Enantioselective synthesis of 2-isoxazolines by a One-flask conjugate addition/oxime-transfer process

Article abstract of DOI:10.1002/chem.200802684

A study was conducted to demonstrate a new catalytic enantioselective method to prepare 5-substituted 2-isoxazolines that applied a small organic catalyst as the source of chirality and enabled the use of different α-β-unsaturated aldehydes as starting materials. The aldehyde and acetone oxime were added to a solution of amine and benzoic acid in toluene at 0 °C. A precooled solution of H₂SO₄ in MeOH was added and the resulting solution was stirred for 15 minutes at 0 °C after stirring the reaction mixture at 0 °C for the indicated period of time. A saturated aqueous solution of NaHCO₃ was added and the mixture was extracted with EtOAc. The combined organic extracts were dried and concentrated, while the residue was purified by flash chromatography to afford the products. The results show that the presence of a strong acid in the reaction mixture induced the formation of similar catalytic species in the reactions.

Full text of DOI:10.1002/chem.200802684

DOI: 10.1002/chem.200802684
Enantioselective Synthesis of 2-Isoxazolines by a One-Flask Conjugate
Addition/Oxime-Transfer Process
Antti Pohjakallio[b] and Petri M. Pihko*[a]
Asymmetric access to 2-isoxazolines, facile precursors to
both aldol- and 1,3-aminoalcohol-type structural subunits, is
an important objective in synthetic methodology.[1] The 1,3-
dipolar cycloaddition between nitrile oxides and alkenes is
by far the most popular method for the synthesis of the 2-
isoxazoline ring. Unfortunately, these cycloadditions are
generally not directly amenable to catalytic asymmetric syn-
thesis.[1] Indeed, the number of catalytic methods used to
prepare the 2-isoxazoline ring system is considerably small-
er, and there seems to be no general solution to the prob-
lem. Although good selectivities have been attained by
using Lewis acids such as ZnII–tartrate complexes[2] or MgII–
bisoxazolines,[3] the substrate scope has been limited owing
to the necessity of having a second metal coordination site
on the alkene substrate. Recently, Kꢀndig and co-workers
reported a ruthenium catalyst that appeared to overcome
this limitation, but the substrate scope was still very restrict-
ed.[4] In addition to cycloadditions, a catalytic asymmetric
cyclization of b,g-unsaturated oximes has been reported,[5]
but only modest selectivities have been attained with this
method so far. Herein, we report a new catalytic enantiose-
lective method to prepare 5-substituted 2-isoxazolines that
applies a small organic catalyst as the source of chirality and
enables the use of different a,b-unsaturated aldehydes as
starting materials.
Scheme 1. Mechanistic hypothesis for the racemic formation of 2-isoxazo-
lines. For previous work see reference [6].
Recently, we disclosed a reaction for the preparation of 3-
unsubstituted 2-isoxazolines in which aniline salts such as 3
(Scheme 1) catalyzed the condensation of a,b-unsaturated
aldehydes and oximes to produce these rare heterocycles.[6]
Access to these ring systems through, for example, 1,3-dipo-
lar cycloadditions, is difficult because fulminic acid[7] or pre-
[a] Prof. P. M. Pihko
generated silyl nitronates[8] are required as the dipolarophile
Department of Chemistry
components.
University of Jyvꢁskylꢁ, 40014 JYU (Finland)
Fax : (+358)14-260-2501
We first hypothesized that this reaction proceeded
through an oxime conjugate addition/transoximation cas-
cade and that the catalyst would accelerate both steps in
this sequence. Support for this order of events was obtained
from 1H NMR spectroscopy experiments. When a mixture
of crotonaldehyde (1a) and anisaldehyde oxime (7) was
treated with the trifluoroacetic acid (TFA) salt of 5, the con-
E-mail: Petri.Pihko@jyu.fi
[b] A. Pohjakallio
Department of Chemistry
Helsinki University of Technology
Kemistintie 1, 02150 Espoo (Finland)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200802684.
3960
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 3960 – 3964

COMMUNICATION
1
version of the aldehyde to 2-isoxazoline (4a) appeared to be
preceded by the formation of the conjugate addition inter-
mediate 8 (Figure 1). Our double iminium catalytic hypothe-
tected in the H NMR spectra. In these cases, it is unlikely
that the complete lack of enantioselectivity could be ex-
plained by the difference in the reaction rates. At this point,
we were uncertain whether our
initial hypothesis, the double
iminium activation mechanism,
was operating at all in this reac-
tion.
