Preparation of β-hydroxyesters from isoxazolines. A selective Ni 0bpy-catalyzed electrochemical method

Article abstract of DOI:10.1016/j.tetlet.2003.09.081

An electrocatalytic method for the reductive N-O cleavage of isoxazolines is described. Ni⁰bpy, generated in situ, was used to promote selective ring opening of 3-methoxy-5-phenylisoxazoline (1a) and 3-methoxy-[4,5]cyclohexylisoxazoline (1b). DMF and NaI were used as solvent and supporting electrolyte, and β-hydroxyesters 2a and 2b were obtained in high yields respectively, after acid hydrolysis. β-Hydroxynitriles 3a and 3b were also identified as side products.

Full text of DOI:10.1016/j.tetlet.2003.09.081

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 44 (2003) 8217–8220
Preparation of b-hydroxyesters from isoxazolines. A selective
Ni0bpy-catalyzed electrochemical method
Viviane F. Caetano, F. W. Joachim Demnitz,* Flamarion B. Diniz, Ronaldo M. Mariz, Jr. and
Marcelo Navarro
Departamento de Qu´ımica Fundamental, CCEN-Universidade Federal de Pernambuco, Cidade Universita´ria CEP: 50670-901,
Recife, PE, Brazil
Received 16 August 2003; revised 1 September 2003; accepted 11 September 2003
Abstract—An electrocatalytic method for the reductive N–O cleavage of isoxazolines is described. Ni0bpy, generated in situ, was
used to promote selective ring opening of 3-methoxy-5-phenylisoxazoline (1a) and 3-methoxy-[4,5]cyclohexylisoxazoline (1b).
DMF and NaI were used as solvent and supporting electrolyte, and b-hydroxyesters 2a and 2b were obtained in high yields
respectively, after acid hydrolysis. b-Hydroxynitriles 3a and 3b were also identified as side products.
© 2003 Elsevier Ltd. All rights reserved.
Isoxazolines have been used as key intermediates in the
synthesis of bioactive molecules.1–4 Heterocyclic ring
opening gives access to many important functional
groups, especially b-hydroxynitriles,2 -ketones3,7,8 and
-esters.4 These occur in many natural products of inter-
est, and are usually obtained by aldol condensation of
ketones and aldehydes often compromising selectivity
however.
protic solvents, NꢀO bond cleavage is observed, fol-
lowed by reduction of the intermediate azomethine,
giving b-hydroxyamines as the principal products. b-
Hydroxyketones, on the other hand, were obtained in
aprotic solvents by prior quaternization of the isoxazo-
lines, followed by electrochemical reductive cleavage.
Although attractive as an alternative to hydrogenolysis,
this direct electrochemical process has some potential
disadvantages such as: elevated reduction potential,
lack of selectivity in the presence of reducible groups
and the use of a two-compartment cell making the
procedure cumbersome.
Torsell5 has developed a synthetic route to b-hydroxy-
ketones via isoxazolines expanding the method to other
b-hydroxycarbonyl compounds as well. Cycloaddition
reactions between readily available olefins and nitrile
oxides furnish excellent yields of the requisite isoxazoli-
nes under very mild conditions.4,6 Tri-substituted olefins
can be used routinely. The major disadvantage of this
approach is the sensitivity of other functionalities to the
catalytic hydrogenolysis typically employed for the sub-
sequent reductive cleavage of the isoxazolines.3,4,7,8 In
spite of these disadvantages, isoxazolines are routinely
employed as precursors in natural product synthesis.
Furthermore alternative methods have been employed
for isoxazoline NꢀO bond cleavage, using transition
metal complexes, for example.7,8
Surov and Lund9,10 described the electrochemical
reduction of isoxazolines (Scheme 1) and isoxazoles. In
* Corresponding author. Present address: Poseidon Pharmaceuticals
A/S, 93 Pederstrupvej, 2750 Ballerup, Denmark. E-mail: jde@
poseidonpharma.com
Scheme 1. Electrochemical NꢀO bond cleavage of isoxazoli-
nes in protic and aprotic media.9,10
0040-4039/$ - see front matter © 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2003.09.081

