Gold-catalyzed stereoselective synthesis of 9-oxabicyclo[3.3.1]nona-4,7- dienes from diverse 1-oxo-4-oxy-5-ynes: A viable formal [4 + 2] cycloaddition on an s-trans-heterodiene framework

Article abstract of DOI:10.1021/ja106493h

We report a highly stereoselective Au-catalyzed synthesis of 9-oxabicyclo[3.3.1]nona-4,7-dienes from diverse 1-oxo-4-oxy-5-ynes. Formation of these highly strained anti-Bredt oxacycles implies the workability of an unprecedented 1,4-dipole of s-trans-meth

Full text of DOI:10.1021/ja106493h

Published on Web 08/23/2010
Gold-Catalyzed Stereoselective Synthesis of
9-Oxabicyclo[3.3.1]nona-4,7-dienes from Diverse 1-Oxo-4-oxy-5-ynes: A Viable
Formal [4 + 2] Cycloaddition on an s-trans-Heterodiene Framework
Tse-Min Teng, Arindam Das, Deepak B. Huple, and Rai-Shung Liu*
Department of Chemistry, National Tsing Hua UniVersity, Hsinchu, Taiwan, ROC
Received July 22, 2010; E-mail: rsliu@mx.nthu.edu.tw
Table 1. Catalytic Activity over Various Metal Catalystsa
Abstract: We report a highly stereoselective Au-catalyzed syn-
thesis of 9-oxabicyclo[3.3.1]nona-4,7-dienes from diverse 1-oxo-
4-oxy-5-ynes. Formation of these highly strained anti-Bredt
oxacycles implies the workability of an unprecedented 1,4-dipole
of s-trans-methylene(vinyl)oxonium. This work reveals the fea-
sibility of a formal [4 + 2] cycloaddition on an s-trans-heterodiene
entry substrate
catalyst (mol %)
condition
productb
framework.
c
1
2
3
4
5
1a
1a
1a
1a
1a
ClAuL (3)/AgSbF6 (3) DCM (25 °C, 2 h)
2a (18%)
ClAuL (3)/AgNTf2 (3) DCM (25 °C, 2 h) 2a (68%)
ClAuL (3)
AgNTf2 (3)
IPrAuCl
DCM (25 °C, 8 h) 1a (88%)
DCM (25 °C, 8 h) 1a (91%)
DCM (25 °C, 1.5 h) 1a (21%),
2a (35%)
[4 + 2] Cycloaddition reactions are powerful tools for construct-
ing complex carbo- and heterocyclic frameworks.1 Starting acyclic
(hetero)dienes A are conformationally flexible and exhibit a rapid
equilibrium between the s-trans and s-cis forms (Scheme 1), but
only the s-cis conformer enables the cycloaddition.2 This rule is
invariable also with rigid s-trans-(hetero)dienes C that are also
inactive toward the reactions. According to our energy estimate,
the selection of bridgehead olefin E (X ) Y ) CH) as the primary
cycloadduct should be a viable route, for which the energy is 34.2
kcal/mol less than that of the normal cycloadduct D.3 Nevertheless,
such an atypical formal cycloaddition (C + olefin f E) has been
entirely ignored. This concept stimulated us to pursue its first
realization for an s-trans-heterodiene framework, i.e., methylene(vi-
nyl)oxonium C (X ) CH, Y ) O+), in this work.
(3)/AgNTf2 (3)
Ph3PAuCl
(5)/AgNTf2 (5)
AuCl3 (5)
6
1a
DCM (25 °C, 2 h) polymerization
7
8
9
1a
1a
1b
DCM (25 °C, 2 h) messy mixtures
toluene (50 °C, 3 h) polymerization
PtCl2/CO (5)
ClAuL (3)/AgNTf2 (3) DCM (25 °C, 12 h) 1b (43%),
c
2b (8%)
a L ) P(t-Bu)2(o-biphenyl), [substrate] ) 0.1 M. b Reported yields
were determined after separation from
a
silica gel column.
c Polymerization was observed for portion of 1a and 1b in entries 1 and
9.
Scheme 3
Scheme 1
at the conformationally rigid s-trans-oxoniums II; their workability
as 1,4-dipoles is the focus of this work.
