Indium-Catalyzed Amide Allylation of N-Carbonyl Imides: Formation of Azaspiro-γ-lactones via Ring Opening-Reclosure

Article abstract of DOI:10.1021/acs.orglett.5b03021

A novel and facile synthesis of azaspiro-γ-lactones with a methylene-lactam framework from N-carbonyl imides is described. Mechanistic investigations provide evidence for a two-step reaction process involving ZnCl₂-promoted addition of β-amido allylindium species followed by an unexpectedly molecular-sieves-mediated ring opening-reclosure concomitantly with the loss of an N-carbonyl unit.

Full text of DOI:10.1021/acs.orglett.5b03021

Letter
pubs.acs.org/OrgLett
Indium-Catalyzed Amide Allylation of N‑Carbonyl Imides: Formation
of Azaspiro-γ-lactones via Ring Opening−Reclosure
Tetsuya Sengoku,† Yusuke Murata,‡ Yuwa Aso,† Ai Kawakami,† Toshiyasu Inuzuka,§ Masami Sakamoto,∥
Masaki Takahashi,† and Hidemi Yoda*,†,‡
†Department of Applied Chemistry, Faculty of Engineering, Shizuoka University, 3-5-1 Johoku, Naka-ku, Hamamatsu 432-8561,
Japan
‡Graduate School of Science and Technology, Shizuoka University, 3-5-1 Johoku, Naka-ku, Hamamatsu 432-8561, Japan
§Division of Instrumental Analysis, Life Science Research Center, Gifu University, 1-1 Yanagido, Gifu 501-1193, Japan
∥Department of Applied Chemistry and Biotechnology, Graduate School of Engineering, Chiba University, 1-33 Yayoi-cho, Inage-ku,
Chiba 263-8522, Japan
S
* Supporting Information
ABSTRACT: A novel and facile synthesis of azaspiro-γ-
lactones with a methylene−lactam framework from N-carbonyl
imides is described. Mechanistic investigations provide
evidence for a two-step reaction process involving ZnCl2-
promoted addition of β-amido allylindium species followed by
an unexpectedly molecular-sieves-mediated ring opening−
reclosure concomitantly with the loss of an N-carbonyl unit.
work of which is found in natural products possessing
ucleophilic allylation of aldehydes and ketones is
Nundoubtedly an important carbon−carbon bond forming significant bioactivities and thus has been attracting much
attention from synthetic chemists as a challenging synthetic
target.11 Here we report a first example of catalytic amide
allylation with N-carbonylimides to synthesize azaspiro-γ-
lactone derivatives.
reaction, and a number of methods including enantioselective
variants have been reported.1 Among a variety of allylating
reagents, allylsilanes,2 allylstannanes,3 and allylboronates4 have
found widespread use in syntheses of bioactive compounds
because of their ready availability and ease of handling. In
recent years, reactions using alternative metal reagents such as
allylindium5 and allylgallium species6 have been developed.
They are generally prepared in situ by oxidative insertion of
low-valent metallic species into allyl halides or transmetalation
with metal halides, allowing for chemoselective transformations
of polyfunctionalized organic compounds.