Interestingly, when a mixture
of aldehyde 1b and oxime 10
was treated with
a catalytic
amount of N-methylaniline, no
reaction could be observed. In-
stead, upon addition of differ-
ent acid catalysts to the reac-
tion mixture, the 2-isoxazoline
product was generated accord-
ing to the plots shown in
Figure 2. The sigmoidal shape
of the conversion plot implies
that a new, potent catalyst is
generated during an induction
period. Both the length of this
induction period and the poten-
cy of the catalytic species
depend on the acid strength,
and a positive correlation be-
Figure 1. Conversion of 1a to 2-isoxazoline in CDCl3 in the presence of the TFA salt of 5. Ar=p-methoxy-
~
&
*
phenyl ( : 1a, : intermediate, and : product).
sis was further supported by Jørgensen, who reported an
iminium-catalyzed asymmetric conjugate addition of oximes
to aldehydes by using catalyst 6,[9] and by Dawson and
Jencks, who described the use of aniline catalysis in oxime
ligation and semicarbazone formation reactions.[10]
However, to our disappointment, we were not able to pro-
duce optically active 2-isoxazolines by using our method and
chiral catalysts 5 or 6.[11] With catalyst 5, only racemic 2-is-
ACHTUNGTRENNUNGoxazolines were obtained when moderately strong acids
such as TFA or diphenylphosphate (DPP) were used as co-
catalysts. In contrast, catalyst 6 was considerably less effec-
tive for this transformation in combination with TFA or
DPP. With weaker acids such as chloroacetic acid as cocata-
lysts, compound 5 and a number of different N-alkylanilines
were able to catalyze the first step of the reaction, the con-
jugate addition reaction of an oxime to the unsaturated al-
dehyde, but the rate of cyclization was retarded.
These results, however, did not explain why the 2-isoxazo-
lines obtained with 5·TFA were almost completely racemic.
We hypothesized that any initially generated enantiomeric
excess (ee) might be lost because the first step is likely to be
reversible and the second step, the cyclization step, is rela-
tively slow compared with the conjugate addition. In this
case, thermodynamic equilibrium is reached with regard to
the first step before the cyclization occurs, and thus, the
product is obtained as a racemic mixture.[12] However, with
aliphatic oximes, the rate of cyclization is faster, and in
some cases no conjugate addition intermediate can be de-
*
Figure 2. Acid-induced catalysis of 2-isoxazoline formation ( : trichloro-
~
&
acetic acid, : diphenyl phosphate, : TFA, &: bis(nitrophenyl)phosphate,
: methanesulfonic acid, : p-TsOH, ꢃ: dichloroacetic acid, and
~
:
^
*
chloroacetic acid).
tween reaction rate and the acid strength was observed.[13]
These results suggested that the presence of a strong acid in
the reaction mixture also induced the formation of similar
catalytic species in the reactions catalyzed by 5, and that
Chem. Eur. J. 2009, 15, 3960 – 3964
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3961

P. M. Pihko and A. Pohjakallio
this unknown species was probably responsible for the pro-
duction of the racemic products.
lines. In the first stage, a mild, enantioselective catalyst
should be used to promote the conjugate addition step with
a hydrolytically labile oxime unit. In the second stage, a
rapid quench with a strong acid stops the conjugate addition
and promotes fast cyclization to the 2-isoxazolines.[15] For
the first stage, we used a modification of the Jørgensen
oxime conjugate addition protocol.[9] Pleasingly, this hypo-
thesis was proven experimentally. Thus, treatment of alde-
hydes with acetone oxime, together with catalytic amounts
of 6 and PhCOOH in toluene at 08C for 3.5 h, and subse-
quent brief treatment of the reaction mixture with either
2.3m H2SO4 in MeOH or a mixture of aqueous HCl in THF
at 08C afforded the desired isoxazolines in moderate yields
and high enantioselectivities (Table 1).[16,17]
Although the identity of the intervening catalyst remains
unknown, the acid experiment also raised the question of
whether the 2-isoxazoline formation was actually dependent
on the acidity of the medium. Accordingly, we analyzed the
effect of different catalysts on the cyclization step. To gener-
ate the delicately unstable conjugate addition intermediates,
we used the commercially available, polystyrene-bound N-
benzylaniline base along with chloroacetic acid as the cata-
lyst. This system enabled us to remove the catalyst compo-
nents together with the majority of the oxime starting mate-
rial from the reaction mixture by simple filtration through a
pad of basic alumina. Although only approximately 25%
conversion to adduct 13 was obtained, it could be reliably
cyclized to the corresponding isoxazoline 4b with different
acid catalysts. Strong acids such as DPP displayed the fastest
rates for the cyclization step (Figure 3).
