An unprecedented stereoselective [2+2] cycloreversion of cyclobutanones

Article abstract of DOI:10.1039/c0cc02694h

A one-pot metathetic ring opening of substituted bicyclo[3.2.0]heptenones to linear polyene ketones is described. The reaction proceeds with conservation of the endocyclic (Z)-double bond. Exo-substituted vinylcyclobutanones result in (4Z,6E)-configured t

Full text of DOI:10.1039/c0cc02694h

View Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
An unprecedented stereoselective [2+2] cycloreversion
of cyclobutanonesw
Xiufang Ji,a Quanrui Wang*a and Andreas Goeke*b
Received 21st July 2010, Accepted 23rd September 2010
DOI: 10.1039/c0cc02694h
A one-pot metathetic ring opening of substituted bicyclo[3.2.0]-
heptenones to linear polyene ketones is described. The reaction
proceeds with conservation of the endocyclic (Z)-double bond.
Exo-substituted vinylcyclobutanones result in (4Z,6E)-configured
trienones, while endo-substituted phenyl derivatives generate
(4Z,6Z)-configured dienones.
Table 1 Screening of reaction parameters for the cycloreversion of
vinyl bicycloheptenonesa
Vinylcyclobutanones are versatile building blocks in preparative
organic chemistry as they undergo a remarkable diversity of
rearrangement reactions.1 In particular, transformations of
easily accessible addition products of cyclopentadiene and
vinylketenes2 have been thoroughly investigated over the past
30 years. For instance, Dreiding and Danheiser reported the
thermal or acid-catalyzed Cope rearrangement of I to bicyclo-
[4.2.1]nonenones II (Scheme 1).3 This reaction was often
accompanied by [1,3]-rearrangements to cyclohexenones III,
or, depending on the substitution pattern, by a [1,2]-migration
to IV.4 Paquette et al. used I in an elegant one-pot alkenylation/
oxy-Cope/methylation sequence for the synthesis of the struc-
tural core V of ophiobolins and other natural products.5
Finally, Danheiser et al. reported the oxyanion-accelerated
[1,3]-rearrangement of in situ reduced I to cyclohexenols VI.6
Similar 2-carbon ring extensions were described by Cohen
et al. who elaborated the method into a synthesis of
(–)-b-selinene.7,8 Importantly, in all these rearrangements,
the cyclopentene ring does not open.
Entry
Reagent
Additive (mol%)
Product (R)
Yield [%]c
1
2
3
4
5
6
7
8b
9
n-BuLi
n-BuLi
n-BuLi
n-BuLi
n-BuLi
n-BuLi
n-BuLi
MeMgBr
KHd
—
3a, (n-Bu)
2a, (n-Bu)
2a, (n-Bu)
2a, (n-Bu)
2a, (n-Bu)
2a, (n-Bu)
3a, (n-Bu)
3b, (Me)
85
15
TMEDA (10)
DMPU (5)
HMPT (5)
DMPU (300)
HMPT (300)
12-C-4 (10)
TMEDA (10)
DMPU (5)
79 (68)d
55
45
51
89
80
—
Decomp.
a
The addition of n-BuLi (1.05 equiv.) to the bicyclic ketone 1a was
performed in THF at ꢀ78 1C , then added the catalyst and refluxed for
b
c
6 h. The reaction was carried at 0 1C, then refluxed for 6 h. Isolated
yield after chromatography. Alcohol 3a was used as substrate.
d
HMPT as complexing agents to yield functionalized polyene
ketones 2 with high chemo- as well as diastereoselectivity
(Table 1). It is important to note that this metathetic cyclo-
reversion is different from a formal retro-[2+2] cycloaddion to
ketone 4; such an oxyanion-promoted fragmentation had
indeed been described by Snider.9 Polyenes of type 2 constitute
an important substructure in pheromones,10 retinoids11 and
complex antibiotics.12 For instance, (3E,5Z)-1,3,5-undecatriene
5 occurs in many natural essential oils, e.g. pineapple and
galbanum.13 The (5Z)-double bond is important for its
extremely low odor threshold and olfactory quality.
Herein, we report a novel reaction of compounds I: the
oxyanion-promoted [2+2]cycloreversion of cyclobutanones 1
triggered by an in situ alkylation with alkyl or aryllithium
reagents in the presence of catalytic amounts of DMPU or
Recently, keto and hydroxy derivatives 6 were also claimed
for its use in the flavor and fragrance industry. These com-
pounds were prepared by Wittig reactions of a,b-unsaturated
aldehydes.14 However, stabilized ylides usually do not react
diastereoselectively and produce considerable amounts of
trans-configured olefins that possess undesirable odor properties.
