Direct palladium-catalyzed C-2 and C-3 arylation of indoles: A mechanistic rationale for regioselectivity

Article abstract of DOI:10.1021/ja043273t

We have recently developed palladium-catalyzed methods for direct arylation of indoles (and other azoles) wherein high C-2 selectivity was observed for both free (NH)-indole and (NR)-indole. To provide a rationale for the observed selectivity ("nonelectrophilic" regioselectivity), mechanistic studies were conducted, using the phenylation of 1-methylindole as a model system. The reaction order was determined for iodobenzene (zero order), indole (first order), and the catalyst (first order). These kinetic studies, together with the Hammett plot, provided a strong support for the electrophilic palladation pathway. In addition, the kinetic isotope effect (KIE^H/D) was determined for both C-2 and C-3 positions. A surprisingly large value of 1.6 was found for the C-3 position where the substitution does not occur (secondary KIE), while a smaller value of 1.2 was found at C-2 (apparent primary KIE). On the basis of these findings, a mechanistic interpretation is presented that features an electrophilic palladation of indole, accompanied by a 1,2-migration of an intermediate palladium species. This paradigm was used to design new catalytic conditions for the C-3 arylation of indole. In case of free (NH)-indole, regioselectivity of the arylation reaction (C-2 versus C-3) was achieved by the choice of magnesium base.

Full text of DOI:10.1021/ja043273t

Published on Web 05/14/2005
Direct Palladium-Catalyzed C-2 and C-3 Arylation of Indoles:
A Mechanistic Rationale for Regioselectivity
Benjamin S. Lane, Meghann A. Brown, and Dalibor Sames*
Contribution from the Department of Chemistry, Columbia UniVersity, 3000 Broadway,
New York, New York 10027
Received November 8, 2004; E-mail: sames@chem.columbia.edu
Abstract: We have recently developed palladium-catalyzed methods for direct arylation of indoles (and
other azoles) wherein high C-2 selectivity was observed for both free (NH)-indole and (NR)-indole. To
provide a rationale for the observed selectivity (“nonelectrophilic” regioselectivity), mechanistic studies were
conducted, using the phenylation of 1-methylindole as a model system. The reaction order was determined
for iodobenzene (zero order), indole (first order), and the catalyst (first order). These kinetic studies, together
with the Hammett plot, provided a strong support for the electrophilic palladation pathway. In addition, the
kinetic isotope effect (KIEH/D) was determined for both C-2 and C-3 positions. A surprisingly large value of
1.6 was found for the C-3 position where the substitution does not occur (secondary KIE), while a smaller
value of 1.2 was found at C-2 (apparent primary KIE). On the basis of these findings, a mechanistic
interpretation is presented that features an electrophilic palladation of indole, accompanied by a 1,2-migration
of an intermediate palladium species. This paradigm was used to design new catalytic conditions for the
C-3 arylation of indole. In case of free (NH)-indole, regioselectivity of the arylation reaction (C-2 versus
C-3) was achieved by the choice of magnesium base.
Introduction
Direct C-H Functionalization in Heteroarenes: Selective
C-Arylation of Indole. Azoles represent important structural
units frequently found in natural products, pharmaceuticals, and
other synthetics. Direct arylation of heteroarenes, achieved via
cross-coupling of (sp2) C-H bonds and haloarenes,1 offers an
attractive alternative to standard cross-coupling methods which
require the establishment of a reactive functionality prior to C-C
coupling (i.e., halogenation, metalation).2 It has been our aim
to develop parallel methodologies capable of furnishing regio-
isomeric products from the same substrate, governed by the
catalytic system employed (Figure 1).3 In this regard, indole
represents a system of particular interest and importance. The
C-2 and C-3 selective methods would be valuable and comple-
mentary to the known N-arylation methodology,4 enabling the
direct access to a series of indole derivatives (Figure 1). In this
paper, we describe the experimental studies that not only
provided a mechanistic rationale for the C-2 arylation reaction
but also led to the development of new conditions for C-3
selective arylation of free (NH)-indoles.
Figure 1. Direct and systematic arylation of indole. Regioselectivity control.
C2-Arylation of Indoles: Effect of N-Substitution. We
have previously reported selective C-arylation of free (NH)-
azole heteroarenes (Scheme 1).3 This method was also effective
for indole substrates, affording C-2 arylation products. The
pivotal step in this process involves the in situ formation of an
indole magnesium salt (N-MgX) in the presence of MgO, which
not only protects the amine functionality but also increases the
nucleophilicity of the azole ring (care must be taken with regard
to the purity of indole and anhydrous conditions).5 This new
C-arylation protocol (Pd/Ph3P/MgO) was highly regioselective;
however, the observed trends were not uniform. Pyrrole and
imidazole afforded products consistent with an electrophilic
substitution mechanism,6,7 while indole provided products
(1) Miura, M.; Nomura, M. Top. Curr. Chem. 2002, 219, 211-41 and
references therein.
(2) Labadie, S. S.; Teng, E. J. Org. Chem. 1994, 59, 4250-4.
(3) (a) Sezen, B.; Sames, D. J. Am. Chem. Soc. 2003, 125, 10580-10585. (b)
Sezen, B.; Sames, D. J. Am. Chem. Soc. 2003, 125, 5274-5275.
(4) (a) Hartwig, J. F.; Kawatsura, M.; Hauck, S. I.; Shaughnessy, K. H.; Alcazar-
Roman, L. M. J. Org. Chem. 1999, 64, 5575-5580. (b) Old, D. W.; Harris,
M. C.; Buchwald, S. L. Org. Lett. 2000, 2, 1403-1406. (c) Klapars, A.;
Antilla, J. C.; Huang, X.; Buchwald, S. L. J. Am. Chem. Soc. 2001, 123,
7727-7729. (d) Grasa, G. A.; Viciu, M. S.; Huang, J.; Nolan, S. P. J.
Org. Chem. 2001, 66, 7729-7737. (e) Lam, P. Y. S.; Clark, C. G.; Saubern,
S.; Adams, J.; Winters, M. P.; Chan, D. M. T.; Combs, A. Tetrahedron
Lett. 1998, 39, 2941-2944.
(5) See Supporting Information for a detailed and updated experimental
procedure.
9
8050
J. AM. CHEM. SOC. 2005, 127, 8050-8057
10.1021/ja043273t CCC: $30.25 © 2005 American Chemical Society

