Cyclohexyne cycloinsertion by an annulative ring expansion cascade

Article abstract of DOI:10.1002/anie.201001137

Figure Presented Cyclohexyne steps into the ring! Cyclohexyne is used for the facile synthesis of polycyclic, medium sized rings. Its insertion into cyclic ketones enables rapid access to densely functionalized building blocks.

Full text of DOI:10.1002/anie.201001137

Communications
DOI: 10.1002/anie.201001137
Ring-Expansion Reactions
Cyclohexyne Cycloinsertion by an Annulative Ring Expansion
Cascade**
Christian M. Gampe, Samy Boulos, and Erick M. Carreira*
Dedicated to Professor John D. Roberts
Cyclohexyne has long captivated the attention of scientists
and has been the focus of many theoretical and experimental
studies.[1,2] The constraints of an alkyne group in a small to
medium sized ring are manifested in its fleeting lifetime and
correspondingly drastically enhanced reactivity. The potential
application of cycloalkynes in organic synthesis has long been
considered attractive. To date, the most widely employed
“cycloalkyne” is 1,2-didehydrobenzene, or benzyne, which
was discovered in 1942 and has recently enjoyed popularity in
the context of complex molecule assembly.[3] Interestingly, the
preparative use of cyclohexyne in the synthesis of useful
building blocks is still lacking. Herein, we describe a direct,
formal cycloinsertion reaction of cyclohexyne (2) into cyclic
ketones 1, to afford medium-sized, fused rings 3 (Scheme 1).
The [2+2] photocycloaddition of two olefins yields a ki-
netically stable cyclobutane (ring strain ca. 26 kcalmolꢀ1).[7,8]
The thermal cycloaddition of small ring cycloalkyne com-
pounds to olefins generates cyclobutenes (ring strain ca. 30
kcalmolꢀ1),[8] which undergo electrocyclic ring opening
(2 minutes at 1808C).[9] In a study involving 6,6-dimethyl-1-
chlorocyclohexene as a cycloalkyne precursor and cyclo-
hexanone enolate, it was noted that the alkoxide adduct is
more prone to electrocyclic ring opening (358C),[6,10] in
analogy to the effect seen with the oxy-Cope rearrange-
ment.[11] In the context of several natural product synthesis
projects, we became interested in examining the chemistry of
cyclohexyne and cyclic ketones, their enolates or silyl enol
ethers. We envisioned a strategy involving a one-pot addition/
electrocyclic-ring-opening cascade. In such a process, cyclo-
hexyne would formally insert into the ketone and generate a
bicyclic ring system, with one of the rings undergoing ring
expansion by two atoms. The insertion reaction of cyclo-
pentanones would furnish bicyclo[5.4.0]undecane systems,
which are found in a variety of terpenoid natural products.
The development of a general cycloinsertion reaction with
cyclohexyne requires mild methods for its selective gener-
ation that prevent 1,2-cyclohexadiene formation. In this
respect, Fujita and co-workers recently described facile
preparation of cyclohexynes from l3 iodanes (4; Scheme 2)
Scheme 1. Assembly of medium-sized polycyclic carbon scaffolds by
cyclohexyne insertion reactions.
Cyclohexyne was first studied and invoked as an inter-
mediate by Roberts and Scardiglia in the substitution reaction
of cyclohexenyl chloride with PhLi.[4] Attempts to generate
and trap cyclohexyne (2) with enolates were thwarted by the
conditions employed, which led to the formation of the
putative 1,2-cyclohexadiene.[5]
A cyclohexyne derivative
could only be generated from 6,6-dimethyl-1-chlorocyclohex-
ene, in the presence of NaNH2 at 358C for 2 days, and trapped
by enolates, albeit in 25–38% yield.[6]
Scheme 2. Scouting experiments for cyclohexyne insertion into 1b.
under mild conditions at 08C.[12] In our scouting experiments,
addition of a precooled solution of 3.0 equivalents of KOtBu
in tetrahydrofuran to a solution of 2.4 equivalents of iodo-
nium 4 and cyclohexanone (1b) in tetrahydrofuran at ꢀ788C
failed to give any adducts (Scheme 2). However, we noted
that when the cold reaction was allowed to warm to room
temperature enone 3b was isolated in 55% yield. The use of
additives (molecular sieves, radical inhibitors) to minimize
side reactions of iodonium 4 did not lead to improved yields.
