Synthesis of 2-isoxazolines: Enantioselective and racemic methods based on conjugate additions of oximes

Article abstract of DOI:10.1002/chem.201000861

The formation of 3-unsubstituted 2-isoxazolines by means of condensation reactions between α,β-unsaturated aldehydes and oximes proceeds readily in the presence of catalytic amounts of anilinium salts. Mechanistically, the process involves a fast conjugate addition of the oxime and a slower intramolecular oxime-transfer reaction. The rate of oxime transfer was found to correlate with the acidity of the catalyst. This finding enabled us to discover an enantioselective process in which the fragile conjugate-addition product generated in the first stage is rapidly cyclized into the stable isoxazo-line under acidic conditions, with conservation of enantiomeric excess. In summary, herein we describe synthetically useful protocols for accessing 3-unsubstituted 2-isoxazolines in both the enantioselective and racemic manner. The mechanism of the condensation reaction catalyzed by the anilinium salt was also investigated by NMR spectroscopy experiments in which the effect of differently substituted aldehydes and oximes as well as water on the reaction rate was studied. The results point to the rate-limiting elimination of water from the 3-hydroxy-2-iso-xazolidine intermediate.

Full text of DOI:10.1002/chem.201000861

FULL PAPER
DOI: 10.1002/chem.201000861
Synthesis of 2-Isoxazolines: Enantioselective and Racemic Methods Based on
Conjugate Additions of Oximes
Antti Pohjakallio,[b] Petri M. Pihko,*[a] and Ulla M. Laitinen[c]
Abstract: The formation of 3-unsubsti-
tuted 2-isoxazolines by means of con-
densation reactions between a,b-unsa-
turated aldehydes and oximes proceeds
readily in the presence of catalytic
amounts of anilinium salts. Mechanisti-
cally, the process involves a fast conju-
in which the fragile conjugate-addition
product generated in the first stage is
rapidly cyclized into the stable isoxazo-
line under acidic conditions, with con-
servation of enantiomeric excess. In
summary, herein we describe syntheti-
cally useful protocols for accessing 3-
unsubstituted 2-isoxazolines in both the
enantioselective and racemic manner.
The mechanism of the condensation re-
action catalyzed by the anilinium salt
was also investigated by NMR spec-
troscopy experiments in which the
effect of differently substituted alde-
hydes and oximes as well as water on
the reaction rate was studied. The re-
sults point to the rate-limiting elimina-
tion of water from the 3-hydroxy-2-iso-
xazolidine intermediate.
gate addition of the oxime and
a
slower intramolecular oxime-transfer
reaction. The rate of oxime transfer
was found to correlate with the acidity
of the catalyst. This finding enabled us
to discover an enantioselective process
Keywords: acidity
·
amines
·
·
cyclization
·
enantioselectivity
isoxazolines
Introduction
hydes to act as electrophiles with uncharged nucleophiles in
a highly asymmetric manner, presumably by forming more
reactive iminium-ion intermediates.[2b,3]
Chiral secondary amine catalysts, often in combination with
various acids, have been shown to be efficient catalysts for
numerous transformations in the recent years.[1] Particularly
significant advances in iminium- as well as enamine-cata-
lyzed enantioselective transformations were made in 2000.
These involved the use of proline and imidazolidinone het-
erocycles as catalysts in aldol, conjugate-addition, and
Diels–Alder reactions.[2] A number of studies have shown
that these catalysts, and particularly the imidazolidinones
and diarylprolinol derivatives, activate a,b-unsaturated alde-
Our initial interest in these catalysts grew from the
demand for the invention of new asymmetric conjugate-ad-
dition protocols for heteroatom nucleophiles, especially
those of O-nucleophiles to unsaturated carbonyl compounds.
The conjugate additions of oxygen-containing nucleophiles
are well-established processes with unsaturated esters and
ketones,[4] even though catalytic asymmetric protocols[5] for
these additions are still rare.[6] The well-documented difficul-
ty associated with these additions is the usually low reactivi-
ty of the hydroxyl functionality that demands kinetic activa-
tion of the reactants. The more fundamental problem, how-
ever, is related to the thermodynamic instability of the prod-
ucts since they readily revert back to the starting materials
under equilibrating conditions. The thermodynamic instabili-
ty of the products is emphasized in the case of unsaturated
aldehydes, which are also prone to acetal and hemiacetal
formation with the alcohol nucleophiles in the presence of
basic or acidic activation of the reaction partners. Not sur-
prisingly, only a few studies on conjugate additions of
oxygen nucleophiles to unsaturated aldehydes,[7] and virtual-
ly no examples of enantioselective catalysis, were published
at the time of the inception of our study. The products of
these additions can be viewed as (protected) aldol products
[a] Prof. P. M. Pihko
Department of Chemistry, University of Jyvꢀskylꢀ
40014 JYU (Finland)
Fax : (+358)14-260-2501
E-mail: Petri.Pihko@jyu.fi
[b] A. Pohjakallio
Department of Chemistry
Aalto University School of Science and Technology
Kemistintie 1, 02150 Espoo (Finland)
[c] U. M. Laitinen
Department of Chemistry, University of Jyvꢀskylꢀ
40014 JYU (Finland)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201000861.
Chem. Eur. J. 2010, 16, 11325 – 11339
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
11325

of acetaldehyde and acceptor aldehydes. These aldol reac-
tions remain extremely challenging and have been successful
only with a limited number of acceptor aldehydes, mostly ar-
omatic aldehydes.[8] With this in mind, we felt that there
would be definite use and need for the development of al-
ternate processes based on conjugate additions.
ty of such species,[16] also steer the reactivity of the nucleo-
phile in favor of the conjugate-addition process.
The reactivity of different oxygen nucleophiles was inves-
tigated by monitoring their reactions with crotonaldehyde in
the presence of the MacMillan imidazolidinone/trifluoroace-
tic acid (TFA) catalyst/co-catalyst system (Table 1).
Since the initiation of our study, both alcohols[9] and
oximes[10] have been successfully used as nucleophiles in
asymmetric addition to enals. However, the aldehyde oxida-
tion state was not preserved in the oxime additions, and the
isolation of the fragile oxime adducts was possible only after
reduction of the aldehydes to the corresponding alcohols.
Oximes and other oxygen nucleophiles have also been used
as components in many creative cascade reactions[11] in
which the initial conjugate-addition products are further
processed in subsequent reactions. Surprisingly, stabilization
of the aldehyde oxidation state by internal oxime transfer
has not been used in asymmetric catalysis. In this paper, we
present the research that resulted in the development of two
new synthetic protocols that utilize the oxime-transfer pro-
cess for the synthesis of 3-unsubstituted 2-isoxazolines both
in the racemic[12] and enantioselective[13] manner (Scheme 1).
The protocols involve two distinct reaction steps, an
oxime conjugate-addition and oxime-transfer reaction,
which, depending on the catalyst, may be conducted sepa-
rately or as a one-flask cascade. These methods are comple-
mentary to the known methods to synthesize these com-
pounds[14] and represent the first catalytic asymmetric
method for the synthesis of 3-unsubstituted 2-isoxazolines.[15]
Table 1. Conjugate addition of crotonaldehyde with different oxygen nu-
cleophiles.[a]
Nucleophile
t [min]
Yield [%][b]
2a
2b
2c
2d
2e
2 f
2g
–
20
20
20
20
0
6
24
26
21
36
28
180
15
Results and Discussion
[a] Cbz=carbobenzyloxy, Troc=carbo(2,2,2-trichloroethyl)oxy. [b] NMR
spectroscopic yields (conversion to product) for 4a–g were determined
by 1H NMR spectroscopy using the catalyst as the internal standard. Con-
ditions: 0.1 mmol 1a, 0.1 mmol 2, 1 mL CDCl3.
Initial studies: Based on literature precedents,[5a] we rea-
soned that the character of the oxygen nucleophile would
play a strong role in the success of iminium-catalyzed conju-
gate-addition reactions to enals. Relatively hard nucleo-
philes, such as alcohols, would not be ideal candidates for
the reaction, since direct 1,2-addition to the iminium carbon
might compete effectively with the desired 1,4-addition.
Thus we hypothesized that the insertion of an a-heteroatom
adjacent to the nucleophilic oxygen atom could, in addition
to its documented effect of raising the relative nucleophilici-
1
Direct monitoring of the reactions by H NMR spectro-
scopic experiments revealed that the conversions to conju-
gate-addition products in the tested conditions were rela-
tively low in all cases. We were unable to isolate these addi-
1
tion products, and they could only be detected by H NMR
spectroscopy, but not by TLC. Nevertheless, the 1H NMR
spectroscopic study revealed
several features that are charac-
teristic to these reactions. First,
as expected,
a
nucleophiles
gave faster reactions and higher
conversions. Second, the con-
versions typically reached a pla-
teau that we expected to repre-
sent the equilibrium of the ad-
dition in the reaction condi-
tions. Third, an exception to
this behavior was observed with
oxime 2g. Under these condi-
Scheme 1. Two protocols for 2-isoxazoline formation presented herein.
11326
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 11325 – 11339

Synthesis of 2-Isoxazolines
FULL PAPER
tions, the reaction proceeded to give 2-isoxazoline 5a and
anisaldehyde 6 as the major products (Figure 1).
Based on this conversion plot, we assumed that isoxazo-
line 5a is generated by means of an intramolecular cycliza-
tion reaction of intermediate 4g. The formation of 4g most
likely proceeds by means of a conjugate addition between
the oxime 2g and the enal 1a catalyzed by an imidazolidino-
nium salt. Interestingly, oxime 2 f was ineffective in this
transformation.[17] As such, a range of oximes was screened
to reveal the structural requirements of the isoxazoline-for-
mation process. These studies are summarized in Figure 2.
In general, aliphatic oximes such as acetone oxime 7a or
pivalaldehyde oxime 7b displayed higher rates towards iso-
xazoline formation when compared to aromatic oximes.
However, in all cases the final yields (determined by NMR
spectroscopy) of the product were typically only around
50%, despite the fact that no significant side products could
1
be identified by H NMR spectroscopy.
Studies with the imidazolidinone catalyst: To improve the
yields, the reaction conditions were optimized. The optimi-
zation was conducted using trans-hex-2-enal and acetone
oxime as the reactants.[18]
First, a temperature screen indicated that optimal yields
could be obtained by running the reaction at 08C. Lowering
the temperature further simply decreased the reactivity but
did not improve the yields,
~
&
Figure 1. Formation of intermediate 4g ( ) and isoxazoline 5a ( ) from
*
crotonaldehyde 1a ( ) and oxime 2g.
whereas at higher temperatures,
faster rates but lower NMR
spectroscopic yields were ob-
served. The side reactions re-
sponsible for erosion of the
yields at higher temperatures
are most likely polymerization
processes since no definite low-
molecular-weight side products
could be detected by NMR
spectroscopic studies.
The reaction medium was
found to have
a significant
impact on the yields. These ex-
periments are summarized in
Table 2. In general, nonpolar
solvents such as toluene afford-
ed the highest rates and yields.
Further optimization studies
revealed that using
a slight
excess amount (110–120 mol%)
of the enal afforded optimal
yields, and lower yields were
obtained when larger excess
amounts of either reagent were
used. Curiously, the order or re-
agent addition seemed to have
a small, but noticeable, effect
on the final yield as well. The
Figure 2. Formation of isoxazoline 5a with different oximes. Conditions: crotonaldehyde 0.1 mmol, oxime
best results were obtained when 0.1 mmol, and catalyst 0.02 mmol in CDCl3. Catalyst was used as the internal standard.
Chem. Eur. J. 2010, 16, 11325 – 11339
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
11327

