6750
J. R. Rizzo et al. / Tetrahedron Letters 49 (2008) 6749–6751
Et3SiH and CF3CO2H in dichloromethane at low temperatures,
developed by Steele9 and Mahadevan,10 thus became attractive
to us for several reasons. First, it avoids the use of alkyl halides, a
class of compounds oftens associated with mutagenic properties,
and hence requires vigorous control in scale-up settings. Secondly,
the method appears to be reasonably efficient, affording 3-alkylin-
dole in one simple step. Indeed, the procedure worked well as
advertized, affording 3-alkylated indoles in moderate to good yield
when aldehydes were used as the alkyl donors. However, yields
were low when ketones had to be used in place of aldehydes.
The authors11 also noted that unsubstituted indole failed to afford
any identifiable products. In our hands, we were able to isolate,
after a prolonged reaction time, 3,1(N)-biscyclopentylindole (8,
R1–R2 = –(CH2)3–) in 38% yield from indole (1) and cyclopentanone
under the Et3SiH/CF3CO2H conditions. Nonetheless, we were
encouraged by the fact that the desired monoalkylated product
compound 2 was also isolated in 30% yield from the same reaction
mixture.
It appeared to us that the shortcomings associated with 2-
unsubstituted indoles in this reductive alkylation reaction are
likely a manifestation of increased reactivity of the product 2
toward further reaction with electrophiles. This is likely due to a
lack of steric hindrance at the 2-position and higher electron den-
sity when an alkyl group is introduced at the 3-position. One way
to address the problem is to control the reactive concentration of
the electrophiles via slow addition of the ketone to the reaction
mixture. However, doing so led to an increased amount of bis-
indolylmethane (7) formation. We reasoned that modulating the
reactivities of the oxonium (3) and the indolium (5) intermediates
might provide a solution toward selective formation of 2, which as
an unsubstituted indole is not stable under the strong acidic and
reductive conditions.7 A quick screen of common acids, both Lewis
and Brønsted, was conducted using the very demanding substrate,
that is, unsubstituted indole (1) and cyclopentanone, as a model
reaction (Eq. 1). Not surprisingly, stronger acids, such as CF3SO3H
and sulfuric acid, uniformly led to decomposition of the indole sub-
strate, as did strong Lewis acids such as ZnCl2 and BF3. Weak acids
(PhCO2H, CH3CO2H) led to recovery of the starting materials. A bal-
ance was found with CCl3CO2H, which afforded an acceptable reac-
tion rate for the reductive alkylation, and resulted in minimal
R1
R1 R2
O
R2
R
R
Et3SiH, Cl3CCO2H
toluene
N
H
N
H
(isolated yield)
O
Me
Me
N
H
N
H
N
H
N
H
65%
71%
46%
59%
F3C
Ph
Et
Et
Et
Me
n-C6H13
Me
N
H
63%
N
H
N
H
N
H
67%
69%
85%
Me
Ph
F
OMe
NO2
N
H
N
H
N
N
H
H
64%
90%
69%
0%
Br
Br
N
H
N
N
H
N
H
Br
H
Br
product decomposition. However, formation of
a substantial
0%
99%
68%
94%
amount of bisalkylated product (8) persisted. This problem was
successfully addressed by a controlled addition of a premixed solu-
tion of ketone or aldehyde with indole to a heated solution of
CCl3CO2H. We were also successful at replacing the chlorinated
reaction solvents (CH2Cl2 or ClCH2CH2Cl) with environmentally
friendlier hydrocarbons (toluene). For example, addition of cyclo-
pentanone and indole to a heated (70 °C) solution of trichloroacetic
acid and triethylsilane in toluene provided the desired 3-cyclopen-
tylindole in 65% isolated yield.11 The improved conditions show
very little formation of the biscyclopentylindole (8) or the bis-
indoylmethane (7). Triethylsilane proved to be the best hydride
donor, as bulkier silanes such as i-Pr3SiH, n-Bu3SiH, and Ph2SiH2
afforded very little product. Other metal hydrides such as
n-Bu3SnH hydride or n-Bu3GeH were ineffective.
Figure 2. Reductive alkylation of indoles with ketones.
convenient. It provides a useful alternative to the existing methods
of making substituted indoles, and proved to be a preferred one for
introducing sec-alkyls into the 3-position without the need for
N-protection. Seasoned medicinal chemists would find these
compounds convenient building blocks for exploring the pharma-
cological effect of structural diversity.
Acknowledgment
The authors wish to thank Dr. Alfio Borghese for demonstrating
this chemistry at multi-kilogram scale.
A reaction scope was briefly investigated (Fig. 2).12 Both ali-
phatic and aromatic ketones worked well, except highly hindered
substrates such as benzophenone (22%) or unstable ones like 4-tet-
rahydropyranone (46%). Substituted indoles, including 2-alkylated
substrates, also gave good yields. However, 4-bromo and 4-nitro-
indole failed to react, probably due to a combination of steric
hindrance and electron deficiency.
In summary, we were able to demonstrate that reductive alkyl-
ation of unsubstituted indoles with ketones can be achieved using
trichloroacetic acid and triethylsilane. The procedure has been
demonstrated at the kilogram scale, and proved to be robust and
References and notes
1. Buchanan, T. M.; Ramadan, N. M.; Aurora, S. Exp. Rev. Neurotherap. 2004, 4,
391–402.
2. Goadsby, P. J. Nat. Rev. Drug Disc. 2005, 4, 741–750.
3. Mannix, L. K.; Files, J. A. CNS Drugs 2005, 19, 951–972.
4. Attaur, R.; Basha, A. Indole Alkaloids. 1997, p 336.
5. Sundberg, R. J. Indoles; Academic Press: New York, 1996.
6. Indoles: The Monoterpenoid Indole Alkaloids, Saxton, J. E., Ed.; The Chemistry of
Heterocyclic Compounds; 1983; Vol. 25, Pt. 4, p 886.
7. Leete, E. J. Am. Chem. Soc. 1959, 81, 6023–6026.
Rizzo, John R.
