Journal of the American Chemical Society
Communication
The development of the polycyclizations in Scheme 1B was
driven by the observation that cyclopropane 3a is converted to
complex polycycle 4a in 10% yield when exposed to
inclusion of Na SO as a desiccant provided a small but
reproducible benefit.
2
4
7
The process has broad scope with respect to the indole unit
(Table 1A). Systems 3c−j possessing electronically diverse
substituents generated polycycles 4c−j in good to excellent
yield. Substitution at C7 results in diminished efficiencies such
that 4k was formed in only 21% yield. The process can be
transferred to substituted aminocyclopropanes (Table 1B); for
example, polycyclization of trans-1,2-disubstituted systems 3l−
n generated 4l−n with excellent levels of diastereo- and
regiocontrol. These features arise from selective cleavage of the
less hindered proximal C−C bond a of 3l−n, followed by
[
%
(
Rh(cod) ]OTf (7.5 mol %), P(3,5-(CF ) C H ) (15 mol
), PhOCH CO H (30 mol %), and CO (1 atm) at 140 °C
2 2
Table 1A). This process can be rationalized by invoking endo-
2 3 2 6 3 3
Table 1. Dearomatizing endo-Polycyclizations of Indole
Systems
8
transfer of cyclopropane stereochemistry to the product. The
starting materials are easily accessed in enantioenriched form,
and this allowed enantiospecific conversion of 3m to 4m
(
3
98.5:1.5 er). Polycyclizations of trisubstituted cyclopropanes
o and 3p resulted in efficient desymmetrization to provide 4o
and 4p with exquisite levels of diastereocontrol (Table 1C).
For nonsymmetrical system 3q, C−C bond activation was
selective for benzylic C−C bond a leading to regioselective
generation of 4q, again with complete diastereocontrol.
The processes in Table 1 validate the conceptual blueprint
in Scheme 1B and also provide a notable contribution to
9
indole dearomatization chemistry. Uniquely, the method
enables the concurrent formation of two C−C bonds at C-2
1
0
(
i.e., “dual C−H functionalization”) such that this position
formally functions as a carbene equivalent (Table 1D). Prior
methods for accessing similar structures require preinstallation
of a substituent at C-2. Other catalytic dual functionalizing
11
indole dearomatizations usually generate new C−C/C−X
12
bonds at both C-2 and C-3.
Polycyclization of 3a in the presence of D O (300 mol %)
2
delivered deuterio-3aa and deuterio-4aa in 32% and 42% yield,
respectively (eq 1 in Scheme 2). For deuterio-3aa, deuterium
incorporation was observed at C2−H (17%) and C3−H
(
43%). Exchange at these positions is dependent on the
the acid additive (see the SI); these data support exchange by
reversible C−H activation of 3a. For deuterio-4aa, deuterium
incorporation was observed at C2−H (27%), C11−H (20%)
a
and C11−H (49%). When the same reaction was run for 72 h,
b
a similar pattern of deuterium incorporation was observed (eq
2
). Deuterium incorporation at C11−H of deuterio-4aa/4ab
b
likely occurs at the stage of 3a because similar levels of
exchange are observed at C3−H of this system. When 4a was
resubjected to the reaction conditions, but in the presence of
D O, deuterium incorporation was observed at C2−H, likely
2
due to enolization (eq 3). Deuterium incorporation at C11−
H of deuterio-4aa/4ab is consistent with syn-carbometalation
a
a
b
c
Run at 140 °C. Run in DCB (0.2 M). [Rh(cod) ]OTf (10 mol %)
2
of the C2−C3 π-system (from VII) prior to protodemetala-
tion. The necessary proton originates from either the acid
additive or C2−H of 3a. When deuterio-3ab was exposed to
standard conditions, the deuterium label was transferred
d
e
was used. Run at 150 °C. Run at 125 °C.
polycyclization of key alkyl-Rh(I) intermediate VII (cf. IV)
onto the C2−C3 π-system of the indole (vide infra). Extensive
studies were undertaken to improve efficiency; ultimately, we
predominantly to C11−H (eq 5). Finally, exposure of
b
equimolar quantities of 3a and deuterio-3ac to standard
found that the combination of [Rh(cod) ]OTf (7.5 mol %)
conditions revealed a small kinetic isotope effect (k /k =
2
H D
and 4-NMe C H CO H (30 mol %) delivers 4a in 81% yield.
1.21), suggesting that C−H cleavage is not turnover limiting
13
2
6
4
2
The inclusion of phosphine ligands resulted in reduced yields,
and neutral Rh-sources were ineffective. The acid additive had
the most profound effect, with 4-Me NC H CO H (30 mol %)
(eq 6).
The experiments above are consistent with a pathway
involving stereoretentive protodemetalation of a C11-alkyl-
Rh(I) species. To provide insight into (a) the mode of C−C
bond activation and (b) the C−C bond forming sequence
from the rhodabicycle (cf. II), we undertook DFT studies
2
6
4
2
Information (SI)). Higher or lower CO pressures offered no
benefits and lower reaction temperatures were ineffective. The
B
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX