6
S. Jansone-Popova et al. / Tetrahedron xxx (2014) 1e10
Table 4
Effects of ring size
reactive to complete the reaction. A para-CF3 group or any number
of meta substituents were readily tolerated.
Alkyl substituents on the alkyne displayed differing reactivity.
The tert-butyl group found in the alkynyl diazoester 50 provided
opportunity for a 1,2-methyl migration to outcompete CeH bond
insertion and form the trimethylvinyl butenolide 51. Similarly,
a 1,2-hydride shift in the n-butyl group of ester 52 generated the
vinylbutenolide 53 as a mixture of cis and trans isomers. This oc-
curred even in pentane as a solvent, which has previously been
shown to discourage 1,2-migrations in carbene cascades.12o,p These
migrations in alkyl alkynes find robust precedent in Padwa’s work,
where similar migrations occur after carbene/alkyne cascade re-
actions.34 One note of interest is that the cis/trans ratios for 53 are
intermediate between those seen for 32 in Hoye’s study in Scheme
2. We may conclude that the early mechanistic steps in this cascade
process are closely related to those seen in previous examples of
the cascade.
Much of the most novel reactivity in this reaction was found in
the variation of the ring found between the alkyne and the diazo-
ester. Due to the ease of substrate synthesis, a number of permu-
tations of the original structure were readily available. A study on
the effects of ring sizes was performed early on to both establish the
scope of the reaction and define the effects of ring conformations
on product outcomes. Because of the conformational demands in
the transition state for CeH bond insertion to form a bridged
bicycloalkane, this analysis was deemed to be an important con-
sideration. As can be seen in Scheme 1, the final step of the cascade
sequence is a transannular CeH bond insertion from intermediate 8
(or its carbenoid mechanistic equivalent). The CeH bond that is
geometrically available for this insertion will be the most likely site
of CeC bond formation for the bridged bicycloalkane.
In the case of our study compound 6, insertion into either of the
CeH bonds highlighted in red in Table 4 provides bicyclo[2.2.1]
heptane 9, albeit as a racemic mixture (entry 1). However, in the
case of the cyclohexyl derivative, insertion into the CeH bond at
either of the 3-positions (see 54a) would form the bicyclo[3.2.1]
octane 55, but insertion at the 4-position would instead form
bicyclo[2.2.2]octane 56. Given our goal to be able to access a broad
selection of bridged ring sizes and connectivities, we need to be
able to access either of those products at will. Fortunately, with no
catalyst or substrate bias, the rates of insertion at those two posi-
tions occurred with a nearly statistical preference (entry 2). The
similarity between insertion rates at those sites will allow for bias
to be introduced either through substituents on the substrate or
through the future synthesis of catalysts with more controlling li-
gands. We note that the silyl-bearing carbon in 55 was found as
a mixture of a- and b-silyl diastereomers, but in all examples in
which an aryl-substituted alkyne was used only one diastereomer
was formed (see Table 6). The reason for the differences in dia-
stereoselectivity are difficult to fully rationalize given the uncertain
nature of the mechanistic intermediate formed from the metal
carbene reacting with the alkyne. A rigid transition state for the
formation of 47 and 49 where the transannular CeH bond insertion
occurs from the midplane of the ring is likely responsible for the
formation of only the endo diastereomer. Greater conformational
flexibility in the cyclohexane ring would allow for attack of the CeH
bond from either side, resulting in a mixture of diastereomers.
The use of seven- and eight-membered rings in the substrates
showed that there is more conformational bias in those systems.
Assuming that the carbenoid derived from the alkyne in cyclo-
heptyl substrate 57 must be in a pseudoaxial position for trans-
annular CeH bond insertion, a likely conformation is illustrated
with the most likely points of CeH bond insertion highlighted in
blue (entry 3). Indeed, the major product bicyclo[4.2.1]nonane 58 is
derived from insertion at those axially-displayed CeH bonds.
Similarly, the major product from the reaction of cyclooctyl
substrate 60 may be predicted. In this case, there are two somewhat
equivalent conformations that put the alkyne pseudoaxial and have
only 2 eclipsing interactions around the ring (entry 4). In the top
case, only one possibility presents itself for CeH bond insertion,
and in the lower case there are two non-equivalent CeH bonds that
appear as targets for insertion. The product distribution reflects this
with 5.5:1 preference for insertion into the red highlighted CeH
bonds to provide bicyclo[4.2.2]decane 61 as the major product.
Given the understanding of the conformations of variously sized
rings with substitution,11 this predictive ability is highly applicable
to design a synthesis of a complex bridged polycycle using this
strategy.
To further demonstrate the ability to predict reactivity based on
stereoelectronic and conformational considerations, we turned our
attention to cyclohexyl substrates with substituents at various
positions around the ring and with differing diastereomeric re-
lationships to the alkyne. Placing a methyl group in the ring’s 4-
position in a trans relationship to the alkyne as shown in 63