In addition to conventional reactions using simple allylmetal
reagents, allylation with functionalized ones has been of great
importance due to broad utility in complex molecule
synthesis.7−9 In this context, we have reported the
enantioselective synthesis of diversely substituted β-amido
homoallyl alcohols 2 as a key precursor of bioactive α-
methylene-γ-butyrolactones via indium-catalyzed allylation of
aldehydes with β-amido allyltributylstannanes 1 (Scheme 1).8
More recently, we have extended this reaction methodology to
a wide range of isatin derivatives and acyclic α-ketoesters, where
extremely high enantioselective synthesis of antineoplastic
spirooxindole and its derivatives has been achieved.9
Apparently, such amide allylation of carbonyl compounds
offers a great advantage in the synthesis of potential drug
candidates; however, no examples using imide derivatives have
been reported in the literature to date.10 Hence, it is a great
challenge to extend our studies with imides 4 in order to
provide new azaspiro-γ-lactone II, the azaspirobicyclic frame-
We initially investigated amide allylation of N-ethoxycarbonyl
maleimide 4a with N-phenyl-β-amido allylstannane 1a in the
presence of an excess amount of 4 Å molecular sieves (MS 4 Å)
in CH2Cl2. When the reaction was conducted with 20 mol % of
Yb(OTf)3 or Sc(OTf)3, no reaction occurred. In contrast, use
of In(OTf)3 provided a new product 5a in 54% yield after 1 day
of stirring at room temperature (Scheme 1c).12 Alternatively,
the reaction with InCl3 (20 mol %) also gave the same product
at an analogous level of product yield (55%). To our surprise,
1H and 13C NMR data for 5a showed that the ethoxycarbonyl
group originating from 4a was completely lost, indicating that
the expected compound I was not formed. High resolution
mass spectral data for 5a (m/z [M]+ 241.0742) supported these
observations, indicating molecular formula C14H11NO3. Assign-
ment with the help of DEPT experiments suggested that 5a
contains two lactam/lactone carbons (δc 167.5 and 169.4 ppm)
and one quaternary hemiaminal carbon (δc 99.0 ppm).
Furthermore, NOE experiments showed that a lactam ring is
fused to an α,β-unsaturated-γ-lactone. Thus, the complete
structure of 5a was revealed to be 8-methylene-6-phenyl-1-oxa-
6-azaspiro[4.4]non-3-ene-2,7-dione as shown in Scheme 1.
Although 5a was not our target molecule I or II, this
Received: October 19, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.5b03021
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
Scheme 1. Indium-Catalyzed Amide Allylation of Carbonyl
Compounds with β-Amido Allylstannanes
Table 1. Metal and Solvent Effects for Azaspiro-γ-lactone
Formation from 4a
entry InCl3 (mol %) additive (mol %)
solvent
t (day) yield (%)
a
1
20
20
20
20
20
20
20
20
20
0
−
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
toluene
1
4
2
2
2
2
1
2
2
3
2
3
1
1
55
72
57
68
24
29
91
71
80
17
37
49
83
91
2
MgCl2 (20)
CaCl2 (20)
FeCl3 (20)
CuCl2 (20)
SnCl4 (20)
ZnCl2 (20)
ZnBr2 (20)
ZnI2 (20)
ZnCl2 (20)
ZnCl2 (10)
ZnCl2 (20)
ZnCl2 (20)
ZnCl2 (20)
3
4
5
6
7
8
9
10
11
12
13
14
10
20
20
20
b
MeCN
c
CH2Cl2
a
The reaction under the given conditions with a prolonged reaction
b
time of 2 days afforded an 80% yield of 5a. Reaction was carried out
with MS 3 Å. Reaction was carried out with 1.2 equiv of 1a.
c
(Table 2, entries 5 and 6). From these results, it should be
emphasized that a wide range of carbonyl functions can serve as
a leaving group. Next, we focused our attention on variation of
substitution of the nitrogen group in 1. When N-aryl (p-tolyl, p-
anis, 4-tert-butylphenyl, 4-chlorophenyl, 2,6-xylyl, and 1-
naphthyl) derivatives 1b−g were treated with 4a under the
given conditions, 5b−g13 were produced in 89−100% yields,
respectively (Table 2, entries 7−12). Alkyl substituents such as
methyl and n-pentyl groups led to substantial loss of reactivity
(55−67% yields) (Table 2, entries 13 and 14), whereas more
sterically demanding tert-butyl group hampered this trans-
formation, resulting in the isolation of amidoallyl adduct 6a in
49% yield (Table 2, entry 15). These results strongly suggested
that 5 would be produced from the amidoallyl intermediate 6
via ring opening−reclosure concomitantly with loss of an N-
carbonyl unit.
