On the preparation of azepinones

Article abstract of DOI:10.1016/j.tetlet.2010.08.003

We report here a simple and efficient preparation of 1H-azepin-5(2H)-ones and their unexpectedly facile isomerization to 1H-azepin-5(4H)-ones under mildly basic reaction conditions.

Full text of DOI:10.1016/j.tetlet.2010.08.003

Tetrahedron Letters 51 (2010) 5325–5327
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
On the preparation of azepinones
Galyna G. Dubinina a, Wesley Y. Yoshida a, William J. Chain a,b,
*
a Department of Chemistry, 2545 McCarthy Mall, University of Hawaii, Honolulu, HI 96822, United States
b The Cancer Research Center of Hawaii, 651 Ilalo Street, Honolulu, HI 96813, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
We report here a simple and efficient preparation of 1H-azepin-5(2H)-ones and their unexpectedly facile
isomerization to 1H-azepin-5(4H)-ones under mildly basic reaction conditions.
Ó 2010 Elsevier Ltd. All rights reserved.
Received 7 July 2010
Accepted 2 August 2010
Available online 10 August 2010
Keywords:
Azepinone
Diels–Alder
Cycloaddition
Isomerization
Ring-expansion
Synthesis
During the course of the total synthesis of a complex alkaloid,
we had need of a substituted 1H-azepin-5(2H)-one. Azepinones
are accessible via the Diels–Alder cycloaddition of electron rich
dienes and 2H-azirine-3-carboxylic acid esters (Scheme 1).1,2 These
reactions proceed under relatively mild conditions and owe their
remarkable success to the combined effects of a highly strained
carbon–nitrogen double bond and an electron-withdrawing
carboxylate group.
OTMS
N
O
N
TMSO
heptane
t-BuO
CO2t-Bu
1
H3CO
OCH3
30%
In the case of unsubstituted 2H-azirine-3-carboxylic acid esters,
it wasreported thatsomecycloadductsunderwenta strain-relieving
scissionof a carbon–nitrogen bond to producemixturesof pyridones
and azepinones.2 For example, treatment of t-butyl 2H-azirine-3-
carboxylate with 1-methoxy-3-trimethylsilyloxybutadiene3 (Dani-
shefsky’s diene) afforded the silyl enol ether cycloadduct 1. Upon
prolonged standing at ambient temperature, the silyl enol ether
decomposed to afford the corresponding azepinone-3-carboxylate
ester 2, presumably via the pathway depicted in Scheme 1. It was
later reported by the same group that treatment of analogous cyc-
loadducts with solutions of tetra-n-butylammonium fluoride (TBAF,
1 M in tetrahydrofuran, 0.3 equiv) afforded the same products.1
We wished to take advantage of this strain-relieving ring-
expansion to generate azepinones like 6 and 7 (Scheme 2). After
much experimentation, we found it most convenient to work with
ethyl 2H-azirine-3-carboxylate (3), which we generated by heating
dilute (0.07 M) solutions of ethyl 2-azidoacrylate2b,4 in dichloro-
methane to 150 °C in a sealed vessel for just over one hour (see
Supplementary data for details).5 The resultant solution of ethyl
O
N
O
H
CO2t-Bu
CO2t-Bu
N
H
H+
2
Scheme 1. Diels–Alder cycloaddition of 2H-azirine-3-carboxylate t-butyl ester
(Alves and Gilchrist, 1998).
2H-azirine-3-carboxylate (3) was charged directly with 1-meth-
oxy-3-trimethylsilyloxybutadiene (0.75 equiv) and heated to
80 °C for 40 min to afford the desired cycloadduct 4 in good yield.
We sought a means of inducing the strain-relieving bond scission
(4 ? 6) as part of a one-pot cycloaddition-ring expansion sequence
to access the azepinone directly. We were surprised to find that
treatmentof crude or purified cycloadducts 4 and 5 withcommercial
solutions of TBAF (conditions previously described by Alves)2d
yielded mixtures of isomeric azepinones. For example, when we
treated the crude cycloadduct 4 with TBAF (1 M in THF, 0.3 equiv),
we observed formation of a mixture of the desired 1H-azepin-
5(2H)-one 6, the isomerized 1H-azepin-5(4H)-one 8 (in which the
* Corresponding author. Tel.: +1 808 956 5795; fax: +1 808 956 5908.
E-mail address: chain@hawaii.edu (W.J. Chain).