The isolated yields of this process appear to be limited by
the competing decomposition of the starting materials in the
process, as well as the Keq value of the initial addition
step.[17] Thus, further conjugation in the aldehyde completely
suppresses the reaction, and no
reaction was observed with cin-
namaldheyde. Steric bulk also
slightly
lowers
the
yield
(Table 1, entry 3). Aside from
the previous examples, the sub-
strate scope seems to be fairly
broad and, generally, all alde-
hydes that withstand the cycli-
zation conditions are viable
substrates. Aldehydes that have
potentially acid-labile PMB or
Cbz groups (Table 1, entries 5,
8, and 9) and unsaturated esters
(Table 1, entry 7) are readily
tolerated. The use of acetone
oxime facilitates rapid cycliza-
tion as well as easy removal of
the ketone byproduct, and re-
sults in high ee values. The al-
dehyde oxidation level is pre-
served during the process, in
contrast with previously report-
ed approaches to asymmetric
conjugate addition of oxygen
nucleophiles.[9a,19]
1
Figure 3. Conversion of intermediate 13 to 4b in the presence of different acid catalysts. H NMR spectroscop-
ic yields are based on the initial concentration of intermediate 13. The acid-induced formation of 13 from the
remaining starting materials 1b and 2 is responsible for conversions that are higher than 100%. The yields
were determined by H MNR spectroscopy ( : p-TsOH, ꢃ: methanesulfonic acid, : diphenylphosphate,
1
In conclusion, after the analy-
sis of the factors responsible for
catalysis of each step of our
oxime-transfer-based synthesis
~
*
^
:
&
*
TFA, : N-methylanilinium DPP, and : dichloroacetic acid).
Importantly, salt 3 turned out to be a less effective cata-
lyst for the cyclization step than any of the acids tested,
whereas N-methylaniline alone was totally inactive. The suc-
cess of the aniline salts as a catalyst in the reaction can be
accounted for by the relatively high acidity of these salts
combined with their ability to catalyze the conjugate addi-
tion step.[14]
of 2-isoxazolines, we have developed a general catalytic pro-
tocol for the asymmetric synthesis of 2-isoxazolines by using
the Jørgensen catalyst (6) as the source of enantioselectivity.
The process considerably extends the usefulness of the origi-
nal racemic process and creates a catalytic entry to a class
of compounds that has not been easily available before. Fur-
ther studies to explore the limits of the process and the
chemistry of the products are underway.
Taken together, these results suggested a two-stage ap-
proach to the catalytic asymmetric synthesis of 2-isoxazo-
3962
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 3960 – 3964

Enantioselective Synthesis of 2-Isoxazolines
COMMUNICATION
Table 1. Substrate screen for the asymmetric formation of 2-isoxazolines.
edged for material and financial support. We thank Dr. Jari Koivisto for
1H NMR spectroscopy assistance and Essi Karppanen, Laura Kolsi,
Hannes Helmboldt, and Esa Kumpulainen for the help in the preparation
of the aldehyde substrates. Esa Kumpulainen is also acknowledged for
help with GC analysis.
Keywords: amines · asymmetric catalysis · cyclization ·
enantioselectivity · isoxazolines
Entry
1
R
Yield [%][a]
39[b]
ee [%]
Me
86
1a
2
3
63
45
91
94
[1] For asymmetric cycloaddition reactions of 2-isoxazolines, see:
a) K. V. Gothelf, K. A. Jørgensen, Chem. Rev. 1998, 98, 863–909;
b) A. B¼doiu, Y. Brinkmann, F. Viton, E. P. Kꢀndig, Pure Appl.
