Catalytic cycloaddition of 2-methyleneaziridines with 1,1-dicyanoalkenes

Article abstract of DOI:10.1021/ol500062a

2-Methyleneaziridine are a good substrate for the catalytic synthesis of cyclopentylidenamines via a [3 + 2] cycloaddition of 1,1-dicyanoalkenes using Bu₂SnI₂ as an effective catalyst. A C-attack from 2-methyleneaziridine yielded the desired products.

Full text of DOI:10.1021/ol500062a

Letter
pubs.acs.org/OrgLett
Catalytic Cycloaddition of 2‑Methyleneaziridines with 1,1-
Dicyanoalkenes
Hiroki Takahashi, Seiji Yasui, Shinji Tsunoi, and Ikuya Shibata*
Research Center for Environmental Preservation, Osaka University, 2-4 Yamadaoka, Suita, Osaka 565-0871, Japan
S
* Supporting Information
ABSTRACT: 2-Methyleneaziridine are a good substrate for the catalytic
synthesis of cyclopentylidenamines via a [3 + 2] cycloaddition of 1,1-
dicyanoalkenes using Bu2SnI2 as an effective catalyst. A C-attack from 2-
methyleneaziridine yielded the desired products.
Table 1. Cycloaddition of 1a with 2a
2-Methyleneaziridines 1 are easy to handle in spite of their ring
strain, and the preparation procedures have been reported so
far.1,2 This compound has been used as a versatile substrate for
synthesis.2,3 The regioselective ring-opening of 1 at C-3 by a
Grignard reagent and the subsequent reaction of the resultant
metalloenamine with an electrophile is a representative
reaction.2b Lewis acid-promoted intramolecular cycloadditions
have been developed.4 Almost conventional procedures are
stoichiometric reactions with metallic reagents. For the catalytic
reaction using 1, transition-metal-catalyzed reactions have been
reported,5,6 in which the initial addition of a metal species to a
CC double bond is a key reaction.6 As main group metal
catalysts, organotin iodides were used in cycloaddition of
epoxides in the preparation of cyclic compounds.7 In the
cycloaddition, the ring-opening of an epoxide by tin iodide is a
key step.8 However, the ring-opening of aziridines are difficult
because of the diminished electronegativity of nitrogen.9 In
contrast, 2-methyleneaziridine 1 seems to cleave easily because
of ring strain,10 which makes it a good candidate for the
substrate of cycloaddition reactions. Also, 2-methyleneaziridine
1 possesses two nucleophilic sites similar to that of amines (N-
attack) and enamine (C-attack) (Scheme 1, eq 1). We recently
a
b
entry
catalyst
yield of 3a (%)
1
2
3
4
5
6
7
8
9
none
Bu2SnI2
LiI
trace
89
76
Bu2SnBr2
LiBr
trace
trace
trace
38
KI
MgI2
ZnI2
27
InI3
53
a
1a (0.8 mmol), 2a (0.5 mmol), THF (1 mL), cat (0.05 mmol).
Determined by H NMR based on 2a.
b
1
using 0.1 equiv of Bu2SnI2 as a catalyst, the cyclic product 3a
was obtained in an 89% yield (entry 2). The initially formed
product 3a was an imine type of adduct that was detected in the
Scheme 1. [3 + 2] Cycloaddition of 2-Methyleneaziridine 1
1
reaction mixture by H NMR. However, cyclopentanone 3a′
was isolated as a stable product because the hydrolysis of 3a
occurred easily during the purification process (Table 1, eq
parentheses). The product 3a was derived from the attack of
the enamine carbon of 1a to the CC bond of 2a. We
examined other metal salts as catalysts (entries 3−9). Using LiI
also gave 3a in a 76% yield (entry 3). Bu2SnBr2 and LiBr
catalysts did not cause the desired reaction (entries 4 and 5).
Using KI as a catalyst also gave no products (entry 6). Using
MgI2 and ZnI2 resulted in low yields (entries 7 and 8). InI3
showed moderate activity (entry 9). Bu2SnI2 has thus far been
the catalyst that afforded the highest yield.
reported the N-attack of 1 in a tin iodide-catalyzed [3 + 2]
cycloaddition with isocyanates leading to nitrogen-hetero-
cycles.11 In view of these backgrounds, we present here the
novel type of [3 + 2] cycloaddition of 1 with activated alkenes
(Scheme 1, eq 2). This reaction could be a catalytic reaction
involving a C-attack from 1 where systems employing catalytic
tin have been established.
Table 2 shows the results of the reactions of 2-
methyleneaziridines 1 with several 1,1-dicyanoalkenes 2
catalyzed by Bu2SnI2 under the optimized conditions. The
Initially, as shown in Table 1, we performed the cyclo-
addition of 1a with benzalmalononitrile (2a). Without a
catalyst, no reaction proceeded at 25 °C for 1 h (entry 1). By
Received: January 8, 2014
Published: February 5, 2014
© 2014 American Chemical Society
1192
dx.doi.org/10.1021/ol500062a | Org. Lett. 2014, 16, 1192−1195

