Ligand-controlled selectivity in the Pd-catalyzed C-H/C-H cross-coupling of indoles with molecular oxygen

Article abstract of DOI:10.1021/acscatal.0c04893

Two indoles linked by a C-C bond have recently emerged as promising scaffolds in medicinal chemistry. Their synthesis, however, involves a multitude of reaction steps. So far, the direct C-H/C-H cross-coupling of two similar heteroaromatics lacking direct

Full text of DOI:10.1021/acscatal.0c04893

pubs.acs.org/acscatalysis
Research Article
Ligand-Controlled Selectivity in the Pd-Catalyzed C−H/C−H Cross-
Coupling of Indoles with Molecular Oxygen
Igor Beckers, Besir Krasniqi, Prashant Kumar, Daniel Escudero, and Dirk De Vos*
Cite This: ACS Catal. 2021, 11, 2435−2444
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: Two indoles linked by a C−C bond have recently
emerged as promising scaffolds in medicinal chemistry. Their
synthesis, however, involves a multitude of reaction steps. So far,
the direct C−H/C−H cross-coupling of two similar heteroar-
omatics lacking directing groups remains particularly challenging.
Transition metals are often added as wasteful sacrificial oxidants to
influence the selectivity in Pd-catalyzed C−H/C−H couplings. In
this work, we report on the selective C−H/C−H cross-coupling of
N-substituted indoles without directing groups, driven by
molecular oxygen under mild conditions. The selectivity of the
C−H activations was studied via kinetic experiments and
computational investigations. A quantitative Hammett study of
the aromatic carboxylate ligands shows that regioselectivity can be
directed by rationally tuning their electronic properties. This ligand-controlled selectivity can be exploited to obtain selective cross-
coupling between two indoles with different substitution patterns, as demonstrated in the synthesis of a potent insulin-like growth
factor inhibitor.
KEYWORDS: homogeneous catalysis, C−H activation, C−C coupling, indole, palladium, oxygen
ndoles are highly important heterocyclic building blocks in
biochemistry. Many bio-active molecules such as the amino-
The oxidative C−H/C−H coupling of indoles has been
pursued as an efficient, one-step route to bis-indole
compounds (Figure 1c). While previously reported Pd-
catalysts allow the direct C−H/C−H coupling of indoles,
transition metals such as Cu(OAc)2 or AgNO3 are used as
oxidants, or excessive Pd-loadings are required.9−11 The
(sub)stoichiometric consumption of transition metals is
expensive and leads to low atom efficiency. This is analogous
to the metal waste generated in C−X/C−M couplings, which
is usually avoided with C−H activation. In addition to their
role as terminal oxidants, the transition-metal complexes were
believed to be essential in various oxidative coupling reactions,
including the C−H/C−H coupling of indoles, to obtain
regioselective C−H activation.12,13 Nevertheless, the previ-
ously reported C−H/C−H cross-coupling reactions require
two strongly distinct (hetero)aromatic scaffolds, one of which
is usually added in large excess.14−20 The C−H/C−H coupling
of two indoles without directing group(s) has so far been
restricted to the homocoupling of identical indole reactants,
I
acid tryptophan or the neurotransmitter serotonin, as well as a
wide range of pharmaceuticals, contain an indole backbone.1
Likewise, two indoles linked by a single C−C bond have
recently emerged as a promising scaffold for medicinal
compounds with anticancer activity (Figure 1a). These were
proven to be excellent inhibitors for various proteins involved
in cell signaling, such as insulin-like growth factors (1),
androgen receptors (2), and the transcription factor p53
(3).2−4 Moreover, bis-indole alkaloids represent a large class of
interesting natural products, occasionally bearing C−C bonds
that directly link two indole rings (4 and 5).5,6
The search for these potent bis-indole compounds and their
subsequent development is, however, hampered by tedious
synthetic procedures (Figure 1b). The conventional C−C
coupling reactions (e.g., Suzuki and Stille) require multiple
preparation and purification steps to obtain halogenated or
metalated indole reactants.7 Unlike their homoaromatic
counterparts, the stability of indolyl halides or boronic acids
is limited. Another conventional route toward bis-indoles
involves the de novo formation of one indole ring. For
example, indolyl aldehydes or ketones can react with
phenylhydrazines in the Fischer indole synthesis.8 However,
this strategy requires functional groups that are consumed in
the ring closure reaction, which also increases the number of
synthetic steps.
Received: November 10, 2020
Revised: January 26, 2021
Published: February 9, 2021
https://dx.doi.org/10.1021/acscatal.0c04893
© 2021 American Chemical Society
ACS Catal. 2021, 11, 2435−2444
2435

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Figure 1. Overview of the conventional routes and the C−H/C−H coupling for the synthesis of bio-active compounds containing the bis-indole
scaffold.
which drastically narrows its applicability for the synthesis of
pharmaceutical compounds.
(C2-C3’ product). A second isomer, 1,1′-dimethyl-3,3′-
bisindole (C3-C3’ product), was detected as a side product
in small quantities. Despite the absence of other transition
metals as the oxidant, the catalyst exhibited a remarkable
selectivity for the C2-C3’ over the C3-C3’ coupling product
(product ratio 11:1). Other possible side products, such as the
C2-C2’ regio-isomer, were not detected. The catalytic activity
was strongly increased by changing the concentration of
AcOH, showing an optimum at the Pd(OAc)2:AcOH ratio of
1:75 (Figure 2a). This is in agreement with our previous
findings in the Pd-catalyzed oxidative C-H/C-H coupling of
benzene.21 The regeneration of Pd(OAc)2 from Pd0 with
oxygen requires two molecules of AcOH, which consecutively
serve as ligands that can assist the C−H bond cleavage via
partial coordination to the Pd(II)-center. However, an
inhibitory effect of AcOH was also observed, which can be
linked to the coordination to the Pd-indolyl intermediates and
In this work, we report on the Pd-catalyzed C−H/C−H
cross-coupling of indoles without directing groups using
molecular oxygen as the terminal oxidant. This catalytic
system enables a single-step and waste-free route to a variety of
bisindole structures, with H2O as the only stoichiometric
byproduct. A high selectivity for bisindoles coupled at the C2
and C3’ positions was obtained. Mechanistic insights into the
reaction were acquired via density functional theory (DFT)
calculations and kinetic isotope studies. By varying the
electronic properties of the carboxylate ligand, the selectivity
of the Pd-catalyst could be steered toward the activation of the
C−H bonds on either the C2 or C3 position. This ligand-
controlled selectivity also enabled the selective C−H/C−H
cross-coupling of indoles, as demonstrated for the synthesis of
IGF-inhibitor 1. Such selective intermolecular C−H/C−H
cross-couplings between two nonidentical yet highly similar
(hetero)aromatic reactants are particularly challenging and
have so far been elusive.
