Palladium-catalyzed ring-contraction and ring-expansion reactions of cyclic allyl amines

Article abstract of DOI:10.1002/anie.201100612

Ring around the rosy: An amino group can act as the leaving group or the nucleophile in a palladium-catalyzed allylic amination. As a result, readily accessible cyclic amines can be either ring contracted or ring expanded (see scheme).

Full text of DOI:10.1002/anie.201100612

Communications
DOI: 10.1002/anie.201100612
Cyclic Amines
Palladium-Catalyzed Ring-Contraction and Ring-Expansion Reactions
of Cyclic Allyl Amines**
Igor Dubovyk, Dmitry Pichugin, and Andrei K. Yudin*
Cyclic allyl amines serve as useful tools in synthesis,[1] and are
found in a vast number of naturally occurring alkaloids.[2]
Preparation of cyclic allyl amines can be achieved by a
number of classical methods such as the addition of allyl
nucleophiles to cyclic imines and nitrones and subsequent
reduction,[3] or Wittig reactions.[4] Such approaches, however,
require harsh reaction conditions and have limited reaction
scope. Milder methods include metal-catalyzed transforma-
tions such as intramolecular oxidative amination,[5] allylic
amination,[6] and intramolecular hydroamination of allenes.[7]
Ring-closing metathesis is perhaps the most convergent
approach that renders itself well to rapid and modular
assembly of a wide range of allyl amines. Herein, we show
that allyl amines, which can be accessed using ring-closing
metathesis and other straightforward methods, are conve-
nient starting points for ring-contraction and ring-expansion
reactions in which the conjugate acid of the nitrogen-
containing nucleophile is enlisted as the leaving group.
Our interest in the field of allyl amine chemistry stems
from earlier studies of regioselectivity in palladium-catalyzed
allylic amination. This work revealed the thermodynamic
origin of linear selectivity in that reaction.[8] It later transpired
that the isomerization had occurred as a result of having
active Pd0 species and acid in solution, and followed the
mechanistic rationale shown in Scheme 1. Although this
intermolecular process is largely undesirable and can be
avoided,[9] it does suggest that an amine can play a dual role
by first acting as the leaving group, and then as the
nucleophile.[10]
allyl amine scaffolds. This method can be strategically applied
to late-stage modifications of complex amine-containing
skeletons by using amine-containing fragments as both
nucleophiles and as leaving group precursors.
Amines that do not bear electron-withdrawing substitu-
ents have long been recognized for their reluctance to
[10]
À
undergo palladium-catalyzed C N bond scission.
In the
course of our earlier studies of intermolecular allylic amina-
tion, we discovered that branched product selectivity is
kinetic in origin when THF is used as the solvent, whereas
linear products are formed as a result of acid-promoted
branched-to-linear isomerization. High levels of branched
selectivity were attained by introducing 1 equivalent of DBU
(1,8-diazabicyclo[5.4.0]undec-7-ene) that prevented isomer-
ization. We recently returned to this system and discovered
that product isomerization was slow under these reaction
conditions, and only 30% of product isomerized after one
week. Unexpectedly, in dichloromethane, branched allyl
amines fully isomerized to form linear products within four
days, even in the presence of DBU.[11] In addition, we have
also observed the effect of solvent on kinetic branched/linear
ratios in allylic amination. High branched regioselectivities
were recorded with THF, as well as THF/dichloromethane
mixtures where there was at least 10% THF, whereas in
dichloromethane both regioisomers were favored equally
throughout the reaction, until full conversion was achieved.[12]
Intrigued by the fact that dichloromethane promotes
formation of the more stable product from both kinetic and
thermodynamic points of view, we wanted to test whether or
not such isomerization can be used to rearrange cyclic allyl
amines. Table 1 shows the results of isomerization-driven ring
construction. We were encouraged that in dichloromethane
We envisioned straightforward access to complex allyl
amines by skeletal isomerizations of readily accessible cyclic
Table 1: Evaluation of palladium-catalyzed isomerization reaction con-
ditions.[a]
Scheme 1. Allylic isomerization promoted by a combination of palla-
dium and protic acid.
Entry
Ligand
Solvent
Additive
(10 mol%)
Conversion [%]
[*] I. Dubovyk, D. Pichugin, Prof. Dr. A. K. Yudin
Davenport Research Laboratories, Department of Chemistry
University of Toronto
1
2
3
4
5
P(OEt)3
P(OEt)3
P(OEt)3
–
THF
–
–
0
50
quant.
CH2Cl2
CH2Cl2
CH2Cl2
THF
80 St. George St., Toronto, ON, M5S3H6 (Canada)
Fax: (+1)416-946-7676
E-mail: ayudin@chem.utoronto.ca
morpholine
morpholine
morpholine
0
0
P(OEt)3
[**] We thank the Natural Science and Engineering Research Council
(NSERC) for financial support.
[a] Reaction conditions: palladium catalyst/ligand/morpholine/sub-
strate/TFA=1:4:10:40:40 in solvent (0.5m) at reflux, overnight.
PMB=p-methoxybenzyl, THF=tetrahydrofuran.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201100612.
5924
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 5924 –5926