We prepared 1-oxo-5-ynes 1a and 1b to test our working
hypothesis using various catalysts (Table 1). Treatment of 1a with
ethoxyethene (3 equiv) and ClAuP(t-Bu)2(o-biphenyl)/AgSbF6 (3
mol %) in dichloromethane (DCM) at 25 °C for 2 h resulted in
complete consumption of the initial 1a, giving 2a in 18% yield
together with a messy unknown mixture (entry 1). To our pleasure,
the use of ClAuP(t-Bu)2(o-biphenyl)/AgNTf2 (3 mol %) greatly
improved the yield of the desired product 2a to 68% (entry 2);
control experiments showed no activity of either ClAuP(t-Bu)2(o-
biphenyl) or AgNTf2 (entries 3 and 4). Decreased activity was found
for IPrAuCl/AgNTf2 [IPr ) 1,3-bis(diisopropylphenyl)imidazol-2-
ylidene], which gave 2a in 35% yield and unreacted 1a with 21%
recovery. Ph3PAuCl/AgNTf2, AuCl3, and PtCl2/CO, each at 5 mol
%, were unsuitable for this catalysis because of a complete
decomposition of initial 1a to polymers or messy unknown mixtures.
We also examined the reaction of substrate 1b and ethoxyethene
using ClAuP(t-Bu)2(o-biphenyl)/AgNTf2 but obtained the desired
product 2b in only 8% yield together with unreacted 1b (43%
recovery). The acyloxy group of 1a appears to be crucial to the
success of this gold catalysis. The 9-oxabicyclo[3.3.1]nona-4,7-
diene framework of 2a was confirmed by an X-ray diffraction study
Scheme 2
In the presence of Au and Pt catalysts, 1-oxo-n-ynes (n ) 3-6)
typically undergo endo-dig cyclizations to give s-cis-methylene(vi-
nyl)oxoniums I, which serve as 1,3- or 1,4-dipoles to react with
alkenes or alkynes.4-7 In contrast, examples of exo-dig cyclizations
are rare and have been restricted to 1-oxo-2-en-4-ynes that give
furyl carbene intermediates.8 We sought an unprecedented 6-exo-
dig cyclization of 1-oxo-5-ynes, as depicted in Scheme 2, aiming
9
10.1021/ja106493h  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 12565–12567 12565

C O M M U N I C A T I O N S
Table 2. Scope of Aromatic 1-Oxo-5-yne Substratesa
Table 4. Scope of Nonaromatic 1-Oxo-5-yne Substratesa
entry
oxoyne
enol ether
t [h]
productsb
1
2
3
4
5
6
7
8
9
X ) Ac, R1 ) H (1a)
X ) Ac, R1 ) Me (1c)
1c
R2 ) n-Bu
R2 ) Et
1
2.5
2
1
1
1
1
1
1.5
1.5
2a′ (62%)
2c (76%)
2c′ (67%)
2d (85%)
2d′ (82%)
2e (82%)
2e′ (94%)
2f (74%)
2g (57%)
2h (42%)
R2 ) n-Bu
R2 ) Et
X ) MOM, R1 ) H (1d)
1d
R2 ) n-Bu
R2 ) Et
X ) MOM, R1 ) Me (1e)
1e
R2 ) n-Bu
R2 ) Et
X ) MOM, R1 ) n-Pr (1f)
X ) Bn, R1 ) Me (1g)
X ) n-Bu, R1 ) Me (1h)
R2 ) Et
10
R2 ) Et
a L ) P(t-Bu)2(o-biphenyl), [oxoyne] ) 0.1 M, enol ether (3 equiv).
b Reported isolated yields were determined after purification on a silica
gel column.
Table 3. Phenyl Effect of 1-Oxo-5-yne Substratesa
a L ) P(t-Bu)2(o-biphenyl), [oxoyne] ) 0.1 M, enol ether (3 equiv).
b Reported isolated yields were determined after purification on a silica
gel column.
entry
oxoyne
t [h]
productsb
Scheme 4
1
2
3
4
5
6
7
X ) OMe, Y ) H (1i)
X ) H, Y ) OMe (1j)
X ) Y ) OMe (1k)
X ) F, Y ) H (1l)
X ) H, Y ) F (1m)
X ) Cl, Y ) H (1n)
X ) H, Y ) Cl (1o)
1
2.5
2
1
1
2i (90%)
2j (86%)
2k (95%)
2l (83%)
2m (91%)
2n (78%)
2o (84%)
1
1
a L ) P(t-Bu)2(o-biphenyl), [oxoyne] ) 0.12 M, enol ether (3 equiv).