Organic Letters
Letter
Table 2. Cycloaddition of 1 with 1,1-Dicyanoalkenes 2
Scheme 2. Reduction of Imino Group of 3
b
c
product 3 yield of 3 yield of 3
a
entry
R
R′
or (3′)
(%)
(%)
1
2
n-Bu (1a) Ph (2a)
a
89
79
80
75
p-NO2C6H4
(2b)
b
3
p-ClC6H4
(2c)
c
93
84
4
5
p-FC6H4 (2d)
d
e
71 (3 h)
52 (3 h)
65
47
Scheme 3. Plausible Catalytic Cycle
p-CH3C6H4
(2e)
6
7
8
9
PMP (2f)
Nphth (2g)
furyl (2h)
f
53
48
48
66
40
g
h
i
50 (3 h)
73
PhCH2CH2
(2i)
46 (24 h)
10
11
12
CH2tol-p
(1b)
Ph (2a)
3j (3a′)
3k (3c′)
3l (3a′)
66 (24 h)
58 (24 h)
76 (3 h)
62
50
73
p-ClC6H4
(2c)
c-Hex (1c) Ph (2a)
a
1 (0.8 mmol), 2 (0.5 mmol), THF (1 mL), Bu2SnI2 (0.05 mmol).
Determined by H NMR based on 2. Isolated yield.
b
c
1
reactions of 1a with 2a−2i proceeded well at 25 °C, and cyclic
products 3a−3i were formed (entries 1−9). In these cases,
various functionalities could be introduced on the aromatic
rings of 2 (entries 1−6). Naphthyl substituted alkene 2g gave
3g (entry 7). Furyl-substituted alkene 2h was also useful and
gave 3h (entry 8). Besides aromatic alkenes, aliphatic 2i was
reactive and gave 3i (entry 9). In a similar manner, a reaction
using N-arylmethyl-substituted substrate 1b underwent the
desired reactions (entries 10 and 11). Even the use of bulky
cyclohexyl-substituted 1c underwent the reaction well (entry
12).
and enones were unreactive because of the lack of electrophilic
nature and were recovered quantitatively. The nucleophilicity of
tin enamine A was insufficient because of electron-withdrawing
ability of the halogen substituent in A. Bu2SnBr2 or LiBr
catalyst showed no catalytic activity (Table 1 entries 4 and 5).
Thus the electron-withdrawing property of bromide substituent
decreases the nucleophilicity of A. Finally, the resultant
nucleophilic Sn−C bond of B causes C-alkylation to afford 4-
cyclopentan-2-imine 3, regenerating the tin iodide catalyst.
We next attempted the use 2-alkylidene aziridine 1d
including an internal (Z)-alkene (Scheme 4). The reactions
with 1,1-dicyanoalkenes, 2a and 2c, proceeded well, and α-
methyl-substituted cyclopentanones 5a and 5b were isolated as
major products, accompanying the minor ones, 6a and 6b. The
major products of 5 were obtained through the tin enamine C
that was derived from the SN2 attack of the Sn−I bond toward
the aziridine ring. On the other hand, the minor products of 6
were formed through the enamine E that was derived from the
SN2′ attack of the Sn−I bond toward the CC bond. The SN2′
attack is known to be a minor process in the ring-opening of a
simple 2-methyleaziridine.15 It is worth noting that major
products 5a and 5b were obtained as single trans-isomers, the
stereochemistry of which was determined by NOE study. Thus,
in the reaction of (Z)-enamine C with 2, passage of the
sterically less-hindered 8-membered cyclic transition state
TS(I) was favored over that of TS(II).16,17 As a result, threo-
adduct D was formed selectively and underwent cyclization to
give 5. Minor products, regioisomers 6a and 6b, were single cis-
isomers, the stereochemistry of which was determined by NOE
study.18
As shown in Tables 1 and 2, although the desired reactions
proceeded well, products 3 had imine groups and were easily
hydrolyzed during the treatment by column chromatography.
Hence, cyclopentanone derivatives 3′ were obtained as pure
products. In the examples shown in Table 2, entries 10 and 11,
the initially obtained 3j and 3k were hydrolyzed to 3a′ and 3c′,
respectively. These isolated products corresponded to entries 1
and 3. In the case of entry 12, the product 3l was also converted
to 3a′. So nitrogen functionalities were lost from the
cycloadducts 3. For the fixation of nitrogen functionalities, we
tried a reduction of the imino groups of 3 (Scheme 2). Without
workup, the resultant reaction mixtures that included 3a, 3j,
12
and 3l were allowed to react with a reductant NaBH(OAc)3
at rt for 18 h. As a result, amines, 4a, 4j, and 4l, respectively,
were obtained as nitrogen-containing products with high
diastereoselectivities.13
Scheme 3 shows a plausible catalytic cycle for the tin-
catalyzed reaction of 1 with 2. Initially, tin enamine A is formed
by the attack of Sn−I toward 2-methyleneaziridine 1.14 Because
of the ring strain of 1, the ring-opening proceeds effectively.
Next, the Michael addition of A to 2 takes place to give adduct
B (C-attack). For the Michael addition, using highly electro-
philic CC bonds like 1,1-dicyano alkenes 2 was essential.
Other alkenes such as monocyanoalkenes, unsaturated esters,
Scheme 5 shows the reaction using C3-substituted 2-
methyleneaziridine 1e. The reactions with 1,1-dicyanoalkenes
2a and 2c proceeded well to give major products 7a and 7b
accompanying minor products 8a and 8b. Contrary to the
above case using 1d, the SN2′ attack of the Sn−I bond toward
1193
dx.doi.org/10.1021/ol500062a | Org. Lett. 2014, 16, 1192−1195