P. M. Pihko et al.
Table 2. Solvent effect on isoxazoline formation.
Table 3. Effect of co-acid on enantiomeric excess using the imidazolidi-
none catalyst.[a]
Solvent[a]
Yield [%] (1 h)[b] Yield [%] (3 h)[b] Yield [%] (22 h)[b]
Acid
Yield [%] (17 h)[b]
0
ee [%][c]
MeCN
MeOH
THF
diethyl ether
toluene
CHCl3
0
0
0
0
20
0
n.d.[d]
0
7
26
18
16
0
12
30
73
61
54
10
40
30
29
5
n.d.
CH2Cl2
[a] Conditions: 0.12 mmol 1b, 0.10 mmol 7a, 0.020 mmol catalyst, 1 mL
solvent. [b] NMR spectroscopic yields (conversion to product) were de-
termined by 1H NMR spectroscopy experiments from aliquots of the re-
action mixture diluted with CDCl3. Catalyst was used as the internal stan-
dard.
13
12
n.d.
n.d.
10
ꢀ10
catalyst and aldehyde were mixed prior to addition of the
oxime. However, gradual addition the oxime did not im-
prove the process.
46
76
34
ꢀ5
8
Surprisingly, under otherwise optimal conditions, the
chiral oxazolidinone catalyst afforded a nearly racemic iso-
xazoline product. Attempts to improve the enantiomeric
excess (ee) values by varying the acid co-catalyst component
did not improve this result: the use of p-toluene sulfonic
acid (pTsOH) raised the ee value to 36%, but NMR spec-
troscopic yields were disappointingly low, and similar results
were obtained with many other co-acids (Table 3). At lower
temperatures (ꢀ208C), somewhat higher ee values (60%)
were obtained with pTsOH 16 and related sulfonic acids
((+)-camphorsulfonic acid, 1- and 2-naphthylsulfonic acids),
but the conversions remained poor (<20%).
36
[a] Conditions: 0.12 mmol 1b, 0.10 mmol 7a, 0.02 mmol catalyst, and
0.02 mmol co-acid, 1 mL toluene. [b] NMR spectroscopic yield (conver-
sion to product), as determined by 1H NMR spectroscopy. Catalyst was
used as the internal standard. [c] Enantiomeric excess was measured with
GCMS. [d] n.d.=not determined.
Development of the racemic process: Having initially failed
to discover the enantioselective process, a catalyst screen
was conducted to find a cheaper and simpler amine catalyst
for the synthesis of 2-isoxazolines in racemic form. Although
the majority of secondary amine iminium-ion catalysts pub-
lished in the literature are chiral, some nonchiral amines
have also been used.[19] The studies with achiral amines are
presented in Table 4. Aliphatic secondary amines (22–27),
secondary anilines (28a–28d), and primary anilines (29a,
29b) were screened. All amines were combined with four
different acids of different pKa. In addition, diphenylphos-
phate (DPP) was tested as it had performed very similary
with TFA in other mechanistically related reactions studied
in our laboratory.[19b]
The amine/acid screen revealed that different secondary
anilines, in combination with DPP, typically afforded superi-
or results. With several amines, the reaction failed to pro-
ceed further from the initial conjugate-addition product
with both weakest (chloroacetic acid) and strongest (metha-
nesulfonic acid) acid co-catalysts. In these cases, the isoxazo-
line formation was suppressed. Based on these results, we
selected the salt of N-methylaniline and DPP for further
studies as both components were cheap and commercially
available.
The other reaction parameters were reoptimized for the
N-methylanilinium diphenylphosphate catalyst. Fine-tuning
the structure of the aliphatic oxime[20] was helpful since
oximes such as 3-pentanone oxime and methyl isopropyl
ketone oxime gave slightly better yields than acetone oxime.
3-Pentanone oxime was thus selected as the oxime source as
it gave the best combination of reaction rate and yield. The
effect of reaction concentration was also evaluated. In gen-
eral, a concentration of 0.2–0.4m appeared to be optimal, as
the lower (<0.1m) concentration resulted in a lower reac-
tion rate, and higher concentration (1m) resulted in a lower
yield. Although increasing the concentration has a positive
impact on the reaction equilibrium and rate, it also acceler-
ates intermolecular side reactions, thereby resulting in lower
overall yields. In some cases, oximes of the initial conjugate
addition products could be detected as side products.
11328
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 11325 – 11339

Synthesis of 2-Isoxazolines
FULL PAPER
Table 4. Racemic catalyst/co-catalyst screen.[a]
1 h: 18% 4h
3 h: 26% 4h
1 h: trace of 5b
3 h: 10% 5b
1 h: 20% 4h
3 h: 15% 4h
1 h: trace of 5b
3 h: trace of 5b
n.d.[b]
1 h: 8% 4h
3 h: 18% 4h
1 h: trace of 5b
3 h: trace of 5b
1 h: 23% 4h
3 h: 17% 4h
1 h: 9% 4h
3 h: 14% 4h
n.d.
3 h: trace of 5b
1 h: 10% 4h
3 h: 13% 4h
1 h: 9% 5b
3 h: 20% 5b
1 h: 4% 4h
3 h: 12% 5b
1 h: trace of 4h
3 h: trace of 4h
1 h: 16% 4h
3 h: trace of 5b
1 h: 6% 4h
3 h: 11% 4h
1 h: 6% 4h
3 h: n.d
1 h: 15% 4h
3 h: 18% 4h
1 h: 18% 4h
3 h: 26% 4h
n.d.
1 h: 16% 4h
3 h: 10% 4h,
15% 5b
1 h: no product
3 h: 5% 4h
1 h: 11% 5b
3 h: 21% 5b
n.d.
n.d.
n.d.
n.d.
1 h : 10% 5b
20 h: 27% 5b
n.d.
n.d.
1 h: 22% 4h
3 h: 16% A,
trace of 5b
1 h: 42% 5b
3 h: 57% 5b
1 h: trace of 5b
3 h: n.d.
1 h: 12% 4h
3 h: 14% 4h
1 h: 55% 5b
3 h: 74% 5b
1 h: 14% 4h
3 h: 17% 4h
1 h: 42% 5b
3 h: 42% 5b
1 h: n.d.
3 h: 22% 5b
1 h: trace of 4h
3 h: 13% 4h
1 h: 53% 5b
3 h: 68% 5b
3 h: no product
3 d: 28% 5b
n.d.
n.d.
n.d.
n.d.
ACHTUNGTRENNUNG(10 mol%)
n.d.
n.d.
n.d.
n.d.
3 h: 33% 5b
1 h: 11% 5b
3 h: 22% 5b
1 h: 44% 5b
3 h: 51% 5b
1 h: 8% 5b
3 h: 40% 5b
1 h: 9% 4h
3 h: 15% 4h
1 h: 54% 5b
3 h: 55% 5b
1 h: 21% 5b
3 h: 36% 5b
1 h: 51% 5b
3 h: 48% 5b
1 h: 60% 5b
3 h: 55% 5b
n.d.
n.d.
[a] Conditions: 0.12 mmol 1b, 0.10 mmol 7a, 0.020 mmol catalyst, 0.020 mmol acid, and 1 mL toluene. Catalysts were used as the internal standards.
[b] n.d. = not determined.
After finding satisfactory conditions for the reaction, the
generality with respect to the a,b-unsaturated aldehydes was
explored (Table 5). These experiments revealed that the re-
action is general to the a,b-unsaturated aldehydes that bear
no further conjugation to the C=C bond. For example, cin-
namaldehyde and (2E,4E)-hexa-2,4-dienal were unreactive,
and the reaction with (E)-methyl 4-oxobut-2-enoate was
slow. Other substrates that bear a g-sp3-carbon atom are
well tolerated. Sensitive substrates (Table 5, entries 4 and 8)
gave reasonable yields, and acid labile functionalities such
as tert-butyldiphenylsilyl ether (TBDPS) and tert-butylcarba-
mate (Boc) (entries 9, 12) readily withstand the reaction
conditions. Notably, sterically demanding substrates that
bear two b-substituents are viable substrates for the reac-
tion; they afford the isoxazolines in good yields (entries 11–
14). However, the tropinone-derived enal (entry 12) gave a
sluggish reaction, presumably due to the presence of a com-
peting tertiary amine group, and addition of more acid did
not help.
Chem. Eur. J. 2010, 16, 11325 – 11339
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
11329

