Nickel-catalyzed, directing-group-assisted [2+2+2] cycloaddition of imine and alkynes

Article abstract of DOI:10.1002/asia.201000564

An aldimine bearing a 3-methyl-2-pyridyl group undergoes [2+2+2] cycloaddition reaction with two alkyne molecules under nickel catalysis to afford a 1,2-dihydropyridine derivative in moderate to good yield. The reaction is likely to involve oxidative cyclization of the imine and alkyne, insertion of another alkyne, and C-N reductive elimination, followed by a 1,5-sigmatropic hydrogen shift. The pyridyl group is proposed to facilitate the reaction by chelation to the aza-nickelacycle intermediates. Copyright

Full text of DOI:10.1002/asia.201000564

DOI: 10.1002/asia.201000564
Nickel-Catalyzed, Directing-Group-Assisted [2+2+2] Cycloaddition of Imine
and Alkynes
Laksmikanta Adak, Wei Chuen Chan, and Naohiko Yoshikai*[a]
Dedicated to Professor Eiichi Nakamura on the occasion of his 60th birthday
Transition-metal-catalyzed [2+2+2] cycloaddition of unsa-
turated molecules represents a powerful and atom-efficient
method for the synthesis of various carbo- and heterocyclic
compounds.[1] Numerous examples of this have been report-
ed for the construction of a benzene ring from three alkyne
Scheme 1. Nickel-catalyzed [2+2+2] cycloaddition of an imine and two
alkynes.
moieties, and a pyridine ring from two alkyne moieties and
one nitrile group. Other unsaturated fragments such as
carbon dioxide,[2] isocyanates,[3] and aldehydes/ketones[4] also
participate in cycloaddition with two alkyne moieties to
afford pyrone, pyridone, and dienone/pyran derivatives.[5]
Although the C=N double bond of an imine can poten-
tially participate in [2+2+2] cycloaddition with two alkyne
moieties to afford a dihydropyridine derivative, such cyclo-
addition reactions are very rare.[6] To our knowledge, a
nickel-catalyzed cycloaddition reaction of alkynes and an N-
sulfonyl imine, which was reported by Ogoshi and co-work-
ers, is the only reaction of some generality.[6a] Herein we
report that an aldimine bearing a 3-methyl-2-pyridyl group[7]
on the nitrogen atom readily participates in nickel-catalyzed
cycloaddition with alkynes. The reaction gives a 1,2-dihydro-
pyridine derivative in moderate to good yield, presumably
through 1,5-hydrogen shift of the initially formed cycload-
duct (Scheme 1). Computational studies indicated that the
pyridyl group serves as a crucial directing group for the se-
lective [2+2+2] cycloaddition.
aza-nickelacycle, and thereby facilitate selective [2+2+2]
coupling of two alkyne moieties and one imino group. Ex-
periments along this hypothesis led us to find that an aryl
imine 1a derived from benzaldehyde and 2-amino-3-methyl-
pyridine efficiently underwent coupling with two molecules
of 4-octyne 2a in the presence of a nickel catalyst generated
from NiCl2 (10 mol%), PMePh2 (20 mol%), and Mn
powder (100 mol%, activated by CF3CO2H) at 1008C in
DMF to afford a 1,2-dihydropyridine 3aa in 82% yield
(Table 1, entry 1).[8] The reaction was accompanied by the
formation of a small amount of a reductive coupling product
4aa.[9] The formation of 3aa may be accounted for by ther-
mal 1,5-sigmatropic hydrogen shift of the initially formed
[2+2+2] cycloadduct (Scheme 1).[10] The reaction of a ben-
zaldimine derived from 2-aminopyridine also gave the corre-
sponding cycloadduct in a slightly lower yield of 67%.
Other imines such as those bearing N-phenyl and N-sulfonyl
groups did not participate in the reaction.
