Platinum-catalyzed cycloisomerization of 1,4-enynes via activation of a sp3-hybridized C-H bond

Article abstract of DOI:10.1039/c0cc00071j

We report the cycloisomerization of 1-alkenyl-1-alkynylcyclopropanes to cyclooctatriene products catalyzed by PtCl₂/CO in hot xylene. In contrast to the reported enyne cycloisomerization, this 1,4-enyne cycloisomerization proceeds via an atypical addition of the allyl carbon to the alkyne in a 6-endo-dig cyclization.

Full text of DOI:10.1039/c0cc00071j

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Platinum-catalyzed cycloisomerization of 1,4-enynes via activation
of a sp3-hybridized C–H bondw
Dhananjayan Vasu, Arindam Das and Rai-Shung Liu*
Received 24th February 2010, Accepted 31st March 2010
First published as an Advance Article on the web 29th April 2010
DOI: 10.1039/c0cc00071j
We report the cycloisomerization of 1-alkenyl-1-alkynylcyclo-
propanes to cyclooctatriene products catalyzed by PtCl2/CO in
hot xylene. In contrast to the reported enyne cycloisomerization,
this 1,4-enyne cycloisomerization proceeds via an atypical addition
of the allyl carbon to the alkyne in a 6-endo-dig cyclization.
Table 1 Catalyst screening and optimized conditions
Temperature
(time/h)
Platinum- and gold-catalyzed cycloisomerizations of enyne
functionalities1,2 have been studied exclusively for 1,5- and
1,6-enynes with very few examples for 1,4-enynes.3 This
synthetic approach provides useful and unusual carbocyclic
products that are not readily prepared by common methods.
The carbocyclic rings resulting from such reactions arise
invariably from a bond connection between the alkenes and
alkynes through various endo- or exo-cyclizations,1,2 as
depicted in Scheme 1. We sought to expand the scope of enyne
cycloisomerizations so as to generate carbocyclic rings
through the addition of the allylic carbon to the tethered
alkyne, as depicted with 1,4-enyne substrate 1a (Table 1).
The mechanistic significance of this novel process is the
activation of the allylic hydrogen with a p-alkyne functionality
that has no precedents.4,5
Entry Catalysta
Additive
Yieldb
1
2
3
4
5
PtCl2/CO (5)
120 1C (2)
120 1C (8)
120 1C (24)
120 1C (24)
—
—
—
—
—
2a (63%)
2a (44%)
1a (88%)
1a (86%)
Messy
PtCl2 (5)
AuCl3 (5)
AuCl (5)
PPh3AuCl (5)/ 120 1C (5)
AgOTf (5)
PtI2(5)
6
120 1C (6)
120 1C (1)
25 1C (3)
120 1C (24)
120 1C (24)
120 1C (24)
—
—
—
Messy
3a0 (63%)
3a (68%)
7
8
9
10
11
HOTf (1)
HOTf (1)
PtCl2/CO (5)
PtCl2/CO (5)
—
2,6-Lutidine (5) 2a (43%)
1a (85%)
1a (89%)
CuBr (20)
—
a
The values in parentheses represent mol% for catalysts and
b
additives. Yields are reported after separation on a silica column.
unreacted 1a in 86–91% under the same conditions.7
We observed no olefin isomerization of 1a in the latter
solvents.
Table 1 shows the optimized conditions for the cyclo-
isomerization of 1,4-enyne 1a. We selected this compound
for the allylic C–H bond activation because its cyclopentylidene
group impedes a conventional 5-endo-dig cyclization toward
the tethered alkyne. With PtCl2/CO (5 mol%)6 as the catalyst,
the cyclization efficiency is highly dependent on the solvent in
which the reaction was performed in a sealed tube. We
obtained desired cyclized compound 2a with 63% yield in
hot p-xylene (120 1C, 2 h), far superior to dichloroethane,
acetonitrile and DMF that gave exclusive recovery of
The use of PtCl2 alone gave a diminished yield of 2a (44%).