8218
V. F. Caetano et al. / Tetrahedron Letters 44 (2003) 8217–8220
The Ni0–bipyridine complex (Ni0bpy) has been
employed as a catalyst in homo- or heterocoupling
reactions of aryl halides.11,12 The process involves a
catalytic cycle with successive reductions of intermedi-
ates,13 which may be performed by using an electro-
chemical system,11 or an easily oxidizable transition
metal,14 as a source of electrons. A reductive cleavage
of the ArꢀX bond (X=Cl or Br) occurs in the first step
involving the Ni0bpy complex:
Although appreciable amounts of b-hydroxynitrile 3a/
3b were formed in the catalytic procedures (entries 1–3,
5), the yields of desired b-hydroxyesters 2a and 2b were
very good compared to the yields of theses products
obtained by Ra–Ni hydrogenolysis4 (62–72% versus
76%-2a and 62% versus 32%-2b). The exclusive forma-
tion of nitrile 3a (entry 7) from the bromoisoxazoline
precursor 1c can be attributed to the better leaving
group ability of bromide as compared to methoxy.
Hence elimination of bromide (R%=Br) from the pro-
posed NiII–chelate intermediate I forms 3a, (see below).
Ni0bpy+Ar-X“NiIIbpyArX
We reasoned that application of such a rationale, in
which a nickel complex acts as the actual electron
source, to the electrochemical reduction of isoxazolines
could result in a much milder protocol, which would
obviate the problems of high reduction potential and
over-reduction to b-hydroxyamines. In the event, the
electrocatalytic isoxazoline NꢀO bond cleavage was
successfully carried out using Ni0bpy in DMF with 0.1
M NaI as supporting electrolyte and a zinc rod as
sacrificial anode in an undivided cell. The reduction,
occurring at around 2.5 V cell potential (−0.4/−0.6V
versus Ag/AgCl cathodic potential), completely avoided
the direct electrochemical reduction of the (presumed)
azomethine intermediate.
If the Ni0bpy complex is indeed the electron source for
the reductive NꢀO cleavage, a reaction employing a
stoichiometric equivalent of the complex in the absence
of current should also work. This is indeed the case,
although the yields are slightly lower (2a, 82%). How-
ever, when one equivalent of the NiIIBr2·bpy complex
and the substrate were subjected concurrently to 2
F/mol in the same way as in the catalytic procedures
(entries 4 and 6), not only did we obtain excellent yields
of reduced products, but the proportions of desired
b-hydroxyesters were much higher.
A possible mechanism is suggested in Scheme 2. A
nickel complex chelate intermediate I is formed by
oxidative insertion of Ni0bpy (formed from the
employed NiII(bpy)Br2 by two-electron reduction) into
the NꢀO bond. This stabilizes the oxy-azomethine
formed after reductive ring cleavage. The continuous
passage of electrons however regenerates Ni0bpy for the
catalytic cycle. A charge of 1.1 to 2.0 F mol−1 was
passed, until complete consumption of the substrate.
The b-OH-esters 2a/b are obtained at the end, after acid
hydrolysis during work-up. We believe that the ratio of
b-OH-ester to b-OH-nitrile products (2:3) for a given
isoxazoline is related to the solvent and the stability of
the hydroxy-azomethine intermediate in the reaction
medium. This may explain why the best results were
obtained when 1 equiv. of Ni0bpy was used as reduc-
tant (entries 4 and 6). In this case the stabilized nickel–
oxyazomethine chelate is not deprived of nickel
entering back into the catalytic cycle. Some reactions
were carried out in acetonitrile (entry 8) and the yield
of b-OH-ester was the same as in DMF. Iron was also
tested as a sacrificial anode (entry 9). In this case the
reaction did not go to completion, stopping at about
25% consumption of starting material and giving a
Table 1 shows the results obtained from electrocatalytic
NꢀO bond cleavage16 of 3-methoxy-5-phenylisoxazoline
(1a), 3-methoxy-[4,5]cyclohexylisoxazoline (1b) and 3-
bromo-5-phenylisoxazoline (1c).
Several Ni0bpy concentrations (7–100%) were tested
and uniformly high reduction yields were obtained.
Table 1. Electrocatalytic isoxazoline NꢀO bond cleavage
using NiIIbpy, DMF, 0.1 M NaI and a sacrificial Zn
anode.16 A constant current of 100 mA was applieda
Entry
Isoxazoline
(56.5 mM)
NiIIbpy (%) Yield (%)d,e 2:3e
1
2
3
4
5
6
1a
1a
1a
1a
1b
1b
1c
1a
1a
7
15
30
100
7
100
7
7
90
99
90
99
98
4:1
1.9:1
2.2:1
13:1
1.7:1
9:1
0:1
3:1
2.3:1
99
7
100
100
23
8b
9c
7
a The reaction was conducted under N2.
b Acetonitrile was used as solvent.
c An iron bar was used as sacrificial anode.
d Isolated yields were uniformly >90%.
e Determined by GC analysis.
Scheme 2. Electrocatalytic reductive NꢀO bond cleavage of
isoxazolines using Ni0bpy as catalyst.