Therefore, we decided to elaborate this cycloreversion reaction
also with respect to a structure–odor correlation study
of new undecatriene analogues.15 While the reaction of 1a
Scheme 1 Rearrangements of bicycloheptenones.
a Department of Chemistry, Fudan University, 220 Handan Road,
Shanghai 200433, China. E-mail: qrwang@fudan.edu.cn;
Fax: (+86)(21) 65641740; Tel: (+86)(21) 65643978
b Givaudan Fragrances (Shanghai) Ltd., 298 Li Shi Zhen Road,
Shanghai 201203, China. E-mail: andreas.goeke@givaudan.com
w Electronic supplementary information (ESI) available: Experimental
procedures, characterization of new compounds and spectra. See DOI:
10.1039/c0cc02694h
c
This journal is The Royal Society of Chemistry 2010
Chem. Commun., 2010, 46, 8845–8847 8845

View Online
with n-BuLi in refluxing THF resulted in the formation of the
trivial adduct 3a (Table 1, entry 1), the cycloreversion of the
intermediate lithium alkoxide only occurred in the presence of
Li-chelating agents. Thus, using TMEDA as an additive led to
trienone 2a in a yield of 15% (Table 1, entry 2).
the ring-opening occurred, the cis-configured double bond
in 1a retained its configuration in the (4Z,6E)-configured
trienone 2a.
Since rate accelerations were achieved by the introduction
of silyloxy groups into Cope systems,17 an attempt to also
facilitate the cycloreversion started from silylether 7 which was
subjected to thermal conditions up to 300 1C. However, only
the Cope rearrangement product 8 was isolated after in situ
hydrolysis (Scheme 2). Apparently, a more dissociated charac-
ter of the intermediate metal alkoxide is necessary for the
cycloreversion to proceed.
This initial result could be improved by using catalytic
amounts of DMPU or HMPT with increased Li-binding
affinities (Table 1, entries 3 and 4).16 The use of stoichiometric
or higher concentrations of DMPU (Table 1, entry 3 vs. 5)
did not further improve yields. Application of the crown
ether 12-C-4 likewise failed to promote the ring opening
(Table 1, entry 7). Furthermore, a magnesium alkoxide
(Table 1, entry 8) generated from the reaction of MeMgBr
to the convex face of 1a also failed to undergo the cyclo-
reversion. However, the stepwise protocol via compound 3a
resulted in a higher yield of ketone 2 (Table 1, entry 3) but
an attempt of performing the same cycloreversion using
KH in refluxing THF led to the decomposition of the starting
material (Table 1, entry 9). Importantly, in all cases in which
Following this surprising finding of the cycloreversion
which contrasts Danheiser’s earlier results of the 1,3-ring
expansion sequence,6 we addressed questions regarding the
scope of this new ring opening (Table 2). The substrates 1a–d
and 9 were synthesized by [p2s+p2a] cycloadditions of cyclo-
pentadiene or 5,5-ethylenecyclopenta-1,3-diene with ketenes
(prepared in situ via dehydrochlorination of the corresponding
a,b-unsaturated acid chlorides using triethylamine).2b,4a
It is well known that compound 9 is formed with high
endo-selectivity while vinyl-substituted cycloadducts 1 are
generated with exo-preference.4b Upon addition of organo-
lithium reagents to substrates 1 and 9 followed by heating of the
resulting intermediates 11, these selectivities were translated
into aryldienones 10 with exclusive (4Z,6Z)-configuration
(Table 2, entries 1–3) and into (4Z,6E)-configured trienones 2
with different substitution patterns (Table 2, entries 4–8).
Scheme 2 Thermal rearrangement of silylether 7.