We then examined the substrate scope of the azaspiro-γ-
lactone formation. Reaction of N-ethoxycarbonyl succinimide,
prepared from succinimide and ethyl chloroformate, with 1b
was examined, but a complex mixture of products was obtained
presumably due to relatively poor electrophilicity of imide
carbonyls. Meanwhile, investigations with substituted malei-
mides 4g and 4h showed that the efficiency of azaspiro-γ-
lactone formation is strongly dependent on the substitution
pattern of imide derivatives. Indeed, monosubstituted mal-
eimide derivative 4g gave a 93% yield of 5j as a 2:1 mixture of
regioisomers, although a longer reaction time was required
(Table 2, entry 16). As for 4h, the reaction failed to give any
spirocyclic product and the corresponding amidoallyl adduct 6b
was predominantly generated in 82% yield (Table 2, entry 17).
Further investigations for extending the reaction scope revealed
that N-ethoxycarbonyl phthalimide 4i reacted with 1a, albeit
sluggishly, leading to an 82% yield of 5k whose structure was
unambiguously confirmed by single crystal X-ray crystallo-
graphic analysis (Table 2, entry 18).14 The reaction of 4i with
other stannyl reagents 1b,c,e,f also successfully gave the
transformation can be regarded as an efficient approach
allowing access to novel azaspiro-γ-lactones with a methyl-
ene−lactam framework. We therefore shifted our focus to
establish the synthetic protocol which will provide new
opportunities to develop these pharmaceutically attractive
molecules.
Our next investigation of azaspiro-γ-lactone formation
focused on the reaction of 4a with 1a in the presence of
InCl3 and various additives. In the absence of additives, 5a was
given in 80% yield even with a prolonged reaction time of 2
days (Table 1, entry 1). Use of MgCl2, CaCl2, CuCl2, FeCl3, or
SnCl4 along with InCl3 proved to be ineffective under
comparable conditions (Table 1, entries 2−6). While ZnBr2
or ZnI2 offered no improvement in the reaction rate (Table 1,
entries 8 and 9), the reaction with ZnCl2 (20 mol %) was
completed within only 1 day to afford the desired product in
91% yield (Table 1, entry 7). The absence of InCl3 or decreased
catalyst loadings drastically deteriorated the reaction efficiency
(Table 1, entries 10 and 11). Further solvent screening showed
that dichloromethane gave the best result in comparison to
toluene and acetonitrile (Table 1, entries 12 and 13). Thus, we
could identify the optimal conditions which involve the use of
20 mol % of InCl3 and ZnCl2 and 0.5 g/mmol of MS 4 Å in
CH2Cl2. The efficient reaction was reproduced under these
conditions even when almost the same equivalent of 1a was
employed (Table 1, entry 14).
Table 2 shows the reaction of variously substituted imides 4
with amido allylstannanes 1. By applying the optimum
conditions used for entry 1, all the N-alkoxycarbonyl
maleimides 4b−d afforded the desired azaspiro-γ-lactone 5a
in good to excellent yields (71−91%), respectively (Table 2,
entries 2−4). Surprisingly, substrates containing N-acetyl and
N-phenylcarbamoyl groups 4e,f could also be tolerable for this
transformation, giving 5a in diminished yields, respectively
B
DOI: 10.1021/acs.orglett.5b03021
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
Table 2. Synthesis of Variously Substituted Azaspiro-γ-
lactones 5
entry
4
R1, R2, R3
1
X
t (day) 5 (yield %)
1
4a
4b
4c
4d
4e
4f
OEt, H, H
OtBu, H, H
OMe, H, H
OBn, H, H
Me, H, H
NHPh, H, H
OEt, H, H
OEt, H, H
OEt, H, H
OEt, H, H
OEt, H, H
OEt, H, H
OEt, H, H
OEt, H, H
OEt, H, H
OEt, Me, H
OEt, Me, Me
−
1a
1a
1a
1a
1a
1a
1b
1c
1d
1e
1f
Ph
Ph
Ph
Ph
Ph
Ph
1
1
1
1
1
4
1
2
3
1
1
1
4
4
3
8
3
3
3
3
3
3
5a (91)
5a (87)
5a (71)
5a (91)
5a (59)
5a (18)
5b (100)
5c (99)
5d (96)
5e (89)
5f (99)
5g (94)
5h (55)
5i (67)
2
3
4
5
6
7
4a
4a
4a
4a
4a
4a
4a
4a
4a
4g
4h
4i
p-tolyl
8
p-anis
9
4-tBu-Ph
4-Cl-Ph
2,6-xylyl
10
11
12
13
14
15
16
17
18
19
20
21
22
a
1g
1h
1i
Nap
Me
npentyl
tBu
Figure 1. Mechanistic investigations of azaspiro-γ-lactone formation.
