formation of 2_R11 (Figure 3). Although pathway d also
proceeds via 2_R, this pathway is not followed because
attack of the pendant acetate on carbon c (Figure 2) requires
a higher deformation energy in the transition structure resulting
in a raised energy barrier. This attack is disfavored because of
the greater distance of the acetate from carbon c. For example,
in 2_R (R ) phenyl) the sp2 oxygen is 5.667 Å away from
carbon c but only 3.895 Å away from carbon a. In the favored
pathway c, we considered two conformations R and R′ of the
starting cyclopropene to be reactive toward ring opening. In
conformation 0_R′, the R group is pointing toward the gem-
dimethyl group, whereas in 0_R it is pointing away from it.
These two conformations undergo initial coordination to the
gold to yield adducts 1_R and 1′_R, which can both lead to
the ring-opened 2_R intermediate. It is of interest to note that
the slightly less stable conformer 0_R has a lower activation
barrier toward ring opening (see top of Figure 3). Following
ring opening, the pendant acetate can undergo 1,2-transfer in
two steps. Attack of the acetate on carbon a (see 12, Figure 2)
occurs first to provide the five-membered ring intermediates
3_R and 3′_R, and then breaking of these five-membered rings
takes place to give the products 4_R and 4′_R (bottom of Figure
3). As illustrated in Figure 1, starting from 2_R there are two
possibilities leading to the Z and E isomers, respectively. Our
calculations show that from 2_R the energy barrier for the
formation of the Z diene 4′_R is significantly lower than that
of the E-diene 4_R. The preference for the pathway leading to
formation of 4′_R can be explained by a steric argument. In
the pathway leading to the formation of 4_R, the R group is
pointing toward the methylbut-2-ene fragment resulting in an
unfavorable steric repulsion; this steric repulsion is absent in
the pathway leading to the formation of 4′_R. It can therefore
be seen that the stereoselectivity of the reaction is controlled
by a steric interaction between the R group and the methylbut-
2-ene fragment (Figure 4). This fits in with the experimental
Z-selectivity (dienes 9b and 9e, Table 2). In both pathways
shown at the bottom of Figure 1, the energy barrier for the
breaking of the five-membered ring (3TS_R and 3TS′_R) is
greater than for its formation (2TS_R and 2TS′_R). For both
R ) phenyl and ethyl, 3TS_R is higher in energy than 3TS′_R,
but the energy difference for R ) ethyl (1.6 kcal mol-1) is
significantly smaller than for R ) phenyl (5.7 kcal mol-1),
resulting in the low Z-selectivity observed for R ) alkyl (see
9g) and the high Z-selectivity observed for R ) aryl (see 9a-e).
The lowered difference in energy between 3TS_R and 3TS′_R
for R ) ethyl is due to the reduced steric effect of the alkyl
group compared to an aryl group. The akyl group has the ability
to adopt a conformation where two hydrogens are pointing
toward the 2-methylbut-2-ene fragment in 3TS_R, resulting in
a less congested and a lower-energy transition state compared
to that of the phenyl group (see Figure S4, Supporting
Information). In conclusion, we have developed an efficient and
stereoselective synthesis of Z-acetoxydienes 9a-g. On the basis
of DFT calculations, we have been able to propose a detailed
mechanism for the formation of these dienes. These calculations
show that ring opening, followed by intramolecular 1,2-acetate
transfer, is the most favorable pathway and that the ring opening
is rate determining.
The mechanistic divergence of cyclopropenylmethyl acetates
6 from propargyl acetates 1 is an important aspect of the
chemistry uncovered here. In Nolan’s elegant computational
study,1b he demonstrated that the gold vinyl carbenoid species
5 and the gold allene species 4 are in rapid equilibrium and
that the exact product obtained is dependent upon a variety of
factors, including precise substrate structure and the nature of the
ligand employed. It is significant that in the case of cyclopro-
penylmethyl acetate 6 reactivity is guided unambiguously
toward a ring-opening pathway due to the cyclic nature of the
starting material. Cyclopropenes 6 also differ from propargyl
acetates in that they afford gold vinyl carbenoid 12, which is
an isomer of gold vinyl carbenoid 5. We anticipate that the
ability to prepare this alternative gold vinyl carbenoid, along
with the mechanistic detail presented here, will stimulate further
research into the development of new gold-catalyzed reactions
of cyclopropenylmethyl acetates. Research is currently under-
way to further develop the synthetic potential of these substrates
and to identify cyclopropenes that will react via the alternative
pathways outlined in Figure 2.
Acknowledgment. This work was supported by an award
from the Research Corporation (CC6929). Financial support
by California State University, Fullerton, is also gratefully
acknowledged. The acquisition of an NMR spectrometer was
funded by NSF CHE-0521665. We also thank the Australian
Research Council for funding and the National Computational
Infrastructure (NCI) and the Tasmanian Partnership for Ad-
vanced Computing (TPAC) for provision of computing.
Figure 4. Optimized geometries for 3TS′_R and 3TS_R (R )
phenyl).
Supporting Information Available: Full computational
details, experimental procedures, and full spectroscopic data
for all new compounds. This material is available free of
results where the more sterically demanding substrates with R
) o-chlorophenyl and R ) o-bromophenyl result in the highest
(11) 2_R is the reactive conformation of structure 12 in Figure 2, and
the IRC calculations show that 1TS_R is directly connected to 2_R.
OL101862U
Org. Lett., Vol. 12, No. 21, 2010
4771