the consecutive protonolysis of the C−Pd bonds.21,22
A
gradual increase in the concentration of other carboxylic acids
has a similar effect on the catalytic activity, as demonstrated
with benzoic acid (BZA) and pivalic acid (PivOH). However, a
lower optimal activity was obtained because of the steric
hindrance of these ligands. Using the optimal concentration of
AcOH in the reaction, high yields of the C2-C3’ product were
obtained for longer reaction times (Figure 2b). The kinetic
profile of the reaction shows a clear first-order dependence for
the 1-methylindole reactant, which may indicate that one of
REACTION OPTIMIZATION
■
Initial results showed that 1-methylindole (0.15 M) can be
effectively coupled using molecular dioxygen (16 bar O2) as
the oxidant, with Pd(OAc)2 (1 mol %) and AcOH (75 mol %)
in DMSO. Mild conditions (40 °C) were crucial for the
selective formation of the 1,1′-dimethyl-2,3′-bisindole product
2436
https://dx.doi.org/10.1021/acscatal.0c04893
ACS Catal. 2021, 11, 2435−2444

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Figure 2. Optimization and kinetics of the C−H/C−H coupling of 1-methylindole with molecular oxygen. Standard reaction conditions: 0.0015 M
Pd(OAc)2, 0.15 M 1-methylindole, 0.1125 M AcOH, 40 °C, 3 h, 16 bar O2, 1 mL DMSO. (a) Influence of AcOH, BZA, and PivOH ligand
concentrations (mol %) on the turnover frequency (h−1) in the formation of 1,1′-dimethyl-2,3′-bisindole. (b) Time profiles of the yield of 1,1′-
dimethyl-2,3′-bisindole (%) produced in the reaction at 1 and 16 bar O2. (c) Influence of the oxygen pressure (bar) on the turnover frequency
(h−1) in the formation of 1,1′-dimethyl-2,3′-bisindole. (d) Time profiles of the yield of 1,1′-dimethyl-2,3′-bisindole (%) at 1 bar O2 at different Pd-
loadings with 0.0015 M Pd(OAc)2 and 0.075, 0.10, 0.15, or 0.20 M 1-methylindole.
the 1-methylindole C−H activations is the rate-limiting step of
the overall catalytic reaction (see Supporting Information 3).
At the reduced oxygen pressure of 1 bar, a slightly lower
reaction rate was observed for initial reaction times, followed
by the deactivation of the catalyst. The effect of oxygen
pressure on the catalytic activity was further investigated
(Figure 2c). Remarkably, no significant decrease in catalytic
activity was observed upon lowering the oxygen pressure to 4
bar. These results show that the reoxidation of Pd0 is fast for
this range of oxygen pressures (4−16 bar) and does not limit
the overall reaction rate. This can be attributed to the use of
dimethylsulfoxide (DMSO) as the solvent: its coordinating
nature could stabilize the monomeric Pd0 species and thus
slow down the formation of inactive Pd0 aggregates.23
Previously, catalytic systems consisting of Pd(II)-complexes
in DMSO have also emerged as promising catalysts for aerobic
alcohol oxidation.24−26 However, at the atmospheric pressure,
the reoxidation step is indeed the rate-limiting step, and this
eventually leads to the formation of inactive Pd0 aggregates.
Nevertheless, over 90% product yield can also be obtained at
the atmospheric oxygen pressure using a Pd(OAc)2 loading of
2 mol % (Figure 2d).
Next, the scope of the indole C−H/C−H coupling reaction
was investigated (Figure 3). The results highlight the
importance of electron density in the indole five-membered
ring for the C−H/C−H coupling. This is consistent with the
electrophilic nature of the Pd-catalyzed C-H activation (see
mechanistic investigation). More in detail, while several indole
reactants bearing electron-donating groups on the N-atom,
such as methyl (6, 94% isolated yield), benzyl (7, 77%), and
methoxymethyl (8, 93%) could be used effectively, no coupling
products were observed for indoles possessing electron-
withdrawing N-substituents, such as 1-acetylindole and 1-
tosylindole. In the case of 1-phenylindole, a neutral substituent,
the reaction occurred at a lower rate (9, 27%). The same trend
was found for indoles substituted at their 5-position: the
electron-donating methoxy substituted compound reacted with
a considerable yield (10, 81%), whereas compounds with
electron-withdrawing substitutions required harsher reaction
conditions that lead to lower product yields. This was
exemplified by the cyano- (11, 36%), fluoro- (12, 29%), and
2437
https://dx.doi.org/10.1021/acscatal.0c04893
ACS Catal. 2021, 11, 2435−2444

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Figure 3. Scope of the C−H/C−H coupling reaction of indoles with molecular oxygen. The yields given correspond to isolated yields. a95 °C, 6
mol % Pd(OAc)2 + 450 mol % AcOH, 16 bar O2. b60°C, 2 mol % Pd(OAc)2 + 150 mol % AcOH, 16 bar O2. c Pd(NO3)2.2H2O (5 mol %) + o-
hydroxybenzoic acid (75 mol %), 3.5 h (41% conversion as determined via GC analysis).
bromo- (13, 38%) substituted compounds. However, the latter
substituent creates the opportunity for a variety of other
subsequent transformations (e.g., Suzuki coupling, Buchwald−
Hartwig amination) to build larger molecular structures.