The other entries in Table 1 summarize the results of the
screening of the reaction conditions. Other monodentate
phosphines such as PPh3, PMe2Ph, and PCy3 afforded cata-
lysts of much lower activities (Table 1, entries 2–4). Biden-
tate phosphine and nitrogen ligands such as 1,3-bis(diphe-
nylphosphino)propane (dppp) and 2,2’-bipyridyl (bpy) com-
pletely stopped the reaction (Table 1, entries 5 and 6). DMF
emerged as the best among the other nonpolar and polar
solvents, including toluene, THF, dioxane, and acetonitrile
(Table 1, entries 7–10). The reaction in acetonitrile afforded
a considerable amount of the reductive coupling product
At the outset of our investigation into imine–alkyne cy-
cloaddition, we envisaged that a chelating substituent on the
nitrogen atom would assist oxidative cyclization of the
alkyne and the imine, stabilize the resulting five-membered
[a] Dr. L. Adak, W. C. Chan, Dr. N. Yoshikai
Division of Chemistry and Biological Chemistry
School of Physical and Mathematical Sciences
Nanyang Technological University
Singapore 637371 (Singapore)
Fax : (+65)6791-1961
E-mail: nyoshikai@ntu.edu.sg
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/asia.201000564.
Chem. Asian J. 2011, 6, 359 – 362
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
359

COMMUNICATION
group to the nickel catalyst (Table 2, entry 4).[12] Aldimines
bearing electron-donating groups at the ortho position react-
ed more smoothly than the para-substituted counterparts, af-
fording the cycloadducts in about 90% yield (Table 2, en-
tries 6 and 7). On the other hand, the reaction became slug-
gish when there was an electron-withdrawing fluorine atom
at the ortho position (Table 2, entry 8). Aldimines derived
from 1-naphthaldehyde and thiophene-2-carbaldehyde also
afforded cycloadducts with 4-octyne in moderate to good
yields (Table 2, entries 9 and 10). The reaction did not toler-
ate the presence of chloro and bromo substituents, probably
because of oxidative addition of the aryl–chloro and aryl–
bromo bonds to the nickel catalyst (data not shown).
Table 1. Nickel-catalyzed [2+2+2] reaction of imine 1a with 4-octyne
2a.[a]
Entry
Ligand
Solvent
Yield [%][b]
3aa
4aa
1
2
3
PMePh2
PPh3
PMe2Ph
PCy3
DMF
DMF
DMF
DMF
(82)
39
56
5
(7)
8
5
4
0
5
dppp
DMF
0
0
Internal aliphatic alkynes other than 4-octyne afforded cy-
cloadducts with the aldimine 1a in good yields (Table 2, en-
tries 11 and 12), while the reaction of diphenylacetylene was
rather sluggish (entry 13). Unsymmetrical alkynes such as 1-
aryl-1-propyne and iso-propylmethylacetylene also partici-
pated in cycloaddition with 1a, albeit in modest yields
(Scheme 2a). The reaction of the former afforded 3ae and
6
7
8
9
bpy
DMF
toluene
THF
dioxane
CH3CN
0
0
0
1
0
PMePh2
PMePh2
PMePh2
PMePh2
21
19
18
29
10
19
[a] The reaction was carried out on
(100 mol%) was activated with CF3CO2H (10–20 mL). [b] GC yields. In
parentheses are shown yields of isolated product.
a 0.5 mmol scale. Mn powder
4aa (Table 1, entry 10).[11] The use of iron or cobalt salts in-
stead of NiCl2 afforded neither the cycloadduct nor the re-
ductive coupling product.