Organic Letters
Letter
Scheme 4. Cycloaddition of 1d with 2a or 2c
Scheme 6. Cycloaddition of 1a with Unsymmetric Alkene 9
In conclusion, we demonstrated the catalytic conversion of 2-
methyleneaziridines 1 to cyclic products 3, 4, 5, 7, and 10 via a
novel type of [3 + 2] cycloaddition. Bu2SnI2 was employed as
an effective catalyst. A C-attack from tin enamines was a key
step in the reaction.
ASSOCIATED CONTENT
* Supporting Information
■
S
General procedures and characterization data of new
compounds, CIF file for 7b. This material is available free of
charge via the Internet at http://pubs.acs.org.
Scheme 5. Cycloaddition of 1e with 2a, 2c
AUTHOR INFORMATION
Corresponding Author
■
*E-mail: shibata@epc.osaka-u.ac.jp.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This research was supported by a Grant-in-Aid for Scientific
Research from the Ministry of Education, Science, Sports, and
Culture and the Naito Foundation.
REFERENCES
■
(1) (a) De Kimpe, N.; De Smaele, D.; Sakonyi, Z. J. Org. Chem. 1997,
62, 2448−2452. (b) Tehrani, K. A.; De Kimpe, N. Curr. Org. Chem.
2009, 13, 854−877.
(2) (a) Shipman, M. Synlett 2006, 3205−3217. (b) Hayes, J. F.;
Shipman, M.; Twin, H. J. Org. Chem. 2002, 67, 935−942.
(3) (a) Hayes, J. F.; Shipman, M.; Twin, H. Chem. Commun. 2000,
1791−1792. (b) Hayes, J. F.; Shipman, M.; Twin, H. Chem. Commun.
2001, 1784−1785. (c) Hayes, J. F.; Shipman, M.; Twin, H. J. Org.
Chem. 2002, 67, 935−942. (d) Margathe, J.-F.; Shipman, M.; Smith, S.
C. Org. Lett. 2005, 7, 4987−4990. (e) Cariou, C. C. A.; Clarkson, G. J.;
Shipman, M. J. Org. Chem. 2008, 73, 9762−9764.
(4) (a) Griffin, K.; Montagne, C.; Hoang, C. T.; Clarkson, G. J.;
Shipman, M. Org. Biomol. Chem. 2012, 10, 1032−1039. (b) Prie, G.;
Prevost, N.; Twin, H.; Fernandes, S. A.; Hayes, J. F.; Shipman, M.
Angew. Chem., Int. Ed. 2004, 43, 6517−6519.
the CC bond was a major process. Thus, the products of 7
were obtained through tin enamine G. However, the minor
products of 8 were obtained through the enamine I that was
derived from the SN2 attack of the Sn−I bond toward the
aziridine ring. The disadvantage of SN2 attack giving I was
caused by the difficulty of a direct ring-opening for vis-
disubstituted rings.19 Major products 7a and 7b were obtained
as single trans-isomers.20 The stereoselectivity in 7 can be
explained in a similar manner giving 5 as illustrated in Scheme
4, because tin enamine G is very similar to C.18
(5) Alper, H.; Hamel, N. Tetrahedron Lett. 1987, 28, 3237−3240.
(6) (a) Oh, B. H.; Nakamura, I.; Yamamoto, Y. Tetrahedron Lett.
2002, 43, 9625−9628. (b) Siriwardana, A. I.; Kathriarachchi, K. K. A.
D. S.; Nakamura, I.; Gridnev, D. I.; Yamamoto, Y. J. Am. Chem. Soc.
2004, 126, 13898−13899. (c) Oh, B. H.; Nakamura, I.; Yamamoto, Y.
J. Org. Chem. 2004, 69, 2856−2858. (d) Kathriarachchi, K. A. D. S.;
Siriwardana, A. I.; Nakamura, I.; Yamamoto, Y. Tetrahedron Lett. 2007,
48, 2267−2270.
When an unsymmetrical alkene 9 was used as an electrophile,
the desired adduct 10 was obtained in 75% yield (Scheme 6).
The reaction was diastereoselective to give a threo isomer. The
diastereoselectivity is supposed to be determinded at the step of
cyclization. Namely, after the C−C bond formation, cyclization
proceeds via TS(III), giving threo-isomer 10 rather than TS(IV)
to an erythro isomer. Thus, the repulsion between NBu and the
stannoxy group would be responsible for the observed
stereochemistry.
(7) Tin-catalyzed reaction using epoxides has been reported.
(a) Shibata, I.; Baba, A.; Iwasaki, H.; Matsuda, H. J. Org. Chem.
1194
dx.doi.org/10.1021/ol500062a | Org. Lett. 2014, 16, 1192−1195