b
1j
−
c
1b
1b
1a
1b
1c
1e
1f
p-tolyl
p-tolyl
Ph
5j (93)
d
−
5k (82)
5l (55)
5m (59)
5n (70)
5o (76)
that the acidic and/or basic nature of molecular sieves16 would
be essential for efficient ring opening−reclosure, for which the
mechanistic pathway is tentatively shown in Figure 1e. By
combining all of these observations, we concluded that this
unique transformation from imide derivatives 4 into azaspiro-γ-
lactones with a methylene−lactam framework 5 involves two
steps, that is, ZnCl2-promoted amide allylation of 4 with
allylindium species generated by transmetalation and sub-
sequent molecular-sieves-mediated ring opening−reclosure
concomitantly with loss of an N-carbonyl unit.
4i
−
−
−
p-tolyl
p-anis
4-Cl-Ph
2,6-xylyl
4i
4i
4i
−
a
b
Nap = 1-naphthyl. Amidoallyl adduct 6a (for structure, see Scheme
1, I; R = COOEt, R′ = H, X = tBu) was obtained as a sole product in
c
49% yield. 5j and its regioisomer 5j′ (R2 = H, R3 = Me, X = p-tolyl)
d
were obtained as an inseparable 2:1 mixture. Amidoallyl adduct 6b
(for structure see Scheme 1, I; R = COOEt, R′ = Me, X = p-tolyl) was
obtained in 82% yield.
In conclusion, we have demonstrated the novel synthesis of
azaspiro-γ-lactones with an exo-methylene−lactam ring via
amide allylation of imide derivatives. This work represents a
unique synthetic entry to provide the new class of azaspirocycle
as well as an obvious expansion in the scope of the amide
allylation and will open new opportunities for future develop-
ment of potential drug candidates. Investigations including
detailed mechanistic aspects and the enantioselective variant of
this molecular-sieves-mediated ring opening−reclosure are
underway in our laboratory.
corresponding products 5l−o in 55% to 76% yields,
respectively.
Finally, mechanistic investigations were undertaken to
understand the role of additives. As in the case of trans-
metalation of allyltributylstannane with InBr3,5d resonances of
exomethylene protons of the sample, prepared by treatment of
1a with InCl3 (1 equiv) for 30 min in dichloromethane-d2, were
significantly shifted downfield relative to those of 1a (Ha and
Hb), suggesting formation of a new allylindium species (Figure
1a,b).15 Subsequent nucleophilic attack of indium species thus
prepared was dramatically promoted by using an equimolar
amount of ZnCl2, resulting in complete consumption of 4a
within 6 h (Figure 1c). We then examined this unexpected
transformation with ring opening−reclosure of 6 into 5. During
the course of the substrate screening described in Table 2, we
fortunately found that the intermediate 6c was isolated in 47%
yield at the reaction time of 1 day. When 6c was treated only
with 0.5 g/mmol of MS 4 Å, the reaction smoothly proceeded
to afford 5k in 84% yield. Use of MS 4 Å and ZnCl2 still
resulted in the predominant formation of the product. In
contrast, reaction in the absence of molecular sieves gave 5k in
poor yields along with recovery of the starting material, despite
employing InCl3 and/or ZnCl2 (Figure 1d). These results show
ASSOCIATED CONTENT
* Supporting Information
■
S
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.or-
glett.5b03021.