P. M. Pihko et al.
Table 5. Synthesis of racemic 2-isoxazolines with N-methylanilinium phosphate catalyst (30).[a]
reaction mixture. As an alterna-
tive to 3-pentanone oxime, cy-
clohexanone oxime gave similar
yields in some cases but the re-
action rates are faster (see
below). The advantage of 3-
pentanone, however, lies in its
slightly higher volatility.
Entry
1
Aldehyde
Product
t [h]
Yield [%][b]
55[c]
6.5
Development of the enantiose-
lective process: The next logical
objective was to return to the
development of the enantiose-
lective version of the reaction,
especially since only a very few
methods for the catalytic syn-
thesis of enantiomerically en-
riched 2-isoxazolines have been
disclosed.[15] In our initial at-
tempts, the MacMillan imidazo-
lidinone catalyst salts typically
gave good conversions but only
very low enantiomeric excess
(Table 3). Based on the hypoth-
esis that the reaction proceeds
in two steps, the conjugate-ad-
dition step and the cyclization
step, we first suspected that the
low enantioselectivity is attrib-
uted to rapid equilibration and
deterioration of the enantiose-
lectivity in the conjugate-addi-
tion step. Equilibration of the
conjugate-addition product with
an imperfect catalyst could
eventually lead to complete rac-
emization of this intermediate
before the cyclization step.
However, we had observed that
the imidazolidinone/TFA cata-
lyst was indeed able to catalyze
the oxygen conjugate-addition
2
3
4
5
6.0
6.5
80[b]
86
15.5
6.5
63
73
6
7
7.0
73
83
14.5
8
15.0
7.5
73
82
83
9
11
15.0
12
13
–
0[e]
85
4.0
14
14.0
69
82
15
16
15.0
no reaction
reaction of
a different but
equally reactive oxygen nucleo-
phile, N,N-diprotected hydrox-
ylamine 2c (Table 1), to croto-
naldehyde under similar condi-
tions with moderate enantio-
meric excess (Scheme 2).[21] As
such, at least at cold tempera-
[a] Bn=benzyl, PMBO=para-methoxybenzyloxy. [b] Yields refer to isolated and chromatographically purified
products. [c] The reaction was performed in CHCl3 using 10 mol% of 30 on a 20 mmol scale using acetalde-
hyde oxime as oxime reactant. Yield refers to distilled product. [d] NMR spectroscopic yield after 6 h is given.
p-Bromoanisole was used as the internal standard. [e] Similar results were obtained without added acid and
with 1 equiv of TFA or acetic acid.
In addition to 3-pentanone oxime, other oximes may be
good alternative reagents in special cases. For the reaction
with crotonaldehyde (Table 5, entry 1), acetaldehyde oxime
was used as the oxime donor as the acetaldehyde by-product
was slowly condensing to crotonaldehyde in the reaction
conditions. Using volatile oxime also enabled the purifica-
tion of the product isoxazoline by direct distillation from the
tures, the conjugate-addition step does not give racemic
products.
During the studies that probed the effect of different acid
and amine combinations to the reaction rate (Table 4), we
had also observed that the acid component alone was able
to promote isoxazoline formation without any added amine
catalyst. When these reactions were closely monitored by
11330
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 11325 – 11339

Synthesis of 2-Isoxazolines
FULL PAPER
ed than initially assumed. To study the cyclization step sepa-
rately, the conjugate-addition product was prepared by using
a polymer-supported N-benzylaniline and chloroacetic acid
co-catalyst. Both catalyst components were removed from
the solution by filtrating the solution through a plug of basic
alumina. This afforded a crude solution that contained the
conjugate-addition product and that could be used to probe
the effect of different acid-catalyst candidates by using
1H NMR spectroscopy (Figure 4).
These results indicated that the cyclization of aldehyde 33
to isoxazoline takes place very rapidly under acidic condi-
tions. No induction period was observed, and the cyclization
appears to be mediated solely by the acids. The N-methyla-
nilinium salt was less effective, and N-methylaniline failed
to promote the reaction at all. These results also explained
why some amine salts failed to produce the isoxazoline
products in the earlier amine/co-acid screen (Table 4).
Amines that are too basic suppress the isoxazoline-cycliza-
tion process by neutralizing the strong acids required for the
process. The decisive factor in the cyclization activity ap-
pears to be the acidity of the amine salt.
Scheme 2. Enantioselective conjugate addition with N-hydroxycarbamate.
NMR spectroscopy, sigmoidal conversion curves were ob-
served (Figure 3). No immediate buildup of the conjugate-
addition product intermediate was observed in these reac-
tions; instead, the rate of the reaction accelerated after an
acid-dependent induction period.[22] As such, the overall re-
action appeared to be catalyzed by a secondary catalyst gen-
erated during the reaction. It is therefore possible that a sec-
ondary catalyst formed during the reaction could also be re-
sponsible for the generation of the racemic products.
Given that the cyclization of the conjugate-addition inter-
mediate proceeds readily under strongly acidic conditions,
we concluded that any enantioselective oxime conjugate-ad-
dition process should be applicable to the asymmetric syn-
thesis of 2-isoxazolines if the unstable conjugate-addition
product could be cyclized effectively. In a timely coinci-
dence, the Jørgensen group had cleverly solved the problem
of asymmetric oxime conjugate addition by using their own
trimethylsilyl (TMS)-diarylprolinol catalyst 34. However,
the products were isolable only after NaBH4 reduction to
the corresponding alcohols. As such, we adapted the condi-
tions used by Jørgensen and co-workers to the conjugate-ad-
dition step. Acetone oxime was used instead of the benzal-
dehyde oxime used in the original method, due to its low
molecular weight and ease of hydrolysis. Several different
acidic conditions for the conversion of the resulting conju-
gate-addition intermediates to 2-isoxazolines were screened.
As summarized in Table 6, a brief exposure of the initial
conjugate-addition reaction mixtures to strongly acidic con-
ditions such as 3.2m H2SO4 in methanol or 1.7m HCl in
aqueous THF resulted in the rapid formation of enantiomer-
ically enriched 2-isoxazolines in good enantioselectivities
and >60% yields. A control experiment with NaBH4
(Table 6, entry 15) indicated that these cyclization conditions
preserved the conjugate-addition product with the same effi-
ciency and ee values as the NaBH4 workup. Although ben-
zaldehyde oxime gave better enantioselectivities (entry 6),
the overall yield was inferior to that obtained with acetone
oxime.[23] The high ee value and the moderate yield of the
product are largely determined in the conjugate-addition
step, and the cyclization step is fast enough to preserve the
initially generated enantiomeric excess when strong acids
are used in protic conditions to cyclize the conjugate-addi-
tion product. Nonprotic cyclization conditions appeared to
deteriorate the ee values. Other catalyst combinations did
not afford better results.[24]
This result also indicated that the catalytic pathway that
leads to isoxazoline formation was perhaps more complicat-
Figure 3. Formation of 2-isoxazoline in the presence of acid catalysts (tri-
~
&
*
chloroacetic acid ( ), diphenyl phosphate ( ), TFA ( ), bis(4-nitrophe-
&
^
^
nyl)phosphate ( ), methanesulfonic acid ( ), p-TsOH ( ), dichloroacetic
~
*
acid ( ), chloroacetic acid ( )). p-Bromoanisole was used as the internal
standard.
Chem. Eur. J. 2010, 16, 11325 – 11339
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
11331

P. M. Pihko et al.
The substrate scope of the
enantioselective
isoxazoline
protocol was examined with a
range of different a,b-unsubsti-
tuted enals (Table 7). Gratify-
ingly, a wide range of function-
alities seemed to tolerate the
acidic cyclization conditions, in-
cluding
para-methoxybenzyl
(PMB) ethers (Table 7, entry 4),
carbamates (entry 8), and unsa-
turated esters (entry 7). Croto-
naldehyde turned out to be
challenging substrate, possibly
due to side reactions. Interest-
ingly, geranial 1j was also a
competent substrate (entry 11);
it afforded the product in aver-
age yield but lower enantiose-
lectivity, probably as a conse-
quence of the smaller energy
difference between the reactive
(E)- and (Z)-iminium ions. Im-
portantly, reactions run at a
slightly larger scale (2 mmol)
and with
a
larger excess
amount of the oxime reagent
(4.5 equiv) and longer reaction
time (5 h) increased the yields
in all cases (entries 2–4, 6–8,
and 11) by around 10–15%.[20]
When the enantioselective
process was scaled up to 2–
5 mmol, we encountered incon-
sistencies in the product ee
values with aldehydes 1c, 1d,
and 1e. After extensive experi-
mentation, including examina-
tion of the reaction conditions,
reagents, reaction times, and
both the chemical and optical
purity of the commercial cata-
lyst, careful purification of the
amine catalyst afforded consis-
tently good enantioselectivities.
The overall yields, however,
were not typically affected by
the impurities.[20]
&
Figure 4. Conversion of intermediate 33 to 5c in the presence of different acids (methanesulfonic acid ( ), p-
~
~
^
*
TsOH ( ), diphenyl phosphate ( ), TFA ( ), dichloroacetic acid ( ), N-methylaniline-DPP (&)). The NMR
spectroscopic yield of 5c, based on the initial concentration of intermediate 33, is reported. p-Bromoanisole
was used as the internal standard. The acid-induced formation of 33 from the remaining starting materials 1c
and 7h is likely responsible for conversions that are higher than 100%.
Table 6. Screen of acidic cyclization conditions and comparison with reductive quench (NaBH4, entry 15).
Entry
R1, R2
Acid[a]
Yield [%][b]
ee [%]
1
2
3
4
5
6
7
8
Me, Me
Et, Et
Me, Me
Et, Et
Me, Me
H, Ph
H, C(CH)3
Me, Me
Me, Me
Me, Me
Me, Me
Me, Me
Me, Me
Me, Me
Me, Me
2.36m H2SO4 (aq)
2.3m H2SO4 (aq)
2m HCl (aq)
2m HCl (aq)
2.3m H2SO4/MeOH
2.3m H2SO4/MeOH
2.3m H2SO4/MeOH
THF/H2O+H2SO4 (2 mmol, 400 mol%)
THF/H2O+pTsOH (2 mmol)
THF/H2O+H3PO4 (2 mmol)
THF/H2O+TFA (2 mmol)
THF/H2O+HCl (2 mmol)
MeSO3H (300 mol%)
Et2O/HCl (dry, 5m)
60
54
48
47
55
14
10
64
52
46
68
65
25
45
60[c]
88
87
91
86
92
98
93
90
92
86
86
92
90
80
91
Mechanistic discussion: We pro-
pose that the overall mecha-
nism for the 2-isoxazoline for-
mation predominantly follows a
pathway along which the conju-
gate-addition product formed in
a fast reaction between alde-
9
10
11
12
13
14
15
NaBH4/MeOH
[a] 1 mL of the solution was used. [b] Isolated yield of 2-isoxazoline. [c] Isolated yield of the corresponding alcohol. hyde and the oxime cyclizes to
11332
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 11325 – 11339