Chem. 2008, 80, 1013–1018; for selected examples of 2-isoxazolines
for the construction of aldol and 1,3-amino alcohol functionalities,
see: c) D. P. Curran, Adv. Cycloaddit. 1998, 1, 129–189; d) A. P. Ko-
zikowski, Acc. Chem. Res. 1984, 17, 410; e) S. Kanemasa, O. Tsuge,
Heterocycles 1990, 30, 719–736; f) J. W. Bode, E. M. Carreira, J.
Org. Chem. 2001, 66, 6410–6424; g) J. W. Bode, N. Fraefel, D. Muri,
E. M. Carreira, Angew. Chem. 2001, 113, 2128–2131; Angew. Chem.
Int. Ed. 2001, 40, 2082–2085; for the use of isoxazolines to generate
b-amino acids, see: h) A. A. Fuller, B. Chen, A. R. Minter, A. K.
Mapp, J. Am. Chem. Soc. 2005, 127, 5376–5383.
[2] a) Y. Ukaji, K. Sada, K. Inomata, Chem. Lett. 1993, 1847; b) M. Shi-
mizu, Y. Ukaji, K. Inomata, Chem. Lett. 1996, 455–456; c) M. Tsuji,
Y. Ukaji, K. Inomata, Chem. Lett. 2002, 1112–1113.
[3] M. P. Sibi, K. Itoh, C. P. Jasperse, J. Am. Chem. Soc. 2004, 126,
5366–5367; for a recent NiII-binaphtyldiimine catalyst, see: H. Suga,
Y. Adachi, K. Fujimoto, Y. Furihata, T. Tsuchida, A. Kakehi, T.
Baba, J. Org. Chem. 2009, 74, 1099–1113.
4
54
52
91
90
5[c]
6[d]
57
47
92
88
7
8[e]
55
59
86
91
[4] Y. Brinkmann, J. M. Reniguntala, R. Jazzar, G. Bernardinelli, E. P.
Kꢀndig, Tetrahedron 2007, 63, 8413–8419.
[5] A. L. Norman, M. D. Mosher, Tetrahedron Lett. 2008, 49, 4153–
4155.
9[e]
[6] A. Pohjakallio, P. M. Pihko, Synlett 2008, 827–830.
[7] R. Huisgen, M. Christl, Chem. Ber. 1973, 106, 3291–3311.
[8] a) K. Torssell, O. Zeuthen, Acta Chem. Scand. Ser. B 1978, 32, 118–
124; b) S. C. Sharma, K. Torssell, Acta Chem. Scand. Ser. B 1979, 33,
379–383; c) N. B. Das, K. B. G. Torssell, Tetrahedron 1983, 39, 2247–
2253.
[a] Yields refer to pure, isolated products. [b] Yield determined by
1H NMR spectroscopy after 3 h.[18] [c] PMP=p-methoxybenzyl.
[d] TBDPS=tert-butyldiphenylsilyl. [e] Cbz=carbobenzyloxy
[9] a) S. Bertelsen, P. Diner, R. L. Johansen, K. A. Jørgensen, J. Am.
Chem. Soc. 2007, 129, 1536–1537; b) for a review of iminium cataly-
sis, see: A. Erkkilꢁ, I. Majander, P. M. Pihko, Chem. Rev. 2007, 107,
5416–5470.
Experimental Section
General procedure for the asymmetric synthesis of 2-isoxazolines with
the MeOH/H2SO4 cyclization conditions: The aldehyde (0.5 mmol,
100 mol%) and acetone oxime (110 mg, 1.5 mmol, 300 mol%) were
added to a solution of amine 6 (29.9 mg, 0.05 mmol, 10 mol%) and ben-
zoic acid (6.1 mg, 0.05 mmol, 10 mol%) in toluene (0.25 mL) at 08C.
After stirring the reaction mixture at 08C for the indicated period of
time, a precooled (08C) solution of H2SO4 (0.21 mL) in MeOH (1 mL,
2.3m H2SO4) was added and the resulting solution was stirred for 15 min
at 08C. A saturated aqueous solution of NaHCO3 (1 mL) was added and
the mixture was extracted with EtOAc (2ꢃ5 mL). The layers were sepa-
rated and, if necessary (as indicated by TLC), the aqueous layer was
completely basified by the addition of a saturated aqueous solution of
NaHCO3, then back-extracted with EtOAc. The combined organic ex-
tracts were dried (Na2SO4), and concentrated. The residue was purified
by flash chromatography to afford the products.