Table 2 Ring opening of substituted bicycle[3.2.0]heptenonesa
Entry
1
Bicyclus (polyene)
R1
Ph
R2
R3
H
R4
t/h
Yield (%)b (6E : 6Z)c
9
Me
n-Bu
2
77
(o2 : 98)
(10a)
9
(10b)
9
(10c)
1a
(2b)
1a
(2c)
1b
(2d)
1c
(2e)
1a
(2f)
2
3
4
5
6
7
8
Ph
Me
H
H
H
H
s-Bu
Ph
1
73
(o2 : 98)
51
Ph
Me
6
(o2 : 98)
Me
Me
Me
Me
Me
vinyl
vinyl
propenyl
vinyl
vinyl
s-Bu
Me
2
71
(498 : 2)
46
1
(498 : 2)
31
(8 : 2)
60e
H
n-Bu
n-Bu
Ph
12
4
d
(8 : 2)
27
H
6
(9 : 1f)
a
The addition of R4Li (1.05 equiv.) to the bicyclic ketone 1 or 9 in the presence of DMPU (0.05 equiv.) was performed in THF at ꢀ78 1C, then
refluxed for the time indicated. Isolated yield after chromatography. E/Z ratio determined by NMR. R3–R3 = CH2–CH2. Isolated yield
f
after distillation. E/Z ratio determined by GC-MS.
b
c
d
e
c
8846 Chem. Commun., 2010, 46, 8845–8847
This journal is The Royal Society of Chemistry 2010

View Online
may cause a carbanion-acceleration of this hydrogen shift.22
Although the outcome of these cycloreversions is in accordance
with an allowed retro-[s2s+s2a] ring opening, the reaction
seems to rather involve a stepwise mechanism via intermediate
Li-complexes 12.7 Its formation corresponds to a reversed allyl
(benzyl) lithium addition18 of initially formed compounds 11.
Such retro-allyl lithium additions were found to be about ten
times faster than the reversal of the corresponding magnesium
alkoxides.18a This also explains why Grignard addition products
of 1 did not at all undergo the ring-opening under standard
reaction conditions (Table 1, entry 8).
For geometrical reasons, this formal [1,7]-H-shift cannot occur
in the (6E)-configured intermediate enolate 13 (Table 2).
In conclusion, we have uncovered a novel metathetic cyclo-
reversion of 7-vinyl and aryl bicyclo[3.2.0]hept-2-en-6-ones
giving rise to linear polyene ketones with high selectivity. This
ring opening offers a fascinating two-step route to olfactory
useful compounds from easily available starting materials.
Notes and references
The subsequent cleavage of the cyclopentenyl ring in 12 to
enolate 13, however, has to be explained by the increased
electrophilic potential of lithium19 (vs. K and Mg); although
species 12 must be regarded as a canonical structure of a
delocalized contact ion pair of allyl/benzyl anions, a tighter
geometrical stabilization of the involved intermediates in
comparison with those that involve K as the counterion are
obvious—independent from the catalytic activity of DMPU.
Intermediate 12 is expected to be highly reactive and short
lived; it rapidly fragments into more stable Li-enolates 13,
before the dynamic properties20 of the allyl anion result in a
loss of stereochemistry.
1 (a) J. E. Baldwin and P. A. Leber, Org. Biomol. Chem., 2008,
6, 36; (b) P. A. Leber and J. E. Baldwin, Acc. Chem. Res., 2002,
35, 279; (c) J. J. Bronson and R. L. Danheiser, Comprehensive
Organic Synthesis, ed. B. M. Trost, I. Flemming and L. A. Paquette,
Pergamon Press, Oxford, GB, 1st edn, 1991, vol. 5, ch 8.3, p. 1016.
2 For reviews see: (a) J. A. Hyatt and P. W. Raynolds, in Organic
Reactions, ed. L. A. Paquette, Wiley, New York, USA, 1st edn,
1994, vol. 45, ch. 2, pp. 159; (b) H. N. C. Wong, K.-L. Lau and
K.-F. Tam, Top. Curr. Chem., 1986, 133, 83–157; (c) D. Bellus and
B. Ernst, Angew. Chem., Int. Ed., 1988, 27, 797.
3 (a) R. Huston, M. Rey and A. S. Dreiding, Helv. Chim. Acta, 1984,
67, 1506; (b) R. Huston, M. Rey and A. S. Dreiding, Helv. Chim.
Acta, 1982, 65, 451; (c) R. L. Danheiser, S. K. Gee and H. Sard,
J. Am. Chem. Soc., 1982, 104, 7670.
4 (a) R. Huston, M. Rey and A. S. Dreiding, Helv. Chim. Acta, 1982,
65, 1563; (b) M. Rey, S. M. Roberts, A. Dieffenbacher and
A. S. Dreiding, Helv. Chim. Acta, 1970, 53, 417.
5 (a) L. A. Paquette, J. A. Colapret and D. R. Andrews, J. Org.