Synthesis of 2-Isoxazolines
FULL PAPER
Table 7. Enantioselective synthesis of different optically active 2-isoxazo-
lines with the diarylprolinol catalyst 34.
elimination of the water molecule from the carbinolamine
intermediate T0 is rate-limiting.[27b] In the intermediate pH
range, the proton-transfer reactions that lead to the carbi-
nolamine intermediate have been shown to have kinetic sig-
nificance. In oxime hydrolysis, on the other hand, the nucle-
ophilic attack of water to the oxime has been shown to limit
the reaction at high pH, whereas the elimination of the hy-
droxylamine limits the reaction rate in acidic conditions.[26c]
These results of these studies are used as the basis for the
following discussion.
Entry
1
Product
Yield [%][a]
39[b]
ee [%]
86
2[c]
72
65
94
The mechanism of the oxime transfer likely involves inter-
mediates that are similar to those observed in the studies of
oxime formation and hydrolysis reactions. In our enantiose-
lective process, we used strongly acidic protic conditions for
the cyclization step. Based on the mechanistic studies men-
tioned above, the overall rate of oxime transfer in these con-
ditions is likely determined by the rates of the elimination
reactions in the proposed intermediates B and C (which cor-
respond to the elimination of hydroxylamine in oxime hy-
drolysis) or by the rates of cyclization of the proposed inter-
mediates B or D (which correspond to the nucleophilic
attack in the formation of T0). This hypothesis is supported
by the cyclization experiments conducted with isolated b-
oxime ketones 35 and 37 that indicated slower cyclization
rates with ketones compared to the aldehyde conjugate-ad-
dition intermediates. Even though a direct comparison of
the cyclization rates in these conditions is not reliable due
to difficulties in standardizing the experimental conditions
for both substrates (aldehyde and ketone), these results
imply that the initial attack of water is not likely to be the
rds under these conditions. Although the reaction is slower
with ketones, it should be noted that the cyclization still pro-
ceeds to give high yields of the corresponding 3-substituted
isoxazolines. As such, the conjugate-addition/oxime-transfer
route is not restricted to aldehydes, and could be used to
prepare 2-isoxazolines from a,b-unsaturated carbonyl com-
pounds (Scheme 4).
3[c]
91–97
4[c]
5
60
52
95
90
6[c]
69
72
94
94
7[c]
8[c]
59
59
92
91
9
10
70
58
90
69
11[c]
[a] Isolated yields. [b] NMR spectroscopic yield. p-Bromoanisole was
used as the internal standard. [c] Reaction conducted on a 2 mmol scale
(2.0 mmol aldehyde, 9.0 mmol oxime, and 0.20 mmol of the chromato-
graphically purified catalyst).
Mechanistically, the process catalyzed by anilinium salt in
which the isoxazoline is formed in aprotic conditions repre-
sents a different case. To the best of our knowledge, the for-
mation and hydrolysis of oximes has not been studied in
aprotic media. The situation differs from aqueous protic
media significantly in two ways. First, water is available only
in catalytic quantities from the aldehyde–iminium equilibri-
um between the catalyst and enal. Second, the reaction con-
ditions are only moderately acidic as the only acid is the cat-
alytic anilinium salt (cf. Figure 4 in which different acids are
compared). Under these conditions, the most likely rate-lim-
iting steps are the formation of B and the dehydration of E.
The possible nucleophilic cyclization (B to C and D to E)
and elimination steps (B to D, C to E) resemble the steps in-
volved in the enamine–iminium equilibrium and are usually
considered to be fast in similar conditions.[28]
2-isoxazoline in the presence of catalysts. The conjugate-ad-
dition reaction of oximes likely involves iminium ions as re-
active intermediates. This mechanism is supposedly opera-
tional in the presence of anilinium salts as well.
The cyclization reaction to 2-isoxazoline appears to be an
intramolecular oxime-transfer reaction that mechanistically
consists of sequential oxime hydrolysis and oxime-formation
reactions (Scheme 3). Related oxime-transfer reactions have
been previously reported for several different substrates in
aqueous acidic conditions.[25]
More OꢂFerrall, Jencks, and Sayer have conducted thor-
ough studies of oxime hydrolysis[26] and formation[27] reac-
tions in aqueous media. Both reactions have been shown to
proceed by means of a carbinolamine intermediate T0
(Scheme 3). For the formation of p-chlorobenzaldehyde O-
methyloxime, the pH of the solution determines the rate-de-
termining step (rds). At low pH (0–2), the attack of the nu-
cleophile to the aldehyde is the rds, and above pH 5, the
We examined the effect of added water and both oxime
and aldehyde structure to the rate of anilinium-salt-cata-
lyzed isoxazoline formation.[20] Upon addition of 10, 20, 50,
and 75 mol% of water to the reaction mixture, a small rate-
Chem. Eur. J. 2010, 16, 11325 – 11339
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
11333

P. M. Pihko et al.
Scheme 3. Proposed mechanistic pathway for the formation of the intermediate A and the subsequent oxime-transfer reaction.
The results of water and oxime structure experiments
would suggest the step A!B as the rate-limiting step. If this
were the case, b,b-disubstitution of the aldehyde would not
be expected to have a significant effect on the rate unless
the rate differences were simply caused by concentration
differences between the reactive intermediates A with differ-
ent aldehydes. To probe these differences, the conjugate-ad-
dition steps between different enals and a particularly
slowly cyclizing oxime, p-nitrobenzaldehyde oxime, were ex-
amined (Figure 5).[30] Although the equilibrium concentra-
tions of the intermediates were higher with exocyclic enals
1r and 1l compared to other substrates,[31] the structurally
very similar pair 1m and 1c behaved differently. The b,b-
disubstituted substrate 1m cyclizes faster than the b-mono-
substituted substrate 1c, although the equilibrium concen-
trations of the intermediates 4m are smaller compared to
4c. It is thus reasonable to assume that the rate-limiting step
occurs later in the cyclization pathway, and the observed ef-
fects with water and different oximes are incorporated into
the overall rate at earlier stages of the sequence, presumably
through equilibration.
Scheme 4. Acid-promoted cyclization of ketone substrates.
accelerating effect was observed in the reaction between
trans-hex-2-enal and 3-pentanone oxime, but no further ac-
celeration could be seen beyond 50 mol% of water, presum-
ably due to saturation of the reaction mixture.[20] The effect
of oxime structure on the reaction rate was evident already
in the preliminary tests of the reaction. Both aldehyde and
ketone oximes proved to be reactive, but increasing the
steric hindrance or introduction of electron-withdrawing
substituents in the para position of benzaldehyde oximes de-
creased the reaction rate.[29] Finally, experiments that com-
pared the reaction rates of 3-pentanone oxime with differ-
ently substituted aldehydes demonstrated a clear accelerat-
ing effect of additional aldehyde b-substitution (Figure 5).
Exocyclic enals such as cyclohexylideneacetaldehyde (1r)
and 1l turned out to be particularly reactive substrates, and
geranial 1j and 1m also reacted faster than structurally simi-
lar 1c.
Based on the established reaction pathways[26,27] (see
above) for the intermolecular oxime formation and hydroly-
sis, the breakdown rates of the tetrahedral intermediates B,
C, or E (Scheme 3) were the likely candidates for the rate-
limiting step of the intramolecular oxime-transfer reaction.
We probed the relative breakdown rates of the tetrahedral
intermediate B or C by comparing the reaction rates of ani-
linium-catalyzed 2-isoxazoline formation between 3-penta-
none oxime, acetone oxime, cyclopentanone oxime, and cy-
clohexanone oxime. Cyclohexanone oxime turned out to be
a superior reagent to all other tested oximes in terms of rate
11334
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 11325 – 11339

Synthesis of 2-Isoxazolines
FULL PAPER
but decreased reactivity in the
ketone elimination step (C!
E). Thus, the increased relative
rate observed with cyclohexa-
none oxime does not support
the slow elimination of ketone
from B or C. If the step A!B
were rate-limiting, the rate dif-
ferences between the cycliza-
tion of the b,b-disubstituted al-
dehyde such as 1m and the b-
monosubstituted aldehyde 1c
would be expected to be rather
small. However, even with cy-
clohexanone oxime, the forma-
tion of 5m (from the b,b-disub-
stituted aldehyde) is almost
twice as rapid as the formation
of 5c (see Figure 6). This result
leaves the rate-limiting break-
down of final intermediate E as
the only possibility that ac-
counts for the differing rates
between both different alde-
hydes and oximes.[32] However,
considering that the evidence
for this is not direct and that
the rate differences between
different substrates and condi-
tions are relatively small, it is
entirely possible that the ener-
gies of cyclization intermediates
Figure 5. NMR spectroscopic yields of conjugate-addition intermediates (4c, 4l, 4m, and 4r) and 2-isoxazo-
lines (5c, 5j, 5l, 5m, and 5r) in the reactions of different enals (1c, 1l, 1m, and 1r; 0.20 mmol, 100 mol%)
with p-NO2-benzaldehyde oxime (7 f; 0.20 mmol, 100 mol%) or 3-pentanone oxime (7h; 0.20 mmol,
100 mol%). p-Bromoanisole or 3,5-bis(trifuoromethyl)bromobenzene were used as the internal standards.
as well as the barriers between them may also be small and,
as a consequence, the overall reaction in different cases may
be determined by several factors that depend on reagents.
Also, other mechanisms that involve the participation of
other nucleophiles, such as the aniline catalyst as the initiat-
ing nucleophile in the cyclization pathway, cannot be exclud-
ed by our present studies.[33]
Conclusion
We have developed a new method for the synthesis of the 3-
unsubstituted 2-isoxazolines that takes advantage of the con-
jugate-addition reaction of oximes with enals. In the pres-
ence of acidic catalysts, the labile conjugate-addition inter-
mediates directly cyclize to thermodynamically more stable
2-isoxazolines. The general strategy was shown to be appli-
cable to the enantioselective synthesis of 3-unsubstituted 2-
isoxazolines by the rapid acid-catalyzed cyclization of enan-
tio-enriched conjugate-addition products. This protocol is
the first general catalytic asymmetric method for the prepa-
ration of 3-unsubstituted 2-isoxazolines. The acid-mediated
cyclization can also be applied to Michael adducts of oximes
and a,b-unsaturated ketones to afford 3,5-disubstituted iso-
xazolines. Finally, the anilinium-catalyzed process displays
Figure 6. NMR spectroscopic yields of 2-isoxazolines (5c, 5j, 5l, 5m, and
5r) and the respective conjugate-addition intermediates (4c–4r) in the
reactions of different enals (1c, 1j, 1l, 1m, and 1r, 0.20 mmol,
100 mol%) with cyclohexanone oxime (7i; 0.20 mmol, 100 mol%). p-
Bromoanisole or 3,5-bis(trifuoromethyl)bromobenzene were used as the
internal standards.
(Figure 6 and Supporting Information). According to
Brown,[31] exocyclic double bonds should be more reactive
in six-membered rings compared to similar acyclic or five-
membered ring systems. As such, cyclohexanone oxime
should show increased reactivity compared to acyclic ketone
oximes in the nucleophilic addition step (A!B, Scheme 3)
Chem. Eur. J. 2010, 16, 11325 – 11339
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
11335

