Enantioselective synthesis of trifluoromethyl-substituted 2-isoxazolines: Asymmetric hydroxylamine/enone cascade reaction

Article abstract of DOI:10.1002/anie.201002065

(Figure Presented) Cuts both ways: The title reaction consists of an addition/cyclization/dehydration sequence and affords the biologically important chiral 3,5-diaryl-5-(trifluoromethyl)-2-isoxazolines 1 in excellent yields with high ee values. The flexi

Full text of DOI:10.1002/anie.201002065

Communications
DOI: 10.1002/anie.201002065
Organocatalysis
Enantioselective Synthesis of Trifluoromethyl-Substituted 2-
Isoxazolines: Asymmetric Hydroxylamine/Enone Cascade Reaction**
Kazutaka Matoba, Hiroyuki Kawai, Tatsuya Furukawa, Akihiro Kusuda, Etsuko Tokunaga,
Shuichi Nakamura, Motoo Shiro, and Norio Shibata*
Heterocycles containing a trifluoromethyl group are repre-
sentatives of a major structure type in agricultural and
medicinal chemistry, and thus the development of a simple
and flexible method to generate a novel trifluoromethylated
heterocyclic system has received much attention.[1] Trifluor-
omethyl-substituted 2-isoxazolines are amongst an important
class of heterocyclic compounds with remarkable biological
activities, and they make interesting synthetic targets.[2] Since
the first discovery of the structurally unique 3,5-diaryl-5-
(trifluoromethyl)-2-isoxazoline derivatives 1 (Figure 1) as
pest control agents in 2004,[3a] the search for new agro-
active component.[3b] Once the importance of the optical
purity of these compounds was verified, the next challenge
was to control the stereochemistry at a quaternary carbon
center bearing a CF3 group. As part of our ongoing research
programs directed to the development of efficient methods
for the asymmetric synthesis of organofluorine compounds,[4]
we disclose herein the synthesis of trifluoromethyl-substituted
2-isoxazolines by a cinchona alkaloid catalyzed asymmetric
hydroxylamine/enone cascade reaction consisting of a con-
jugate addition/cyclization/dehydration sequence (Scheme 1).
Figure 1. Structures of 3,5-diaryl-5-(trifluoromethyl)-2-isoxazolines 1
and biologically attractive derivatives of types A and B.
chemicals and veterinary medicines has focused largely on
this skeleton,[3] despite their diverse structural variants.[2] Thus
far, more than 20000 compounds have been synthesized and
patented in the past 5 years (the results of substructure
searching using 1 with Scifinder), and many promising drug
candidates have been disclosed including an antiparasiticide
compound of either type A or B (Figure 1).[3] More interest-
ingly, recent biological evaluation of the optically active
type B compound (X6 = CONHCH2COCH2CF3), obtained by
HPLC methods using a chiral stationary phase, revealed that
the R enantiomer of B is inactive and the S enantiomer is the
Scheme 1. Enantioselective synthesis of trifluoromethyl-substituted 2-
isoxazolines 1 by an asymmetric hydroxylamine/enone cascade reac-
tion.