Experimental details and characterization data for all new
compounds (PDF)
X-ray structure of 5k (CIF)
AUTHOR INFORMATION
Corresponding Author
■
*E-mail: tchyoda@ipc.shizuoka.ac.jp.
C
DOI: 10.1021/acs.orglett.5b03021
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
refinement cycles on F2, the model converged at R1 = 0.0379 (I >
2ρ(I)), wR2 = 0.0969, and GOF = 1.061 for 3171 reflections and 199
parameters (CCDC deposition number 1431150).
Notes
The authors declare no competing financial interest.
(15) The resonance assignable to allylic protons of the indium
species (δ 2.63 ppm) was also shifted downfield relative to that of 1a
(δ 2.05 ppm) (Figure S2, Supporting Information).
(16) (a) Adinolfi, M.; Barone, G.; Iadonisi, A.; Schiattarella, M.
Tetrahedron Lett. 2003, 44, 4661−4663. (b) Sen, S. E.; Zhang, Y.;
Roach, S. L. J. Org. Chem. 1996, 61, 9534−9537. (c) Asakura, N.;
Hirokane, T.; Hoshida, H.; Yamada, H. Tetrahedron Lett. 2011, 52,
534−537.
ACKNOWLEDGMENTS
■
This work was supported in part by a Grant-in-Aid for Scientific
Research from the Ministry of Education, Culture, Sports,
Science and Technology, Japan.
REFERENCES
■
(1) Yamamoto, Y.; Asao, N. Chem. Rev. 1993, 93, 2207−2293.
(2) (a) Ramachandran, P. V.; Nicponski, D. R.; Gagare, P. D. In
Comprehensive Organic Synthesis, 2nd ed.; Knochel, P., Ed.; Elsevier:
Amsterdam, 2014; Vol. 2, pp 79−124. (b) Chabaud, L.; James, P.;
Landais, Y. Eur. J. Org. Chem. 2004, 2004, 3173−3199. (c) Barbero, A.;
Pulido, F. J. Acc. Chem. Res. 2004, 37, 817−825. (d) Barbero, A.;
Pulido, F. J.; Carmen Sanudo, M. Beilstein J. Org. Chem. 2007, 3, 16.
(3) (a) Ramachandran, P. V.; Nicponski, D. R.; Gagare, P. D. In
Comprehensive Organic Synthesis, 2nd ed.; Knochel, P., Ed.; Elsevier:
Amsterdam, 2014, Vol. 2, pp 129−135. (b) Barbero, A.; Pulido, F. J.
Chem. Soc. Rev. 2005, 34, 913−920.
(4) (a) Ramachandran, P. V.; Gagare, P. D.; Nicponski, D. R. In
Comprehensive Organic Synthesis, 2nd ed.; Knochel, P., Ed.; Elsevier:
Amsterdam, 2014, Vol. 2, pp 1−71. (b) Hoffmann, R. W. Pure Appl.
Chem. 1988, 60, 123−130. (c) Huo, H.−X.; Duvall, J. R.; Huang, M.−
Y.; Hong, R. Org. Chem. Front. 2014, 1, 303−320.