Similar results were obtained in the homocoupling of 1-
methylindole-5-carboxylic acid (14, 34%). In the presence of a
strong electron-withdrawing nitro-substituent, the reaction led
to the formation of side products instead (see Supporting
Information 4). By further increasing the electron density of 1-
methylindole with two methoxy groups on both the 5- and 6-
positions, the reaction led to the formation of the
corresponding C2-C3’ coupling product under mild con-
ditions, but a decrease in the reaction rate was observed (15,
23%). Remarkably, 1-methyl-7-azaindole (16, 53%) could be
successfully coupled despite the presence of a coordinating,
pyridine-like N-heteroatom. These 7-azaindoles are highly
interesting heterocycles in medicinal chemistry.27 Moreover,
the reaction could be used successfully for the coupling of 5,6-
dihydro-4H-pyrrolo[3,2,1-ij]quinoline, which forms the basic
carbon skeleton of the lilolidine alkaloids (17, 76%). The
reaction was found to be ineffective in the coupling of other
heteroaromatics, such as benzofuran and benzothiophene. By
changing the Pd(II)-precursor and ligand in the reaction, the
C3-C3’ coupling isomer of 1-methylindole (18, 37%) could
also be obtained (see below).
electrophiles. Many classical organic reactions that involve
electrophilic substitutions (e.g., Friedel−Crafts acylation and
alkylation, halogenations) occur readily on the C3 position of
the indole reactant. This is explained by the formation of a
carbocation intermediate, which is facilitated via the
delocalization of the resulting positive charge on the adjacent
nitrogen atom. On the other hand, the C−H bond at the C2
position is preferred in deprotonation by nucleophiles. More
pronounced C2 selectivity has been observed in Pd-catalyzed
C−H activations that involve the deprotonation of the C−H
bond in the transition state, that is, via the concerted-
metalation deprotonation (CMD) mechanism with acetate
ligands.28,29 In the extreme case, the treatment of 1-
methylindole with superbases (e.g., n-butyllithium) in the
absence of transition metals gives rise to the deprotonation of
the C−H bond at the C2 position as well.30
While Pd(OAc)2 forms trimeric aggregates in the solid state
and in nonpolar solvents, DMSO tends to form monomeric
Pd(OAc)2 because of its coordinating and polar nature.31−36
To understand how the catalyst gives rise to the regioselectivity
experimentally observed for the C−H bond activation, the
mechanism was modeled with DFT calculations. For the latter
studies, the PBE0 hybrid functional was chosen, as suggested
in a benchmark study on Pd-based catalysts.37,38 Furthermore,
the calculations account for dispersion and solvent effects, as
well as for scalar relativistic effects (see Supporting Information
6).39−43 The calculated Gibbs free energy diagram shows that
the Pd(OAc)2 (I1) first performs C−H activation via a CMD
pathway (leading to TS1) (Figure 4). This step occurs
preferentially at the C2 position of the 1-methylindole reactant
because of the proximity of a basic acetate ligand that can
deprotonate the C−H bond in the transition state. By
accepting a proton, the acetate turns from a negatively charged
MECHANISTIC INVESTIGATION
■
Interestingly, the seemingly simple Pd-catalyst gives rise to the
selective C2-C3’ coupling of N-substituted indoles. This
implies that the catalytic cycle involves the sequential
activation of two different C−H bonds. On the one hand,
the C3 position of indoles is known to be favored by
2438
https://dx.doi.org/10.1021/acscatal.0c04893
ACS Catal. 2021, 11, 2435−2444

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Figure 4. Computed (PBE0-D3BJ/cc-pVTZ-DK) Gibbs free energy diagram for the Pd-catalyzed C−H/C−H coupling of 1-methylindole with C2-
C3’ selectivity. Gibbs free energies are given in kcal/mol, and relevant bond distances are given in Å.
bidentate ligand (I1) into a weakly coordinating AcOH moiety
(I2). The electrophilic activation of a second 1-methylindole
reactant occurs at the coordination site of the AcOH ligand
rather than dissociating the anionic acetate (TS2). These
results are in agreement with the computational experiments
This mechanistic pathway is further supported by several
kinetic isotope experiments (Figure 5a). A kinetic isotope
effect (KIE) of 2.4 and 1.0 was measured for 1-methylindole
deuterated at its C2 and C3 positions, respectively (see
Supporting Information 5). The first KIE of 2.4 is consistent
with the C−H activation at the C2 position via CMD.46 The
latter value of 1.0 observed for the C−H activation at the C3
position may be in agreement with the formation of a
carbocation intermediate as the rate-determining step because
this step does not involve the cleavage of a C−H bond.47,48
When AcOH is replaced with AcOD-d4 under the standard
reaction conditions, the negative effect of high acid
concentrations on the catalytic activity disappears (Figure
5b). The higher O−D bond strength of AcOD-d4 compared to
that of nondeuterated AcOH hinders its deprotonation.49 The
experiment indicates that this limits the protonolysis of the
Pd−C bond formed in the first C−H activation step and hence
counteracts the reverse reaction (from I2 to I1).
́
previously reported by Sedlak and co-workers, showing that
the presence of electron-donating carboxylate ligands disfavors
the formation of σ-complexes with Pd(II) in aromatic C−H
activation.44 TS2 involves a ligand-substitution reaction, where
the electron-rich indole acts as the nucleophilic species and
AcOH as the “leaving group”. This step exhibits the highest
activation energy in the overall catalytic cycle and is therefore
proposed to be the rate-determining step, which is consistent
with the first-order kinetics in 1-methylindole (at 16 bar of
oxygen pressure) and the increased reactivity of electron-rich
indoles observed in the substrate scope.