0040-4039/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2010.08.003

5326
G. G. Dubinina et al. / Tetrahedron Letters 51 (2010) 5325–5327
Table 1
Isomerization of azepinones
R
OTMS
O
N
TMSO
Δ
R
O
R
EtO
O
R
CO2Et
CH2Cl2
reagent
H3CO
N
CO2Et
CO2Et
OCH3
4, R = H, 65%
5, R = CH3, 59%
3
THF, 23 °C
N
H
N
6, R = H
8, R = H
10, R = CH3
H
7, R = CH3
silica gel
Entry
Substrate
R
Reagenta
Conversionb (%)
1
2
3
4
5
6
7
8
6
7
6
7
6
7
6
7
6
7
6
7
H
CH3
H
CH3
H
CH3
H
CH3
H
CH3
H
TBAFc
TBAF
100
100
100
70
50
<5
O
O
O
R
R
NaOEt
NaOEt
CO2Et
CO2Et
CO2Et
d
O
NEt3
N
N
N
H
d
NEt3
H
H3C
PPTS
PPTS
TsOH
TsOH
TsOHf
TsOHf
15
8, R = H
10, R = CH3
6, R = H
69% (45% one-pot)
7, R = CH3
OCH3
0
11e
0
9
9
10
11
12
86% (51% one-pot)
20
Decomp.
CH3
Scheme 2. Diels–Alder cycloaddition of 2H-azirine-3-carboxylate ethyl ester (this
a
work).
Experiments conducted on 0.1 mmol scale in 1 mL THF with 1 equiv reagent.
See Supplemental data for details.
b
Determined by 1H NMR analysis after 20 h reaction time.
c
Determined by 1H NMR analysis after 3.5 h reaction time.
double bond has migrated from C3–C4 to C2–C3), and the Michael
adduct 9 incorporating 4-methoxy-3-buten-2-one in a 1:1:1.3 ratio
(1H NMR analysis).6 We have evidence to suggest the Michael adduct
arises from the conjugate addition of the cycloadduct at nitrogen to
4-methoxy-3-buten-2-one, followed by the strain-releasing bond-
scission of the resultant cation.7 Separating each component of this
mixture by column chromatography proved quite tedious, thus pre-
cluding the use of this procedure for a one-pot azepinone synthesis.
Treatment of the purified cycloadduct 4 with TBAF (0.3 equiv)
yielded the desired 1H azepin-5(2H)-one 6 (17%), and the isomer-
ized 1H-azepin-5(4H)-one 8 as the major product (60%). In the
presence of TBAF in any amount, mixtures of products were ob-
served, however we found the stoichiometry was critical to our
success. The desired azepinone isomer 6 was observed and isolated
upon treatment of the cycloadduct 4 with 0.3 equiv TBAF, provided
the reaction was closely monitored by TLC and purified immedi-
ately upon consumption of the starting material. However, when
the same reaction was conducted with 1.0 equiv TBAF under other-
wise identical conditions, it was difficult to obtain any of the
desired isomer. The rate of fluoride-mediated azepinone isomeri-
zation is competitive with fluoride-mediated ring expansion;
complete consumption of the cycloadduct was observed after
approximately 3 h at room temperature, which is sufficient time
for nearly complete azepinone isomerization to occur. We cannot
rule out the possibility that trace amounts of hydroxide in our
TBAF solutions were responsible for the isomerization, and under
otherwise identical reaction conditions, other milder sources of
fluoride (e.g., TBAF/AcOH, Et3NÁ3HF, HFÁpyridine, HFÁacetonitrile,
CsF) did not efficiently induce the desired ring expansion (<10%
conversion was observed in all cases).
After careful study, we noted that the isomerization of 6 and 7
to 8 and 10, respectively, occurred under a variety of basic condi-
tions. Our results are summarized in Table 1. Conversion of 6 to
8 was complete within 3.5 h in the presence of 1 equiv TBAF. Con-
version of 6 to 8 in the presence of stoichiometric sodium ethoxide
proceeded at a much slower rate (complete conversion after
ꢀ20 h). Azepinone 7 bearing a C4 methyl group is generally more
resistant to isomerization under the conditions we examined,
and required much longer reaction times to achieve full conver-
sion. Both 6 and 7 proved stable to acidic conditions; very little
isomerization was observed upon treatment of 6 with PPTS or p-
TsOH for several hours.8 We have not observed conversion of 8
and 10 back to 6 and 7 under any conditions we have examined.
It is also noteworthy that thermal isomerization of 6 or 7 does
d
Two equivalent NEt3 employed.
e
Determined by 1H NMR analysis after 15 h reaction time.
f
Experiment conducted at 60 °C for 6 h.
not occur at an appreciable rate in the absence of base, though
we did observe isomerization of 6 to 8 after prolonged storage
on the bench at ambient temperature.