b Reported isolated yields were determined after purification on a silica
gel column.
of its hydroxy derivative 3a (X ) OH);9 this anti-Bredt structure
confirms a [4 + 2] cycloaddition to s-trans-oxonium II. Notably,
the route to the anti-Bredt [4 + 2] cycloadduct is not attainable
with s-cis-oxonium I (n ) 1, Scheme 2) and related benzopyrili-
ums.5 The utility of this synthesis is manifested by its facile access
to bioactive molecules G1-G3 (Scheme 3), which exhibit activity
in the central nervous system and HIV-1 inhibitory effects.10
The value of this catalysis is highlighted by its successful
extension to nonaromatic 1-oxo-4-oxy-5-ynes 4a-f, which gave
oxacyclic frameworks 5a-f of various classes (Table 4). For
substrates 4a-d bearing tunable R1 (R1 ) H, Me), OX (X ) Ac,
MOM), and Y (Y ) H, Me) groups, the same gold catalysis
delivered products 5a-d in 63-78% yield (entries 1-4). For
bicyclic oxoalkyne 4e, gold catalysis gave the desired product 5e
in 76% yield; the stereochemistry of 5e was confirmed with X-ray
diffraction measurements.9 We prepared acyclic oxoalkyne 4f,
which also gave the expected compound 5f in 63% yield.
Shown in Scheme 4 are additional new frameworks that are
readily available from this gold catalysis. We prepared aldehydes
4g/4i and ketone 4h bearing fused cyclopentenyl and cycloheptenyl
rings, respectively, and their corresponding cycloadducts 5g/5i and
5h were obtained very efficiently (yields >81%) and stereoselec-
tively.
As depicted in Scheme 5, we studied gold-catalyzed catalysis on
substrate 1p bearing a noncoordinating siloxy group, and its reaction
with ethoxyethene gave compound 6 rather than the desired oxacyclic
cycloadducts such as 2. Compound 6 was formed with 45% deuterium
(X ) 0.45D) at the trans-vinyl hydrogen position in the presence of
D2O (1 equiv); this information supports the intermediacy of gold-
containing s-trans-oxonium II′ through a 6-exo-dig mode. Accordingly,
the oxy groups, including OMOM, OAc and OTMS, facilitate the
6-exo-dig cyclization via an electron-withdrawing effect on the alkyne
rather than through metal coordination. We also prepared methoxy
derivative 1q, which gave complex oxacyclic species 7 in 36% yield;
We prepared various aldehyde- and ketone-containing substrates
1a-h bearing alterable 4-oxy groups to assess the scope of this
catalysis (Table 2); each resulting product 2a′-h was obtained as
a single diastereomer despite its complicated molecular framework.
Entries 1-3 show additional examples of the applicability of this
new synthesis not only to ketone substrate 1c but also to butoxy-
ethene; the desired oxacycles 2a′-c′ were obtained in 62-76%
yield. Particularly notable are the high yields (74-94%) of products
2d-f produced from 1-oxo-5-ynes 1d-f bearing a methoxymethyl
(MOM) ether group (entries 4-8). The effect of the 4-oxy group
is also reflected by substrates 1g and 1h bearing benzyl and n-butyl
ethers, respectively, which gave the corresponding products 2g and
2h in moderate yields (57 and 42%; entries 9 and 10).
We also prepared substrates 1i-o to examine the effects of their
phenyl substituents (Table 3); only one diastereomeric product was
obtained in all cases. In entries 1-3, excellent product yields
(86-95%) were obtained for oxacyclic compounds 2i-k bearing
a methoxy group at the phenyl C(4) or C(5) position. The
workability with substrates 1l-o bearing fluoro and chloro sub-
stituents, which produced the expected products 2l-o in good yields
(78-91%; entries 4-7), reflects the wide scope of this catalysis.