We speculated that under these conditions cyclohexanone
(1b) was only partially deprotonated, thus limiting the
amount of reactive species present. The use of strong amide
bases (lithium diisopropylamide, lithium tetramethylpiperi-
dide) did not lead to product formation.
[*] C. M. Gampe, S. Boulos, Prof. Dr. E. M. Carreira
Laboratorium fꢀr Organische Chemie, ETH Zꢀrich, HCI H335
Wolfgang-Pauli-Strasse 10, 8093 Zꢀrich (Switzerland)
Fax: (+41)44-632-1328
E-mail: carreira@org.chem.ethz.ch
Homepage: http://www.carreira.ethz.ch
[**] This research was supported by the ETH and the Swiss National
Science Foundation (200020-119838). A Kekulꢁ scholarship was
provided by the Fonds der Chemischen Industrie (to C.M.G.). We
are grateful to Dr. W. B. Schweizer for the X-ray crystallographic
analysis.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201001137.
4092
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 4092 –4095

Angewandte
Chemie
A study of the deprotonation of pinacolone with hindered
alkoxide bases was conducted by Brown.[13] It was noted that
in the equilibrium KOtBu + pinacoloneÐHOtBu + K-pina-
colonate Keq = 6.7, whereas Keq = 57 when KOCEt3 is used.
Consequently, we decided to examine the use of KOCEt3 as
the base in the ring-expansion reaction. Treatment of 1b with
2.5 equivalents of KOCEt3 and 1.5 equivalents of iodonium 4
afforded enone 3b in 70% yield. Cyclobutenol 5b was
isolated as a co-product from the reaction in 6% yield. It
could be shown that 5b undergoes smooth conversion into 3b
under the conditions of the cycloinsertion reaction (KOtBu,
THF), implicating the cyclobutenol adduct as the primary
reaction product.[14]
With these conditions in hand, the substrate scope of this
ring insertion reaction was further examined (Table 1).
Cyclopentanone, -hexanone, -heptanone, and -octanone
underwent cycloinsertion to give fused 7-6, 8-6, 9-6 and 10-6
bicyclic ketones, respectively, in 51–76% yield (Table 1,
entries 1–4). In certain cases (Table 1, entries 1, 3, 5, and 8)
deconjugated enones were partially formed under the stan-
dard reaction conditions, and isomerization with NaOMe in
methanol provided the corresponding conjugated isomers.[14]
Cyclooctanone (1d; Table 1, entry 4) exclusively yielded
deconjugated enone 3d as a mixture of double bond isomers.
Having investigated simple cycloalkanones, we sought to
examine more complex structures. Nopinone 1e (Table 1,
entry 5) participated smoothly in the cycloinsertion reaction
to give 3e in 74% yield. Hajos–Parrish ketone derivative 1 f
provided 3 f along with a cyclobutenol, which was converted
into 3 f under basic conditions at ambient temperature for a
total yield of 66% (Table 1, entry 6).[14]
O-Benzylestrone 1g and dihydrocholesterone 1h
(Table 1, entries 7 and 8) directly provided D- and A-ring
dihomologues 3g and 3h, respectively. Notably, ring insertion
occurs selectively into the thermodynamic enolate of 1h.[15]
The described reaction therefore enables access to unprece-
dented steroidal scaffolds. Menthone (1i; Table 1, entry 9)
selectively adds 2 through the trisubstituted enolate, provid-
ing cyclobutenol 5i, which is, however, reluctant to undergo
base-induced ring opening. This result suggests that a
substituent at the C4-position (iPr in 5i) halts the cascade at
the stage of the cyclobutenol adduct.
Sandresolide A (8) as a target offers the opportunity to
investigate more highly substituted and densely functional-
ized cyclopentanones (Scheme 3).[16] In particular, the appli-
cation of the cycloinsertion strategy to sandresolide A would
require the use of a Ca-methyl-substituted ketone (6, R1 =
Me). In this regard, the result obtained with menthone (1i)
was of some concern.