P. M. Pihko et al.
phosphate and N-methylaniline. Diphenylphosphate (100 mol%) was dis-
solved in Et2O (used as much as needed to ensure complete dissolution
of DPP) at room temperature. The solution was cooled on an ice bath
and N-methylaniline (105 mol%) was added. Stirring the solution on an
ice bath resulted in precipitation of the salt that was filtered, washed well
with Et2O, and dried in vacuum. The salt was used as catalyst without fur-
ther purification. For analytical purposes, the salt was recrystallized from
interesting substituent effects. These suggest that that the
rate-limiting step in these processes might be the dehydra-
tion of the 3-hydroxy-2-isoxazolinidine intermediate.
1
Experimental Section
EtOH/Et2O. White solid; m.p. 78.5–79.58C; H NMR (400 MHz, CDCl3):
d=11.34 (brs, 2H), 7.30–7.19 (m, 13H), 7.10–7.0 (m, 2H), 2.70 ppm (s,
3H); 13C NMR (100 MHz, CDCl3): d=152.6, 152.5, 139.3, 129.7, 129.3,
127.6, 123.6, 121.7, 120.3, 120.25, 37.1 ppm; elemental analysis calcd (%)
for C19H20NO4P: C 63.9, H 5.6, N 3.9; found: C 64.0, H 5.6, N 3.7.
General: All moisture-sensitive reactions were carried out under an
argon atmosphere in flame-dried glassware unless otherwise noted. Prep-
arative 2-isoxazoline reactions were conducted in screw-cap glass vials or
small flasks by using dry toluene as the reaction solvent, but without ex-
clusion of air or moisture. The reaction temperatures refer to the temper-
atures of the cooling or heating baths.
General synthetic procedure for racemic 2-isoxazolines using N-methyla-
nilinium diphenylphosphate (30) catalyst:[20] Catalyst 30 (36 mg,
0.10 mmol, 20 mol%) was dissolved in toluene (2.5 mL) at room temper-
ature and the solution was cooled to 08C. a,b-Unsaturated aldehyde
(0.60 mmol, 120 mol%) was added, followed by 3-pentanone oxime
(0.50 mmol, 100 mol%) (or cyclohexanone oxime; see notes in the Sup-
porting Information). After the reaction was completed (based on con-
sumption of the aldehyde or oxime, indicated by TLC or 1H NMR spec-
troscopy) the reaction mixture was diluted with Et2O (15 mL) and the re-
sulting solution was washed with saturated NaHCO3 (5 mL) and 5%
aqueous oxalic acid (2ꢃ5 mL). Both basic and acidic aqueous phases
were then back-extracted with Et2O (6 mL, 2ꢃ5 mL, respectively). The
combined organic extracts were washed with brine (10 mL), dried
(Na2SO4), and concentrated carefully in vacuo, and the resulting residue
was purified by column chromatography.
When needed, nonaqueous reagents were transferred under argon by sy-
ringe or cannula and dried prior to use. THF and toluene used were ob-
tained either by passing deoxygenated solvents through activated alumina
columns (MBraun SPS-800 Series solvent purification system) or by dis-
tillation over sodium/benzophenone. Other solvents and reagents were
used as obtained from suppliers unless otherwise noted. Analytical TLC
was performed using Merck silica gel F254 (230–400 mesh) plates and an-
alyzed by UV light (254 or 366 nm) and by staining upon heating with
standard vanillin, permanganate, ninhydrin, or cerium-phosphomolybdic
acid solutions. For silica gel chromatography, the flash chromatography
technique was used, with Merck silica gel 60 (230–400 mesh) and p.a.
grade solvents unless otherwise noted.
The 1H and 13C NMR spectra were recorded in CDCl3 using a Bruker
Avance 400 (1H: 399.98 MHz; 13C: 100.59 MHz) spectrometer. The chem-
ical shifts are reported in ppm relative to CHCl3 (d=7.26 ppm), C2H2Cl4
(d=5.98 ppm), or CD3CN (d=1.94 ppm) for 1H NMR spectroscopy and
(d=77.0, 73.7, 1.32 ppm) for 13C NMR spectroscopy. In general, the
NMR spectroscopic yields of the conjugate-addition products and the in-
termediates were determined by 1H NMR spectroscopic experiments by
using the integrals of the aldehyde peak (d=9.55–9.50 ppm, 1H) for the
starting material, the aldehyde peak (d=9.8 ppm, 1H) for the intermedi-
ate, and the 5-H (dꢁ4.7–4.6 ppm, 1H) or 4-H (dꢁ3.1–3.0 and 2.6–
2.5 ppm, both 1H) peaks for the isoxazoline. The integrals were com-
pared against an internal standard (20 mol%, p-bromoanisole (d=7.4
(2H), 6.8 (2H), 3.8 ppm (s, 3H)) or 3,5-bis(trifluoromethyl)bromoben-
zene (d=8.0 (2H), 7.8 ppm (1H)). In the experiments described in
Tables 1–4, the integrals of the starting materials and products were cali-
brated with catalyst peaks (for example, tert-butyl peak of 3 at d=0.75–
0.80 ppm).
Compound 5g (mixture of diastereomers 1:1): Prepared as described in
the general procedure. Starting materials: a,b-unsaturated aldehyde
(103 mL, 1=1.04, 0.60 mmol, 120 mol%), 3-pentanone oxime (55 mL, 1=
0.92, 0.50 mmol, 100 mol%). Reaction time: 14.5 h. Eluent: gradient 8–
30% methyl tert-butyl ether (MTBE) in hexanes. Yield: 80 mg (83%),
yellow oil; Rf =0.4 (50% EtOAc in hexanes); 1H NMR (CDCl3,
400 MHz): diastereomer 1: d=7.10 (t, J=1.8 Hz, 1H), 5.90 (d, J=3 Hz,
1H), 5.86–5.83 (m, 1H), 4.46 (dtd, J1 =5.3 Hz, J2 =7.9 Hz, J3 =10.5 Hz,
1H), 3.06–2.94 (m, 1H), 2.96 (ddd, J1 =1.8 Hz, J2 =10.5 Hz, J3 =17.4 Hz,
1H), 2.49 (ddd, J1 =1.8 Hz, J2 =7.9 Hz, J3 =17.4 Hz, 1H), 3.44 (s, 3H),
2.25 (d, J=0.9 Hz, 1H), 1.87 (ddd, J1 =5.3 Hz, J2 =8.1 Hz, J3 =13.8 Hz,
1H), 1.80 (ddd, J1 =5.3 Hz, J2 =9.2 Hz, J3 =13.8 Hz, 1H), 1.27 ppm (d,
J=7.0 Hz, 3H); diastereomer 2: d=7.10 (t, J=1.8 Hz, 1H), 5.87–5.82
(m, 2H), 4.46 (m, 1H), 3.03 (ddd, J1 =1.8 Hz, J2 =10.4 Hz, J3 =17.2 Hz,
1H), 3.05–2.90 (m, 1H), 2.58 (ddd, J1 =1.9 Hz, J2 =8.2 Hz, J3 =17.3 Hz,
1H), 2.25 (d, J=0.9 Hz, 3H), 2.12 (td, J1 =7.0 Hz, J2 =13.9 Hz, 1H), 1.68
(ddd, J1 =6.6 Hz, J2 =7.3 Hz, J3 =13.9 Hz, 1H), 1.29 ppm (d, J=7.0 Hz,
3H); 13C NMR (CDCl3, 100 MHz): d=157.4, 157.0, 150.4, 150.3, 146.0,
145.9, 105.7, 105.6, 105.1, 104.4, 77.2, 76.5, 41.8, 40.81, 40.78, 40.6, 30.7,
29.9, 20.1, 18.8, 13.5, 13.45 ppm; FTIR for the mixture of diastereomers
The enantiomeric excesses of the products were determined by HPLC
using chiral stationary phases and the corresponding racemic samples as
references. Melting points were determined in open capillaries using a
Gallenkamp melting-point apparatus and are uncorrected. IR spectra
were recorded using a Perkin–Elmer Spectrum One FTIR spectrometer.
Optical rotations were obtained using a Perkin–Elmer 343 polarimeter.
High-resolution mass spectrometric data were obtained using a Micro-
Mass LCT Premier Spectrometer by the Analytical Services of the De-
partment of Chemistry.
(film): n˜ =2968, 2824, 1601, 1567, 1453, 1436, 1220, 1020, 840, 784 cmꢀ1
;
HRMS (ESI+): m/z calcd for [C11H15NO2+Na]: 194.1181; found:
194.1178.
Compound 5j: Prepared as described in the general procedure. Starting
materials: a,b-unsaturated aldehyde (105 mL, 1=0.87, 0.60 mmol,
120 mol%), 3-pentanone oxime (55 mL, 0.50 mmol, 100 mol%). Reaction
time: 15 h. Eluent: gradient 10–20% EtOAc in hexanes. Yield: 70 mg
(83%), colorless oil; Rf =0.34 (20% EtOAc in hexanes); 1H NMR
(400 MHz, CDCl3): d=7.01 (t, J=1.7 Hz, 1H), 5.07 (m, 1H), 2.81 (dd,
J1 =1.7 Hz, J2 =17.5 Hz, 1H), 2.65 (dd, J1 =1.7 Hz, J2 =17.5 Hz, 1H), 2.03
(m, 2H), 1.67 (s, 3H), 1.66–1.62 (m, 2H), 1.59 (s, 3H), 1.34 ppm (s, 3H);
13C NMR (100 MHz, CDCl3): d=145.6, 132.0, 123.5, 84.5, 45.5, 39.9, 25.6,
25.4, 22.9, 17.6 ppm; FTIR (film): n˜ =2971, 2928, 2858, 1673, 1600, 1439,
Materials: All commercial reagents, unless otherwise noted, were used
without further purification. 2,2,6,6-Tetramethylpiperidine N-oxide
(TEMPO) was purified by sublimation. All oximes used in this study
were prepared in our laboratory and purified by distillation or recrystalli-
zation from hydrocarbon solvents. When benzaldehyde oxime is referred
to, only the E isomer (separated from the Z isomer by column chroma-
tography) was used.
1376, 1296, 1112, 870 cmꢀ1
;
HRMS (ESI+): m/z calcd for
Geranial 1j and 2-cyclohexylideneacetaldehyde 1r used in the mechanis-
tic studies were prepared from corresponding allylic alcohols as described
by Koskinen and Kumpulainen.[34] The preparation and full characteriza-
tion data of 2-isoxazolines not reported here have been reported in the
preceding communications.[12,13] For the preparation and characterization
data of other compounds involved, see the Supporting Information.
[C10H17NO+H]: 168.1388; found: 168.1387.
Compound 5l: Prepared as described in the general procedure except
that 100 mol% of both starting materials were used. Starting materials:
a,b-unsaturated aldehyde (113 mg, 0.53 mmol, 106 mol%), 3-pentanone
oxime (55 mL, 0.50 mmol, 100 mol%). Reaction time: 4 h. Eluent: gradi-
ent 30–40% EtOAc in hexanes. Yield: 105 mg (85%), white solid; m.p.
67–688C; Rf =0.32 (60% EtOAc in hexanes); 1H NMR (400 MHz,
N-Methylanilinium diphenylphosphate (30): The catalyst was most con-
veniently used as a salt that was prepared from commercial diphenyl-
11336
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 11325 – 11339

