Full Paper
analog Cloxacillin.[9] The nitrogen–oxygen bond in the isoxa-
zole ring is relatively weak and, similarly to electron deficient
triazoles, its scission via hydrogenation[10] or other metal-medi-
ated reactions,[11] reveals the isoxazole ring as a latent 1,3-di-
carbonyl compound.[12] Therefore, unsurprisingly, isoxazoles
and their subsequent transformations, have been used in the
synthesis of complex molecules.[13] Although several methods
for the synthesis of isoxazoles are available, the 1,3-dipolar cy-
cloaddition of corresponding nitrile oxides and alkynes is par-
ticularly attractive due to the direct formation of the C3ꢀC4
carbon–carbon bond.[14]
ate improvement of both yield and regioselectivity. However,
formation of furoxan, a consequence of nitrile oxide dimeriza-
tion, remained an undesired side reaction.[32] In stark contrast,
electron-deficient halogenated alkynes, as illustrated by 1-
bromo-dimethyl-propiolamide (1), readily participated in
[CpRuCl(cod)]-mediated catalysis. Details of their reactivity with
organic nitrile oxides and azides are described below.
Results and Discussion
The reaction of bromoalkyne 1 and nitrile oxide (1.25 equiv)
obtained from hydroximoyl chloride 2 in the presence of
Hꢁnig’s base at ambient temperature produced a regioisomeric
mixture of bromoisoxazoles favoring the 5-halogenated isomer
by approximately 4:1 (3b/3a) with a combined isolated yield
of approximately 35% (Scheme 2A). The addition of a catalytic
After having developed a direct copper-catalyzed route for
the synthesis of 5-iodo-1,2,3-triazoles from 1-iodoalkynes,[15] we
were naturally curious about a similar ruthenium-mediated
process, which could perhaps also engage 1-chloro and 1-bro-
moalkynes in cycloaddition reactions. The direct regiospecific
generation of haloazoles would be an important advantage of
this method, as these compounds are known pharmaco-
phores,[16] as well as useful synthetic intermediates from which
a variety of derivatives can be obtained by manipulation of the
carbon–halogen bond.[17] In addition to the above-mentioned
copper-catalyzed methods, which encompass 1-bromo and 1-
iodoalkynes,[15,18a–c] an iridium-catalyzed process to access 4-
bromo-triazoles was recently developed.[18d] Halotriazoles can
also be accessed by trapping reactive triazolyl intermediates,
including 5-triazolyl-copper,[19] 5-aluminum,[20] 5-bismuth,[21] or
4-triazolyl magnesium bromide[22] analogues with electrophilic
halide reagents. Alternatively, nucleophilic displacement of
diazo,[23] silyl,[24] or stannyl[25] groups directly attached to the tri-
azole ring by a halide anion can be used in the synthesis of
halotriazoles.
Although the direct halogenation of the isoxazole ring was
reported by Claisen over a century ago,[26] the reaction condi-
tions are harsh and require the use of strong acids at elevated
temperature. Some of these limitation have been addressed
using electrophilic cyclization of 2-alkyn-1-one O-methyl
oximes, which provides 4-substituted isoxazoles containing
halides and selenium functional groups.[27] The 1,3-dipolar cy-
cloaddition between terminal alkynes and bromonitrile oxides
is a means to provide 3-haloisoxazoles in good yield with high
regioselectivity.[28] However, examples of high-yielding reac-
tions involving 1-haloalkynes and nitrile oxides remain remark-
ably scarce.[29] For example, thermal cycloadditions of nitrile
oxides and 1-haloalkynes often either require alkynes substitut-
ed with highly activated hypervalent iodine species[30] or
a large excess of one of the reactants[29c] and/or controlled ad-
dition of the nitrile oxide component.[29b] Despite these precau-
tions, these reactions still often result in mixtures of regioiso-
mers with low to moderate yields.[16a,31]
Scheme 2. [CpRuCl(cod)]-catalyzed cycloaddition of 1-bromodimethylpropio-
lamide (1) with 4-chloro-N-hydroxybenzimdoyl chloride (2) (A) and pheneth-
yl azide (4) (B). Isolated yields after column chromatography are shown. Re-
1
gioselectivity was determined using a combination of H NMR spectroscopy
and GCMS or LCMS analysis. [a] 1 (0.35 mmol), 2 (1.25 equiv), DIPEA
(1.4 equiv), 1,4-dioxane, RT, 3 h. [b] 1 (3 mmol), 2 (1.25 equiv), DIPEA
(1.4 equiv), THF, RT, 1 h. [c] 1 (0.17 mmol), 4 (1.1 equiv), [D8]toluene, 908C,
6 d. [d] 1 (3.2 mmol), 4 (1.25 equiv), MeCN, RT, 1 h. [e] 100% conversion
1
monitored by H NMR spectroscopy.
amount of [CpRuCl(cod)] transformed this reaction into a com-
pletely regiospecific process that furnished the 4-bromoisoxa-
zole product 3a in good yield.[33] A slight excess of hydroximo-
yl chloride was optimal, and [CpRuCl(cod)] catalyst loading
could be reduced to as little as 3 mol%. Further optimization
revealed that this reaction was air tolerant and 1,4-dioxane, di-
chloroethane, dichloromethane, and THF (at 0.1–0.4m concen-
tration of the reactants) were identified as preferred reaction
solvents. Ambient reaction temperature was optimal as dimeri-
zation of nitrile oxide noticeably increased at elevated temper-
ature. Hꢁnig’s base (diisopropylethylamine) was more effective
than triethylamine, primarily due to the tendency of triethyla-
mine to cross-react with the alkyne component by Michael ad-
dition.[34] Experimentally, a typical reaction protocol dictates
the addition of [CpRuCl(cod)] followed within seconds by the
addition of Hꢁnig’s base.
To begin, we examined the room temperature reaction of
(1-chloroethynyl)benzene and nitrile oxide precursor, 4-chloro-
N-hydroxybenzimdoyl chloride (2). This reaction was predicta-
bly poor in the absence of a catalyst, resulting in low yields (<
15%) of the desired chloroisoxazole. Disappointingly, [Cp*RuCl-
(cod)], an efficient RuAAC catalyst, proved completely ineffec-
tive. However, replacement of the pentamethylcyclopentadien-
yl ligand with the cyclopentadienyl ligand resulted in a moder-
&
&
Chem. Eur. J. 2014, 20, 1 – 11
2
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!