Communication
(R1 =R2 =H) provided the desired bisindole 5a in about 40%
yield (based on NMR spectroscopy). On the other hand, reac-
tion of 4b (R1 =Me, R2 =H) or 4c (R1 =Bn, R2 =H) with 3a af-
forded a complex reaction mixture.[14] The most promising
result was obtained by reaction of 3a with 4d (R1 =Ac, R2 =H)
affording bisindole 5b in 43% isolated yield. Therefore, the re-
action between 3a and 4d was chosen for further survey of
the reaction conditions (Table 1). The key results are summar-
scope of o-alkynylanilines 3, both aromatic and aliphatic sub-
stituents attached on the Csp carbon atom (R5 =aryl or alkyl)
were well tolerated. The reaction was insensitive to the elec-
tronic properties of the aniline moieties of the two starting
materials. When N-methyl-N-alkylanilines 3 (R4 =alkyl) were
used as substrates, the reaction delivered the N-demethylated
bisindoles selectively (5ab, 5ac). Finally, a variety of functional
groups including fluoride, chloride, bromide, hydroxyl, amide,
and ester were compatible with the reaction conditions to pro-
vide bisindoles with a handle for further transformations.
The indolyl N-protecting groups can be removed under stan-
dard conditions (Scheme 2). Hydrolysis of the N-acetyl bisin-
Table 1. Survey of conditions for the cyclizative cross-coupling reaction.
Entry Pd [equiv]
Additive [equiv] Time [h] Yield [%] (ratio)[b]
1
2
3
4
5
6
7
8
Pd(OAc)2 (0.025)
nBu4NI (0.25)
nBu4NBr (0.25)
nBu4NI (0.25)
nBu4NBr (0.25)
nBu4NBr (0.5)
LiBr (1.0)
KBr (0.5)
nBu4NCl (0.5)
nBu4NBr (0.5)
24
33
52
26
23
44
44
24
21
21
21
59 (2.5:1)
–
Pd(OAc)2 (0.05)
Pd/C (0.025)
Pd/C (0.025)
Pd/C (0.025)
Pd/C (0.05)
Pd/C (0.025)
Pd/C (0.025)
[Na2PdCl4] (0.05)
41 (2:1)
69 (3:1)
77 (5:1)[c]
67 (5:1)
72 (4:1)
n.d.[d]
31 (0.8:1)
40 (1: 1)
50 (1:1)
9
10
11
[Pd(PPh3)2Cl2] (0.05) nBu4NBr (0.5)
[Pd2(dba)3] (0.025) nBu4NBr (0.5)
[a] Conditions: 3a (0.05 mmol), 4d (0.15 mmol), Pd catalyst, additive,
HOAc (1.0 equiv), DMSO (0.5 mL), air (1 atm), 808C, 700 rpm. [b] NMR
yields of 5b and ratios of 5b to 2b (in parentheses) were calculated
based on 1H NMR spectra using CH2Br2 as the internal standard. [c] 70%
isolated yield of 5b. [d] 5b was not detected.
ized as follows: a) nBu4NI was a better additive than nBu4NBr
when Pd(OAc)2 was used as a catalyst (Table 1, entries 1 vs. 2);
b) Pd/C was an excellent pre-catalyst and interestingly,
nBu4NBr was a better additive (Table 1, entry 4) than nBu4NI
(Table 1, entry 3) in this case. LiBr and KBr were also effective
additives (Table 1, entries 6, 7), while addition of nBu4NCl com-
pletely inhibited the desired transformation (Table 1,
entry 8);[15] c) other PdII (Table 1, entries 9, 10) and Pd8 species
(Table 1, entry 11) can also catalyze the reaction, albeit with re-
duced catalytic efficiency relative to Pd/C. Overall, the opti-
mized conditions consisted of heating a DMSO (c=0.1m) solu-
tion of 3a (0.05 mmol) and 4d (3.0 equiv) in the presence of
Pd/C (0.025 equiv, 10 wt% on activated charcoal), nBu4NBr
(0.5 equiv), and HOAc (1.0 equiv) at 808C under air (sealed
tube). Under these reaction conditions, the cyclizative cross-
coupling product 5b was isolated in 70% yield together with
2b, a homo-dimer of 3a, in 14% yield (5b/2b=5:1). We note
that there are only a few examples of using Pd/C as a pre-cata-
lyst for PdII-catalyzed oxidative transformations.[16]
Scheme 2. Selective N-deprotection of bisindoles 5.
dole 5b under basic conditions delivered 2,3’-bisindole 5a in
86% yield,[17] while cleavage of the N-Boc-bisindole 5o under
acidic conditions gave 5a in 89% yield.[18] The N-methoxycar-
bonylethyl protecting group in 5ac was selectively removed
under basic conditions through a retro-Michael reaction to
afford 10 in 66% yield.[19]
There are two catalytic cycles according to our working hy-
pothesis: formation of 3-alkynylindoles 8 and their subsequent
conversion to bisindoles 5 (see Scheme 1). To gain mechanistic
insights on this novel transformation, the reaction between 3a
and 4d was carefully monitored. As shown in Figure 1, cycliza-
tive alkynylation of 3a with 4d took place rapidly to deliver 3-
alkynylindole 8b. Only after the total consumption of aniline
3a (6 h) was the cyclization of 8b initiated to afford, after an-
other 16 h, the bisindole 5b. This kinetic data clearly suggest-
ed a temporal separation of the two catalytic cycles.[20] The
aminopalladation of N,N-dimethyl-2-(p-tolylethynyl)aniline (3a)
is apparently much faster than that of the acetanilide 8b due
presumably to the higher nucleophilicity of the aniline nitro-
gen.[21] We note here that in DMSO, acetic acid is not acidic
With the optimum conditions in hand, the scope of the reac-
tion was next examined (Table 2). These conditions were appli-
cable not only to N-acetyl-o-ethynylanilines, but also to N-ben-
zoyl- (5n) and N-Boc-o-ethynylanilines with N-Boc-o-ethynylani-
lines being in general the best substrates. With respect to the
&
&
Chem. Eur. J. 2015, 21, 1 – 5
2
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!