Because the carbocation formed upon the electrophilic
attack of Pd(II) at the C3 position is stabilized via resonance
between the N1 and C2 atoms, the carbocation intermediate
(I3) also explains the high selectivity for C-H activation at the
C3 position. The calculated geometry of I3 is consistent with
previous Pd(indolium) species isolated and characterized by
Yamamoto et al.45 Next, the aromaticity of the indole ring is
restored upon deprotonation in a fast, separate step (TS3),
leading to the formation of a bis(indolyl) complex (I4) that
can undergo reductive elimination to form the C2-C3’
coupling product.
LIGAND-CONTROLLED REGIOSELECTIVITY
■
The mechanistic investigation highlights the important role of
the acetate ligand. It is protonated in the first C−H activation
via CMD, followed by its dissociation upon the formation of a
carbocation intermediate. It is also expected that the ligands
exert an indirect effect on the C−H activation by altering the
electrophilicity of the Pd(II)-metal center. We envisaged that
the electronic properties (e.g., basicity) of the carboxylate
2439
https://dx.doi.org/10.1021/acscatal.0c04893
ACS Catal. 2021, 11, 2435−2444

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
as the catalyst, the benzoate ligand showed a good selectivity
(11:1) for the formation of the C2-C3’ isomer as major
coupling product, with the C3-C3’ isomer obtained as the only
side product.
The regioselectivity could be further increased with more
electron-rich benzoate ligands. In contrast, the presence of
electron-withdrawing groups on the benzoate ligand promoted
the formation of the C3-C3’ coupling product instead. These
results clearly show that the selectivity of the first C−H
activation via CMD can be altered by the choice of the
carboxylate ligand. More basic benzoate ligands favor the initial
C−H activation at the C2 position, whereas less-basic ligands
exert higher selectivity for the more electron-rich C3 position.
The second C−H activation step, however, occurs invariably at
the C3 position because the SEAr mechanism lacks such ligand-
assisted deprotonation in its rate-determining step.
A linear correlation between the regioselectivity of para-
substituted benzoate ligands and their respective Hammett
sigma constants was found. This confirms that the selectivity of
the C−H activation via CMD can be controlled by
simultaneously changing the basicity of the carboxylate ligand,
required for proton abstraction, and the electrophilicity of the
Pd(II) center for electrophilic metalation. This way, the
selectivity of the catalyst could be increased toward either the
more labile C−H bond at the C2 position or the more
electron-rich C−H bond at the C3 position. Nevertheless,
simultaneous metalation of the carbon atom also remains
crucial in the CMD mechanism. In the absence of assistance by
the metal center, only strong bases (e.g., n-BuLi) are capable of
deprotonating the C−H bond of 1-methylindole. It was
previously postulated by Fagnou et al. that the CMD
mechanism fills a “spectrum” between the C−H activation
mechanisms of electrophilic metalation and nucleophilic
deprotonation.50 This suggests that the respective contribu-
tions of metalation and deprotonation in the CMD transition
state may vary depending on the nature of the C−H bond and
by extension also on the electronic properties of the ligand and
metal catalyst. Recently, computational studies on a variety of
transition-metal catalysts were performed by Carrow and co-
workers that support this continuous character of the CMD
mechanism.51 The current results, however, confirm exper-
imentally that this concept can be used to obtain a high degree
of control over the regioselectivity of C−H activation via the
rational choice of the carboxylate ligand.
Figure 5. Mechanistic experiments on the C−H/C−H coupling of 1-
methylindole. (a) Time profiles of the reaction with 1-methylindole-
d1 (C2), 1-methylindole-d1 (C3), and nondeuterated 1-methylindole
as the reactant. Reaction conditions: 0.0015 M Pd(OAc)2, 0.15 M
(deuterated) 1-methylindole, 0.1125 M AcOD-d4, 40 °C, 16 bar O2.
(b) Influence of AcOH and AcOD-d4 ligand concentration (mol %)
on the turnover frequency (h−1). Reaction conditions: 0.0015 M
Pd(OAc)2, 0.15 M 1-methylindole, 40 °C, 3 h, 16 bar O2.
While for some ortho-substituted ligands, the selectivity
follows the Hammett relationship (see Supporting Information
7), o-hydroxybenzoic acid (salicylic acid) and o-methoxyben-
zoic acid give rise to the exceptional selectivities of 1:1 and
26:1 respectively. This is attributed to the secondary effects on
the electronic structure that are not explained by a simple
Hammett relationship. Intramolecular hydrogen bonding in o-
hydroxybenzoic acid is known to significantly increase the
acidity of the carboxylic acid group, lowering the selectivity for
C−H activation at C2.52 The electron-dense substituent in o-
methoxybenzoic acid is expected to exert an opposite effect,
leading to higher selectivities toward C−H activation at the C2
position.
ligands could be altered to influence the selectivity of the
reaction. A wide range of monosubstituted benzoate ligands
were tested, and the regioselectivity of the resulting Pd-catalyst
was examined (Figure 6a). The electronic properties of the
ligands were finely tuned in a controlled way using electron-
donating or electron-withdrawing groups, either in the ortho,
meta, or para-position. The relationship between the electronic
structure of the ligand and the catalytic performance was
unveiled in a Hammett study. We found this quantitative
investigation very useful to understand and predict the effect of
ligand modifications on the selectivity of the C−H activation
reaction.
