Angewandte
Chemie
perform this reaction with halides as the leaving group were
unsuccessful.
Next, we tested differently substituted alkynyl sulfonates
(Table 2). With ether derivatives instead of the alkyl tethers,
longer reaction times were observed, and the yields were only
moderate (entries 1 and 2). As the presence of the ether
linkage breaks the symmetry of the cyclopentylidene ring, the
products were obtained as mixtures of diastereoisomers,
favoring the E isomers (assigned by 1H,1H NOESY NMR
spectroscopy). The corresponding substrates with a sulfona-
mide group in the chain were also investigated. Whereas no
reactivity was observed with the tosyl-substituted substrate 1i
(entry 3), switching to the brosylate (OBs = para-bromoben-
zenesulfonate) leaving group delivered the desired product 2j
in 83% yield as a 4:1 mixture of diastereoisomers (entry 4).
Based on the higher reactivity that was observed for the
brosylate substrates, we performed further studies with this
leaving group. As a next step, we investigated substrates with
an additional substituent at the propargylic position. A
phenethyl substituent not only increased the reaction rate
and yield, but also led to the exclusive formation of the
sterically less hindered E isomer (entry 5). It is noteworthy
that the competing CH insertion reaction was not observed.
A completely different picture was obtained with a directly
attached phenyl group. Most probably because of the highly
activated propargylic/benzylic ether position, decomposition
of the starting material was observed under the reaction
conditions (entry 6). The branched alkyl substrate 1m, which
does not bear additional heteroatoms, was also investigated.
The cyclized product 2m was obtained in good yield, but
again as a mixture of diastereoisomers (entry 7). Next, we
evaluated whether this transformation would be possible for
systems with aromatic moieties as part of the tether.
Unfortunately, substrate 1n, which bears a tether with six
carbon atoms, was not converted into the analogous annu-
lated cyclopentylidene, and only incomplete conversion was
observed, along with an unselective transformation (entry 8).
Whereas the expected reactivity was restored by increasing
the chain length, problems still occurred with unfunctional-
ized substrate 1o (entry 9). Aside from incomplete conver-
sion, the final product turned out to be fairly unstable.
Fortunately, the introduction of an oxygen atom into the
phenolic substrates 1p–1s delivered the stable products 2p–
2s, which could be isolated in good yields (entries 10–13).
Crystal structure analysis of compound 2p confirmed the
constitution of the target compound as well as the E geometry
of the exocyclic double bond.[7,10]
Scheme 2. Transformation of tosylate 1a in the presence of
[IPrAu(propynyl)] (3a; 5 mol%). Ts=tosyl=4-toluenesulfonyl.
An optimization of the reaction conditions[7] revealed that
gold propyne acetylide 3a was superior to the corresponding
methyl or phenyl gold complexes as well as to [IPrAuOH].[8]
Next, we performed a ligand screening; unsaturated NHC(15)
ligand 3b[9] turned out to be the best candidate (NHC = N-
heterocyclic carbene; NHC(15) bears a cyclopentadecyl sub-
stituent on one of the nitrogen atoms). Different solvents
were screened, but the yields dropped significantly in other
reaction media. A control experiment without any catalyst
showed no conversion; the addition of catalytic amounts of p-
TsOH also led to no reaction. A radical pathway can be
excluded as no drop in reactivity was observed when 2,6-
bis(1,1-dimethylethyl)-4-methylphenol (BHT) was added as
a radical inhibitor.[7]
With the optimized conditions in hand (5 mol% 3b,
benzene, reflux), we set out to investigate the scope of this
transformation. First, we tested different sulfonate moieties.
As shown in Table 1, sulfonates with aromatic substituents
Table 1: Gold-catalyzed cyclization of hexynyl sulfonates.[a]
Entry
R
Starting material
Product
Yield [%]
1
2
3
4
5
6
p-MeC6H4
p-BrC6H4
p-NO2C6H4
p-MeOC6H4
mesityl
1a
1b
1c
1d
1e
1 f
2a
2b
2c
2d
2e
2 f
92
88
53
84
93
72
Me
[a] All reactions were performed in benzene (150 mm) at reflux with 3b
Finally, we tested different chain lengths for the open-
chain substrates as well (entries 14–17). The formation of six-
membered rings was possible, but longer reaction times were
required than for the formation of the corresponding cyclo-
pentylidene rings. For both leaving groups, allenes 4t and 4u
were isolated as side products in small amounts (entries 14
and 15). Derivative 1v did not undergo cyclization; instead,
allene 4v, along with the HOBs addition product 5v,[11] was
obtained (entry 16). The structure of 5v could be confirmed
by X-ray crystal structure analysis.[7,10] A similar result was
obtained for butynyl derivative 1w (entry 17). In this case, the
analogous HOTs addition product 5w was isolated together
(5 mol%). All yields refer to isolated products.
delivered the corresponding products in good to excellent
yields, only a p-NO2 substituent resulted in a slightly lower
yield (entry 3). Mesylate (OMs = methanesulfonate) as an
aliphatic substituent was also suitable, and the corresponding
product 2 f could be obtained in 72% yield (entry 6).
For product 2b, single crystals that were suitable for X-ray
crystallography[7,10] could be obtained; therefore, the struc-
tural assignment of a cyclopentylidene ring with an exocyclic
double bond could be confirmed. Unfortunately, attempts to
Angew. Chem. Int. Ed. 2014, 53, 3854 –3858
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3855