Table 2 summarizes the scope of the present reaction. Al-
dimines derived from various benzaldehyde derivatives par-
ticipated in cycloaddition with 4-octyne to afford the corre-
sponding dihydropyridines in moderate to high yields
(Table 2, entries 1–8). Among aldimines substituted at the
para position, the one bearing an electron-withdrawing tri-
fluoromethyl group gave the highest yield of 82% (Table 2,
entries 1–3). A 4-cyano substituent caused the reaction to be
sluggish, probably because of the coordination of the cyano
Scheme 2. Products of [2+2+2] cycloaddition of 1a with unsymmetrical
alkynes (a) and 1,6- and 1,7-diynes (b). The reaction was performed
under the conditions in Table 1, entry 1, for 15–24 h. The amount of the
diyne substrate was 1.5 equiv.
Table 2. Nickel-catalyzed [2+2+2] cycloaddition of imines and alkynes.[a]
3af as the only detectable regioisomers. Note that the regio-
chemistry of the C1/C2 position agrees with the general
trend for nickel-mediated reductive coupling of such alky-
nes.[9a–c,13] On the other hand, the latter substrate gave a
mixture of two isomers (C1/C2 position, ca. 1:1) owing to
modest regioselectivity of oxidative cyclization of the ali-
phatic internal alkyne.[14,15] In both cases, perfect regioselec-
tivity of the C3/C4 position was observed, which should
have been controlled by the steric factors in alkyne insertion
into the aza-nickelacycle intermediate (see below).[16] 1,6-
and 1,7-Diynes underwent 1:1 coupling with 1a to afford bi-
cyclic adducts 3ah and 3ai in modest yields. Note that, when
the cycloaddition reaction was sluggish, the rest of the aldi-
mine decomposed into unidentifiable products, while the
alkyne could be recovered. Terminal alkynes such as phenyl-
acetylene and 1-octyne underwent trimerization and failed
to give cycloadducts with 1a.
Entry
R1
R2
t [h]
Yield [%][b]
1
2
3
4
5
6
7
8
4-CF3C6H4 (1b)
4-tBuC6H4 (1c)
4-MeOC6H4 (1d)
4-NCC6H4 (1e)
3-MeOC6H4 (1 f)
2-MeOC6H4 (1g)
2-MeC6H4 (1h)
2-FC6H4 (1i)
1-naphthyl (1j)
2-thienyl (1k)
Ph (1a)
n-C3H7 (2a)
n-C3H7 (2a)
n-C3H7 (2a)
n-C3H7 (2a)
n-C3H7 (2a)
n-C3H7 (2a)
n-C3H7 (2a)
n-C3H7 (2a)
n-C3H7 (2a)
n-C3H7 (2a)
C2H5 (2b)
24
15
24
15
24
15
24
24
15
15
15
15
20
82 (3ba)
72 (3ca)
59 (3da)
36 (3ea)
62 (3 fa)
89 (3ga)
92 (3ha)
38 (3ia)
76 (3ja)
39 (3ka)
81 (3ab)
90 (3ac)
14 (3ad)
9
10
11
12
13
Ph (1a)
Ph (1a)
n-C4H9 (2c)
Ph (2d)
[a] The reaction was carried out on a 0.5 mmol scale. [b] Yields of isolat-
ed product.
360
www.chemasianj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 2011, 6, 359 – 362

Nickel-Catalyzed [2+2+2] Cycloaddition
To gain insight into the reaction pathway of the nickel-
catalyzed part of the present reaction, we performed DFT
calculations on a model reaction.[17,18] We chose PMe3, 2-
butyne, and an imine derived from acetaldehyde and 2-pyri-
dylamine as the model ligand and substrates. In light of the
report of Ogoshi and co-workers,[6a] an h2 complex between
[NiACHTUNGTRENNUNG(PMe3)2] and the imine (CP1, Py=2-pyridyl) was used as
the initial catalytic intermediate (Scheme 3).[19]
Figure 1. Optimized structures of CP3, CP5, and TS3 (see Scheme 3).
Bond lengths are in ꢁ.