We observed that the electronic properties of the ligands
indeed exert a strong influence on the regioselectivity of the
C−H activation (Figure 6b). The regioselectivity was
investigated by determining the ratio of the C2-C3’ and C3-
C3’ coupling products after the reaction time of 3 h, which
corresponds to the ratio of the initial reaction rates for C2-C3’
and C3-C3’ coupling. Similar to the reactions with Pd(OAc)2
In line with these results, the selectivity for the C3-C3’
coupling isomer could be further increased by weakly
coordinating ligands that render the Pd-complex even more
electrophilic, such as sulfate (SO42−), trifluoroacetate (TFA−)
−
or nitrate (NO3 ). According to the pKa of the respective
conjugate acids, the electrophilicity of the corresponding
2440
https://dx.doi.org/10.1021/acscatal.0c04893
ACS Catal. 2021, 11, 2435−2444

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Figure 6. Screening of aromatic carboxylate ligands and Pd-catalyst precursors for the regioselective C−H/C−H coupling of 1-methylindole. The
given error bars represent the calculated standard error (SI 7A). Reaction conditions: 0.0015 M Pd(OAc)2, 0.15 M 1-methylindole, 0.1125 M
carboxylic acid, 40 °C, 3 h, 16 bar O2, 1 mL DMSO. (a) Influence of the aromatic carboxylate ligand on the turnover frequency (h−1) and
regioselectivity (ratio 1,1′-dimethyl-2,3′-bisindole : 1,1′-dimethyl-3,3′-bisindole).*Mean value over three independent runs. (b) Plot of the
quantitative structure−activity relationship between the Hammett σ constant and the regioselectivity on a logarithmic scale. (c) Influence of
different Pd-catalyst precursors, in combination with AcOH or salicylic acid (75 mol %) ligand, on the regioselectivity. Reaction conditions: 0.0045
M Pd(NO3)2.2H2O, 0.15 M 1-methylindole, 0.1125 M AcOH or salicylic acid, 40 °C, 3 h, 16 bar O2, 1 mL of DMSO.
Pd(II) complexes increases in the order AcO− < SO4
<
also enable the cross-coupling of electronically biased indoles.
On the one hand, the SEAr mechanism prefers more electron-
rich indole reactants (e.g., 5,6-dimethoxy-1-methylindole). On
the other hand, the selectivity of the CMD step can be directed
toward the more labile C−H bonds in neutral (e.g., 1-
methylindole) or electron-poor indoles (e.g., 1-methyl-7-
azaindole) by increasing the basicity of the carboxylate ligand.
Cross-coupling between 1-methyl-5,6-dimethoxyindole and 1-
methylindole was achieved with ortho-methoxybenzoic acid as
the ligand, which showed the highest selectivity for the C2-C3’
coupling product in 1-methylindole homocoupling (Figure 7).
In contrast to the previous C−H/C−H cross-couplings, this
reaction does not require a large excess of one of the
(hetero)aromatic reactants.12−20 The product coupled at the
C2 position of 1-methylindole and at the C3’ position of 1-
methyl-5,6-dimethoxyindole was obtained as the major
product when the respective ratios of 2:1 or 1.5:1 were used
(see Supporting Information 8). Using the latter ratio of indole
reactants, the formation of cross-coupling products was favored
by the factors 1.4 and 1.3, in comparison to the respective
homocoupling of 5,6-dimethoxy-1-methylindole and 1-meth-
ylindole. The isomer (19) was obtained in 48% isolated yield
based on the 1-methyl-5,6-dimethoxyindole limiting reactant.
1-Benzylindole (20, 37%) and the electron-poor 1-benzyl-7-
2−
TFA− < NO3 . The trend in C3 selectivity is maintained when
these Pd-complexes are used as catalyst precursors (Figure 6c).
PdSO4 and Pd(TFA)2 led to an improved selectivity for the
C3-C3’ bisindole product. When combined with o-hydrox-
ybenzoic acid as the additional ligand, the selectivity could be
reversed toward C3-C3’ coupling. While a mixture of both the
C2-C3’ and C3-C3’ isomers was still obtained with these Pd-
sources, Pd(NO3)2.2H2O led to a complete inversion of the
selectivity toward the C3-C3’ isomer as the major product. The
effect of changing the basicity of the carboxylate ligand (from
AcOH to o-hydroxybenzoic acid) diminished in the presence
of more weakly coordinating anions, which indicates that the
electrophilic metalation becomes dominant, presumably due to
an open coordination sphere of Pd(II). This result can be
interpreted as an “electrophilic” CMD or eCMD, previously
defined by Carrow and co-workers.51
−
C−H/C−H CROSS-COUPLING OF INDOLES
■
So far, we have only considered how to control the
regioselectivity between the C2 and C3 positions in the
coupling of identical indole reactants. We envisaged that the
framework of two different C−H activation mechanisms could
2441
https://dx.doi.org/10.1021/acscatal.0c04893
ACS Catal. 2021, 11, 2435−2444

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Suzuki coupling.2 Three intermediate synthetic steps are
avoided in the direct C−H/C−H cross-coupling of indoles.
Moreover, H2O is produced as a byproduct instead of
stoichiometric metal waste. This opens a synthetic route
toward bioactive bisindole compounds with unparalleled atom-
and step-economy.
CONCLUSIONS
■
The selective Pd-catalyzed C−H/C−H cross-coupling of N-
substituted indoles with molecular oxygen was developed in
this work. The method strongly reduces the number of
reaction steps and waste generated in the synthesis of bisindole
scaffolds with interesting medicinal properties. Fundamental
insights into the selectivity of the indole C−H activation were
obtained with DFT calculations and kinetic isotope studies. A
Hammett investigation of the ligands allowed the prediction of
the regioselectivity for 1-methylindole C−H activation, which
enables the design of a specific, selective catalyst for an
envisaged synthetic application of C−H/C−H coupling.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acscatal.0c04893.
■
sı
Figure 7. Investigated scope of C-H/C-H indole cross-couplings
between electron-rich and electron-poor or neutral indoles. The yields
given correspond to isolated yields based on the 5,6-dimethoxy-1-
methylindole limiting reactant. aReaction conditions for 21: Pd-
(OAc)2 (10 mol %)*, ortho-methoxybenzoic acid (250 mol %), 0.38
M 1-benzyl-7-azaindole, 0.38 M 1-methyl-5,6-dimethoxyindole*, 1
mL DMSO*, 40 °C, 6 h, 16 bar O2. *Catalyst and reactant added
gradually to maintain three-fold excess of 1-benzyl-7-azaindole
concentration (see Supporting Information 8).
The Supporting Information is available free of charge
on the ACS Publications website.supporting_info.docx
(Experimental Section and Supporting Data).