AuCl and AuCl3 each at 5 mol% loading led to exclusive
recovery of 1a whereas PPh3AuCl/AgOTf gave a messy
mixture of products in hot xylene. This observation again reflects
the unsuitability4a of gold catalysts to activate the C–H bond
induced with a p-alkyne. Under the same conditions, PtI2
(5 mol%) also provided a messy mixture of products (entry 6),
whereas HOTf (1 mol%) in hot xylene (120 1C, 1 h) gave
distinct product 3a0 in 63% yield (entry 7), but 5-cyclopropyl-
1,3-hexadiene 3a in 68% yield near 25 1C. Added 2,6-lutidine
(5 mol%) gave a decreased yield (43%) of desired 2a, whereas
CuBr8 (20 mol%) completely inhibited the reaction (entries 9
and 10). Thermal activation alone failed to give a traceable
amount of 2a (entry 11).
We prepared various 1-alkenyl-1-arylalkynylcyclopropanes9
1b–1i and 1j–1l to assess the generality of such a bicyclo-
[6.3.0]undecatriene synthesis (Table 2). Notably, resulting
products 2b–2i and 2j–2l have varied olefin positions, according
to 1H NMR analysis. The electronic effect of 4-phenyl
substituent X has a pronounced effect on product yield. For
X = Me, F and Cl derivatives (1b–1d), we obtained desired
products 2b–2d in 51–61% yields. This platinum-catalyzed
cycloisomerization is inapplicable to 1,4-enyne 1e bearing a
para-methoxy group, which we recovered exclusively even
after a long period (24 h). The cycloisomerization works well
Scheme 1
Department of Chemistry, National Tsing Hua University, Hsinchu,
30013, Taiwan (ROC). E-mail: rsliu@mx.nthu.edu.tw;
Fax: (+886) 3-5711082
w Electronic supplementary information (ESI) available: Experimental
details and characterization data and 1H and 13C NMR spectra of
compounds 1a–8. CCDC 755454. For ESI and crystallographic data in
CIF or other electronic format see DOI: 10.1039/c0cc00071j
ꢀc
This journal is The Royal Society of Chemistry 2010
Chem. Commun., 2010, 46, 4115–4117 | 4115

View Article Online
Table 2 Catalytic synthesis of bicyclo[6.3.0]undecatrienes
Compound
(yield)b
Entry
Enynea
Time/h
1
2
3
4
X = Me (1b)
X = F (1c)
X = Cl (1d)
X = OMe (1e)
X = C(QO)Me
(1f)
1.5
1
1
24
2b (61%)
2c (51%)
2d (55%)
1e (75%)
5
2.5
2f (80%)
Scheme 2
6
7
8
X = NO2 (1g)
X = CO2Et (1h)
X = CN (1i)
4.5
5
5
2g (79%)
2h (84%)
2i (72%)
allylic C–H activation proceeds more efficiently with a cyclo-
pentylidene than a cyclohexylidene ring. Again, we observed
enhanced yields (81–84%) of products 5b–5e bearing an
electron-withdrawing ester, acyl, cyano and nitro moieties at
the 4-phenyl position.
The preceding cyclization might occur with an initial
platinum-catalyzed olefin isomerization of 1,4-enyne 1j to
1,5-enyne 1j0 (Scheme 2) that ultimately gave desired bicyclo-
[6.3.0]undecatriene 2j through a 6-endo-dig cyclization. To test
this hypothesis, we prepared 1,5-enyne 1j0 from a separate
route; its treatment with PtCl2/CO in xylene at room tempera-
ture leads to a distinct cycloisomerization/oxidation cascade,
giving 1,2-dihydrocyclobutabenzene 6 in 73% yield.
9
10
11
X = H (1j)
X = Me (1k)
X = Cl (1l)
2.5
2
2
2j (67%)
2k (61%)
2l (71%)
a
b
[Enyne] = 0.5 M. Products are reported after separation on a silica
column.
1
The structure of compound 6 was determined by H-NOE
and confirmed by X-ray crystallographic study.9 The absence
of desired 2j excludes the participation of 1,5-enyne 1j0 as an
intermediate for initial 1,4-enyne 1j. We envisage that the
formation of bicyclic benzene 6 likely proceeds from inter-
mediate B, of which the cyclopropyl and platinum-dienyl
groups activate oxidation of the methylene protons with
protons or residual oxygen. This proposed mechanism
provides a convincing rationale for the HOTf-catalyzed
carbocyclization that gave distinct products 3a and 3a0
(Table 1, entries 7–8).
with enynes 1f–1i bearing an electron-withdrawing group such
as acyl, nitro, ester and cyano, giving corresponding products
2f–2i in satisfactory yields. On comparison of their respective
yields, we found the cycloisomerization of 1,4-enynes 1j–1l
bearing a 1-indanylidene moiety to be more efficient than of
their cyclopentylidene analogues 1a, 1b and 1d.
