acyclic carbonates in the presence of a catalytic amount of
[PdCl(π-allyl)]2/phosphine complexes. The protocol allowed
the regioselectivity of the reaction to be effectively con-
trolled.
The allylic substitution between indole 1a and rac-1,3-
diphenylprop-2-enyl methyl carbonate (2a) was taken into
account as the model reaction and a brief screening of
reaction conditions were undertaken.
Then the generality of the reaction was taken into account;
excellent tolerance toward sterically demanding indoles (1b
and 1c), which smoothly reacted with 2a to give 3b and 3c
in high yields (78%, 92% respectively) and in excellent
regioselectivity (Table 2). Moreover, in the presence of both
Table 2. Screening of Indoles in the Pd-catalyzed Allylic
Alkylation with 2aa
The data collected in Table 1 clearly proved that, in the
presence of 5 mol % of [PdCl(π-allyl)]2/dppe8 as the catalytic
Table 1. Pd-Catalyzed Alkylation of Indole 1a by Allylic
Substitutiona
entry
indole
product
yieldb (%)
1
2
3
4
5
6
7
1b
1c
1d
1e
1f
3b
3c
3d
3e
3f
78 (72)
92 (87)
62 (60)
87 (82)
68 (62)
15 (8)
entry
solvent/base
allylic compd convnb (%) yieldc (%)
1
2
3
4
5
6
DMF/Cs2CO3
THF/K2CO3
THF/Li2CO3
CH2Cl2/Li2CO3
CH2Cl2/Li2CO3
CH2Cl2/Li2CO3
2a
2a
2a
2a
2b
2a
62
75
25
70
<10
58
55 (7)
61 (19)
19 (-)
65 (8)
1g
1h
3g
3h
a All the reactions were carried out under nitrogen atmosphere in CH2Cl2
at reflux. Reaction time 24 h. Indole/carbonate/[Pd]/dppe/Li2CO3 ratio: 1/2/
0.05/0.11/2. b Determined by HPLC analysis of the crude material. Isolated
yields are given in parentheses.
52 (31)d
a All the reactions were carried out under nitrogen atmosphere in CH2Cl2
at reflux unless otherwise specified. Reaction time 24 h. Indole/carbonate/
[Pd]/dppe/base ratio: 1/2/0.05/0.11/2. b Determined by HPLC analysis of
the crude material and related to the 3a formation. c Isolated yields. Isolated
yields of the N/C-dialkylated product are given in parentheses. d The reaction
was carried out without base.
indoles bearing electron-releasing (1d) and electron-with-
drawing groups (1e,f) the reaction with 2a proceeded
effectively yielding the C-alkylated adduct (62-87%) in high
regioselectivity (entries 3-5). Finally, the disappointing
results obtained with N-Me- (1g) and N-SO2Ph-indole (1h)
(entries 6 and 7, Table 2) suggest the pivotal role played by
N-metalloindole species as active nucleophiles in the present
allylic alkylation.
system, indole 1a (1 equiv) coupled smoothly with allyl
carbonate 2a9 (2 equiv), and the combined use of a low-
coordinating solvent (DCM) and Li2CO3 as the base drove
the reaction course toward the exclusive formation of the
thermodynamic C-attack (70%) with the double-alkylated
adduct being the only side product (8%, entry 4). Interest-
ingly, no kinetic N-allylic indole was observed under these
conditions. On the other hand, the use of strongest bases
(i.e., Cs2CO3, K2CO3) and more coordinating solvents (i.e.,
DMF, THF) also promotes the alkylation of 1a in satisfactory
yields (entries 1 and 2). However, the reaction conversion
as well as the regioselectivity of the process were inferior
respect to the use of Li2CO3 in DMC.
Allyl carbonates as alkylating precursors as well as the
base proved to be essential in order to guarantee optimal
chemical outcomes. In fact, while the use of allylic acetate
(2b) afforded 3a just in traces (<10%, entry 5), the absence
of base drops the regioselection by enhancing the formation
of the side N/C-dialkylated product (entry 6).10
Then, we turned our attention to the alkylation of 1a (1
equiv) in the presence of less reactive pent-2-enyl methyl
carbonate 2c (2 equiv). However, under the previously
utilized conditions (CH2Cl2, dppe, Li2CO3), a mixture of
C-alkylated 4 and N-alkylated product 5 (3.7:1) was isolated
in moderate yield (62%, Scheme 1). In accord with our
intention to optimize a metallo-catalyzed allylic alkylation
protocol that allows allylindoles to be prepared regioselec-
tively,11 we investigated the influence of the reaction
(9) Trost, B. M.; Fraisse, P. L.; Ball, Z. T. Angew. Chem., Int. Ed. 2002,
41, 1059-1061.
(10) As a peculiarity of carbonate 2a, we observed that mild acidic
conditions (solvent, silica, etc.) also significantly promote the allylic
alkylation of 1a predominantly at the C3 position (i.e., CH2Cl2 in the
presence of silica, reflux 24 h, convn >90%). On the other hand, carbonates
2c-e and 12 (vide infra), less prone to originate carbocationic species respect
to 2a, did not undergo the present indole alkylation in any extent under the
above-mentioned conditions. For a related study, see: Bisaro, F.; Prestat,
G.; Vitale M.; Poli, G. Synlett 2002, 1823-1826.
(8) Several commercially available Pd complexes were tested in the
model reaction, however the use of [PdCl(π-allyl)]2 provided the highest
yields and selectivity. For instance, by using [Pd2dba3]‚CHCl3, a significant
amount of 1,4-addition of indole to the dba (dibenzylidenacetone) was
detected. For a related case, see: Bandini, M.; Cozzi, P. G.; Giacomini,
M.; Melchiorre, P.; Selva, S.; Umani-Ronchi, A. J. Org. Chem. 2002, 67,
3700-3704.
(11) For a related study concerning the synthesis of 3-allylindoles by
Pd-catalyzed cyclization of alkynyltrifluoroacetanilidenes, see: Cacchi, S.;
Fabrizi, G.; Pace, P. J. Org. Chem. 1998, 63, 1001-1011.
3200
Org. Lett., Vol. 6, No. 18, 2004