AUTHOR INFORMATION
Corresponding Author
■
Dirk De Vos − Department of Microbial and Molecular
Systems, Centre for Membrane Separations, Adsorption,
Catalysis and Spectroscopy for Sustainable Solutions
(cMACS), KU Leuven, Leuven 3001, Belgium; orcid.org/
0000-0003-0490-9652; Email: dirk.devos@kuleuven.be
azaindole (21, 32%) could be effectively coupled with 5,6-
dimethoxy-1-methylindole at the expected C2 and C3’
positions, respectively. Moreover, a step-efficient benzyl
deprotection under aerobic conditions in DMSO may be
performed after the coupling reaction, upon the addition of
KOtBu as the strong base.53
This demonstrates the synthetic utility of the C−H/C−H
cross-coupling of indoles toward the synthesis of medicinal
bisindole compounds (Figure 8). The previously reported
route toward the insulin-like growth factor inhibitor (1)
involves the formation of the iodinated indole reactants, either
via direct electrophilic halogenation at the C3 position or
preliminary lithiation at the C2 position, followed by the
borylation of one indole reactant and the actual Pd-catalyzed
Authors
Igor Beckers − Department of Microbial and Molecular
Systems, Centre for Membrane Separations, Adsorption,
Catalysis and Spectroscopy for Sustainable Solutions
(cMACS), KU Leuven, Leuven 3001, Belgium; orcid.org/
0000-0002-4363-4893
Besir Krasniqi − Department of Microbial and Molecular
Systems, Centre for Membrane Separations, Adsorption,
Catalysis and Spectroscopy for Sustainable Solutions
(cMACS), KU Leuven, Leuven 3001, Belgium
Figure 8. Comparison of the conventional medicinal chemistry route toward a bisindole IGF inhibitor (adapted from Nemecek et al.2) with the
direct C-H/C-H coupling of indoles. (A) Iodination with I2, (B) borylation with B(OMe)3, (C) lithiation with BuLi followed by iodination with I2,
(D) Suzuki coupling with Pd(PPh3)4 in DMF, 110 °C. (E) Direct C−H/C−H cross-coupling of indoles developed in this work.
2442
https://dx.doi.org/10.1021/acscatal.0c04893
ACS Catal. 2021, 11, 2435−2444

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
(11) Kumar, K. S.; Meesa, S. R.; Naikawadi, P. K. Palladium-
Catalyzed [2+2+2] Annulation via Transformations of Multiple C-H
Bonds: One-Pot Synthesis of Diverse Indolo[3,2-a]carbazoles. Org.
Lett. 2018, 20, 6079−6083.
Prashant Kumar − Department of Chemistry, Quantum
Chemistry and Physical Chemistry Division, KU Leuven,
Leuven 3001, Belgium; orcid.org/0000-0002-0432-2790
Daniel Escudero − Department of Chemistry, Quantum
Chemistry and Physical Chemistry Division, KU Leuven,
Leuven 3001, Belgium; orcid.org/0000-0002-1777-8578
(12) Stuart, D. R.; Villemure, E.; Fagnou, K. Elements of
Regiocontrol in Palladium-Catalyzed Oxidative Arene Cross-Cou-
pling. J. Am. Chem. Soc. 2007, 129, 12072−12073.
(13) Potavathri, S.; Dumas, A. S.; Dwight, T. A.; Naumiec, G. R.;
Hammann, J. M.; DeBoef, B. Oxidant-Controlled Regioselectivity in
the Oxidative Arylation of N-acetylindoles. Tetrahedron Lett. 2008, 49,
4050−4053.
(14) Stuart, D. R.; Fagnou, K. The Catalytic Cross-Coupling of
Unactivated Arenes. Science 2007, 316, 1172−1175.
Complete contact information is available at:
https://pubs.acs.org/10.1021/acscatal.0c04893
Notes
The authors declare no competing financial interest.
(15) Campbell, A. N.; Meyer, E. B.; Stahl, S. S. Regiocontrolled
aerobic oxidative coupling of indoles and benzene using Pd catalysts
with 4,5-diazafluorene ligands. Chem. Commun. 2011, 47, 10257−
10259.
(16) Li, R.; Jiang, L.; Lu, W. Intermolecular Cross-Coupling of
Simple Arenes via C−H Activation by Tuning Concentrations of
Arenes and TFA. Organometallics 2006, 25, 5973−5975.
(17) Dwight, T. A.; Rue, N. R.; Charyk, D.; Josselyn, R.; DeBoef, B.
C−C Bond Formation via Double C−H Functionalization: Aerobic
Oxidative Coupling as a Method for Synthesizing Heterocoupled
Biaryls. Org. Lett. 2007, 9, 3137−3139.
(18) Hull, K. L.; Sanford, M. S. Catalytic and Highly Regioselective
Cross-Coupling of Aromatic C−H Substrates. J. Am. Chem. Soc. 2007,
129, 11904−11905.
(19) Brasche, G.; García-Fortanet, J.; Buchwald, S. L. Twofold C−H
Functionalization: Palladium-Catalyzed Ortho Arylation of Anilides.
Org. Lett. 2008, 10, 2207−2210.
(20) Malakar, C. C.; Schmidt, D.; Conrad, J.; Beifuss, U. Double C−
H Activation: The Palladium-Catalyzed Direct C-Arylation of
Xanthines with Arenes. Org. Lett. 2011, 13, 1378−1381.
(21) Beckers, I.; Henrion, M.; De Vos, D. E. The Dual Effect of the
Acetate Ligand on the Mechanism of the Pd-Catalyzed C-H/C-H
Coupling of Benzene. ChemCatChem 2019, 12, 90−94.
(22) Wang, D.; Izawa, Y.; Stahl, S. S. Pd-Catalyzed Aerobic
Oxidative Coupling of Arenes: Evidence for Transmetalation between
Two Pd(II)-Aryl Intermediates. J. Am. Chem. Soc. 2014, 136, 9914−
9917.