2
À
to the nickel center in CP3 and CP5 (Ni N : 2.17 ꢁ and
2.05 ꢁ, respectively) regardless of the strain caused by the
formation of four-membered chelation structure. This che-
Scheme 3. Outline of reaction pathway for Ni-catalyzed [2+2+2] cycload-
dition of N-pyridylimine and alkynes. All energies are given in kcalmolÀ1
values in parentheses refer to energies relative to CP1.
;
lating pyridyl group dissociates from the nickel center upon
2
À
À
C N reductive elimination (see TS3; Ni N : 2.85 ꢁ). Upon
1
À
going from CP5 to TS3, the C N bond forms to a consider-
1
1
À
À
À
Scheme 3 shows the outline of the reaction pathway ob-
tained for the model system. It is essentially the same as the
one proposed by Ogoshi and co-workers. The h2 complex
CP1 can reversibly undergo exchange of PMe3 with 2-
butyne to afford a complex CP2. Oxidative cyclization of
CP2 takes place with an activation energy of 16.3 kcalmolÀ1,
and leads to a five-membered aza-nickelacycle CP3. The
aza-nickelacycle CP3 is largely stabilized by internal coordi-
nation of the pyridyl group. Compared with this process, for-
mation of nickelacyclopentadiene through oxidative cycliza-
tion of two 2-butyne molecules was calculated to be kineti-
cally and thermodynamically less favorable (see the Sup-
porting Information). Note that, however, intramolecular
oxidative cyclization of two alkyne moieties may be an alter-
native reaction pathway for the reaction of the diynes
(Scheme 2b).
able extent (C N : 1.86 ꢁ), while the Ni C and Ni N bond
lengths change little.
In summary, we have reported on a nickel-catalyzed
[2+2+2] cycloaddition reaction of an imine bearing a N-pyr-
idyl group and two alkyne units. This reaction represents a
rare example of metal-catalyzed cycloaddition reactions in-
volving imines as substrates. The directing effect of the pyr-
idyl group was supported by DFT calculations. We consider
that such directing-group strategies would also be effective
for other cycloaddition and coupling reactions involving oxi-
dative cyclization as the key elementary step.
Acknowledgements
We thank National Research Foundation, Singapore (NRF-RF-2009-05
to N.Y.) and Nanyang Technological University for generous financial
support.
Insertion of another molecule of 2-butyne takes place via
an intermediate CP4 to afford a seven-membered nickelacy-
cle CP5. This process is facile (DE° =9.0 kcalmolÀ1) and sig-
À
nificantly exothermic. C N reductive elimination of the
nickelacycle CP5, which requires a sizable activation energy
Keywords: alkynes
calculations · heterocycles · nickel
· cycloaddition · density functional
(26.5 kcalmolÀ1), affords a cycloadduct that is complexed
with {NiACHTUNGTRENNUNG(PMe3)} (i.e., CP6). In accordance with the experi-
mental observations of Ogoshi and co-workers,[6a] reductive
elimination is likely to be the rate-limiting step of the reac-
tion, which is followed by a 1,5-hydrogen shift to afford the
final product.
[1] For recent reviews, see: a) S. Saito, Y. Yamamoto, Chem. Rev. 2000,
100, 2901–2915; b) J. A. Varela, C. Saꢂ, Chem. Rev. 2003, 103,
3787–3801; c) G. D. Henry, Tetrahedron 2004, 60, 6043–6061; d) S.
Kotha, E. Brahmachary, K. Lahiri, Eur. J. Org. Chem. 2005, 4741–
4767; e) Y. Yamamoto, Curr. Org. Chem. 2005, 9, 503–519; f) P. R.
Chopade, J. Louie, Adv. Synth. Catal. 2006, 348, 2307–2327; g) K.
Tanaka, Synlett 2007, 1977–1993; h) B. Heller, M. Hapke, Chem.