This platinum catalysis is extensible to the synthesis of
bicyclo[6.4.0]dodecatrienes 5a–5e through the cycloisomerization
of 1,4-enynes 4a–4e bearing a cyclohexylidene group (Table 3).
For unsubstituted substrate 4a, its corresponding product 5a
was obtained in 56% yield, less than that (67%) of bicyclo-
[6.3.0]undecatriene 2j. This information indicates that the
Eqn (1) depicts a crucial control experiment to assist in the
understanding of the reaction mechanism. For 1,4-enyne 7, its
PtCl2-catalyzed cycloisomerization proceeds even under
ambient conditions (25 1C, 8 h), giving ring-expansion product
8 in 78% yield. Such a low temperature excludes the involvement
of a Brønsted acid such as PtCl2(H2O) to effect an olefin
isomerization.7 The ease of this cycloisomerization is attributed
to a small angle y = 111.71 between the interacting alkene and
alkyne substituents.10 In contrast, the corresponding angle for
1-alkenyl-1-alkynylcyclopropane 1j is as large as 116.01,
rendering the cyclization difficult. More importantly, the
formation of compound 8 involves a 1,2-migration of its cyclo-
pentyl group toward the alkynyl rather than alkenyl group.11
Table 3 Catalytic synthesis of bicyclo[6.4.0]dodecatriene
Entry
Enynea
Time/h
Compound (yield)b
1
2
3
4
5
X = H (4a)
4
5a (56%)
5b (84%)
5c (83%)
5d (81%)
5e (83%)
X = CO2Me (4b)
X = COMe (4c)
X = CN (4d)
X = NO2 (4e)
5.3
5.3
5.0
5.3
ð1Þ
a
b
[Enyne] = 0.5 M. Products are reported after separation on a silica
column.
Scheme 3 shows a plausible mechanism involving a p-alkyne-
activated 1,6-hydrogen shift of 1,4-enynes. This mechanism is
ꢀc
This journal is The Royal Society of Chemistry 2010
4116 | Chem. Commun., 2010, 46, 4115–4117

View Article Online
See: (a) X. Shi, D. J. Gorin and F. D. Toste, J. Am. Chem. Soc.,
2005, 127, 5802; (b) A. Buzas and F. Gagosz, J. Am. Chem. Soc.,
2006, 128, 12614; (c) N. Marion, S. Diez-Gonzalez, P. De Fremont,
A. R. Noble and S. P. Nolan, Angew. Chem., Int. Ed., 2006, 45,
3647.
4 Metal-catalyzed alkyne-induced hydrogen shifts can be initiated by
p-alkyne4 or metal vinylidene intermediates,5 see: (a) I. D. Jurberg,
Y. Odabachian and F. Gagosz, J. Am. Chem. Soc., 2010, 132, 3543;
(b) P. A. Vadola and D. Sames, J. Am. Chem. Soc., 2009, 131,
16525; (c) D. Shikanai, H. Murase, T. Hada and H. Urabe, J. Am.
Chem. Soc., 2009, 131, 3166; (d) S. Yang, Z. Li, X. Jian and C. He,
Angew. Chem., Int. Ed., 2009, 48, 3999; (e) J.-J. Lian, C.-C. Lin,
H.-K. Chang, P.-C. Chen and R.-S. Liu, J. Am. Chem. Soc., 2006,
128, 9661.
Scheme 3
5 (a) M. Tobisu, H. Nakai and N. Chatani, J. Org. Chem., 2009, 74,
5471; (b) A. Odedra, S. Datta and R.-S. Liu, J. Org. Chem., 2007,
72, 3289; (c) G. B. Bajracharya, N. K. Pahadi, I. D. Gridnev and
Y. Yamamoto, J. Org. Chem., 2006, 71, 6204; (d) S. Datta,
A. Odedra and R.-S. Liu, J. Am. Chem. Soc., 2005, 127, 11606.
6 For the original use of PtCl2/CO, see: (a) A. Furstner, P. W. Davies
and T. Gress, J. Am. Chem. Soc., 2005, 127, 8244; (b) A. Furstner
and P. W. Davies, J. Am. Chem. Soc., 2005, 127, 15024.