(23) Wang, D.; Weinstein, A. B.; White, P. B.; Stahl, S. S. Ligand-
Promoted Palladium-Catalyzed Aerobic Oxidation Reactions. Chem.
Rev. 2018, 118, 2636−2679.
(24) Peterson, K. P.; Larock, R. C. Palladium-Catalyzed Oxidation of
Primary and Secondary Allylic and Benzylic Alcohols. J. Org. Chem.
1998, 63, 3185−3189.
(25) Steinhoff, B. A.; Fix, S. R.; Stahl, S. S. Mechanistic Study of
Alcohol Oxidation by the Pd(OAc)2/O2/DMSO Catalyst System and
Implications for the Development of Improved Aerobic Oxidation
Catalysts. J. Am. Chem. Soc. 2002, 124, 766−767.
(26) Steinhoff, B. A.; Stahl, S. S. Mechanism of Pd(OAc)2/DMSO-
Catalyzed Aerobic Alcohol Oxidation: Mass-Transfer-Limitation
Effects and Catalyst Decomposition Pathways. J. Am. Chem. Soc.
2006, 128, 4348−4355.
ACKNOWLEDGMENTS
■
The research leading to these results has received funding from
the NMBP-01-2016 Program of the European Union’s
Horizon 2020 Framework Program H2020/2014-2020/
under grant agreement no [720996]. DDV is grateful to KU
Leuven for support in the frame of the CASAS Metusalem
project and to Research Foundation Flanders (FWO) for
project funding (G0781118 N, G0D0518 N). The computa-
tional resources and services used in this work were provided
by the VSC (Flemish Supercomputer Center), funded by the
FWO and the Flemish Government − department EWI. We
́
also thank Susana García Abbellan for assistance with the
experimental work.
REFERENCES
■
(1) Kaushik, N. K.; Kaushik, N.; Attri, P.; Kumar, N.; Kim, C. H.;
Verma, A. K.; Choi, E. H. Biomedical Importance of Indoles.
Molecules 2013, 18, 6620−6662.
(2) Nemecek, C.; Metz, W. A.; Wentzier, S.; Ding, F.-X.; Venot, C.;
Souaille, C.; Dagallier, A.; Maignan, S.; Guilloteau, J.-P.; Bernard, F.;
Henry, A.; Grapinet, S.; Lesuisse, D. Design of Potent IGF1-R
Inhibitors Related to Bis-azaindoles. Chem. Biol. Drug Des. 2010, 76,
100−106.
(3) Munuganti, R. S. N.; Hassona, M. D. H.; Leblanc, E.; Frewin, K.;
Singh, K.; Ma, D.; Ban, F.; Hsing, M.; Adomat, H.; Lallous, N.; Andre,
C.; Jonadass, J. P. S.; Zoubeidi, A.; Young, R. N.; Guns, E. T.; Rennie,
P. S.; Cherkasov, A. Identification of a Potent Antiandrogen that
Targets the BF3 Site of the Androgen Receptor and Inhibits
Enzalutamide-Resistant Prostate Cancer. Chem. Biol. 2014, 21,
1476−1485.
(4) Neochoritis, C. G.; Wang, K.; Estrada-Ortiz, N.; Herdtweck, E.;
Kubica, K.; Twarda, A.; Zak, K. M.; Holak, T. A.; Domling, A. 2,3’-
̈
Bis(1’H-indole) heterocycles: New p53/MDM2/MDMX antagonists.
Bioorg. Med. Chem. Lett. 2015, 25, 5661−5666.
(5) Norton, R. S.; Wells, R. J. A series of chiral polybrominated
biindoles from the marine blue-green alga Rivularia firma. J. Am.
Chem. Soc. 1982, 104, 3628−3635.
(6) Mee, S. P. H.; Lee, V.; Baldwin, J. E.; Cowley, A. Total synthesis
of 5,5′,6,6′-tetrahydroxy-3,3′-biindolyl, the proposed structure of a
potent antioxidant found in beetroot (Beta vulgaris). Tetrahedron
2004, 60, 3695−3712.
(27) Sharma, N.; Anurag. 7-Azaindole Analogues as Bioactive Agents
and Recent Results. Mini-Rev. Med. Chem. 2019, 19, 727−736.
(28) Potavathri, S.; Pereira, K. C.; Gorelsky, A. P.; Pike, A.; LeBris,
A. P.; DeBoef, B. Regioselective Oxidative Arylation of Indoles
Bearing N-alkyl Protecting Groups: Dual C-H Functionalization via a
Concerted Metalation-Deprotonation Mechanism. J. Am. Chem. Soc.
2010, 132, 14676−14681.
(7) Cacchi, S.; Fabrizi, G. Synthesis and Functionalization of Indoles
Through Palladium-catalyzed Reactions. Chem. Rev. 2005, 105,
2873−2920.
(8) Robinson, B. The Fischer Indole Synthesis. Chem. Rev. 1963, 63,
373−401.
(29) Ackermann, L. Carboxylate-Assisted Transition-Metal-Cata-
lyzed C−H Bond Functionalizations: Mechanism and Scope. Chem.
Rev. 2011, 111, 1315−1345.
(9) Liang, Z.; Zhao, J.; Zhang, Y. Palladium-Catalyzed Regioselective
Oxidative Coupling of Indoles and One-Pot Synthesis of Acetoxylated
Biindolyls. J. Org. Chem. 2010, 75, 170−177.
(30) Sundberg, R. J.; Russell, H. F. Syntheses with N-Protected 2-
Lithioindoles. J. Org. Chem. 1973, 38, 3324−3330.
(10) Li, Y.; Wang, W.-H.; Yang, S.-D.; Li, B.-J.; Shi, Z.-J. Oxidative
dimerization of N-protected and free indole derivatives toward 3,3′-
biindoles via Pd-catalyzed direct C−H transformations. Chem.
Commun. 2010, 46, 4553−4555.
(31) Stephenson, T. A.; Morehouse, S. M.; Powell, A. R.; Heffer, J.