In the context of the sandreso-
lide project, we had enol silanes 9a–
Table 1: Reaction of ketones (1) with iodonium compound 4.[a]
b in hand and decided to examine
whether silyl enol ethers would also
participate in the cycloinsertion
process (Scheme 4). In the course
of optimizing the reaction, we iden-
tified conditions prescribing the use
of 3.0 equivalents of KOtBu,
2.4 equivalents of iodonium 4, and
1.2 equivalents of H2O. Ca-unsub-
stituted enol silane 9a underwent
smooth cycloaddition to give 10a in
76% yield.[14] As was previously
observed with the unsubstituted
cyclobutenols, base-induced ring-
opening provided 11a. Substrate
9b, which incorporated a Ca-Me
group, successfully engaged in the
cycloaddition reaction to furnish
cyclobutenol 10b in 80% yield. As
with 5i, it did not undergo ring
opening, even under forcing condi-
tions.
We decided to target cyclobute-
nol adducts that contained a leaving
group at the C4-position to enable
fragmentation as an alternative
pathway for ring opening (see
10c).[17] However, caution might be
warranted as the corresponding
enolate could be susceptible to
elimination to the enone under the
reaction conditions.[18] Neverthe-
Entry
Substrate
Product
Yield [%]
1
2
3
n=1: 1a
n=2: 1b
n=3: 1c
3a
3b
3c
67[b]
76[c]
64[b]
4
5
1d
1e
3d
3e
51
74[b]
6
7
1 f
3 f
66[c]
1g
3g
54[d]
8
9
1h
1i
3h
5i
58[b]
52
[a] Reaction conditions: ketone 1 (0.5 mmol), 4 (1.5 equiv), KOCEt3 (2.5 equiv), THF (25 mL), ꢀ788C to
RT. [b] Combined yield of 3 and its deconjugated isomer. [c] Cyclobutene adduct isolated and opened in
successive step.[14] [d] 10% starting material recovered.
Angew. Chem. Int. Ed. 2010, 49, 4092 –4095
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 4093

Communications
ꢀ
the structure is stereoelectronically optimally set up for C1
C4 bond rupture.[23]
A detailed discussion of the reaction mechanism exceeds
the scope of this communication; however, some general
comments can be made. The action of alkoxide base on
ketone 1 or silylenolether 9 generates enolate 12 (Scheme 5),
which subsequently undergoes syn addition to cyclohexyne
(2). The unsubstituted cycloadducts 5a–h and 10a (R1 = H)
undergo rapid, conrotatory, electrocyclic ring-opening. This
process is facilitated by the oxy substituent, which presumably
undergoes torquoselective outward rotation to give dienolate
13.[10,22] The reluctance of certain substrates (5i, 10b, R1 ¼ H)
to undergo electrocyclic opening can be understood by the
additional steric congestion that arises from either conrota-
tory mode.
Scheme 3. Ca-substituted ketone 6 in the construction of the sandreso-
lide (8) core.
Scheme 4. Cycloaddition reactions with enol ethers 9a–c. Reagents
and conditions: a) 4 (2.4 equiv), KOtBu (3.0 equiv), H2O (1.2 equiv),
THF, ꢀ788C to RT; b) KOtBu (0.5 equiv), [18]crown-6 (0.5 equiv), THF,
RT; c) KHMDS (1.1 equiv), [18]crown-6 (0.5 equiv), THF, RT.
KHMDS=potassium hexamethyldisilazide, SEM=2-(trimethylsilyl)-
ethoxymethyl, TMS=trimethylsilyl, sm=starting material,
brsm=based on recovered starting material.
Scheme 5. Proposed working model for the cycloalkyne insertion.
In summary, we report the first cycloinsertion reaction of
cyclohexyne (2) into cyclic ketones. This transformation is
comprised of consecutive annulation and ring-expansion
reactions. Facile derivatization of cyclic structures is achieved,
which enables rapid access to polycyclic medium-sized rings,
from a collection of simple and more complex cyclic ketones
(cycloalkanones, estrone, cholesterone, and densely function-
alized cyclopentanones). The cycloadducts of unsubstituted
enolates readily undergo base-induced electrocyclic ring-
opening reactions. The surprising participation of a b-alkoxy
enolate in the cycloaddition affords a product that is set up for
ring-opening fragmentation. Interesting insight was obtained
by the analysis of the structural characteristics of cyclobutenol
5 f. The X-ray crystal structure provides experimental vali-
dation for the increased reactivity of p-donor-substituted
cyclobutenes, which had earlier been theorized in computa-
tional studies. The cyclohexyne insertion provides an intrigu-
ing simplifying transformation for medium-sized rings. The
intermediate cyclobutenes may also find further applications
as they are amenable to a host of other manipulations. Studies
into the reactions of cyclohexyne and its derivatives are
ongoing and will be reported in due course.
less, we were pleasantly surprised to find that when 9c was
subjected to the reaction conditions adduct 10c was isolated
in 83% yield. Treatment of this cyclobutenol with 1.1 equiv-
alents of KHMDS and 18-crown-6 in tetrahydrofuran at room
temperature afforded 11c in 51% yield (76% brsm).