tetrahydroazepine 1a gave the corresponding 2-prenyl pyrro-
lidine 1b in 50% yield after 8 hours (Table 1, entry 2).
Trifluoroacetic acid (TFA; 1 equiv) was employed in order
to activate the amine. Interestingly, the addition of 10 mol%
of morpholine pushed the reaction to completion (Table 1,
entry 3). We think that in the presence of TFA, morpholine
acts as a buffer and enables controlled protonation of the
starting material. In addition, it can assist in the activation of
the palladium catalyst. Importantly, replacing morpholine
with N-methyl morpholine gave no change to the reaction
efficiency. With no ligand present,
Fused heterocycles were shown to work as substrates
during ring contraction (Table 2, entries 7 and 8), although
competing allylic rearrangement of phenol ethers was also
observed under the reaction conditions (Table 2, entry 7).
Nonetheless, both 7a and 8a rearranged fully at 408C in
dichloromethane. Significantly, the substrates containing
seven- and eight-membered rings led to ring contraction in
high yields (Table 2, entries 9–12). As can be seen (Table 2,
entries 1, 9, and 10), alleviation of ring strain overrides the
substitution pattern of the olefin as the driving force. In cases
the rearrangement of 1a provided
Table 2: Synthesis of cyclic allyl amines via ring rearrangement.[a]
no conversion (Table 1, entry 4). In
addition, uncatalyzed olefin isomer-
ization was not observed when
either 1a or 1b were used as starting
materials. Switching the solvent to
THF with or without the addition of
morpholine gave no conversion
(Table 1, entries 1 and 5). The reac-
tion with 5 mol% of TFA (with
morpholine added) gave little con-
version.
Entry Substrate
Product
Yield [%][b]
Entry Substrate Product
Yield [%][b]
1
97
8
78
The substrate scope of the reac-
tion is shown in Table 2. It includes
ring-contraction reactions of seven-
and eight-membered allylic amines
and a ring-expansion of allylic aze-
tidine (Table 2, entry 13). All of the
starting materials were prepared
using metathesis protocols and
other standard reactions.[12] When
the optimized reaction conditions
(Table 1, entry 3) were applied to
2a, no conversion to the corre-
sponding pyrrolidine 2b was
observed even when the solvent
was switched to dichloroethane at
608C (Table 2, entry 2). Placing
substituents in a 2,4 relation and
carrying out the reaction in dichloro-
ethane did improve conversion; 3a
produced 3b with a 2:1 diastereose-
lectivity in modest yield, with the
unreacted starting material making
up for the mass balance (Table 2,
entries 3). Likewise, 4a gave the
corresponding pyrrolidine 4b with
both vinyl and isopentyl substituents
in a trans relation. When the sub-
stituents were placed in a 2,5 relation
in a larger ring, 6b was produced in
good yield and 3:1 diastereoselectiv-
ity (Table 2, entry 6). Because the
reaction proceeds under the thermo-
dynamic control, the product ratio is
proportional to the difference in the
ground-state stability of the possible
diastereomers.
94
2[c]
0
9
(E/Z=3.6:1)
3[c]
53 (2:1)
10
93
97
4[c]
50 (2:1)
11
92
5[c]
68 (1:1)
71 (3:1)
12
(E/Z=4.6:1)
6
13[d]
96
7
83 (1:1)
[a] Reaction conditions: palladium catalyst/ligand/morpholine/substrate/TFA=1:4:10:40:40 in CH2Cl2
(0.5m) at reflux, overnight. [b] Yield of isolated product. [c] Palladium catalyst/ligand/morpholine/
substrate/TFA=1:4:10:20:20 in DCE (0.5m) at reflux, overnight. [d] Palladium catalyst/ligand/
morpholine/substrate/TFA=1:4:10:100:100 in CH2Cl2 (0.2m) at room temperature, overnight. Bn=
benzyl, DCE=1,2-dichloroethane.
Angew. Chem. Int. Ed. 2011, 50, 5924 –5926
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5925