Synthesis of 2-Isoxazolines
FULL PAPER
CDCl3): d=7.09 (t, J=1.8 Hz, 1H), 3.67 (td, J1 =4.5 Hz, J2 =13.6 Hz,
2H), 3.34 (ddd, J1 =3.3 Hz, J2 =10.2 Hz, J3 =13.6 Hz, 2H), 2.71 (d, J=
1.8 Hz, 2H), 1.78 (m, 2H), 1.61 (ddd, J1 =4.5 Hz, J2 =10.2 Hz, J3 =
13.5 Hz, 2H), 1.43 ppm (s, 9H); 13C NMR (100 MHz, CDCl3): d=154.6,
145.8, 81.8, 49.6, 45.8, 40.9, 35.5, 28.3 ppm; FTIR (film): n˜ =2976, 2934,
2871, 1694, 1603, 1473, 1454, 1422, 1366, 1269, 1246, 1234, 1176, 1149,
869 cmꢀ1; HRMS (ESI+): m/z calcd for [C12H20N2O3+Na]: 263.1372;
found: 263.1371.
(ddd, J1 =6.9 Hz, J2 =9.3 Hz, J3 =13.9 Hz, 1H), 2.62 (ddd, J1 =1.8 Hz, J2 =
7.8 Hz, J3 =17.4 Hz, 1H), 2.01 (dddd, J1 =5.6 Hz, J2 =7.8 Hz, J3 =9.3 Hz,
J4 =13.6 Hz, 1H), 1.83 ppm (dddd, J1 =5.2 Hz, J2 =6.9 Hz, J3 =9.5 Hz,
J4 =13.6 Hz, 1H). Analysis data corresponds to that published in the lit-
erature.[13] Enantiomeric purity was determined using HPLC analysis
with a Daicel Chiralcel OD column (25 cm) along with a precolumn
(5 cm). Eluent: 10% 2-propanol in hexanes; flow rate: 0.7 mLminꢀ1; l=
230 nm; major isomer: tr =19.03 min, minor isomer: tr =20.87 min. Con-
version of (R)-5c to a known b-hydroxynitrile allowed the determination
of the absolute configuration by chemical correlation.[13] The absolute ste-
reochemistry of other compounds was assigned by analogy.
Compound 5m: Prepared as described in the general procedure except
that 100 mol% of both starting materials were used. Starting materials:
a,b-unsaturated aldehyde (102 mg, 0.50 mmol, 100 mol%), 3-pentanone
oxime (55 mL, 0.50 mmol, 100 mol%). Reaction time: 14 h. Eluent: gradi-
ent 30–18% hexanes in CH2Cl2. Yield: 76 mg (69%), tan oil that solidi-
fies upon standing; m.p. 50–518C; Rf =0.5 (60% EtOAc in hexanes);
1H NMR (400 MHz, CDCl3): d=7.13–7.08 (m, 2H), 7.04 (t, J=1.8 Hz,
1H), 6.85–6.80 (m, 2H), 3.78 (s, 3H), 2.83 (dd, J1 =1.8 Hz, J2 =17.5 Hz,
1H), 2.69 (dd, J1 =1.8 Hz, J2 =17.5 Hz, 1H), 2.66–2.60 (m, 2H), 1.93 (m,
2H), 1.41 ppm (s, 3H); 13C NMR (100 MHz, CDCl3): d=157.8, 145.7,
133.65, 129.1, 113.8, 84.3, 55.2, 45.7, 42.1, 29.6, 25.5 ppm; FTIR (film): n˜ =
2969, 2932, 2836, 2059, 1884, 1611, 1583, 1513, 1459, 1440, 1376, 1301,
Compound (S)-5d: Prepared as described in the general procedure.
Starting materials: a,b-unsaturated aldehyde (323 mL, 1=1.09, 2.0 mmol,
100 mol%), acetone oxime (0.66 g, 9.0 mmol, 450 mol%). Reaction time:
5.2 h. Eluent: gradient: 20–50% EtOAc in hexanes. Yield: 229 mg
(60%), pale yellow oil; Rf =0.25 (50% EtOAc in hexanes); [a]D =
+109.0 (c=1, CH2Cl2, 95% ee); 1H NMR (CDCl3, 400 MHz): d=7.37–
7.27 (m, 5H), 7.12 (t, J=1.8 Hz, 1H), 4.71 (ddt, J1 =5.1 Hz, J2 =7.2 Hz,
J3 =10.8 Hz, 1H), 4.58 (s, 2H), 3.58 (dd, J1 =5.1 Hz, J2 =10.4 Hz, 1H),
3.52 (dd, J1 =5.0 Hz, J2 =10.4 Hz, 1H), 3.04 (ddd, J1 =1.8 Hz, J2 =
10.8 Hz, J3 =17.6 Hz, 1H), 2.90 ppm (ddd, J1 =1.8 Hz, J2 =7.2 Hz, J3 =
17.6 Hz, 1H). Analysis data corresponds to that published in the litera-
ture.[13] The enantiomeric purity was determined by HPLC analysis using
a Daicel Chiralpak IC column. Eluent: 10% 2-propanol in hexanes; flow
rate: 0.7 mLminꢀ1; l=230 nm; major isomer: tr =33.9 min, minor isomer:
tr =38.7 min.
1245, 1178, 1035, 869, 822 cmꢀ1
; HRMS (ESI+): m/z calcd for
[C13H17NO2+Na]: 242.1157; found: 242.1152.
Compound 5n: Compound 30 (36 mg, 0.10 mmol, 20 mol%) was added
to
a
solution of a,b-unsaturated aldehyde 1n[35] (85 mg, 0.60 mmol,
120 mol%) in toluene (2.5 mL) at room temperature. After the catalyst
had dissolved (2 min), the reaction mixture was cooled to 08C and 3-pen-
tanone oxime (55 mL, 0.50 mmol, 100 mol%) was added. After 15 h, the
reaction mixture was diluted with Et2O (15 mL) and the resulting solu-
tion was washed with saturated NaHCO3 (5 mL) and 5% aqueous oxalic
acid (2ꢃ5 mL). Both basic and acidic aqueous phases were back-extract-
ed with Et2O (10 mL, 2ꢃ5 mL, respectively) and again with EtOAc (2ꢃ
10 mL, both) to recover the relatively polar product from the aqueous
phase. The combined organic extracts were washed with brine (10 mL),
dried (Na2SO4), and concentrated to approximately 1 mL. Purification of
the residue by column chromatography (50–60% MTBE in hexanes) af-
forded product 7n (65 mg, 82%) as a colorless oil. Rf =0.21 (70%
MTBE in hexanes); 1H NMR (400 MHz, CDCl3): d=7.10 (t, J=1.8 Hz,
1H), 4.60–4.50 (m, 1H), 3.66 (s, 3H), 3.08 (ddd, J1 =1.8 Hz, J2 =10.5 Hz,
J3 =17.5 Hz, 1H), 2.62 (ddd, J1 =1.8 Hz, J2 =7.3 Hz, J3 =17.5 Hz, 1H),
2.52–2.39 (m, 2H), 1.88 ppm (m, 2H); 13C NMR (100 MHz, CDCl3): d=
173.3, 145.8, 77.2, 51.6, 40.5, 30.1, 29.9 ppm; FTIR (film): n˜ =3076, 2954,
1736, 1601, 1439, 1362, 1262, 1201, 1176, 845 cmꢀ1; HRMS (ESI+): m/z
calcd for [C7H11NO3+Na]: 180.0637; found: 180.0643.
Compound (R)-5e: Prepared as described in the general procedure.
Starting materials: a,b-unsaturated aldehyde (300 mL, 1=0.92, 2.0 mmol,
100 mol%), acetone oxime (0.66 g, 9.0 mmol, 450 mol%). Reaction time:
5.2 h. Eluent: gradient 20–25% EtOAc in hexanes. Yield: 203 mg (66%),
colorless oil, Rf =0.38 (50% MTBE in hexanes); [a]D =+124.3 (c=1.2,
CH2Cl2, 91% ee); 1H NMR (CDCl3, 400 MHz): d=7.08 (t, J=1.8 Hz,
1H), 4.23 (ddd, J1 =6.9 Hz, J2 =8.8 Hz, J3 =10.8 Hz, 1H), 2.92 (ddd, J1 =
1.8 Hz, J2 =10.8 Hz, J3 =17.5 Hz, 1H), 2.69 (ddd, J1 =1.8 Hz, J2 =8.8 Hz,
J3 =17.5 Hz, 1H), 1.90–1.83 (m, 1H), 1.79–1–70 (m, 2H), 1.69–1.56 (m,
2H), 1.47 (m, 1H), 1.30–1.09 (m, 3H), 1.08–0.91 ppm (m, 2H). Analysis
data corresponds to that published in the literature.[13] Enantiomeric
purity was determined by HPLC analysis using a Daicel Chiralpak AS
column (25 cm) along with a precolumn (5 cm). Eluent: 10% 2-propanol
in hexanes; flow rate: 0.7 mLminꢀ1
; l=230 nm; major isomer: tr =
19.8 min, minor isomer: tr =16.9 min.
Compound (R,E)-5h: Prepared as described in the general procedure.
Starting materials: a,b-unsaturated aldehyde (364 mg, 2.0 mmol,
100 mol%), acetone oxime (0.66 g, 9.0 mmol, 450 mol%). Reaction time:
5.3 h. Eluent: gradient 40–50% EtOAc in hexanes. Yield: 280 mg (71%),
colorless oil; Rf =0.23 (30% MTBE in hexanes); [a]D =+105.2 (c=1,
CH2Cl2, 94% ee); 1H NMR (CDCl3, 400 MHz): d=7.11 (t, J=1.8 Hz,
1H), 6.93 (td, J1 =7.0 Hz, J2 =15.6 Hz, 1H), 5.82 (td, J1 =1.6 Hz, J2 =
15.6 Hz, 1H), 4.51 (ddt, J1 =4.7 Hz, J2 =7.4 Hz, J3 =10.5 Hz, 1H), 3.73 (s,
3H), 3.05 (ddd, J1 =1.8 Hz, J2 =10.5 Hz, J3 =17.4 Hz, 1H), 2.60 (ddd, J1 =
1.8 Hz, J2 =7.9 Hz, J3 =17.4 Hz, 1H), 2.25 (m, 2H), 1.72–1.47 ppm (m,
4H). Analysis data corresponds to that published in the literature.[13]
Enantiomeric purity was determined by HPLC analysis using a Daicel
Chiralpak AS column (25 cm) along with a precolumn (5 cm). Eluent:
10% 2-propanol in hexanes; flow rate: 0.7 mLminꢀ1; l=220 nm; major
isomer: tr =75.0 min, minor isomer: tr =61.5 min.
General synthetic procedure for the enantioselective synthesis of 2-isoxa-
zolines:[20] (S)-a,a-[Bis(3,5-bistrifluoromethyl)phenyl]-2-pyrrolidine meth-
anol trimethylsilyl ether (0.20 mmol, 10 mol%, purified by column chro-
matography) and benzoic acid (0.20 mmol, 10 mol%) were dissolved in
toluene (1 mL) at room temperature. a,b-Unsaturated aldehyde
(2.0 mmol, 100 mol%) was added to this solution at 08C (an ice–water
bath was used), and after 1 min, acetone oxime (9.0 mmol, 450 mol%)
was then added. After stirring the reaction mixture for the indicated
period of time, a precooled (08C) freshly prepared solution of concen-
trated H2SO4 (0.84 mL, 15.9 mmol, 800 mol%), methanol (3.8 mL), and
water (0.2 mL) was added and the resulting solution was stirred vigorous-
ly for 15–20 min at 08C. Saturated aqueous NaHCO3 (5 mL) was added
and the resulting mixture was extracted with EtOAc (2ꢃ15 mL). The
combined organic extracts were washed with saturated NaHCO3
(10 mL), dried (Na2SO4), and concentrated. The residue was purified by
column chromatography (EtOAc/hexanes).
Compound (R)-5j: Prepared as described in the general procedure. Start-
ing materials: a,b-unsaturated aldehyde (304 mg, 2.0 mmol, 100 mol%),
acetone oxime (0.66 g, 9.0 mmol, 450 mol%). Reaction time 5.2 h.
Eluent: gradient: 10–25% EtOAc in hexanes. Yield: 195 mg (58%).
[a]D =+47.9 (c=1, CH2Cl2, 69% ee) All other spectral data were identi-
cal to those of the racemic product. The enantiomeric purity was deter-
mined by HPLC analysis using a Daicel Chiralcel OD column (25 cm)
along with a precolumn (5 cm). Eluent: 1% 2-propanol in hexanes; flow
rate: 0.7 mLminꢀ1; l=220 nm; major isomer: tr =14.5 min, minor isomer:
tr =16.1 min.
Compound (R)-5c: Prepared as described in the general procedure.
Starting materials: a,b-unsaturated aldehyde (320 mL, 1=1.0, 2.0 mmol,
100 mol%), acetone oxime (0.66 g, 9.0 mmol, 450 mol%). Reaction time:
5.2 h. Eluent: gradient: 20–25% EtOAc in hexanes. Yield: 253 mg
(72%), colorless oil; Rf =0.35 (40% EtOAc in hexanes); [a]D =+138.9
(c=1, CH2Cl2, 94% ee); 1H NMR (CDCl3, 400 MHz): d=7.31–7.27 (m,
2H), 7.21–7.18 (m, 3H), 7.12 (t, J=1.8 Hz, 1H), 4.52 (ddt, J1 =5.2 Hz,
J2 =7.8 Hz, J3 =10.5 Hz, 1H), 3.04 (ddd, J1 =1.8 Hz, J2 =10.5 Hz, J3 =
17.4 Hz, 1H), 2.80 (ddd, J1 =5.6 Hz, J2 =9.5 Hz, J3 =13.9 Hz, 1H), 2.72
Compound (R)-5i: Prepared as described in the general procedure
except that the following cyclization procedure and workup were con-
Chem. Eur. J. 2010, 16, 11325 – 11339
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
11337