Optically active 2-isoxazolines are mostly constructed by
the asymmetric 1,3-dipolar cycloaddition of nitrile oxides to
olefins,[5a–h] the asymmetric cyclization of b,g-unsaturated
oximes,[5i] stereoselective ring-closure reaction through
oximes of chiral Michael adducts of thiophenol to chalco-
nes,[5j] and proline-catalyzed conjugate addition/oxime-trans-
fer reactions of unsaturated aldehydes.[5k] These methods,
however, were not applicable to the synthesis of 1 because of
the limitations in substrate specificity, and indeed, the
catalytic enantioselective synthesis of trifluoromethyl-substi-
tuted 2-isoxazolines 1 has not been reported. We therefore
started to investigate the asymmetric version of the nonchiral
2-isoxazoline-formation reaction.[3c] We first examined the
reaction of (E)-4,4,4-trifluoro-1,3-diphenylbut-2-en-1-one
(2a) with hydroxylamine in the presence of NaOH and a
catalytic amount of N-3,5-bis(trifluoromethylbenzyl) quinidi-
nium bromide (3a) as a chiral phase-transfer catalyst in
toluene at À308C (Table 1). The trifluoromethylated 2-
isoxazoline (R)-1a was formed in 89% ee, although the
yield was 24% (Table 1, entry 1). Preliminary results encour-
aged us to change the solvent, which additionally improved
[*] K. Matoba, H. Kawai, T. Furukawa, A. Kusuda, E. Tokunaga,
Dr. S. Nakamura, Prof. N. Shibata
Department of Frontier Materials, Graduate School of Engineering,
Nagoya Institute of Technology
Gokiso, Showa-ku, Nagoya 466-8555 (Japan)
Fax: (+81)52-735-5442
E-mail: nozshiba@nitech.ac.jp
Dr. M. Shiro
Rigaku Corporation
3-9-12 Mastubara-cho, Akishima Tokyo 196-8666 (Japan)
[**] Support was provided by KAKENHI (21390030, 22106515). The
corresponding author thanks the JSPS for financial support of this
work in part through a Joint Research Project between Japan and
France.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201002065.
5762
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 5762 –5766

Angewandte
Chemie
Table 2: Substrate scope.
the yields under the same catalyst conditions. Whereas THF,
DME, MeCN, and EtOH did not effectively improve both the
yield and ee value (Table 1, entries 2–5), chlorinated solvents
such as CH2Cl2, (ClCH2)2, CHCl3, and CH2(CH2Cl)2 proved to
Entry
2
R1
R2
3
1
Yield [%] [a] ee [%][b]
Table 1: Optimization of reaction conditions.[a]
1
2b 3-MeC6H4
2c 4-MeC6H4
2d 4-MeOC6H4 Ph
2e 4-FC6H4
2 f 4-ClC6H4
2g 4-BrC6H4
Ph
Ph
3a 1b 83
3a 1c 83
3a 1d 99
3a 1e 92
3a 1 f 97
3a 1g 99
3a 1h 94
3a 1i 92
3a 1j 96
3a 1k 99
92
94
92
91
91
90
92
82
91
93
91
88
90
91
91
91
2[c]
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Ph
Ph
Ph
Entry
Base
Solvent
3
Yield [%][b]
ee [%][c]
2h 2-naphthyl Ph
1
2
3
4
5
6
7
8
10n NaOH
10n NaOH
10n NaOH
10n NaOH
10n NaOH
10n NaOH
10n NaOH
10n NaOH
10n NaOH
5n LiOH
toluene
THF
DME
3a
3a
3a
3a
3a
3a
3a
3a
3a
3a
3a
3a
3a
3b
24
16
12
36
14
80
99
92
71
34
38
88
n.r.
89
89(R)
76(R)
67(R)
2(R)
62(R)
85(R)
90(R)
89(R)
92(R)
91(R)
92(R)
92(R)
–
2i Ph
2j Ph
2k Ph
2l Ph
2m Ph
2n Ph
2o Ph
2p Ph
2q Ph
2-MeC6H4
3-MeC6H4
4-MeC6H4
EtOH
4-MeOC6H4 3a 1l 88
MeCN
CH2Cl2
(ClCH2)2
CH2(CH2Cl)2
CHCl3
CHCl3
CHCl3
CHCl3
CHCl3
4-FC6H4
4-ClC6H4
4-BrC6H4
2-naphthyl 3a 1p 99
cyclohexyl 3a 1q 80
3a 1m 91
3c 1n 96
3c 1o 99
9
10
11
12
13
14
[a] Yield of isolated product. [b] Determined by HPLC methods using a
chiral stationary phase. [c] (CH2Cl)2 was used as a solvent.
10n KOH
10n CsOH
10n Na2CO3
10n CsOH
CHCl3
88(S)
[a] All reactions performed at 0.10 mmol substrate in 1.1 mL of solvent.
[b] Yield of isolated product. [c] Determined by HPLC methods using a
chiral stationary phase. DME=1,2-dimethoxyethane, n.r.=no reaction,
THF=tetrahydrofuran.