V. F. Caetano et al. / Tetrahedron Letters 44 (2003) 8217–8220
8219
lower proportion of b-OH-ester. Probably iron forms a
stable complex with the isoxazoline impeaching the
Ni0bpy catalytic effect.
4651; (e) De Amici, M.; Magri, P.; De Micheli, C.;
Cateni, F.; Bovara, R.; Carrea, G.; Riva, S.; Casalone, G.
J. Org. Chem. 1992, 57, 2825; (f) Torssell, K. B. G.;
Hazell, A. C.; Hazell, R. G. Tetrahedron 1985, 41, 5569.
2. (a) Bacher, E.; Demnitz, F. W. J.; Hurni, Th. Tetrahedron
1997, 53, 14317; (b) Zhengqing, Y.; Khalil, M. A.; Dong-
Hoon, K; Lee, H. J. Tetrahedron Lett. 1995, 36, 3303.
3. (a) Curran, D. P. Tetrahedron Lett. 1983, 24, 3443; (b)
Curran, D. P. J. Am. Chem. Soc. 1982, 104, 4024; (c)
Curran, D. P. J. Am. Chem. Soc. 1983, 105, 5826; (d)
Kozikowski, A. P.; Stein, P. D. J. Am. Chem. Soc. 1982,
104, 4023; (e) Asaoka, M.; Mukuta, T.; Takei, H. Tetra-
hedron Lett. 1981, 22, 735; (f) Wollenberg, R. H.;
Goldstein, J. E. Synthesis 1980, 757; (g) Kozikowski, A.
P.; Adamczyk, M. Tetrahedron Lett. 1982, 23, 3123.
4. (a) Nishi, T.; Morisawa, Y. Heterocycles 1989, 29, 1835;
(b) Caldirola, P.; Ciancaglione, M.; De Amici, M.; De
Micheli, C. Tetrahedron Lett. 1986, 27, 4647.
A complete mechanistic study and determination of
ideal reaction parameters for this reaction is under way.
3-Methoxy-5-phenylisoxazoline (1a):4 1H NMR (300
MHz, CDCl3): l 7.23–7.32 (m, 5H), 5.54 (dd, J=10.9
Hz, 1H), 3.88 (s, 3H), 3.24 (dd, J=16.10 Hz, 1H), 2.91
(dd, J=16.9 Hz, 1H). MS: m/e 177 (39), 132 (100), 105
(36), 77 (30).
3-Methoxy-[4,5]cyclohexylisoxazoline (1b):4 1H NMR
(300 MHz, CDCl3): l 4.54 (dt, J=11, 9 Hz, 1H), 3.86
(s, 3H), 2.87 (dt, J=11.9 Hz, 1H), 1.9–1.2 (m, 8H). MS:
m/e 155 (6), 124 (10), 69 (68), 43 (100).
3-Bromo-5-phenylisoxazoline (1c):4 1H NMR (300 MHz,
CDCl3): l 7.36–7.46 (m, 5H), 5.7 (dd, J=11, 9 Hz, 1H),
3.66 (dd, J=17, 11 Hz, 1H), 3.25 (dd, J=17, 9 Hz,
1H). MS: m/e 225/227 (58), 128 (39), 115 (32), 105
(100).
5. Torsell, K.; Zenthen, O. Acta Chem. Scand. Ser. B 1978,
32, 118.
6. (a) Christl, M.; Huisgen, R. Chem. Ber. 1973, 106, 3345
and references cited therein; (b) Bianchi, G.; De Micheli,
C.; Gandolfi, R.; Gru¨nanger, R.; Finzi, P. V.; de Pava, O.
V. J. Chem. Soc., Perkin Trans. 1 1973, 1148 and refer-
ences cited therein; (c) Larsen, K. E.; Torssell, K. B. G.
Tetrahedron 1984, 40, 2985.
Methyl 3-phenyl-3-hydroxypropanoate (2a):17 oil, 1H
NMR (300 MHz, CDCl3): l 7.24–7.18 (m, 5H), 4.98
(dd, J=4, 9 Hz, 1H), 3.69 (s, 1H), 3.50 (s, 3H), 2.56
(ddd, J=4, 9, 16 Hz, 2H). MS: m/e 180 (27), 163 (28),
120 (25), 107 (100), 79 (85), 77 (61).
7. Baraldi, P. G.; Barco, A.; Benetti, S.; Manfredini, S.;
Simoni, D. Synthesis 1987, 276.
8. Guarna, A.; Guidi, A.; Goti, A.; Brandi, A.; de Sarlo, F.
Synthesis 1989, 175.
9. Surov, I.; Lund, H. Acta Chem. Scand., Ser. B 1986, 40,
831.
10. Lund, H. Acta Chem. Scand. 1959, 13, 249.
11. Ne´dele´c, J.-Y.; Pe´richon, J.; Troupel, M. Top. Curr.
Chem. 1997, 185, 147.
1
Methyl hexahydrosalicylate (2b):18 oil, H NMR (300
MHz, CDCl3): l 4.18 (dt, J=11 Hz, 1H), 3.68 (s, 3H),
2.52 (dt, J=11 Hz, 1H), 1.9–1.2 (m, 8H). MS: m/e 158