We have obtained a crystal structure of 5 f and its analysis
is revealing (Figure 1).[14,19] The C1 C4 single bond in the
ꢀ
cyclobutenol is significantly stretched to 1.596 ꢀ, as com-
pared to 1.573 ꢀ in a previously reported unsubstituted
cyclobutene.[20] Concomitantly, the C1 O single bond is
ꢀ
shortened by 0.022 ꢀ, when compared to an unstrained
tertiary alcohol.[21] These observations suggest a hyperconju-
ꢀ
gative interaction of the oxygen lone pair with the C1 C4 s*
orbital. This phenomenon was predicted computationally by
Houk and Rondan[22a] as being pivotal for the drastically
lowered activation barrier for the electrocyclic ring-opening
reaction of p-donor-substituted cyclobutenes.[10,22] Therefore,
Experimental Section
General procedure: A solution of KOCEt3 (1.25 mmol) in THF
(5 mL) was allowed to stream down the walls of the flask over
5 minutes into a precooled (ꢀ788C) solution of ketone 1 (0.5 mmol)
and iodonium 4 (0.75 mmol) in THF (20 mL). The mixture was stirred
Figure 1. Representation of the crystal structure of 5 f with selected
bond lengths in ꢂngstrꢃms (hydrogen atoms omitted for clarity).
Ellipsoids set at 50% probability.
4094 www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 4092 –4095

Angewandte
Chemie
at ꢀ788C for 30 minutes, brought to RT over 25 minutes, and
partitioned between phosphate buffer (1 molLꢀ1, pH 7, 20 mL) and
Et2O (20 mL). After extraction of the aqueous phase with Et2O (2 ꢁ
20 mL) the combined organic phases were dried over Na2SO4 and the
solvent removed in vacuum. Column chromatography on silica gel,
eluting with pentane/Et2O, gave the desired pure products.
[8] For the calculation of ring strain, see: a) K. B. Wiberg, Angew.
Chem. 1986, 98, 312 – 322; Angew. Chem. Int. Ed. Engl. 1986, 25,
312 – 322; b) P. R. Khoury, J. D. Goddard, W. Tam, Tetrahedron
2004, 60, 8103 – 8112.
[9] a) L. Fitjer, U. Kliebisch, D. Wehle, S. Modaressi, Tetrahedron
Lett. 1982, 23, 1661 – 1664; b) L. Fitjer, S. Modaressi, Tetrahedron
Lett. 1983, 24, 5495 – 5498.
[10] For reviews on the influence of p-donor substituents on
pericyclic reactions, see: a) N. D. Epiotis, Angew. Chem. 1974,
86, 825 – 855; Angew. Chem. Int. Ed. Engl. 1974, 13, 751 – 780;
b) K. N. Houk, Acc. Chem. Res. 1975, 8, 361 – 369.
[11] D. A. Evans, A. M. Golob, J. Am. Chem. Soc. 1975, 97, 4765 –
4766.
[12] a) M. Fujita, Y. Sakanishi, W. H. Kim, T. Okuyama, Chem. Lett.
2002, 908 – 909; b) M. Fujita, W. H. Kim, Y. Sakanishi, K.
Fujiwara, S. Hirayama, T. Okuyama, Y. Ohki, K. Tatsumi, Y.
Yoshioka, J. Am. Chem. Soc. 2004, 126, 7548 – 7558; c) T.
Okuyama, M. Fujita, Acc. Chem. Res. 2005, 38, 679 – 686.
[13] C. A. Brown, J. Chem. Soc. Chem. Commun. 1974, 680 – 681.
[14] For further details, see the Supporting Information.
[15] Y. Mazur, F. Sondheimer, J. Am. Chem. Soc. 1958, 80, 5220 –
5229.
[16] A. D. Rodriguez, C. Ramirez, I. I. Rodriguez, Tetrahedron Lett.
1999, 40, 7627 – 7631.
[17] a) A. Eschenmoser, A. Frey, Helv. Chim. Acta 1952, 35, 1660 –
1666; b) C. A. Grob, W. Baumann, Helv. Chim. Acta 1955, 38,
594 – 610.