1h, 1p, and 1q in high yields with ee values of 91%–92%
(Table 2, entries 7, 15, and 16). The absolute stereochemistry
of (R)-1g was clearly determined by X-ray analysis (Figure 2)
be better solvents, providing 1a with up to 92% ee (Table 1,
entries 6–9).
A subsequent survey of bases (Table 1,
entries 10–13) revealed that CsOH was the base of choice
with regards to both enantioselectivity and yield (Table 1,
entry 12). We also discovered that the analogous ammonium
bromide, 3b, derived from quinine showed excellent enantio-
selectivity for giving 1a with the opposite (S) stereochemistry
(Table 1, entry 14). Screening of the cinchona alkaloids
ascertained that the use of 3a displayed the highest enantio-
selectivity in the formation of 1a (see Table S1 in the
Supporting Information).
With optimal reaction conditions in hand, the scope of the
asymmetric hydroxylamine/enone cascade reaction was
explored using a variety of substrates selected to establish
the generality of the process. By using a catalytic amount of
3a, all substrates afforded good to excellent yield and high to
excellent enantioselectivity (Table 2). A series of trifluoro-
methylated enone derivatives, 2b—m, having a variety of
substituents such as bromo, chloro, fluoro, methyl, and
methoxy, on their aromatic rings were nicely converted into
the corresponding products 1b–m in good yields with 82%–
94% ee (Table 2, entries 1–12). Catalyst 3c was a better
catalyst for a couple of substrates such as 2n and 2o,
furnishing 1n and 1o, respectively, in excellent yields with
high enantioselectivity (Table 2, entries 13 and 14). Sterically
demanding naphthyl-substituted enones 2h and 2p, and
enolizable aliphatic aryl enone 2q were also compatible
with the reaction conditions, affording the respective products
Figure 2. X-ray crystallographic analysis of (R)-1g. The thermal ellip-
soids are shown at 50% probability.[10]
and all the other products were tentatively assigned by
analogy with 1g. Notably, the enantioselectivity was dramat-
ically changed when the cascade reaction of 2a with hydroxyl-
amine was performed using the methylether catalyst 3e to
give (S)-1a with 34% ee (Scheme 2a). This result indicates
the hydrogen bond of the OH group is crucial for asymmetric
induction. Furthermore, a similarly high enantioselectivity of
(R)-1a with 85% ee was observed in the presence of the
cinchonine-type catalyst 3 f instead of the quinidinium
catalyst 3a (Scheme 2b). Consequently, the methoxy group
on the quinoline ring of 3a does not effect on the enantio-
meric excess of the products.
Angew. Chem. Int. Ed. 2010, 49, 5762 –5766
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5763

Communications
of 2 is difficult to obtained because of the viscous
character of 2. The three-dimensional conformation of
2a was therefore generated through calculations with
MOPAC PM3 method (Figure 3b). It is of great interest
that the two phenyl substituents and the carbonyl s plane
of 2a are found to be nonplanar, that is, not p conjugated.
This phenomena is surprisingly similar to the case of the
highly enantioselective epoxidation of enones using the
ammonium salt of a cinchona alkaloid reported by Corey
and co-workers.[8] On the basis of Coreyꢀs speculation on
the reaction mechanism using the charge-accelerated
catalysis by a rigid quaternary ammonium salt,[8] we
postulated a similar transition-state structure for the
Scheme 2. Effects of the hydroxy group and methoxy group on 3 for
asymmetric induction.
This isoxazoline-formation reaction consists of two dis-
tinct steps, Michael addition and imine formation reactions.
The possible reaction routes, a and b, are shown in Scheme 3.