We chose to use the stability of the azepinones to acidic reac-
tion conditions to our advantage to achieve our goal of a one-pot
azepinone synthesis. While the cycloadducts are indeed isolable
by careful column chromatography (see Supplementary data for
details), we found that prolonged exposure to silica gel cleanly
and efficiently induced the ring expansion.9 After the cycloaddition
was observed to be complete (TLC analysis) in our optimized pro-
cedure, silica gel was added directly to the crude reaction mixture.
The resultant suspension was stirred at ambient temperature
(23 °C) and then filtered to afford the 1H-azepin-5(2H)-one 6 in
good yield (45%) as a single isomer (determined by 1H NMR anal-
ysis of the crude product). This strategy worked well for the
cycloaddition-ring expansion sequence of both 1-methoxy-3-trim-
ethylsilyloxybutadiene and 1-methoxy-3-trimethylsilyloxypenta-
1,3-diene3a,10 with ethyl 2H-azirine-3-carboxylate.
We reasoned that removing the C3-carboxylate might affect the
position of the double bond. To that end, we prepared the azepa-
none 1211 (Scheme 3) from the commercially available N-Boc-4-
piperidone (11). We were able to introduce the desired unsatura-
tion via oxidation of 12 with 2-iodoxybenzoic acid (IBX) in the pres-
ence of 4-methoxypyridine N-oxide to afford the azepinone 13.12
However, treatment of this substrate with either of two different
fluoride sources induced complete isomerization of the double
bond in addition to removal of the 2-(trimethylsilyl)ethyl carba-
mate protecting group to give azepinone 14. The C3–C4 double
bond isomer was not detected. We prepared the corresponding
azepanone-4-carboxylate 15 utilizing the same strategy but we
were unable to introduce the desired unsaturation selectively to
give 16.
In summary, we have developed an efficient one-pot prepara-
tion of 1H-azepin-5(2H)-ones and noted an unexpectedly facile
isomerization of the C3–C4 double bond to C2–C3. The isomeriza-
tion occurs readily under basic reaction conditions, regardless of
substitution, and appears to be irreversible. We are currently
studying the synthesis of new azepinones with a variety of substi-
tution patterns, the effects of basic and acidic reaction conditions,

G. G. Dubinina et al. / Tetrahedron Letters 51 (2010) 5325–5327
5327
N2
References and notes
CO2Et
1.
O
O
O
IBX
MPO
BF3•OEt2
Et2O, 97%
1. (a) Anderson, D. J.; Hassner, A. Synthesis 1975, 483–495; (b) Bhullar, P.;
Gilchrist, T. L.; Maddocks, P. Synthesis 1997, 271–272.
2. (a) Alves, M. J.; Gilchrist, T. L. J. Chem. Soc., Perkin Trans. 1 1998, 299–303; (b)
Alves, M. J.; Gilchrist, T. L. Tetrahedron Lett. 1998, 39, 7579–7582; (c) Alves, M.
J.; Fortes, A. G.; Costa, F. T. Tetrahedron 2006, 62, 3095–3102; (d) Alves, M. J.;
Fortes, A. G.; Costa, F. T.; Duarte, V. C. M. Tetrahedron 2007, 63, 11167–11173.
3. (a) Danishefsky, S.; Kitahara, T.; Yan, C. F.; Morris, J. J. Am. Chem. Soc. 1979, 101,
6996–7000; (b) Danishefsky, S.; Yan, C.-F.; Singh, R. K.; Gammill, R. B.; McCurry,
P. M., Jr.; Fritsch, N.; Clardy, J. J. Org. Chem. 1979, 101, 7001–7008.
4. (a) Kakimoto, M.; Kai, M.; Kondo, K. Chem. Lett. 1982, 525–526; (b) Ghosh, S. K.;
Verma, R.; Ghosh, U.; Mamdapur, V. R. Bull. Chem. Soc. Jpn. 1996, 69, 1705–
1711.
DMSO
80 °C
25%
2. 4N HCl, Δ
N
N
N
3. triphosgene
Teoc
Boc
Teoc
TMS
HO
11
12
13
K2CO3, THF
33%, 2 steps
CsF, DMF
90°C
TBAF, THF
23 °C
or
85%
96%
5. The thermal electrocyclic ring closure of vinyl azides to azirines is often
accompanied by dimerization to give pyrroles, polymerization, and other
unproductive decomposition. Generating the azirine quickly at high
temperature and high dilution minimizes these issues and provides material
sufficient for use without further purification or isolation.