Chem., 1985, 50, 201; (b) L. A. Paquette, D. R. Andrews and
J. P. Springer, J. Org. Chem., 1983, 48, 1147.
6 R. L. Danheiser, C. Martinez-Davila and H. Sard, Tetrahedron,
1981, 37, 3943.
7 (a) T. Cohen, M. Bhupathy and J. R. Matz, J. Am. Chem. Soc.,
1983, 105, 520; (b) M. Bhupathy and T. Cohen, J. Am. Chem. Soc.,
1983, 105, 6978.
8 For these and other transformations of cyclobutanes, see:
J. C. Namyslo and D. E. Kaufmann, Chem. Rev., 2003, 103, 1485.
9 B. B. Snider and M. Niwa, Tetrahedron Lett., 1988, 29, 3175.
10 (a) O. Krebs and R. J. K. Taylor, Org. Lett., 2005, 7, 1063;
(b) P. T. Bell and W. A. Donaldson, J. Nat. Prod., 1992, 55, 1669.
11 B. H. Lipshutz, G. C. Clososki, W. Chrisman, D. W. Chung and
D. B. Ball, Org. Lett., 2005, 7, 4561.
This assumption of a reactive and short-lived allyl lithium
species is also fuelled by the conversion of less substituted
endo-1d (Scheme 3). The reaction takes exclusively the route of
the 1,3-shift (a-attack via 120b) leading to the tertiary bicyclic
alcohol 14. The bond order and electron density of this
less substituted derivative is diminished at the g-position20
rendering a less reactive species which undergoes isomeri-
zation as well as rotation around the pentanoyl group.
In comparison to the formation of compounds 10
from endo-9, a strikingly different result was observed in the
transformation of vinyl-substituted cyclobutanone endo-1a to
(E,E,Z)-configured trienone 16 as described in Scheme 4. The
diastereoselective shift of the triene unit into the b-position of
the carbonyl group can be explained by a [1,7]-H-shift in
postulated intermediate 15.21 The neighboring enolate
12 S. D. Rychnovsky, Chem. Rev., 1995, 25, 2021.
13 For a review on fragrance chemistry, see: P. Kraft, J. A. Bajgrowicz,
C. Denis and G. Frater, Angew. Chem., Int. Ed., 2000, 39, 2.
´
14 (a) K. Nakanishi, Y. Okubo, N. Tomita, T. Maeda and
N. Miyazawa, JP 4143683; (b) K. Nakanishi, Y. Okubo,
N. Tomita, N. Miyazawa and T. Maeda, JP 4057639.
15 For odor descriptions, see the Supporting Informationw.
16 T. Mukhopadhyay and D. Seebach, Helv. Chim. Acta, 1982, 65,
385.
17 C. Schneider, Synlett, 2001, 1079.
18 The reversible character of allylmetal additions to carbonyl com-
pounds is documented: (a) R. A. Benkeser, M. P. Siklosi and
E. C. Mozdzen, J. Am. Chem. Soc., 1978, 100, 2134; (b) P. Jones
and P. Knochel, J. Org. Chem., 1999, 64, 186; (c) M. Iwasaki,
S. Hayashi, K. Hirano, H. Yorimitsu and K. Oshima, J. Am.
Chem. Soc., 2007, 129, 4463.
19 (a) M. Schlosser and S. Strunk, Tetrahedron, 1989, 45, 2649;
(b) E. Moret and M. Schlosser, Tetrahedron Lett., 1984, 25, 4491.
20 For detailed studies of the structure and dynamic behaviour of allyl
lithium species, see: (a) G. Fraenkel and F. Qiu, J. Am. Chem. Soc.,
2000, 122, 12806; (b) G. Fraenkel and F. Qiu, J. Am. Chem. Soc.,
1997, 119, 3571.
Scheme 3 Cycloreversion of endo-1d.
21 A helical arrangement of an antarafacial 1,7-H-shift was described
in the conversion of previtamin D3 to vitamin D3: C. A. Hoeger,
A. D. Johnston and W. H. Okamura, J. Am. Chem. Soc., 1987,
109, 4690, However, this cannot explain the (3E,5E,7Z)-configura-
tion in triene 16.
22 For anion-accelerated 1,5-H-shifts see: I. V. Alabugin,
M. Manoharan, B. Breiner and F. D. Lewis, J. Am. Chem. Soc.,
2003, 125, 9329.
Scheme 4 Ring-opening of endo-1a followed by a 1,7-H-shift.
c
This journal is The Royal Society of Chemistry 2010
Chem. Commun., 2010, 46, 8845–8847 8847