P.; Wilkinson, G. Carboxylates of palladium, platinum, and rhodium,
and their adducts. J. Chem. Soc. 1965, 0, 3632−3640.
2443
https://dx.doi.org/10.1021/acscatal.0c04893
ACS Catal. 2021, 11, 2435−2444

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
(32) Evans, J.; O’Neill, L.; Kambhampati, V. L.; Rayner, G.; Turin,
S.; Genge, A.; Dent, A. J.; Neisius, T. Structural characterisation of
solution species implicated in the palladium-catalyzed Heck reaction
by Pd K-edge X-ray absorption spectroscopy: palladium acetate as a
catalyst precursor. J. Chem. Soc., Dalton Trans. 2002, 10, 2207−2212.
(33) Adrio, L. A.; Nguyen, B. N.; Guilera, G.; Livingston, A. G.; Hii,
K. K. (. M.). Speciation of Pd(OAc)2 in ligandless Suzuki−Miyaura
reactions. Catal. Sci. Technol. 2012, 2, 316−323.
(52) Dunn, G. E.; Penner, T. L. Effect of Intramolecular Hydrogen
Bonding on the Relative Acidities of Substituted Salicylic Acids in
Benzene Solution. Can. J. Chem. 1967, 45, 1699−1706.
(53) Haddach, A. A.; Kelleman, A.; Deaton-Rewolinski, M. V. An
Efficient Method for the N-debenzylation of Aromatic Heterocycles.
Tetrahedron Lett. 2002, 43, 399−402.
(34) Carole, W. A.; Colacot, T. J. Understanding Palladium Acetate
from a User Perspective. Chem. − Eur. J. 2016, 22, 7686−7695.
(35) Diao, T.; White, P.; Guzei, I.; Stahl, S. S. Characterization of
DMSO Coordination to Palladium(II) in Solution and Insights into
the Aerobic Oxidation Catalyst, Pd(DMSO)2(TFA)2. Inorg. Chem.
2012, 51, 11898−11909.
́
̌
́
(36) Vana, J.; Hanusek, J.; Sedlak, M. Bi and trinuclear complexes in
palladium carboxylate-assisted C−H activation reactions. Dalton
Trans. 2018, 47, 1378.
(37) Steinmetz, M.; Grimme, S. Benchmark study of the
performance of density functional theory for bond activations with
(Ni, Pd)-based transition-metal catalysts. ChemistryOpen 2013, 2,
115−124.
(38) Adamo, C.; Barone, V. Toward reliable density functional
methods without adjustable parameters: The PBE0 model. J. Chem.
Phys. 1999, 110, 6158−6170.
(39) Grimme, S.; Ehrlich, S.; Goerigk, L. Effect of the damping
function in dispersion corrected density functional theory. J. Comput.
Chem. 2011, 32, 1456−1465.
(40) Marenich, A. V.; Cramer, C. J.; Truhlar, D. G. Universal
solvation model based on solute electron density and on a continuum
model of the solvent defined by the bulk dielectric constant and
atomic surface tensions. J. Phys. Chem. B. 2009, 113, 6378−6396.
(41) Pritchard, B. P.; Altarawy, D.; Didier, B.; Gibson, T. D.;
Windus, T. L. New basis set exchange: An open, up-to-date resource
for the molecular sciences community. J. Chem. Inf. Model. 2019, 59,
4814−4820.
̀
(42) Pérez−García, P. M.; Daru, A.; Scheerder, A. R.; Lutz, M.;
Harvey, J. N.; Moret, M.-E. Oxidative Addition of Aryl Halides to a
Triphosphine Ni (0) Center to Form Pentacoordinate Ni (II) Aryl
Species. Organometallics 2020, 39, 1139−1144.
(43) Reiher, M. Douglas−Kroll−Hess Theory: a relativistic
electrons-only theory for chemistry. Theor. Chem. Acc. 2006, 116,
241−252.
́
́
́
́
(44) Vana, J.; Bartacek, J.; Hanusek, J.; Roithova, J.; Sedlak, M. C.-H.
Functionalizations by Palladium Carboxylates: The Acid Effect. J. Org.
Chem. 2019, 84, 12746−12754.
(45) Yamamoto, K.; Kimura, S.; Murahashi, T. σ-π Continuum in
Indole-Palladium(II) Complexes. Angew. Chem., Int. Ed. 2016, 55,
5322−5326.
(46) Shilov, A. E.; Shul’pin, G. B. Activation of C−H Bonds by
Metal Complexes. Chem. Rev. 1997, 97, 2879−2932.
(47) Simmons, E. M.; Hartwig, J. F. On the Interpretation of
Deuterium Kinetic Isotope Effects in C-H Bond Functionalizations by
Transition-Metal Complexes. Angew. Chem., Int. Ed. 2012, 51, 3066−
3072.
(48) Melander, L. Introduction of Substituents in the Aromatic
Nucleus. Exploration of its Mechanism by means of Isotopic
Hydrogen. Acta Chem. Scand. 1949, 3, 95−96.
(49) McConnell, R. M.; Godwin, W. E.; Stanley, B.; Green, M. S.
Acidity Studies of Deuterated Acids and Bases Commonly Used as
Buffers in NMR Studies. J. Ark. Acad. Sci. 1997, 51, 135−140.
(50) Gorelsky, S. I.; Lapointe, D.; Fagnou, K. Analysis of the
Concerted Metalation-Deprotonation Mechanism in Palladium-
Catalyzed Direct Arylation Across a Broad Range of Aromatic
Substrates. J. Am. Chem. Soc. 2008, 130, 10848−10849.
(51) Wang, L.; Carrow, B. P. Oligothiophene Synthesis by a General
C−H Activation Mechanism: Electrophilic Concerted Metalation−
Deprotonation (eCMD). ACS Catal. 2019, 9, 6821−6836.
2444
https://dx.doi.org/10.1021/acscatal.0c04893
ACS Catal. 2021, 11, 2435−2444