cyclization in comparable time to In(OTf)3, but with a
significant decrease in yield (entry 4). Inspired by its
successful use in Sc(OTf)3- and In(OTf)3-catalyzed Frie-
delꢀCrafts acylations14 and Nazarov cyclizations,8 Li-
ClO4 was explored as an additive. However, when 1
equiv of LiClO4 was used with either In3þ salt, very low
conversion was observed. Thus, In(OTf)3 at 5 mol %
loading in dichloromethane at room temperature was
chosen as the final protocol.
Table 2. Heteroaromatic Homo-Nazarov Cyclizationsa
With optimized conditions for the cyclization in hand, a
diverse set of heteroaryl substrates was examined to de-
termine the reaction scope and limitations (Table 2). In
direct comparison to the 2-thienyl cyclopropane 3a which
afforded 4a in 86% yield (entry 1), the 3-thienyl substrate
3c provided cyclohexanone 4c in 73% yield (entry 3). In
contrast, the 3-furanyl substrate 3d (entry 4) provided its
product 4d in a higher yield (73% vs 67%) than that of
2-substituted counterpart 3b15 (entry 2). The 2-indolyl
cyclopropane 3e afforded the cyclized product 4e in 63%
yield (entry 5), and the 3-indole 3f gave product 4f in 91%
yield (entry 7). Similarly, the 2-benzofuran 3g cleanly
cyclized to give 4g in 61% yield (entry 6) and the 3-benzo-
furan 3h afforded the desired product 4h in 71% yield
(entry 8). This is noteworthy because benzofuran had been
unsuccessful as a homo-Nazarov substrate due to issues
with competing polymerization.10
To date, only 2,3-ring-fused heteroaromatics had been
synthesized using homo-Nazarov cyclizations. As an ex-
pansion of the scope of this reaction, we were interested in
trying to synthesize 3,4-ring-fused compounds as well. We
envisioned employing a 3-substituted substrate with the
2-position blocked, as in the 2-bromo thienyl cyclopropane
3i. With the 2-position unavailable, the only site for electro-
philic attack would be the 4-position. Thus, cyclization of 3i
should produce the 3,4-fused heteroaryl cyclohexanone 4i.
As anticipated, when subjected to the reaction conditions, 3i
generated the desired product 4i in 56% yield (entry 9).
Next, other donor substituents about the cyclopropane
were examined (Table 3). Phenyl derivatives 3j and 3k
provided the desired ring-fused cyclohexanones 4j and 4k
in 81% and 83% yield, respectively (entry 1).16 When R-
methyl styrene was used to generate cyclopropane 3l,
product 4l was obtained in 71% yield (entry 3). Similarly,
the indanyl substrate 3m gave tetracycle 4m in 87% yield
(entry 4). Inspired by Yadav’s report, silyl derivative 3n
was synthesized and cyclized to afford 4n in 72% yield
(entry 5). When the donor group on the cyclopropane was
changed to an oxygen donor (as in 3p, derived from
dihydropyran), no cyclization was observed (entry 7).
Even at higher catalyst loading or elevated temperatures,
a Reactions run with 1 equiv of substrate 3 and 5 mol % In(OTf)3 in
CH2Cl2 at 25 °C and complete within 6 h. b Isolated yields after column
chromatography. c Diastereoselectivities as determined by crude 1H
NMR. d Reaction performed in 1,2-dichloroethane at 80 °C. e 2:1
Mixture of keto and enol forms.
only startingmaterialwas recovered. Thislackofreactivity
can be rationalized if no chelation event was occurring at
the two carbonyls. To support this hypothesis, we looked
at the lowest energy conformers of 3o based on MMFF
calculations.17 In each of these conformers, the two carbo-
nyl oxygens are anti to one another, thus eliminating the
ability to generate the requisite six-membered chelate.
Without this activation, cyclopropane ring-opening does
not occur. Furthermore, this effect seems to arise from the
stereoelectronic influence of the methyl ester on the con-
formation of the fused pyran ring.18 This effect can
1
arguably be seen in the H NMR spectrum of 3o where
the hydrogen at the fused ring junction adjacent to the
oxygen is located at 6.5 ppm, which represents a ∼3 ppm
downfield shift from the analogous pyranyl derivative
without the ester.9,19 This downfield shift suggests the
strong influence of the electron-withdrawing ester on
the conformation of the fused ring. This hypothesis
was further supported when the acyclic ether 3p (derived
from ethyl vinyl ether) was subjected to the standard
reaction conditions and the substituted benzothiophene 5
was obtained in 51% yield (entry 8). This product
(14) (a) Chapman, C. J.; Frost, C. G.; Hartley, J. P.; Whittle, A. J.
Tetrahedron Lett. 2001, 42, 773. (b) Kawada, A.; Mitamura, S.;
Kobayashi, S. Chem. Commun. 1996, 183. (c) Hachiya, I.; Moriwaki,
M.; Kobayashi, S. Tetrahedron Lett. 1995, 36, 409.
(15) The 2-furanyl derivative 3c readily cyclized at room tempera-
ture, but the yields (∼28%) were fairly low due to byproduct formation.
To circumvent this issue, 3c was refluxed in DCE. No byproduct
formation was observed.
(17) MMFF calculations were run using Trident software available
from Wave function, Inc.
(18) It has been previously shown that cyclization readily occurs for
the analogous pyran derivative without the methyl ester (see ref 9).
(19) See Supporting Information.
(16) We have observed that heating is required to achieve efficient
cyclization for phenyl rings without substitution and those substituted
with electron-withrawing groups.
1954
Org. Lett., Vol. 13, No. 8, 2011