COMMUNICATION
studies, which suggest that electron-rich and -deficient ben-
zoic acids might take two different reaction pathways due to
their difference in reactivity towards decarboxylation.
Initially, we examined the reaction of N-acetylindole 1
with 2,4-dimethoxybenzoic acid (2a; 2 equiv) in DMF at
palladation on indole. As expected, the addition of propion-
ic acid (0.25 equiv) gave rise to an improvement in yield
(Table 1, entry 8). However, as the amount of propionic acid
increased, the yield of the desired product decreased be-
cause too much propionic acid slowed decarboxylation. In-
terestingly, an increase in the amount of 2,4-dimethoxyben-
zoic acid not only promoted the conversion of indole but
also improved the yield (Table 1, entry 10). Gratifyingly, fur-
ther optimization disclosed that the treatment of N-acetylin-
dole with 2,4-dimethoxybenzoic acid (5 equiv) in dioxane
(1.5 mL) at 808C in the pres-
808C in the presence of PdACTHNUTRGNE(UNG TFA)2 (5 mol%, TFA=trifluor-
oacetate) as a catalyst, Ag2CO3 (2 equiv) as an oxidant and
DMSO (3.5 equiv) as a ligand (1:70 ratio of Pd to DMSO),
and observed that this reaction offered the desired C2-aryla-
tion product in 9% yield (Table 1, entry 1). Under otherwise
ence of Pd
Ag2CO3
ACHTUNGTRENNUNG
Table 1. Optimization studies for Pd/Ag-catalyzed arylation of indoles(1) with 2,4-dimethoxybenzoicACTHNUTRGNEUNG
(2a).[a]
(1.5 equiv), and propionic acid
(0.25 equiv) resulted in almost
total conversion of indole, af-
fording the desired product in
78% yield with 16.4:1 C2/C3
ratio (Table 1, entry 13). The
effect of an N-protecting group
on the reactivity of indole has
also been investigated under
the optimal conditions. N-Piva-
loylindole generated the corre-
Entry Pd
G
2a
[equiv]
R[b] Solvent
Ligand
([equiv])
EtCOOH
[equiv]
Yield
[%][c]
Ratio (3/
4)[d]
[mol%]
1
5
2
2
2
2
2
2
2
2
2
4
5
5
5
5
5
5
5
5
Ac DMF
DMSO (3.5)
DMSO (3.5)
DMSO (3.5)
DMSO (3.5)
DMSO (7.0)
DMSO (1.0)
TMSO (1.0)
TMSO (1.0)
TMSO (1.0)
TMSO (1.0)
TMSO (1.0)
TMSO (1.5)
9
39
15
4
>99:1
8.0:1
>99:1
>99:1
10:1
2
5
Ac dioxane
Ac toluene
Ac NMP
3
5
4
5
5
5
Ac dioxane
Ac dioxane
Ac dioxane
Ac dioxane
Ac dioxane
Ac dioxane
Ac dioxane
Ac dioxane
26
43
49
57
49
65
33
70
78
0
sponding product in
comparable to that of N-acetyl-
indole with a lower selectivity
(2.1:1 C2/C3 ratio), and N-
tosylindole was less reactive
a yield
6
5
9.7:1
7
5
11.4:1
16.9:1
20.4:1
19.2:1
>99:1
17.4:1
16.4:1
AHCTUNGTRENNUNG
8
5
0.25
1
9
5
10
11
12
13
14
15
16
17
18
5
5
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.25
ACHTUNGTRENNUNG
than N-acetylindole, providing
a moderate yield, whereas free
N-H indole, N-methylindole,
7.5
7.5
7.5
7.5
7.5
7.5
7.5
Ac dioxane[e] TMSO (1.5)
H
dioxane[e] TMSO (1.5)
CH3 dioxane[e] TMSO (1.5)
Piv dioxane[e] TMSO (1.5)
<5
72
42
trace
and
N-tert-butyloxycarbonyl
2.1:1
>99:1
(Boc) indole gave no product.
With a set of optimized con-
ditions in hand (Table 1,
entry 13), we evaluated the
scope of substituents on the N-
acetylindole. As shown in
Table 2, both electron-with-
Ts
dioxane[e] TMSO (1.5)
Boc dioxane[e] TMSO (1.5)
[a] Reaction conditions: indole (0.20 mmol, 1.0 equiv), Ag2CO3 (2 equiv), solvent (1.0 mL), 24 h, 808C.
[b] Ac=acetyl, Piv=pivaloyl, Ts=tosyl, Boc=tert-butoxycarbonyl. [c] Isolated yields. [d] Determined by GC–
MS. [e] Dioxane (1.5 mL).
identical conditions, screening solvents showed that dioxane
is the solvent of choice (Table 1, entry 2–4). The reaction
conducted in dioxane exhibited a good selectivity at the 2-
position of indole with an 8.0:1 C2/C3 ratio. Increasing the
amount of DMSO to seven equivalents retarded the decar-
boxylation process and led to a decrease in yield (Table 1,
entry 5). In contrast, when the amount of DMSO was re-
duced to one equivalent (1:20 ratio of Pd to DMSO), 2,4-di-
methoxybenzoic acid smoothly underwent decarboxylation
and provided an improved yield (Table 1, entry 6). Tetra-
methylene sulfoxide (TMSO) was observed to be more ef-
fective than DMSO (Table 1, entry 7). Although the use of
one equivalent of DMSO or TMSO led to complete decar-
boxylation of 2a, only moderate yields were obtained due to
incomplete conversion of indole. We speculated that the in-
troduction of additional carboxylic acids, such as propionic
acid, to the reaction system would facilitate the electrophilic
drawing and -donating substituents were compatible, afford-
ing synthetically useful yields in most cases. Importantly,
high selectivity can be obtained for 2-arylindoles. Although
substituents on the indole core can influence the yields of
this reaction, no clear trend was observed for the depend-
ence of the reaction yield on the nature of the substituents.
For example, the electron-withdrawing fluoro-substituent at
the C5-position provided a good yield (Table 2, 3 f), whereas
electron-withdrawing nitro and ester groups at the 5-posi-
tion of the indole core led to lower yields (Table 2, 3j, 3k).
The position of the substituent on the indole core was also
observed to affect both the reaction yield and selectivity, as
illustrated by the reactions of three isomers of methoxy-sub-
stituted N-acetylindole (Table 2, 3c–e). Notably, the reac-
tions were tolerant of halogen substituents (Table 2, 3 f–i),
providing the complementary platform for further transfor-
mations by Pd0-catalyzed cross-coupling reactions.
Chem. Eur. J. 2010, 16, 5876 – 5881
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5877