Figure 3. a) X-Ray crystallographic structure of catalyst 3 f generated
[11]
from the cif file of 3 f.
b) Three-dimensional arrangement of 2a
optimized by using the MOPAC PM3 method.
production of (R)-1a catalyzed by 3, that is, the specific three-
dimensional arrangement of cation 3, 2a, and the hydroxyl-
amine anion (Figure 4). The free hydroxy group in 3 captures
substrate 2a, presumably by intermolecular hydrogen-bond
formation to the oxygen atom in 2a. Only the arrangement as
shown in the Figure 4a should be acceptable because of the
very limited space available by the sterically demanding and
rigid structure of 3. The hydroxylamine anion generated by
the base is sandwiched between 3 f and 2a, and associates
strongly with the nitrogen cation of the catalyst 3, as a nearest
neighbor, through ion pairing. The hydroxylamine
approaches from the Re face of 2a. Both the hydrogen
bonding between the prochiral substrate 2 and catalyst 3, and
the ionic interaction between the ammonium cation 3 and
Scheme 3. The possible reaction routes, a or b, for 1a. Reaction
conditions: 10n CsOH (3.3 equiv), 3a (10 mol%), CHCl3, À30 to
À108C, 10–20 h.
The first hypothesis (route a) is that the oxygen anion of
hydroxylamine adds in Michael fashion to the unsaturated
ketone 2a to furnish 4, with a subsequent intramolecular ring-
closure cascade from the nitrogen atom onto ketone moiety.
Then the elimination of water generates the isoxazoline 1a.
An alternative (route b) is that the reaction proceeds through
an imine formation first to produce the intermediate oxime 5;
this then ring closes in a second step by intramolecular
conjugated addition. To identify the mechanism, the oxime 5
was isolated individually and then we attempted to cyclize it
under the same reaction conditions in a separate experiment.
However, no isoxazoline 1a was detected. This result can also
be supported by Baldwinꢀs rules of guidelines for ring-closing
reactions that the 5-endo-trig is disfavored.[6] Thus, route b
does not operate here. In contrast, route a consists of the ring-
closure reaction in a second step through a favored 5-exo-trig
process.[6] Accordingly, intermediate 4 is difficult to isolate
because of the rapid cyclization of 4 into 1a.
To understand the high enantioselectivity observed for the
present reaction catalyzed by the ammonium salts of cinchona
alkaloids, the three-dimensional arrangement of the starting
enones 2 and either the catalyst 3a or 3 f is required. Although
the X-ray crystallographic structure of catalyst 3 f has been
disclosed by us[7] (Figure 3a), the conformational information
Figure 4. Proposed transition-state model from 2a to (R)-1a catalyzed
by 3 f. a) CPK model based on the X-ray crystallographic structure of
3 f and calculated structure of 2a. b) Schematic presentation of the
transition-state structure.
5764
www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 5762 –5766

Angewandte
Chemie
medical Applications, Elsevier, Amsterdam, 1993; e) J. T. Welch,
hydroxylamine anion should realize the highly enantioselec-
tive reaction (Figure 4b). The possibility of the aromatic p–p
interactions between 2a and 3 might be ruled out because the
nonaromatic substrate 2p was also converted into product 1p
with high enantiomeric excess.
S. Eswarakrishman, Fluorine in Bioorganic Chemistry, Wiley,
New York, 1991; f) R. Filler, Y. Kobayashi, Biomedicinal Aspects
of Fluorine Chemistry, Elsevier Biomedical Press and Kodansya
Ltd, Amsterdam, 1982; g) V. M. Muzalevskiy, A. V. Shastin, E. S.
Balenkova, G. Haufe, V. G. Nenajdenko, Synthesis 2009, 3905 –
3929; h) A. W. Erian, J. Heterocycl. Chem. 2001, 38, 793 – 808;
i) P. Lin, J. Jiang, Tetrahedron 2000, 56, 3635 – 3671; j) M. J.
Silvester, Adv. Heterocycl. Chem. 1994, 59, 1 – 38; k) J. T. Welch,
Tetrahedron 1987, 43, 3123 – 3197.
