Synthesis of indazole-N-oxides via the 1,7-electrocyclisation of azomethine ylides

Article abstract of DOI:10.1016/S0040-4039(01)00892-9

The first example of the 1,7-electrocyclisation of a non-stabilised azomethine ylide, e.g. 2, onto a nitro group, to give a 1,2,6-oxadiazepine intermediate, e.g. 4, is reported. Subsequent ring contraction results in the formation of the indazole-N-oxides 5, 12, and 14.

Full text of DOI:10.1016/S0040-4039(01)00892-9

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 42 (2001) 5081–5083
Synthesis of indazole-N-oxides via the 1,7-electrocyclisation of
azomethine ylides
Miklo´s Nyerges,a,* Imre Fejes,a Andrea Vira´nyi,a Paul W. Groundwaterb and La´szlo´ Tokea
3
aResearch Group of the Hungarian Academy of Sciences, Department of Organic Chemical Technology,
Technical University of Budapest, PO Box 91, H-1521 Budapest, Hungary
bInstitute of Pharmacy and Chemistry, School of Sciences, University of Sunderland, Sunderland SR1 3SD, UK
Received 20 April 2001; revised 17 May 2001; accepted 24 May 2001
Abstract—The first example of the 1,7-electrocyclisation of a non-stabilised azomethine ylide, e.g. 2, onto a nitro group, to give
a 1,2,6-oxadiazepine intermediate, e.g. 4, is reported. Subsequent ring contraction results in the formation of the indazole-N-
oxides 5, 12, and 14. © 2001 Elsevier Science Ltd. All rights reserved.
In the search for new syntheses of the pyrrolo[3,2-
c]quinoline ring system of the bradykinin receptor
antagonist martinelline alkaloids1 we have studied the
formation and reactions of the non-stabilised azome-
thine ylide 2 formed in the reaction of o-nitrobenzalde-
hyde 1 with sarcosine in refluxing benzene. To our
surprise, in spite of the presence of a large excess of
active dipolarophiles, such as ethyl acrylate or methyl
Scheme 1.
Keywords: cycloadditions; electrocyclic reactions; indazoles; isoquinolines; nitro compounds.
* Corresponding author.
0040-4039/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(01)00892-9

5082
M. Nyerges et al. / Tetrahedron Letters 42 (2001) 5081–5083
vinyl ketone, we could not observe any trace of the
expected cycloadducts 3 in the H NMR spectrum of
the crude reaction mixture. However, two products, an
indazole-N-oxide 5 and an oxazolidine 7 were isolated
after chromatographic separation (in 40 and 43%
yields, respectively) and their structures confirmed by
spectroscopic analysis.
linium chloride 105 by dehydrohalogenation with tri-
ethylamine (Scheme 3).6 In the presence of N-phenyl-
maleimide a 2:1 mixture of the cycloadduct 11 (as single
1
1
isomer, proven by H NOE experiments) and indazole-
N-oxide 12 was obtained, while in the absence of the
dipolarophile the N-oxide 12 was formed in quantita-
tive yield. In contrast to the example described above,
in this case the aldehyde arising from the fragmentation
(which in this example is attached to the indazole
component) is not sufficiently reactive to act as a
dipolarophile in a 1,3-dipolar cycloaddition process in
competition with the 1,7-electrocyclisation.
The formation of these two compounds is probably due
to the fragmentation of the unstable intermediate 4
shown in Scheme 1, in which the decarboxylative
condensation2 of o-nitrobenzaldehyde 1 and sarcosine
is followed by a 1,7-electrocyclisation3 of the non-sta-
bilised azomethine ylide 2. As such, this represents the
first 1,7-electrocyclisation of an azomethine ylide onto a
nitro group. The seven-membered ring of 4 subse-
quently undergoes a ring contraction, resulting in the
elimination of formaldehyde and the production of
2-methyl-2H-indazol-N-oxide 5. The formaldehyde by-
product is then able to react with the excess sarcosine
present in the reaction mixture, resulting in the forma-
tion of azomethine ylide 6. This dipole could then react
with the other starting material, o-nitrobenzaldehyde 1,
to yield the 3-methyl-5-aryloxazolidine 7 as the second
product.
With regard to the proposed mechanism we performed
the next series of experiments with 6,7-diethoxy-3,4-
dihydro-1-(2-nitrophenyl)-N-substituted-isoquinolinium
bromides (prepared from the corresponding halide and
3,4-dihydroisoquinoline7) (Scheme 4). In all cases the
isoquinoline fused indazole-N-oxide was formed. In
one case (13c R=CO2CH3) the competitive formation
of the 1,3-dipolar cycloadduct 15 as a single isomer
1
(proven again by H NOE experiments) was observed
(ratio of 14:15 is approx. 3:1) due to the high reactivity
of the electron-deficient CꢀO double bond of the by-
product aldehyde.
A similar process has been reported for the 1,7-electro-
cyclisation of the azomethine imine 8, leading to the
formation of the benzotriazole-N-oxide 9 (Scheme 2).4
Acknowledgements
This work was financially supported by the National
Foundation for Science and Research, Hungary
(OTKA Project No. F 029198 and T 032221). N.M.
We next chose to form the azomethine ylides from 6,7-
dimethoxy - 3,4 - dihydro - N - (2 - nitrobenzyl)isoquino-
Scheme 2.
Scheme 3. Reagents and conditions: (i) N-phenylmaleimide, Et3N, MeOH, rt; (ii) Et3N, MeOH, rt.

M. Nyerges et al. / Tetrahedron Letters 42 (2001) 5081–5083
5083
Scheme 4. Reagents and conditions: (i) Et3N, MeOH, rt.
thanks the Hungarian Academy of Sciences for a
Bolyai J. fellowship.
3. (a) Groundwater, P. W.; Nyerges, M. Adv. Heterocyclic
Chem. 1999, 73, 97; (b) Arany, A.; Bendell, D.; Ground-
water, P. W.; Garnett, I.; Nyerges, M. J. Chem. Soc.,
Perkin Trans. 1 1999, 2605.
4. Grashey, R. Angew. Chem., Int. Ed. Engl. 1962, 1, 158.
References
5. Shamma, M.; To
6. Fejes, I.; Nyerges, M.; To
2000, 56, 623.
3
ke, L. Tetrahedron 1975, 31, 1991.
3
ke, L.; Pak, C. S. Tetrahedron
1. Fejes, I.; Nyerges, M.; To3 ke, L. Tetrahedron Lett. 2000,
40, 7951.
´
7. Cortes, E. C.; Romere, E. C.; Ramirez, G. F. J. Hetero-
cyclic Chem. 1994, 31, 1425.
2. Tsuge, O.; Kanemasa, S.; Ohe, M.; Takenaka, S. Bull.
Chem. Soc. Jpn. 1987, 60, 4079.
.