(5) (a) Paquette, L. A.; Mitzel, T. M.; Issac, M. B.; Crasto, C. F.;
Schomer, W. W. J. Org. Chem. 1997, 62, 4293−4301. (b) Chan, T. H.;
Yang, Y. J. Am. Chem. Soc. 1999, 121, 3228−3229. (c) Hirashita, T.;
Daikoku, Y.; Osaki, H.; Ogura, M.; Araki, S. Tetrahedron Lett. 2008, 49,
5411−5413. (d) Yasuda, M.; Haga, M.; Baba, A. Organometallics 2009,
28, 1998−2000. (e) Behr, J.−B.; Hottin, A.; Ndoye, A. Org. Lett. 2012,
14, 1536−1539. (f) Suzuki, I.; Kiyokawa, K.; Yasuda, M.; Baba, A. Org.
Lett. 2013, 15, 1728−1731.
(6) (a) Araki, S.; Horie, T.; Kato, M.; Hirashita, T.; Yamamura, H.;
Kawai, M. Tetrahedron Lett. 1999, 40, 2331−2334. (b) Tsuji, T.;
Usugi, S.; Yorimitsu, H.; Shinokubo, H.; Matsubara, S.; Oshima, K.
Chem. Lett. 2002, 31, 2−3. (c) Hayashi, S.; Hirano, K.; Yorimitsu, H.;
Oshima, K. Org. Lett. 2005, 7, 3577−3579.
(7) (a) Yamaguchi, R.; Mochizuki, K.; Kozima, S.; Takaya, H. Chem.
Lett. 1994, 1809−1812. (b) Thomas, E. J. Chem. Rec. 2007, 7, 115−
124. (c) Burns, N. Z.; Baran, P. S. Angew. Chem., Int. Ed. 2008, 47,
205−208.
(8) (a) Suzuki, T.; Sengoku, T.; Takahashi, M.; Yoda, H. Tetrahedron
Lett. 2008, 49, 4701−4703. (b) Suzuki, T.; Atsumi, J.; Sengoku, T.;
Takahashi, M.; Yoda, H. J. Organomet. Chem. 2010, 695, 128−136.
(9) (a) Murata, Y.; Takahashi, M.; Yagishita, F.; Sakamoto, M.;
Sengoku, T.; Yoda, H. Org. Lett. 2013, 15, 6182−6185. (b) Takahashi,
M.; Murata, Y.; Yagishita, F.; Sakamoto, M.; Sengoku, T.; Yoda, H.
Chem. - Eur. J. 2014, 20, 11091−11100. (c) Takahashi, M.; Murata, Y.;
Ishida, M.; Yagishita, F.; Sakamoto, M.; Sengoku, T.; Yoda, H. Org.
Biomol. Chem. 2014, 12, 7686−7689.
(10) For the allylation of imide derivatives with conventional
allyltributylstannanes, see: Araki, S.; Shimizu, T.; Johar, P. S.;
Butsugan, Y. J. Org. Chem. 1991, 56, 2538−2542.
(11) Sinibaldi, M. E.; Canet, I. Eur. J. Org. Chem. 2008, 2008, 4391−
4399.
(12) N-Phenyl maleimide instead of an N-carbonyl unit gave no
reaction upon treatment with 1a in the presence of In(OTf)3 or InCl3.
(13) The NMR spectra of azaspiro-γ-lactone 5g derived from 4a and
1g displayed two sets of signals (ca. 2:1), indicating the element of
chirality attributed to axial chirality must be present in the structure.
Their rotational barrier was so low that a diastereomerically pure
sample gave an almost 1:1 mixture of atropisomers after heating at 40
°C in CHCl3 for 3 h.
(14) Crystal data for 5k (also see Figure S1 in the Supporting
Information): monoclinic, space group P21/n, a = 9.2686(15) Å, b =
16.156(3) Å, c = 9.4109(16) Å, V = 1403.3(4) Å3, Z = 4, ρ = 1.379 Mg
m−3, μ(Cu Kα) = 0.095 mm−1, T = 173 K; in the final least-squares
D
DOI: 10.1021/acs.orglett.5b03021
Org. Lett. XXXX, XXX, XXX−XXX