Communications
[3] a) I. Delso, T. Tejero, A. Goti, P. Merino, Tetrahedron 2010, 66,
1220; b) K. P. Kaliappan, P. Das, S. T. Chavan, S. G. Sabharwal, J.
Org. Chem. 2009, 74, 6266; c) F. Cardona, G. Moreno, F. Guarna,
P. Vogel, C. Schuetz, P. Merino, A. Goti, J. Org. Chem. 2005, 70,
6552; d) L. E. Overman, R. M. Burk, Tetrahedron Lett. 1984, 25,
5739; e) J. L. Stavinoha, P. S. Mariano, A. Leone-Bay, R.
Swanson, C. Bracken, J. Am. Chem. Soc. 1981, 103, 3148.
[4] a) A. Fuerstner, J. W. J. Kennedy, Chem. Eur. J. 2006, 12, 7398;
b) G. A. Molander, J. A. C. Romero, Tetrahedron 2005, 61, 2631.
[5] a) G. T. Rice, M. C. White, J. Am. Chem. Soc. 2009, 131, 11707;
b) K. J. Fraunhoffer, M. C. White, J. Am. Chem. Soc. 2007, 129,
7274; c) R. M. Trend, Y. K. Ramtohul, B. M. Stoltz, J. Am.
Chem. Soc. 2005, 127, 17778.
where a disubstituted olefin could be formed, the E geometry
is preferred (Table 2, entries 9 and 12).
In summary, we have demonstrated the use of acid-
promoted palladium-catalyzed isomerization in the skeletal
rearrangement of cyclic amines. As complex cyclic allyl
amines can be prepared by a multitude of methods, our study
should open up new retrosynthetic possibilities. One might be
particularly excited about nontrivial disconnections that
would involve functional group interconversions by allyl
amine dehydrogenation as the first retrosynthetic step.
[6] a) J. F. Teichert, M. Faꢁanꢂs-Mastral, B. L. Feringa, Angew.
Chem. 2011, 123, 714; Angew. Chem. Int. Ed. 2011, 50, 688; b) C.
Shi, I. Ojima, Tetrahedron 2007, 63, 8563; c) B. Ferber, G.
Prestat, S. Vogel, D. Madec, G. Poli, Synlett 2006, 13, 2133; d) K.
Ito, S. Akashi, B. Saito, T. Katsuki, Synlett 2003, 12, 1809; e) B.
Ferber, S. Lemaire, M. M. Mader, G. Prestat, G. Poli, Tetrahe-
dron Lett. 2003, 44, 4213; f) Y. Hamada, I. Kunimune, O. Hara,
Heterocycles 2002, 56, 97; g) A. R. Katritzky, J. Yao, B. Yang, J.
Org. Chem. 1999, 64, 6066; h) B. M. Trost, M. J. Krische, R.
Radinov, G. Zanoni, J. Am. Chem. Soc. 1996, 118, 6297.
[7] a) B. D. Stubbert, T. J. Marks, J. Am. Chem. Soc. 2007, 129, 4253;
b) L. Ackermann, R. G. Bergman, R. N. Loy, J. Am. Chem. Soc.
2003, 125, 11956; c) T. E. Mueller, M. Beller, Chem. Rev. 1998,
98, 675; d) M. Meguro, Y. Yamamoto, Tetrahedron Lett. 1998, 39,
5421; e) M. Al-Masum, M. Meguro, Y. Yamamoto, Tetrahedron
Lett. 1997, 38, 6071; f) D. N. A. Fox, T. Gallagher, Tetrahedron
1990, 46, 4697; g) R. Kinsman, D. Lathbury, P. Vernon, T.
Gallagher, J. Chem. Soc. Chem. Commun. 1987, 243; h) S.
Arseniyadis, J. Goroe, Tetrahedron Lett. 1983, 24, 3997.
[8] I. D. G. Watson, A. K. Yudin, J. Am. Chem. Soc. 2005, 127,
17516.
Experimental Section
Allyl palladium chloride dimer (0.01 mmol, 3.6 mg) was added to a
flame-dried vial equipped with a stir bar. The complex was dissolved
in anhydrous dichloromethane (0.80 mL) under nitrogen. A mixture
of triethyl phosphite (0.04 mmol, 6.6 mg) and morpholine (0.05 mmol,
4.36 mg) were added to the reaction mixture, and it was stirred for
5 min. Then 1a (0.40 mmol, 98 mg) was dissolved in anhydrous
dichloromethane (1.0 mL) and was added to the reaction mixture,
followed by the addition of TFA (0.40 mmol, 45.6 mg). The reaction
vial was sealed and stirred at reflux for 10 h. The reaction mixture was
cooled to RT, washed with saturated solution of sodium bicarbonate,
extracted with dichloromethane, dried with sodium sulfate and
concentrated. The crude material was purified by flash chromatog-
raphy on silica gel (hexanes/EtOAc = 9:1, Rf = 0.36) to give 1b
(0.39 mmol, 96 mg, 97%) as an orange oil.
Received: January 24, 2011
Published online: May 12, 2011
Keywords: allyl amines · allylic rearrangement · isomerization ·
.
[9] I. Dubovyk, I. D. G. Watson, A. K. Yudin, J. Am. Chem. Soc.
2007, 129, 14172.
[10] B. M. Trost, M. Osipov, G. Dong, J. Am. Chem. Soc. 2010, 132,
15800.
[11] Unpublished results.
[12] See the Supporting Information.
palladium · synthetic methods
[1] a) P. A. Evans, J. E. Robinson, J. D. Nelson, J. Am. Chem. Soc.
1999, 121, 6761; b) B. Weiner, A. Baeza, T. Jerphagnon, B. L.
Feringa, J. Am. Chem. Soc. 2009, 131, 9473.
[2] E. Steglich, B. Fugmann, S. Lang-Fugmann in ROMPP Ency-
clopedia Natural Products, Vol. 1 (Eds: H. Baltes, W. Gꢀpel, J.
Hesse), Thieme, Stuttgart, 1997.
5926 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 5924 –5926