P. M. Pihko et al.
ducted. After the indicated reaction time, a cooled solution of 10m HCl
(0.8 mL) in THF (3 mL)/H2O (1 mL) was added to the reaction mixture.
The resulting solution was stirred for 20 min and diluted with H2O
(8 mL). The mixture was extracted with EtOAc (2ꢃ20 mL). Combined
organic layers were washed with saturated NaHCO3 (10 mL) and brine
(10 mL), dried (Na2SO4), and concentrated. The residue was purified by
column chromatography. Starting materials: a,b-unsaturated aldehyde
(733 mg, 2.0 mmol, 100 mol%), acetone oxime (0.66 g, 9.0 mmol,
450 mol%). Reaction time: 5.2 h. Eluent: gradient 15–25% EtOAc in
hexanes. Yield: 530 mg (69%), colorless oil; Rf =0.23 (40% MTBE in
hexanes); [a]D =+50.7 (c=1, CH2Cl2, 94% ee); 1H NMR (CDCl3,
400 MHz): d=7.70–7–64 (m, 4H), 7.46–7.35 (m, 6H), 7.10 (t, J=1.8 Hz,
1H), 4.60–4.40 (m, 1H), 3.67 (t, J=6.3 Hz, 2H), 3.01 (ddd, J1 =1.8 Hz,
J2 =10.5 Hz, J3 =17.3 Hz, 1H), 2.58 (ddd, J1 =1.8 Hz, J2 =8.0 Hz, J3 =
17.3 Hz, 1H), 1.74–1.38 (m, 6H), 1.05 ppm (s, 9H). Analysis data corre-
sponds to that published in the literature.[13] Enantiomeric purity was de-
termined by HPLC analysis using a Daicel Chiralpak AS column (25 cm)
along with a precolumn (5 cm). Eluent: 2% 2-propanol in hexanes; flow
rate: 0.8 mLminꢀ1; l=220 nm; major isomer: tr =16.0 min, minor isomer:
tr =14.0 min.
[2] a) B. List, R. A. Lerner, C. F. Barbas III, J. Am. Chem. Soc. 2000,
122, 2395–2396; b) N. A. Paras, D. W. C. MacMillan, J. Am. Chem.
Soc. 2001, 123, 4370–4371.
[3] M. Marigo, J. Franzen, T. Poulsen, W. Zhuang, K. A. Jørgensen, J.
Am. Chem. Soc. 2005, 127, 6964–6965.
[4] a) P. B. Kisanga, P. Ilankumaran, B. M. Fetterly, J. G. Verkade, J.
Org. Chem. 2002, 67, 3555–3560; b) I. C. Stewart, R. G. Bergman,
F. D. Toste, J. Am. Chem. Soc. 2003, 125, 8696–8697; c) T. C. Wab-
nitz, J. P. Spencer, Org. Lett. 2003, 5, 2141–2144; d) X. Xiong, C.
Ovens, A. W. Pilling, J. W. Ward, D. J. Dixon, Org. Lett. 2008, 10,
565–567.
[5] a) C. D. Vanderwal, E. N. Jacobsen, J. Am. Chem. Soc. 2004, 126,
14724–14725; b) A. Carlone, G. Bartoli, M. Bosco, F. Pesciaioli, P.
Ricci, L. Sambri, P. Melchiorre, Eur. J. Org. Chem. 2007, 5492–
5495; c) C. M. Reisinger, X. Wang, B. List, Angew. Chem. 2008, 120,
8232–8235; Angew. Chem. Int. Ed. 2008, 47, 8112–8115; d) X. Lu,
B. Sun, B. Cindric, L. Deng, J. Am. Chem. Soc. 2008, 130, 8134–
8135.
[6] For a recent review of oxy-conjugate additions: C. F. Nising, S.
Brꢀse, Chem. Soc. Rev. 2008, 37, 1218–1228.
Compound (S)-5n: Prepared as described in the general procedure at a
0.5 mmol scale with the R enantiomer of the catalyst. Starting materials:
a,b-unsaturated aldehyde (68 mL, 1=1.05, 0.50 mmol, 100 mol%), ace-
tone oxime (0.110 g, 1.50 mmol, 300 mol%). Reaction time: 4 h. Eluent:
gradient: 50–60% MTBE in hexanes. Yield: 55 mg (70%). [a]D =ꢀ129.2
(c=1, CH2Cl2, 90% ee). All other spectral data were identical to those of
the racemic product. The enantiomeric purity was determined by HPLC
analysis using a Daicel Chiralcel AD column (25 cm) along with a precol-
[7] W. Langenbeck, R. Sauerbier, Chem. Ber. 1937, 70, 1540–1541.
[8] a) S. E. Denmark, T. J. Bui, J. Org. Chem. 2005, 70, 10190–10193;
b) Y. Hayashi, T. Itoh, S. Aratake, H. Ishikawa, Angew. Chem. 2008,
120, 2112–2114; Angew. Chem. Int. Ed. 2008, 47, 2082–2084;
c) M. B. Boxer, H. Yamamoto, J. Am. Chem. Soc. 2007, 129, 2762–
2763; d) B. J. Albert, H. Yamamoto, Angew. Chem. 2010, 122, 2807–
2809; Angew. Chem. Int. Ed. 2010, 49, 2747–2749.
[9] a) T. Kano, Y. Tanaka, K. Maruoka, Tetrahedron 2007, 63, 8658–
8664. For a racemic version of the reaction, see: T. Kano, Y. Tanaka,
K. Maruoka, Tetrahedron Lett. 2006, 47, 3039–3041.
umn (5 cm). Eluent: 1% 2-propanol in hexanes; flow rate: 0.8 mLminꢀ1
;
l=230 nm; major isomer: tr =74.4 min, minor isomer: tr =63.2 min.
Compound (S)-5p: Prepared as described in the general procedure with
the R enantiomer of the catalyst. Starting materials: a,b-unsaturated al-
dehyde (647 mg, 2.0 mmol, 100 mol%), acetone oxime (0.66 g, 9.0 mmol,
450 mol%). Reaction time: 5.3 h. Eluent: gradient 30–50% EtOAc in
hexanes. Yield: 375 mg (55%), pale yellow oil; Rf =0.17 (70% MTBE in
[10] S. Bertelsen, P. Diner, R. L. Johansen, K. A. Jørgensen, J. Am.
Chem. Soc. 2007, 129, 1536–1537.
[11] a) T. Govender, L. Hojabri, F. M. Moghaddam, P. I. Arvidsson, Tet-
rahedron: Asymmetry 2006, 17, 1763–1767; b) H. Li, J. Wang, T. E-
Nunu, L. Zu, W. Jiang, S. Wei, W. Wang, Chem. Commun. 2007,
507–509; c) H. Sundꢄn, I. Ibrahem, G.-L. Zhao, L. Eriksson, A.
Cꢅrdova, Chem. Eur. J. 2007, 13, 574–581; d) E. Reyes, G. Talavera,
H. L. Vicario, D. Badꢆa, L Carillo, Angew. Chem. 2009, 121, 5811–
5814; Angew. Chem. Int. Ed. 2009, 48, 5701–5704; e) B. Tan, Z. Shi,
P. J. Chua, Y. Li, G. Zhong, Angew. Chem. 2009, 121, 772–775;
Angew. Chem. Int. Ed. 2009, 48, 758–761; f) F.-L. Zhang, A.-W. Xu,
Y.-F. Gong, M.-H. Wei, X.-L. Yang, Chem. Eur. J. 2009, 15, 6815–
6818.
1
hexanes); [a]D =ꢀ63.1 (c=1, CH2Cl2, 92% ee); H NMR (858C, C2D2Cl4,
400 MHz): d=7.41–7.21 (m, 10H), 7.05 (t, J=1.7 Hz, 1H), 5.22 (s, 2H),
4.54 (ABq, JA,B =15.5 Hz, Dn=24.7 Hz), 4.45 (m, 1H), 3.49–3.33 (m,
1H), 2.97 (dd, J1 =10.4 Hz, J2 =17.2 Hz, 1H), 2.55 (dd, J1 =7.6 Hz, J2 =
17.2 Hz, 1H), 1.85 ppm (m, 2H). Analysis data corresponds to that pub-
lished in the literature.[13] Enantiomeric purity was determined by HPLC
analysis using a Daicel Chiralpak AS column (25 cm) along with a pre-
column (5 cm). Eluent: 15% 2-propanol in hexanes; flow rate:
0.7 mLminꢀ1; l=220 nm; major isomer: tr =57.6 min, minor isomer: tr =
49.3 min.
[12] A. Pohjakallio, P. M. Pihko, Synlett 2008, 827–830.
[13] A. Pohjakallio, P. M. Pihko, Chem. Eur. J. 2009, 15, 3960–3964.
[14] a) R. Huisgen, M. Christl, Chem. Ber. 1973, 106, 3291–3311; b) K.
Torssell, O. Zeuthen, Acta Chim. Scand. B 1978, 32, 118–124;
c) S. C. Sharma, K. Torssell, Acta Chem. Scand. B 1979, 33, 379–
383; d) N. B. Das, K. B. G. Torssell, Tetrahedron 1983, 39, 2247–
2253; e) R. Jacquier, J.-L. Olive, C. Petrus, F. Petrus, Tetrahedron
Lett. 1975, 16, 2337–2340; f) L. Pennicott, S. Lindell, Synlett 2006,
463–465.
[15] Other asymmetric approaches to 2-isoxazolines: a) Y. Ukaji, K.
Sada, K. Inomata, Chem. Lett. 1993, 1847; b) M. Shimizu, Y. Ukaji,
K. Inomata, Chem. Lett. 1996, 455–456; c) J. D. Toker, P. Went-
worth Jr. , Y. Hu, K. N. Houk, K. D. Janda, J. Am. Chem. Soc. 2000,
122, 3244–3245; d) M. Tsuji, Y. Ukaji, K. Inomata, Chem. Lett.
2002, 1112–1113; e) M. P. Sibi, K. Itoh, C. P. Jasperse, J. Am. Chem.
Soc. 2004, 126, 5366–5367; f) H. Suga, Y. Adachi, K. Fujimoto, Y.
Furihata, T. Tsuchida, A. Kakehi, T. Baba, J. Org. Chem. 2009, 74,
1099–1113; g) Y. Brinkmann, R. J. Madhushaw, R. Jazzar, G. Ber-
nardinelli, E. P. Kꢇndig, Tetrahedron 2007, 63, 8413–8419.
[16] For a review, see: E. Buncel, I.-K. Um, Tetrahedron 2004, 60, 7801–
7825.
Acknowledgements
TEKES (National Agency for Technology and Innovation), the Graduate
School of Organic Chemistry and Chemical Biology, the Finnish Cultural
Foundation, and the Academy of Finland (project nos. 123813 and
217224) are acknowledged for material and financial support. We thank
Dr. Jari Koivisto for NMR spectroscopy assistance and Essi Karppanen,
Laura Kolsi, Dr. Hannes Helmboldt, and Dr. Esa Kumpulainen for their
help in preparation of aldehyde substrates. Dr. Esa Kumpulainen and Dr.
Meryem Benohoud are also acknowledged for their assistance with GC
and HPLC analysis, respectively.
[1] a) A. Berkessel, H. Grçger, Asymmetric Organocatalysis: From Bio-
mimetic Concepts to Application in Asymmetric Synthesis, Wiley-
VCH, Weinheim, 2005; b) G. Lelais, D. W. C. MacMillan, Aldrichim.
Acta 2006, 39, 79–87; c) A. Erkkilꢀ, I. Majander, P. M. Pihko, Chem.
Rev. 2007, 107, 5416–5570; d) S. Mukherjee, J. W. Yang, S. Hoff-
mann, B. List, Chem. Rev. 2007, 107, 5471–5569.
[17] We could not detect N-hydroxypyrroles as products under these re-
action conditions. See reference [11e].
[18] trans-Hex-2-enal was chosen as the test substrate because its reactiv-
ity resembles the reactivity of other enals better than crotonalde-
11338
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 11325 – 11339