7 We obtained recovery yields of starting 1a for the following
solvents in a sealed tube (120 1C, 24 h), DMF (91%), CH3CN
(86%) and 1,2-dichloroethane (90%). This information suggests
that olefin isomerization, as exemplified by 1j - 1j0 is not a facile
process using PtCl2/CO alone.
proposed based on an activity/structure relationship and a
control experiment. We propose that the initial step involved a
p-alkyne-activated hydride shift,12 giving alkenylplatinum
species E. This process is accelerated by an electron-with-
drawing group at the aryl group because of a stabilization of
the negative charge developing on the platinum center.13
Intramolecular cyclization of species E via an attack of
alkenylplatinum at the allyl cation forms cyclohexenyl carbenoid
species F, further inducing an expansion of the cyclo-
propyl ring.
8 The enhancement of CuBr on benzyl C–H bond activation was
reported by He and co-workers, see ref. 4d.
The support of this cyclopropyl shift is provided by the
isomerization of 1,4-enyne 7 to compound 8. Species G
contains a strained cyclobutene fragment, and is prone to a
rapid olefin isomerization to give thermodynamically stable
species H, catalyzed by platinum or via a thermal 1,5-hydrogen
shift. A ring opening of species H through a retro 6-p-electro-
cyclization forms bicyclo[6.3.0]undecatriene I that ultimately
produces the observed products 2a–2d and 2f–2i.
9 Preparation details for 1-alkenyl-1-alkynylcyclopropanes are
provided in the ESIw; CCDC 755454.
10 Optimizations of the molecular structures of compounds 1j and 7
were performed using B3LYP/6-31G* opt.
11 The mechanism of formation of compound 8 is also rationalized
according to the pathway below. According to the mechanism
(Scheme 3) starting species 7 will form compound J that is prone to
oxidation in solution, as activated by the two phenyl groups, to
give the observed product 8 ultimately.
In summary, we report the cycloisomerization of 1-alkenyl-
1-alkynylcyclopropanes to cyclooctatriene products catalyzed
by PtCl2/CO in hot xylene. In contrast to reported enyne
cycloisomerization, this 1,4-enyne cycloisomerization proceeds
via an atypical addition of the allyl carbon to the alkyne via a
6-endo-dig cyclization. Control experiments exclude the
involvement of Brønsted acid and a prior alkene isomerization.
On the basis of experimental data, we propose a carbocyclization
involving a p-alkyne activated 1,6-hydride shift as the key step.
In a control experiment, we discovered that such a carbo-
cyclization can proceed at room temperature if the allyl C–H
bond is near its tethered functionality. We believe that this
original observation will assist the design of novel catalytic
reactions.
.
12 We have performed a deuterium-labeling experiment to elucidate
the reaction mechanism. We prepared d-4c that gave desired d-5c
with deuterium at the three olefin protons following the same
catalytic sequence. However, this observation is probably mean-
ingless because the olefin protons of compound 5c undergo deute
rium exchange with external D2O.
The authors wish to thank the National Science Council,
Taiwan for supporting this work.
Notes and references
1 Leading references for gold catalyzed 1,n-enyne cycloisomerizations:
(a) E. Jimenez-Nunez and A. M. Echavarren, Chem. Rev., 2008, 108,
3326; (b) E. Jimenez-Nunez and A. M. Echavarren, Chem. Commun.,
2007, 333; (c) L. Zhang, J. Sun and S. A. Kozmin, Adv. Synth. Catal.,
2006, 348, 2271.
2 (a) A. Furstner and P. W. Davies, Angew. Chem., Int. Ed., 2007, 46,
3410; (b) D. J. Gorin and F. D. Toste, Nature, 2007, 446, 395;
(c) A. S. K. Hashmi, Angew. Chem., Int. Ed., 2005, 44, 6990.
3 All reported 1,4-enynes comprised a tethered acetate that is prone
to a 1,3-shift to give allenyl acetates; such gold-catalyzed carbo-
cyclizations actually proceed through an allene-ene cyclization.
.
13 We observed no activity for substrate 1e (Table 2, entry 4) because
its 4-methoxyphenyl substituent will destabilize the alkenyl-
platinum functionality of hypothetical intermediate E.
ꢀc
This journal is The Royal Society of Chemistry 2010
Chem. Commun., 2010, 46, 4115–4117 | 4117