Organic Letters
Letter
1986, 51, 2177−2184. (b) Shibata, I.; Yoshimura, N.; Baba, A.;
Matsuda, H. Tetrahedron Lett. 1992, 33, 7149−7152.
(8) Synthetic uses of organotin compounds; for example, see:
(a) Baba, A.; Shibata, I.; Yasuda, M. In Comprehensive Organometallic
Chemistry III; Knochel, P., Ed.; Elsevier: Oxford, 2007; Vol. 9, Chapter
8, pp 341−380. (b) Pereyre, M.; Quintard, P. J.; Rahm, A. In Tin in
Organic Synthesis; Butterworth: London, 1987; pp 261−285.
(c) Davies, A. G. In Organotin Chemistry; VCH: Weinhelm, 1997;
pp 166−193.
(9) Aziridine chemistry: (a) Sweeney, J. B. Chem. Soc. Rev. 2002, 31,
247−258. (b) Singh, G. S.; D’hooghe, M.; De Kimpe, N. Chem. Rev.
2007, 107, 2080−2135. (c) Pearson, W. H.; Lian, B. W.; Bergmeier, S.
C. In Comprehensive Heterocyclic Chemistry II; Padwa, A., Ed.;
Pergamon Press: New York, 1996; Vol. 1A, pp 1−96.
(10) Bachrach, S. W. J. Phys. Chem. 1993, 97, 4996−5000.
(11) Takahashi, H.; Yasui, S.; Tsunoi, S.; Shibata, I. Eur. J. Org. Chem.
2013, 40−43.
(12) (a) Abdel-Magid, A. F.; Maryanoff, C. A.; Carson, K. G.
Tetrahedron Lett. 1990, 31, 5595−5598. (b) Abdel-Magid, A. F.;
Carson, K. G.; Harris, B. D.; Maryanoff, C. A.; Shah, R. D. J. Org.
Chem. 1996, 61, 3849−3862. (c) Kim, H. O.; Carrol, B.; Lee, M. S.
Synth. Commun. 1997, 27, 2505−2515.
(13) We could determine the stereochemistry of products as 1,3-trans
by NOE study for 4j.
(14) In an equimolar mixture of 1a with Bu2SnI2 the measurement of
1H NMR resulted in a complex mixture, perhaps because of the fact
that intermediate A was too labile to detect.
(15) (a) Ince, J.; Shipman, M.; Ennis, D. S. Tetrahedron Lett. 1997,
38, 5887−5890. (b) Hayes, J. F.; Shipman, M.; Twin, H. J. Org. Chem.
2002, 67, 935−942.
(16) For eight membered cyclic transition state: (a) Oare, D. A.;
Heathcock, C. H. J. Org. Chem. 1990, 55, 157−172. (b) Shibata, I.;
Yasuda, K.; Y. Tanaka, Y.; Yasuda, M.; Baba, A. J. Org. Chem. 1998, 63,
1334−1336.
(17) Acyclic antiperiplanar transition states are inconsiderable. See
Supporting Information.
(18) The reason for the diastereoselective formation of minor
products 6, 8 is discussed in the Supporting Information.
(19) Organotin iodide catalyzed the cycloaddition of epoxides with
isocyanates could not applied to the case using vic-disubstituted rings
because of difficulty of ring opening.7
(20) The stereochemistry of 7b is determined by X-ray crystal
analysis (CCDC 972461).
1195
dx.doi.org/10.1021/ol500062a | Org. Lett. 2014, 16, 1192−1195