gave a good yield (entry 7). The reaction rate was consider-
ably lower when using aniline (entry 8) or secondary amine
(entry 9), probably due to a lower nucleophilicity and a steric
hindrance, and the reaction did not proceed at all in the
presence of tertiary amines (entries 10 and 11). The other
nucleophiles (t-BuOK or Bu3P) did not afford 5a, leading
to the formation of rather complicated mixture (entries 12
and 13). Under optimal conditions, the effects of substituents
on the 1-aryl group were examined (entries 14-18). It was
confirmed that nitrogen extrusion occurred in nearly quan-
titative yields, regardless of electron-donating or -withdraw-
ing substituents. The only exception was the case of the
2-tolyl substrate 4c (entry 15), which gave the indenone 5c
in 47% yield after a prolonged reaction time. The discontinu-
ity of the reactivity observed here is considered to originate
in the steric hindrance of the ortho substituent, suggesting
that the key interaction between diazepine and amine leading
to nitrogen extrusion takes place at or around the 1-position.
Scheme 3
However, gentle heating of this solution led to conversion to
indenone 5, suggesting the thermal instability of adduct 6,
probably due to ready release of the stable nitrogen molecule.8
In this connection, the same adduct formation was likely to
occur in a less polar C6D6 solvent at room temperature (Figure
1c,d), although thermolysis gave a somewhat complex mixture
in this case (vide supra). After extrusion of nitrogen, BnNH2
was regenerated to release the indenone product 5 and achieve
a catalytic cycle (Scheme 3). Although examples of nitrogen
extrusion from a 2,3-benzodiazepine nucleus have been found
in the literature, these reactions have only been described as
minor reaction modes of several degradation and isomerization
paths under thermolytic or photolytic conditions.9 It is important
to note that indenone 5 did not produce its adduct form 7 in a
reversible manner. Thus, the amine-adduct formation at ambient
temperature is a novel characteristic of diazabenzotropone
derivatives 4.10
With the observation of unique reactivity of diazaben-
zotropones 4 toward nucleophiles, our interest turned to their
biological behaviors. Their highly electron-deficient character
was associated with the ability of a strong interaction with
electron-donating substances in living cells. In fact, we could
observe an instantaneous reaction of 4a with N-Boc-cysteine
methyl ester as an example of a biomolecule component, in a
1H NMR analysis. Although this did not guarantee a bioactivity
of diazabenzotropones, we considered that examinations of
biological activities of 4 may be meaningful. We have recently
reported intracellular oxidative stress-triggered apoptosis of
human lymphoma cells, which is more sharply induced by small
molecule natural products (macrosphelides) having a higher
oxidation state (i.e., higher electron deficiency) rather than the
relatively lower oxidation state.11 Aside from the structural
irrelevancy to these natural compounds, the diazabenzotropones
Figure 1.
1H NMR spectra of 4a before and after addition of
benzylamine.
The seven-membered ring of the 2,3-benzodiazepin-5-one
derivatives 4 may be highly electron-deficient because of the
presence of two carbonyl groups and two electronegative
nitrogen atoms. This is likely to make these compounds highly
(8) Recently, decomposition of 1,2,3-triazolines with loss of nitrogen to
form aziridines has been reported; see: Kim, S.; Lee, Y. M.; Lee, J.; Lee, T.;
Fu, Y.; Song, Y.; Cho, J.; Kim, D. J. Org. Chem. 2007, 72, 4886–4891.
(9) (a) Reid, A. A.; Sood, H. R.; Sharp, J. T. J. Chem. Soc., Perkin
Trans. 1 1976, 362–366. (b) Munro, D. P.; Sharp, J. T. J. Chem. Soc., Perkin
Trans. 1 1980, 1718–1723. (c) Reid, A. A.; Sharp, J. T.; Murray, S. J.
J. Chem. Soc., Chem. Commun. 1972, 827.
1
reactive toward nucleophilic reagents. Actually, H NMR
analyses of 4a (+ BnNH2) demonstrated quantitative formation
of an addition compound (Figure 1). Addition of a slight excess
of BnNH2 to a solution of 4a in CD3OD resulted in the complete
disappearance of 4a, new benzyl methylene protons with a
geminal coupling (J ) 15 Hz) were seen (3.5-3.7 ppm), and
the methylene protons of ethyl ester lost equivalence (Figure
1a,b). We presumed that these spectral changes were caused
by the formation of addition compound 6 (Scheme 3).7 This
transformation was completed instantaneously at room tem-
perature, and no changes were observed after a further 1 h.
(10) Less electron deficient derivatives 8 and 9, which were prepared
by NaBH4 reduction of 4a (see the Supporting Information), did not
participate in the nitrogen extrusion reaction under the optimal conditions.
(11) (a) Ahmed, K.; Zhao, Q.-L.; Matsuya, Y.; Yu, D.-Y.; Salunga, T. L.;
Nemoto, H.; Kondo, T. Int. J. Hyperthermia 2007, 23, 353–361. (b) Ahmed,
K.; Zhao, Q.-L.; Matsuya, Y.; Yu, D.-Y.; Feril, L. B., Jr; Nemoto, H.;
Kondo, T. Chem. Biol. Interact. 2007, 170, 86–99. (c) Matsuya, Y.;
Kawaguchi, T.; Ishihara, K.; Ahmed, K.; Zhao, Q.-L.; Kondo, T.; Nemoto,
H. Org. Lett. 2006, 8, 4609–4612.
(7) The adduct 6 is likely to be stable only in a solution. Attempts to
isolate 6 ended in failure because regeneration of the diazepine 4 took place
during the isolation process.
Org. Lett., Vol. 11, No. 6, 2009
1363