We next focused on the asymmetric hydroxylamine/enone
cascade reaction of multiply substituted substrate 2r. Under
the optimized reaction conditions using catalyst 3a, the
corresponding trifluoromethyl-substituted 2-isoxazoline 1r
with an R configuration was obtained in 99% yield with
75% ee. Upon addition of catalyst 3c instead of 3a to the
reaction, the enantioselectivity of the 2-isoxazoline was
improved by as much as 13% ee to provide 1r in 99% yield
with 88% ee. As expected, (S)-1r with 81% ee was obtained
in excellent yield from the asymmetric hydroxylamine/enone
cascade reaction by using the pseudoenantiomer catalyst 3d.
The 2-isoxazoline 1r could be a key intermediate for the
synthesis of chiral type B compounds (Scheme 4).[3g–j]
[2] For example: a) M. L. Quan, C. D. Ellis, A. Y. Liauw, R. S.
Alexander, R. M. Knabb, G. Lam, M. R. Wright, P. C. Wong,
R. R. Wexler, J. Med. Chem. 1999, 42, 2760 – 2773; b) V. Kumar,
R. Aggarwal, S. P. Singh, J. Fluorine Chem. 2006, 127, 880 – 888;
c) P. Bravo, L. Bruchꢁ, M. Crucianelli, A. Farina, S. V. Meille, A.
Merli, P. Seresini, J. Chem. Res. Synop. 1996, 348 – 349.
[3] For example: a) T. Mita, Y. Kudo, T. Mizukoshi, H. Hotta, K.
Maeda, S. Takii, WO2004018410, 2004; b) Y. Ozoe, M. Asahi, F.
Ozoe, K. Nakahira, T. Mita, Biochem. Biophys. Res. Commun.
2010, 391, 744 – 749; c) M. Yaosaka, T. Utsunomiya, Y. Mor-
iyama, T. Matsumoto, K. Matoba, WO2009001942, 2009; d) K.
Matoba, WO 2009063910, 2009. type A.; e) G. P. Lahm, W. L.
Shoop, M. Xu, WO2007079162, 2007; f) J. K. Long, T. P. Selby,
M. Xu, WO2009045999, 2009; type B.; g) T. Mita, T. Kikuchi, T.
Mizukoshi, M. Yaosaka, M. Komoda, WO2005085216, 2005;
h) W. Zambach, P. Renold, WO2009049846, 2009; i) P. Renold,
W. Zambach, P. Maienfisch, M. Muehlebach, WO2009080250,
2009; j) T. Murata, Y. Yoneta, J. Mihara, K. Domon, M.
Hatazawa, K. Araki, E. Shimojo, K. Shibuya, T. Ichihara, U.
Goergens, A. Voerste, A. Becker, E. Franken, K. Mueller,
WO2009112275, 2009.
[4] a) N. Shibata, S. Mizuta, H. Kawai, Tetrahedron: Asymmetry
2008, 19, 2633 – 2644; b) N. Shibata, T. Ishimaru, S. Nakamura, T.
Toru, J. Fluorine Chem. 2007, 128, 469 – 483; c) N. Shibata, T.
Furukawa, D. S. Reddy, Chem. Today 2009, 27, 38 – 42; d) N.
Shibata, S. Mizuta, T. Toru, J. Synth. Org. Chem. Jpn. 2008, 66,
215 – 228; e) N. Shibata, J. Synth. Org. Chem. Jpn. 2006, 64, 14 –
24.
Scheme 4. Synthesis of key type B compounds (see Figure 1).
[5] a) Y. Ukaji, K. Sada, K. Inomata, Chem. Lett. 1993, 1847 – 1850;
b) M. Shimizu, Y. Ukaji, K. Inomata, Chem. Lett. 1996, 455 –
456; c) M. Tsuji, Y. Ukaji, K. Inomata, Chem. Lett. 2002, 1112 –
1113; d) M. P. Sibi, K. Itoh, C. P. Jasperse, J. Am. Chem. Soc.
2004, 126, 5366 – 5367; e) H. Suga, Y. Adachi, K. Fujimoto, Y.
Furihata, T. Tsuchida, J. Org. Chem. 2009, 74, 1099 – 1113; f) A.
Ros, E. Alvarez, H. Dietrich, R. Fernꢂndez, J. M. Lassaletta,
Synlett 2005, 2899 – 2904; g) Y. Brinkmann, R. J. Madhushaw, R.