Soc. Rev. 2007, 36, 1085–1094; i) N. Agenet, O. Busine, F. Slowinski,
Figure 1 shows the structures of the aza-nickelacycle inter-
À
mediates (CP3 and CP5) and the C N reductive elimination
TS (TS3). The pyridyl nitrogen atom is tightly coordinated
Chem. Asian J. 2011, 6, 359 – 362
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemasianj.org
361

N. Yoshikai et al.
COMMUNICATION
V. Gandon, C. Aubert, M. Malacria, Org. React. 2007, 68, 1–302;
j) T. Shibata, K. Tsuchikama, Org. Biomol. Chem. 2008, 6, 1317–
1323; k) B. R. Galan, T. Rovis, Angew. Chem. 2009, 121, 2870–2874;
Angew. Chem. Int. Ed. 2009, 48, 2830–2834.
[9] a) S. J. Patel, T. F. Jamison, Angew. Chem. 2003, 115, 1402–1405;
Angew. Chem. Int. Ed. 2003, 42, 1364–1367; b) S. J. Patel, T. F. Jami-
son, Angew. Chem. 2004, 116, 4031–4034; Angew. Chem. Int. Ed.
2004, 43, 3941–3944; c) M. Ohashi, O. Kishizaki, H. Ikeda, S.
Ogoshi, J. Am. Chem. Soc. 2009, 131, 9160–9161; d) A. Barchuk,
M.-Y. Ngai, M. J. Krische, J. Am. Chem. Soc. 2007, 129, 8432–8433;
e) M.-Y. Ngai, A. Barchuk, M. J. Krische, J. Am. Chem. Soc. 2007,
129, 12644–12645; f) M. Z. Chen, G. C. Micalizio, Org. Lett. 2009,
11, 4982–4985.
[10] The putative initial [2+2+2] cycloadduct could not be detected even
when the reaction was quenched at the early stage. We speculate
that 1,5-hydrogen shift takes place more quickly than cycloaddition.
[11] Control experiments suggested that the proton source for the reduc-
tive coupling product 4aa was moisture in the solvent or CF3CO2H.
[12] T. A. Ates¸in, T. Li, S. Lachaize, J. J. Garcꢃa, W. D. Jones, Organome-
tallics 2008, 27, 3811–3817, and references therein.
[2] a) D. Walther, H. Schçnberg, E. Dinjus, J. Sieler, J. Organomet.
Chem. 1987, 334, 377–388; b) T. Tsuda, S. Morikawa, R. Sumiya, T.
Saegusa, J. Org. Chem. 1988, 53, 3140–3145; c) J. Louie, J. E. Gibby,
M. V. Farnworth, T. N. Tekavec, J. Am. Chem. Soc. 2002, 124,
15188–15189; d) T. N. Tekavec, A. M. Arif, J. Louie, Tetrahedron
2004, 60, 7431–7437; e) Y. Kishimoto, I. Mitani, Synlett 2005, 2141–
2146.
[3] a) R. A. Earl, K. P. C. Vollhardt, J. Org. Chem. 1984, 49, 4786–4800;
b) Y. Yamamoto, H. Takagishi, K. Itoh, Org. Lett. 2001, 3, 2117–
2119; c) H. A. Duong, M. J. Cross, J. Louie, J. Am. Chem. Soc. 2004,
126, 11438–11439; d) Y. Yamamoto, K. Kinpara, T. Saigoku, H. Ta-
kagishi, S. Okuda, H. Nishiyama, K. Itoh, J. Am. Chem. Soc. 2005,
127, 605–613; e) K. Tanaka, A. Wada, K. Noguchi, Org. Lett. 2005,
7, 4737–4739; f) T. Kondo, M. Nomura, Y. Ura, K. Wada, T. Mitsu-
do, Tetrahedron Lett. 2006, 47, 7107–7111; g) K. Tanaka, Y. Takaha-
shi, T. Suda, M. Hirano, Synlett 2008, 1724–1728.
[13] P. Liu, P. McCarren, P. H.-Y. Cheong, T. F. Jamison, K. N. Houk, J.