[10] a) A. Dirksen, T. M. Hackeng, P. E. Dawson, Angew. Chem. 2006,
118, 7743–7746; Angew. Chem. Int. Ed. 2006, 45, 7581–7584; b) for
a pioneering report of aniline-catalyzed transamination reactions,
see: E. H. Cordes, W. P. Jencks, J. Am. Chem. Soc. 1962, 84, 826–
831; c) for a related cyclization that proceeds through nitrilium ions,
see: S. D. Lepore, M. R. Wiley, J. Org. Chem. 2000, 65, 2924–2932.
[11] Attempts to convert 1a to the corresponding 2-isoxazoline by using
acetone oxime with 20 mol% of catalyst 5 together with various
acids resulted in racemic products at room temperature. When the
reaction was conducted at À208C and sulfonic acids (MsOH or p-
TsOH) were used as cocatalysts, moderate (ca. 60%) ee values were
obtained, but the yields were below 20%.
[12] It should be noted that a perfectly enantioselective catalyst could
not be responsible for the nearly complete racemization of the inter-
mediate. However, equilibration with a less-than-perfect catalyst
will unavoidably result in the erosion of enantiomeric purity.
[13] The yield of the 2-isoxazoline product significantly decreased when
1
strong acids were used. H NMR spectroscopic analysis of the crude
reaction mixtures indicated that the only significant side products
appeared to be polymeric. No discrete side-product peaks could be
discerned in the 1H NMR spectra and the decomposition of 1b was
most reliably detected with the use of an internal standard (see the
Supporting Information).
Acknowledgements
TEKES (National Agency for Technology and Innovation) and the Grad-
uate School of Organic Chemistry and Chemical Biology are acknowl-
Chem. Eur. J. 2009, 15, 3960 – 3964
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3963

P. M. Pihko and A. Pohjakallio
[14] Apparently, the salts resulting from the combination of relatively
weak bases, such as anilines or imidazolidinones, with moderately
strong acids, such as TFA or diphenylphosphoric acid, are acidic
enough to activate cyclization. Stronger bases, such as pyrrolidine or
its derivatives, form salts that are weaker acids and thus poor cata-
lysts for the cyclization process.
[17] The THF/H2O/HCl system typically displayed slightly higher yields
of the isoxazoline product. However, the isolated products were
often contaminated with minor impurities that complicated the
HPLC analysis.
[18] The yield of 4a is largely limited by competing polymerization of 1a
in the reaction conditions.
[15] Previously reported examples of acid-promoted oxime transfer reac-
tions require harsher conditions and longer reaction times, see:
a) P. A. Jacobi, M. Martinelli, E. C. Taylor, J. Org. Chem. 1981, 46,
5416–5417; b) G. M. Shutske, J. Org. Chem. 1984, 49, 180–183.
[16] Other catalysts were also screened for this transformation. The di-
phenyl analoge of 6 (Hayashi catalyst see: Y. Hayashi, H. Gotoh, T.
Hayashi, M. Shoji, Angew. Chem. 2005, 117, 4284–4287; Angew.
Chem. Int. Ed. 2005, 44, 4212–4215) afforded the product in similar
yield and enantioselectivity (63% yield, 90% ee). In contrast, the
MacMillan catalyst 5·ClCH2COOH gave a significantly lower yield
and enantioselectivity (40% yield, 45% ee).
[19] Examples of asymmetric conjugate additions of oxygen nucleophiles
to enals are rare. Addition of salicylaldehydes to enals preserves the
aldehyde oxidation state but does not preserve the original function-
ality, see: a) T. Govender, L. Hojabri, F. M. Moghaddam, P. I. Ar-
vidsson, Tetrahedron: Asymmetry 2006, 17, 1763; b) H. Sundꢄn, I.
Ibrahem, G.-L. Zhao, L. Eriksson, A. Cordova, Chem. Eur. J. 2007,
13, 574–581.
Received: December 19, 2008
Published online: March 4, 2009
3964
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 3960 – 3964