[18] We believe that the reaction proceeds through the enolate that is
generated by H2O-promoted desilylation under the basic
reaction conditions.
[19] CCDC 767284 (5 f) contains the supplementary crystallographic
data for this paper. These data can be obtained free of charge
from The Cambridge Crystallographic Data Centre via www.
ccdc.cam.ac.uk/data_request/cif.
[20] F. Allen, Acta Crystallogr. Sect. B 1984, 40, 64 – 72.
[21] Average values for tert-alcohols, as taken from: CRC Handbook
of Chemistry and Physics 89th ed. (Ed.: D. R. Lide), CRC Press
Taylor & Francis Group, Boca Raton, 2008, pp. 9-1 – 9-16.
[22] a) N. G. Rondan, K. N. Houk, J. Am. Chem. Soc. 1985, 107,
2099 – 2111; b) B. K. Carpenter, Tetrahedron 1978, 34, 1877 –
1884.
[23] Only two X-ray crystal structures of cyclobutenols have been
reported to date, and no correlation between structure and
reactivity has been drawn, see: a) J. Suffert, B. Salem, P. Klotz, J.
Am. Chem. Soc. 2001, 123, 12107 – 12108; b) B. Salem, J. Suffert,
Angew. Chem. 2004, 116, 2886 – 2890; Angew. Chem. Int. Ed.
2004, 43, 2826 – 2830.
Received: February 24, 2010
Published online: April 29, 2010
Keywords: cyclohexynes · fused-ring systems ·
.
medium-ring compounds · ring expansion · strained molecules
[1] Monographs on cycloalkynes: a) R. W. Hoffmann, Dehydroben-
zene and Cylcoalkynes, Academic Press, New York, 1967,
pp. 317 – 357; b) R. Gleiter, R. Merger in Modern Acetylene
Chemistry (Eds.: P. J. Stang, F. Diederich), VCH, Weinheim,
1995, pp. 285 – 319; c) H. Hopf, Classics in Hydrocarbon
Chemistry, Wiley-VCH, Weinheim, 2000, pp. 156 – 160; for a
review, see: A. Krebs, J. Wilke, Top. Curr. Chem. 1983, 109, 189 –
233.
[2] For theoretical studies on cyclohexyne, see: a) S. Olivella, M. A.
Pericas, A. Riera, A. Sole, J. Org. Chem. 1987, 52, 4160 – 4163;
b) R. P. Johnson, K. J. Daoust, J. Am. Chem. Soc. 1995, 117, 362 –
367; c) I. Yavari, F. Nasiri, H. Djahaniani, A. Jabbari, Int. J.
Quantum Chem. 2006, 106, 697 – 703.
[3] Seminal publication: a) G. Wittig, Naturwissenschaften 1942, 30,
696 – 703; b) J. D. Roberts, H. E. Simmons, L. A. Carlsmith,
C. W. Vaughan, J. Am. Chem. Soc. 1953, 75, 3290 – 3291; for
recent applications in synthesis, see: c) R. L. Danheiser, A. L.
Helgason, J. Am. Chem. Soc. 1994, 116, 9471 – 9479; d) U. K.
Tambar, B. M. Stoltz, J. Am. Chem. Soc. 2005, 127, 5340 – 5341;
for a recent review, see: e) R. Sanz, Org. Prep. Proced. Int. 2008,
40, 215 – 291.
[4] F. Scardiglia, J. D. Roberts, Tetrahedron 1957, 1, 343 – 344.
[5] a) P. Caubere, J. J. Brunet, Tetrahedron 1971, 27, 3515 – 3526;
b) P. Caubere, J. J. Brunet, Tetrahedron 1972, 28, 4835 – 4845;
c) P. Caubere, J. J. Brunet, Tetrahedron 1972, 28, 4847 – 4857;
d) P. Caubere, J. J. Brunet, Tetrahedron 1972, 28, 4859 – 4869.
[6] The 6,6-gem-dimethyl substitution prevents competitive forma-
tion of the allene, see: B. Fixari, J. J. Brunet, P. Caubere,
Tetrahedron 1976, 32, 927 – 934.
[7] For a review of [2+2] photocyclization/cyclobutene fragmenta-
tions in synthesis, see: W. Oppolzer, Acc. Chem. Res. 1982, 15,
135 – 141.
Angew. Chem. Int. Ed. 2010, 49, 4092 –4095
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 4095