9
12566 J. AM. CHEM. SOC. VOL. 132, NO. 36, 2010

C O M M U N I C A T I O N S
Scheme 5
Scheme 6
Scheme 7a
edented 1,4-dipole of s-trans-methylene(vinyl)oxonium II. Neverthe-
less, in view of the highly diastereoselective outcome, the resulting
cycloadducts arise from an initial [3 + 2] cycloaddition of R-carbonyl
ylide II′′ followed by a ring expansion. The concept of a formal [4 +
2] cycloaddition on an s-trans-heterodiene should be helpful in the
design of new synthetic methods.
Acknowledgment. The authors thank the National Science
Council, Taiwan, for support of this work.
Supporting Information Available: Experimental procedures,
characterization data for new compounds, and X-ray crystallographic
data (CIF) for compounds 3a and 5e. This material is available free of
charge via the Internet at http://pubs.acs.org.
References
(1) For reviews, see: (a) Nicolaou, K. C.; Snyder, S. A.; Montagnon, T.;
Vassilikogiannakis, G. Angew. Chem., Int. Ed. 2002, 41, 1668. (b) Pindur,
U.; Lutz, G.; Otto, C. Chem. ReV. 1993, 93, 741. (c) Boger, D. L. Chem.
ReV. 1986, 86, 781. (d) Boger, D. L. J. Heterocycl. Chem. 1996, 33, 1519.
(e) Boger, D. L. Tetrahedron 1983, 39, 2869.
(2) (a) Houk, K. N.; Lin, Y.-T.; Brown, F. K. J. Am. Chem. Soc. 1986, 108,
554. (b) Goldstein, E.; Beno, B.; Houk, K. N. J. Am. Chem. Soc. 1996,
118, 6036.
(3) The calculations were performed using the B3LYP method with the
LANL2DZ basis set (B3LYP/LANL2DZ).
(4) For reviews, see: (a) Patil, N. T.; Yamamoto, Y. Chem. ReV. 2008, 108,
3395. (b) Abu Sohel, S. Md.; Liu, R.-S. Chem. Soc. ReV. 2009, 38, 2269.
(5) For s-cis-oxoniums I, their [4 + 2] cycloaddition intermediates V do not
give anti-Bredt oxacycles VI; these hypothetical intermediates undergo
facile rearrangement to dihydronaphthalenes5a,b or 8-oxabicyclo[3.2.1]-
octanes.6 See: (a) Asao, N.; Kasahara, T.; Yamamoto, Y. Angew. Chem.,
Int. Ed. 2003, 42, 3504. (b) Asao, N.; Aikawa, H.; Yamamoto, Y. J. Am.
Chem. Soc. 2004, 126, 7458. (c) Hsu, Y.-C.; Ting, C.-M.; Liu, R.-S. J. Am.
Chem. Soc. 2009, 131, 2090. (d) Shu, X.-Z.; Zhao, S.-C.; Ji, K.-G.; Zheng,
Z.-J.; Liu, X.-Y.; Liang, Y.-M. Eur. J. Org. Chem. 2009, 117.
a Reagents: (i) Pd-C/H2 (1 atm), 1:1 1,4-dioxane/DCM, 25 °C, 16 h,
82% (ii) m-CPBA (1.5 equiv), NaHCO3 (3 equiv), DCM, 1.5 h, 76% (iii)
ICI (1.1 equiv), wet DCM, 1 h, 0 °C, 63%.
its formation may imply a gold-containing carbene intermediate III,
as depicted in Scheme 6.
It is possible that compounds 2 are produced from either an initial
[3 + 2] cycloaddition to R-carbonyl ylide II′′ 6 or from a [4 + 2]
cycloaddition on s-trans-oxonium II′.5 We envisage that the [3 +
2] path would enable the oxy group to approach the enol ether
closely, rendering great diastereocontrol of the cycloadducts. As
shown in structure II′′, the enol ether approaches the carbonyl ylide
from the less-hindered endo face and away from the proximate oxy
group (OR) to give gold carbenium III. The formation of compound
7 from 1-oxy-5-yne 1q might imply intermediate III, whose carbene
functionality would activate a methoxy C-H insertion.11 An
alternative stepwise [4 + 2] cycloaddition would make it difficult
to rationalize the stereodirecting effect of the oxy group.