Am. Chem. Soc. 2010, 132, 2050–2057, and references therein.
[14] For example, see: E. A. Colby, T. F. Jamison, J. Org. Chem. 2003, 68,
156–166.
[4] a) T. N. Tekevac, J. Louie, Org. Lett. 2005, 7, 4037–4039; b) K.
Tanaka, Y. Otake, A. Wada, K. Noguchi, M. Hirano, Org. Lett.
2007, 9, 2203–2206; c) K. Tsuchikama, Y. Yoshinami, T. Shibata,
Synlett 2007, 1395–1398; d) Y. Otake, R. Tanaka, K. Tanaka, Eur. J.
Org. Chem. 2009, 2737–2747.
[5] For cycloaddition of other unsaturated fragments, see: a) Y. Yama-
moto, H. Takagishi, K. Itoh, J. Am. Chem. Soc. 2002, 124, 28–29;
b) K. Tanaka, A. Wada, K. Noguchi, Org. Lett. 2006, 8, 907–909;
c) D. D. Young, A. Deiters, Angew. Chem. 2007, 119, 5279–5282;
Angew. Chem. Int. Ed. 2007, 46, 5187–5190; d) D. D. Young, J. A.
Teske, A. Deiters, Synthesis 2009, 3785–3790.
[6] a) S. Ogoshi, H. Ikeda, H. Kurosawa, Angew. Chem. 2007, 119,
5018–5020; Angew. Chem. Int. Ed. 2007, 46, 4930–4932; b) S.
Ogoshi, H. Ikeda, H. Kurosawa, Pure Appl. Chem. 2008, 80, 1115–
1125; c) P. A. Wender, T. M. Pedersen, M. J. C. Scanio, J. Am. Chem.
Soc. 2002, 124, 15154–15155.
[15] H. A. Malik, G. J. Sormunen, J. Montgomery, J. Am. Chem. Soc.
2010, 132, 6304–6305.
[16] Regioselectivity of alkyne insertion into a nickelacycle is known to
be susceptible to the steric environment. For example, see: Y. Yoshi-
no, T. Kurahashi, S. Matsubara, J. Am. Chem. Soc. 2009, 131, 7494–
7495.
[17] The calculations were performed at the B3LYP level using basis sets
consisting of LANL2DZ for Ni and 6-31G(d) for the rest. See the
Supporting Information for more details.
[18] For computational studies on nickel-catalyzed reductive coupling re-
actions of alkynes, see: a) H. P. Hratchian, S. K. Chowdhury, V. M.
Gutiꢄrrez-Garcꢃa, K. K. D. Amarasinghe, M. J. Heeg, H. B. Schlegel,
J. Montgomery, Organometallics 2004, 23, 4636–4646; b) P. R.
McCarren, P. Liu, P. H.-Y. Cheong, T. F. Jamison, K. N. Houk, J.
Am. Chem. Soc. 2009, 131, 6654–6655, and Ref. [13].
[19] Coordination of the pyridyl nitrogen atom was calculated to be
much less favorable than h2 coordination of the C=N bond. Note
also that the pyridyl nitrogen atom and the C=N bond could not si-
multaneously coordinate to the nickel center.
[7] For the use of a 3-methyl-2-pyridyl group as a directing group for
À
C H bond activation of aldimines, see: a) J. W. Suggs, J. Am. Chem.
Soc. 1979, 101, 489; b) C.-H. Jun, C. W. Moon, D.-Y. Lee, Chem.
Eur. J. 2002, 8, 2423–2428; c) Y.-J. Park, J.-W. Park, C.-H. Jun, Acc.
Chem. Res. 2008, 41, 222–234, and references therein.
[8] The quantity of Mn powder could be reduced to 30 mol%, with a
slight decrease in the product yield.
Received: August 13, 2010
Published online: October 26, 2010
362
www.chemasianj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 2011, 6, 359 – 362