Synthesis of 2-Isoxazolines
FULL PAPER
hyde, which was found exceptionally reactive in both the desired
conjugate addition and unwanted side reactions.
4291–4293; c) N. Zotova, A. Moran, A. Armstrong, D. G. Black-
mond, Adv. Synth. Catal. 2009, 351, 2765–2769.
[19] a) J. W. Yang, M. T. Hechavarria Fonseca, B. List, Angew. Chem.
2004, 116, 6829–6832; Angew. Chem. Int. Ed. 2004, 43, 6660–6662;
b) A. Erkkilꢀ, P. M. Pihko, M.-R. Clarke, Adv. Synth. Catal. 2007,
349, 802–804.
[29] The substituent effects were studied based on the initial rate data
for the aldehyde and 2-isoxazoline products. See the Supporting In-
formation.
[30] The formation of 2-isoxazolines cannot be observed with p-NO2-
benzaldehyde oxime, within the timeframe of the experiment. Al-
though these experiments cannot measure the equilibrium concen-
trations of the actual substrates, it should be noted that the concen-
trations of the intermediates with 3-pentanone and cyclohexanone
oximes, viable substrates for the cyclization reaction, do follow the
trend [6c]>[6m] or [6c]>[6j] during the reaction (see the Support-
ing Information).
[31] H. C. Brown, J. H. Brewster, H. Shechter, J. Am. Chem. Soc. 1954,
76, 467–474.
[32] Computational studies to address the full mechanism have been ini-
tiated and will be reported separately.
[33] We have tried to isolate possible reaction intermediates that have
sufficient lifetimes by silylation experiments, yet so far without suc-
cess. In an attempt to isolate reaction intermediates, silylation with
tert-butyldimethylsilyl triflate (TBSOTf) afforded the silylated dia-
stereomeric oxime hemiacetals derived from the conjugate-addition
product (39/40, see the Supporting Information). These hemiacetals
can also be detected by NMR spectroscopy during the reaction.
[34] E. T. T. Kumpulainen, A. M. P. Koskinen, Chem. Eur. J. 2009, 15,
10901–10911.
[35] Prepared by Wittig olefination of commercially available aldehydes
with (1,3-dioxolan-2-yl-methyl)triphenylphosphonium bromide. See:
R. M. de Figueiredo, R. Berner, J. Julis, T. Liu, D. Tꢇrp, M. Christ-
mann, J. Org. Chem. 2007, 72, 640–642 and reference [13] for exam-
ples.
[20] See the Supporting Information for details.
[21] Although the size of the nucleophile was considerably larger, the
rates of the conjugate addition were similar in both cases.
[22] See the comment in the Supporting Information.
[23] The main difficulty is the purification of the products from excess
benzaldehyde oxime.
[24] See the Supporting Information for the catalyst/co-catalyst screen.
[25] a) G. M. Shutske, J. Org. Chem. 1984, 49, 180–183; b) S. D. Lepore,
M. R. Wiley, J. Org. Chem. 2000, 65, 2924–2932; c) S. D. Lepore,
M. R. Wiley, Org. Lett. 2003, 5, 7–10.
[26] a) J. M. Sayer, P. J. Conlon, J. Am. Chem. Soc. 1980, 102, 3592–
3600; b) R. A. More OꢂFerrall, D. M. OꢂBrien, G. Murphy, Can. J.
Chem. 2000, 78, 1594–1612; c) R. A. More OꢂFerrall, D. OꢂBrien, J.
Phys. Org. Chem. 2004, 17, 631–640.
[27] a) W. P. Jencks, J. Am. Chem. Soc. 1959, 81, 475–481; b) S. Rosen-
berg, S. M. Silver, J. M. Sayer, W. P. Jencks, J. Am. Chem. Soc. 1974,
96, 7986–7998; c) J. M. Sayer, B. Pinsky, A. Schonbrunn, W. Wash-
tien, J. Am. Chem. Soc. 1974, 96, 7998–8009.
[28] If the formation of iminium ions or enamines under typical process-
es catalyzed by an ammonium salt would limit the rates, one would
expect the overall rates of these processes to be independent of the
nucleophiles or electrophiles used. With highly reactive nucleophiles
or electrophiles, the enamine-/iminium-formation reactions can
become rate-limiting, see: a) N. Halland, M. A. Lie, A. Kjærsgaard,
M. Marigo, B. Schiøtt, K. A. Jørgensen, Chem. Eur. J. 2005, 11,
7083–7090; b) S. P. Mathew, M. Klussmann, H. Iwamura, D. H.
Wells, A. Armstrong, D. G. Blackmond, Chem. Commun. 2006,
Received: April 6, 2010
Published online: August 19, 2010
Chem. Eur. J. 2010, 16, 11325 – 11339
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
11339