Importantly, this catalysis requires 1-oxo-5-ynes 1 bearing an
oxy group12,13 to give the desired oxacycles 2, with R ) MOM or
Ac being more efficient than R ) TMS. We hypothesize that MOM
assists a 1,2-alkyl migration to form stable oxonium species IV,
which is subsequently convertible to the formal “[4 + 2] cycload-
duct” 2; we disparage the bridgehead oxonium IV′ because of its
highly strained skeleton (see Scheme 1, species D). Verification of
this hypothesis needs additional work in the future.
(6) For selected examples of [3 + 2] cycloadditions, see: (a) Kusama, H.; Ishida,
K.; Funami, H.; Iwasawa, N. Angew. Chem., Int. Ed. 2008, 47, 4903. (b)
Li, G.; Huang, X.; Zhang, L. J. Am. Chem. Soc. 2008, 130, 6944. (c) Oh,
C. H.; Lee, J. H.; Lee, S. J.; Kim, J. I.; Hong, C. S. Angew. Chem., Int. Ed.
2008, 47, 7505. (d) Oh, C. H.; Lee, S. M.; Hong, C. S. Org. Lett. 2010,
12, 1308.
(7) For other reactions of s-cis-oxonium I, see: (a) Liu, F.; Yu, Y.; Zhang, J.
Angew. Chem., Int. Ed. 2009, 48, 5505. (b) Jin, T.; Yamamoto, Y. Org.
Lett. 2007, 9, 5259. (c) Jin, T.; Yamamoto, Y. Org. Lett. 2008, 10, 3137.
(8) Miki, K.; Uemura, S.; Ohe, K. Chem. Lett. 2005, 34, 1068.
(9) X-ray crystallographic data for compounds 3a and 5e are provided in the
Supporting Information.
(10) (a) Wu¨nsch, B.; Zott, M.; Ho¨fner, G. Arch. Pharm. 1992, 325, 733. (b)
Wu¨nsch, B.; Zott, M.; Ho¨fner, G. Arch. Pharm. 1992, 325, 823. (c) Maurin,
C.; Bailly, F.; Cortelle, P. Curr. Med. Chem. 2003, 10, 1795. (d) Dupont,
R.; Jeanson, L.; Mouscadet, J.-F.; Cotelle, P. Bioorg. Med. Chem. Lett.
2001, 11, 3175.
(11) (a) Hashmi, A. S. K.; Scha¨fer, S.; Wo¨lfle, M.; Gil, C. D.; Fischer, P.;
Laguna, A.; Blanco, M. C.; Gimeno, M. C. Angew. Chem., Int. Ed. 2007,
46, 6184. (b) Horino, Y.; Yamamoto, T.; Ueda, K.; Kuroda, S.; Toste, F. D.
J. Am. Chem. Soc. 2009, 131, 2809. (c) Cui, L.; Peng, Y.; Zhang, L. J. Am.
Chem. Soc. 2009, 131, 8394. (d) Bhunia, S.; Liu, R.-S. J. Am. Chem. Soc.
2008, 130, 16488.
(12) For internal alkyne 9a, we observed a distinct 7-endo-dig cyclization to
give indenyl ketone 10a in 45% yield. The steric interaction between gold
and n-butyl destabilizes intermediate II′. For the acetoxy derivative 9b,
we obtained 10b stereoselectively from an initial 1,3-acetoxy shift followed
by a [4 + 2] cycloaddition on benzopyrilium (see ref 13).
Scheme 7 shows the use of this catalysis for a stereoselective
synthesis of highly oxygenated molecules. Stereocontrolled func-
tionalization of 2a′ at the bridgehead olefin was readily achieved
via (i) Pd/C hydrogenation and (ii) m-CPBA epoxidation from the
open face, giving 8a and 8b in 82 and 76% yield, respectively.
Treatment of 2a′ with ICl in wet CH2Cl2 gave hemiketal 8c (63%)
as a single diastereomer.
In summary, we have reported a highly stereoselective Au-catalyzed
synthesis of 9-oxabicyclo[3.3.1]nona-4,7-dienes from diverse 1-oxo-
5-ynes bearing an indispensable 4-oxy group. Formation of these highly
strained anti-Bredt oxacycles reveals the workability of an unprec-
(13) Teng, T.-M.; Liu, R.-S. J. Am. Chem. Soc. 2010, 132, 9298.
JA106493H
9
J. AM. CHEM. SOC. VOL. 132, NO. 36, 2010 12567