(1), 127 (7), 130 (30), 87 (100), 81 (31), 55 (36).
3-Phenyl-3-hydroxypropanecarbonitrile (3a):19 oil, 1H
NMR (300 MHz, CDCl3): l 7.26–7.20 (m, 5H), 4.80 (t,
J=6, 3 Hz, 1H), 3.60 (s, 1H), 2.54 (d, J=6 Hz, 2H).
MS: m/e 148 (16), 130 (6), 118 (9), 107 (68), 79 (100), 51
(34).
12. Cassol, T. M.; Demnitz, F. W. J.; Navarro, M.; Neves, E.
A. Tetrahedron Lett. 2000, 41, 8203.
13. Amatore, C.; Jutand, A. Organometallics 1988, 7, 2203.
14. Iyoda, M.; Otsuka, H.; Sato, K.; Nisato, N.; Oda, M.
Bull. Chem. Soc. Jpn. 1990, 63, 80.
15. Uchino, M.; Asagi, K.; Yamamoto, A.; Ikeda, S. J.
1
2-Cyanocyclohexanol (3b):20 oil, H NMR (300 MHz,
Organomet. Chem. 1975, 84, 93.
CDCl3): l 4.10 (dt, J=10 Hz, 1H), 2.51 (dt, J=10 Hz,
1H), 1.9–1.2 (m, 8H). MS: m/e 125 (10), 99 (9), 85 (40),
71 (72), 57 (67), 43 (100).
16. The controlled current preparative electrolyses were car-
ried out in undivided cells of 15 mL capacity. A zinc rod
of 0.8 cm diameter and nickel foam (10 cm×4 cm, Nitech)
were used as sacrificial anode (immersed >1 cm in solu-
tion) and working electrode, respectively (Ni foil or bar
can also be used). The Ni electrode can be reused about
20 times, after cleaning with a 6 M HCl solution after
each application. The same solution was used to clean the
anode. For experiments involving Ni0bpy as catalyst, the
precursor NiII(bpy)Br2 was prepared separately according
to the literature.15 The electrolytic cell was charged under
nitrogen with 15 mL DMF containing 100 mM NaI. The
isoxazoline (56.5 mM) was mixed with NiII(bpy)Br2 in 5
mL of solvent and then added to the cell. A constant
current (100 mA) was applied until full consumption of
the starting reagent (Q=175). (In experiments using 1
equiv. of Ni complex a charge of 320 C was passed, and
stirring continued for 2 h before work-up). The solvent
was removed at reduced pressure and the residue was
dissolved in diethyl ether, washed thoroughly with five
portions of 0.2 M HCl solution, dried with Na2SO4,
Acknowledgements
We would like to acknowledge CNPq (Brazilian Organ)
for financial support. R. M. Mariz Jr. and V. F.
Caetano thank CNPq/PIBIC for undergraduate
scholarships.
References
1. (a) Kozikowski, A. P. Acc. Chem. Res. 1984, 17, 410; (b)
Levina, I. S.; Mortikova, E. I.; Kamernitzky, A. V.
Synthesis 1974, 562; (c) Vyas, D. M.; Chiang, Y.; Doyle,
T. W. Tetrahedron Lett. 1984, 25, 487; (d) Caldirola, P.;
De Amici, M.; De Micheli, C. Tetrahedron Lett. 1986, 27,

8220
V. F. Caetano et al. / Tetrahedron Letters 44 (2003) 8217–8220
tate (3:2).
filtered and concentrated in vacuo. The product ratios
were determined by GC analysis using a Varian 3380
chromatograph, fitted with a 30 m capillary CP-SPL5CB
Chrompack column, using a 60–200°C temperature ramp
(20°C min−1). The mixture of products was purified by
silica gel column chromatography using hexane/ethyl ace-
17. Fleming, I.; Winter, S. B. D. J. Chem. Soc., Perkin Trans.
1 1998, 17, 2687.
18. Blancou, H.; Casadevall, E. Tetrahedron 1976, 32, 2907.
19. Huisgen, R.; Christl, M. Chem. Ber. 1973, 106, 3291.
20. Guillanton, L.; Lamant, F. Nouv. J. Chim. 1978, 2, 157.