6. 4-Methoxy-3-buten-2-one is present in small amounts in either commercially
available or freshly prepared 1-methoxy-3-trimethylsilyloxybutadiene, and is
undoubtedly produced upon treatment of our crude reaction mixtures with
TBAF.
O
O
O
CO2Et
CO2Et
N
N
N
H
Teoc
Teoc
14
15
16
7. Treatment of either the azepinones 6 or 8 with 4-methoxy-3-buten-2-one in
the presence of TBAF produces Michael adduct
9 in only trace amounts.
Scheme 3. Preparation of azepinones by unsaturation.
Formation of this side product is completely suppressed by employing the
azirine in excess and the diene as limiting reagent in the cycloaddition reaction.
8. Direct treatment of the cycloadducts with p-TsOH resulted in ring opening
products incorporating tosylate. See Supplementary data. An analogous
reaction was reported with hydrochloric acid. See Ref. 2c
and the use of azepinones as nucleophiles at nitrogen and at car-
bon. These studies will be disclosed in due course.
9. Silica gel is mentioned once in the discussion of a prior report as inducing the
desired transformation, but no experimental procedure was provided. See Ref.
2d. General procedure for the one-pot cycloaddition-ring expansion—ethyl 5-oxo-
2,5-dihydro-1H-azepine-3-carboxylate (6): A solution of ethyl 2-azidoacrylate.2b,4
(1.41 g, 10.0 mmol, 1 equiv) in dichloromethane (150 mL, 0.07 M) was heated to
150 °C in a sealed tube for 75 min, then was cooled to 23 °C whereupon 1-
methoxy-3-trimethylsilyloxybutadiene.3 (1.30 g, 7.50 mmol, 0.75 equiv) was
added. The resultant orange solution was heated to 80 °C for 40 min, then was
cooled to 23 °C whereupon 10 g silica gel was added. The resultant suspension
was stirred at 23 °C for 18 h, then was filtered. The silica gel was washed with
ethyl acetate (2 Â 30 mL) and the combined filtrates were concentrated. The
residue was purified by flash column chromatography (50% ethyl acetate–
hexanes) to give the azepinone 6 (610 mg, 45% yield) as a yellow semisolid. TLC
(ethyl acetate): Rf = 0.23 (UV, KMnO4). 1H NMR (500 MHz, CDCl3), d: 7.13 (d,
J = 2.0 Hz, 1H), 7.07 (dd, J1 = 8.1 Hz, J2 = 6.8 Hz, 1H), 5.97 (br, 1H), 5.30 (d, J = 8.4
Acknowledgments
Financial support from the University of Hawai’i and the Cancer
Research Center of Hawai’i is gratefully acknowledged. The authors
thank Professors Tius, Hemscheidt, Williams, Navarro, and Vicic
and their respective groups (UH) for generous donations of equip-
ment and chemicals, and for helpful discussions. The authors thank
W. Niemczura (UH) for mass spectrometry analyses and Professor
David Horgen and Mr. J. Reinicke (Hawai’i Pacific University) for
NMR assistance.
Hz, 1H), 4.26 (q, J = 7.1 Hz, 2H), 4.14 (d, J = 5.0 Hz, 2H), 1.32 (t, J = 7.1 Hz, 3H). 13
NMR (125 MHz, CDCl3), d: 189.8, 165.7, 147.8, 142.1, 129.8, 104.6, 61.8, 42.7,
14.1. FTIR (NaCl, thin film), cmÀ1: 3221, 2985, 1712, 1639. HRMS (EI+): Calcd for
C9H11NO3 [M+H]+: 181.0739. Found: 181.0736.
C
Supplementary data
10. We prepared 1-methoxypent-1-en-3-one by a one-step procedure: Burger, M.
T.; Still, W. C. J. Org. Chem. 1996, 61, 775–777.
11. (a) Dolle, R. E.; Le Bourdonnec, B.; Chu, G.-H. WO2006/105442, 2006.; (b)
Igarashi, J.; Kobayashi, Y. Tetrahedron Lett. 2005, 46, 6381–6384.
12. Nicolaou, K. C.; Montagnon, T.; Baran, P. S. Angew. Chem., Int. Ed. 2002, 41, 993–
996.
Supplementary data (detailed experimental procedures and
spectral data) associated with this article can be found, in the on-
line version, at doi:10.1016/j.tetlet.2010.08.003.