by molecular modeling (RHF/3-21G*//
RHF/3-21G*). Preliminary examination of the reaction
of 3a with benzonitrile oxide has shown that the transi-
tion state leading to the syn isomer is in fact slightly
lower in energy (2.36 kcal/mol) than the transition state
leading to the anti. These results are consistent with the
experimental findings of the syn as the major isomer.
Further investigation of the entire reaction pathway at
higher levels of theory, as well as with other sub-
stituents is currently underway
While the diastereoselectivities of these cycloadditions
are good, they also provide the first examples of the
diastereoselective cycloadditions of nitrile oxides with
chiral b-substituted homoallylic alcohols. They also
provide a basis from which to further improve the
diastereoselectivity. Efforts are underway to evaluate
the effects of reaction conditions, structure of the nitrile
oxide and other additives on the diastereoselectivity. In
addition, we are currently trying to extend this method-
ology to cyclic homoallylic alcohols as well as b,g-
unsaturated carboxylic acids.
Acknowledgements
7. Bode, J. W.; Fraefel, N.; Muri, D.; Carreira, E. M.
Angew. Chem. Int. Ed. 2001, 40, 2082–2085.
8. Wang, Z.; Zha, Z.; Zhou, C. Org. Lett. 2002, 4, 1683–
Thanks to Penn State Erie, The Behrend College for
start-up funds and Allegheny College, Department of
Chemistry for use of their 400 MHz NMR.
1685 and supplementary material.
9. For representative procedure, see: Rose, C. B.; Taylor, S.
K. J. Org. Chem. 1974, 39, 578–580.
1
0. (a) Morken, J. P.; Didiuk, M. T.; Hoveyda, A. H. J. Am.
Chem., Soc. 1993, 115, 6997–6998 and supplementary
material; (b) Ent, H.; de Koning, H.; Speckamp, W. N. J.
Org. Chem. 1986, 51, 1687–1691 and supplementary
material; (c) Takekawa, Y.; Shishido, K. J. Org. Chem.
2001, 66, 8490–8503.
References
1. Curran, D. P. J. Am. Chem. Soc. 1983, 105, 5826–5833
and references cited therein.
2
. (a) Jager, V.; Muller, V. Tetrahedron Lett. 1952, 23,
11. Panek, J. S.; Beresis, R. T. J. Am. Chem. Soc. 1993, 115,
7898–7899 and supplementary material.
4
777–4780; (b) Jager, V.; Buss, V. Liebigs Ann. Chem.