Jazzar, Tetrahedron 2007, 63, 8413 – 8419; h) C. D. Davies, S. P.
Marsden, E. S. E. Stokes, Tetrahedron Lett. 2000, 41, 4229 – 4233;
i) A. L. Norman, M. D. Mosher, Tetrahedron Lett. 2008, 49,
4153 – 4155; j) M. Zielinska-Błajet, R. Kowalczyk, J. Skarzewski,
Tetrahedron 2005, 61, 5235 – 5240; k) A. Pohjakallio, P. M. Pihko,
Chem. Eur. J. 2009, 15, 3960 – 3964.
[6] J. E. Baldwin, J. Chem. Soc. Chem. Commun. 1976, 734 – 736.
[7] H. Kawai, A. Kusuda, S. Nakamura, M. Shiro, N. Shibata,
Angew. Chem. 2009, 121, 6442 – 6445; Angew. Chem. Int. Ed.
2009, 48, 6324 – 6327.
[8] a) E. J. Corey, F. Xu, M. C. Noe, J. Am. Chem. Soc. 1997, 119,
12414 – 12415; b) E. J. Corey, Y. Bo, J. Busch-Petersen, J. Am.
Chem. Soc. 1998, 120, 13000 – 13001; c) E. J. Corey, M. C. Noe, F.
Xu, Tetrahedron Lett. 1998, 39, 5347 – 5350; d) M. Horikawa, J.
Busch-Petersen, E. J. Corey, Tetrahedron Lett. 1999, 40, 3843 –
3846; e) E. J. Corey, F.-Y. Zhang, Angew. Chem. 1999, 111, 2057 –
2059; Angew. Chem. Int. Ed. 1999, 38, 1931 – 1934; f) E. J. Corey,
F.-Y. Zhang, Org. Lett. 1999, 1, 1287 – 1290.
In summary, we have developed the first catalytic
enantioselective synthesis of trifluoromethyl-substituted
2-isoxazolines by a cinchona alkaloid-catalyzed asymmetric
hydroxylamine/enone cascade reaction consisting of conju-
gate addition/cyclization/dehydration reactions to provide the
medicinally important trifluoromethylated compounds.[3d]
Wide substrate generality, the flexibility to generate either
the S or R enantiomers, and high levels of enantioselectivity
have been achieved.[9]
˙
Received: April 7, 2010
Revised: May 27, 2010
Published online: July 7, 2010
Keywords: asymmetric catalysis · fluorine · heterocycles ·
.
organocatalysis · phase-transfer catalysis
[1] a) V. A. Petrov, Fluorinated Heterocyclic Compounds: Synthesis
Chemistry, and Applications, Wiley, Hoboken, NJ, 2009; b) P.
Kirsch, Modern Fluoroorganic Chemistry, Wiley-VCH, Wein-
heim, 2004; c) I. Ojima, J. R. MacCarthy, J. T. Welch, Biomedical
Frontiers of Fluorine Chemistry, American Chemical Society,
New York, 1996; d) R. Filler, Y. Kobayashi, L. M. Yagupolskii,
Organofluorine Compounds in Medicinal Chemistry and Bio
[9] Although the wisdom in using chloroform as a solvent with a
strong base such as CsOH might be questionable since dichlor-
ocarbene can be formed, an unexpected side reaction was not
Angew. Chem. Int. Ed. 2010, 49, 5762 –5766
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5765

Communications
detected at all. Importantly, we have ascertained that environ-
[10] CCDC 763815 ((R)-1g) contains the supplementary crystallo-
graphic data for this paper. These data can be obtained free of
charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif).
mentally benign trifluorotoluene can be used as a solvent in this
transformation without any loss of the yields and enantioselec-
tivity.
[11] CCDC 738896 (3 f) contains the supplementary crystallographic
data for this paper. These data can be obtained free of charge
from The Cambridge Crystallographic Data Centre via www.
ccdc.cam.ac.uk/data_request/cif).
5766
www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 5762 –5766