C-2 and C-3 Arylation of Indoles
A R T I C L E S
Scheme 1. Selective C-2 Arylation of (NH)-Indole via in Situ
Table 1. . C-2 Arylation of (NR)-Indoles (Substitution on Nitrogen)
Formation of Indole Magnesium Salts
entry
R
yield %a
entry
R
yield %a
1
2
3
CH3
Bn
Pr
88 (54)b
81
92
4
5
6
Ph
68
55
0
p-(CN)-C6H4
SO2Ph, SO2CH3, COCH3
a All values based on isolated yields. b Chlorobenzene used as donor and
dicyclohexylphenylphosphine as a ligand at 150 °C.
Scheme 2. The Key Reaction Pathways in the C-2 Arylation of
1-Methylindole
compatible with a broad array of functional groups.8 The effect
of N-substitution is pertinent to the discussion below and
deserves a note. N-Alkyl substituents including larger groups
such as iso-propyl or benzyl as well as N-aryl groups were well
tolerated, giving high yields of the corresponding 2-phenyl
products (Table 1). In contrast, N-acetyl- and N-methansulfonyl-
indole were completely inert, clearly indicating the need for
high electron density in the azole-ring. In this paper, we provide
a new mechanistic insight for the productive pathway, the
reaction between the aryl-palladium(II) intermediate and indole
(Scheme 2).
stemming from phenylation at the R-position to the nitrogen
atom (“nonelectrophilic” regiochemistry, Scheme 1). To inves-
tigate the basis of the observed regioselectivity, we first
examined whether the magnesium metal plays an important role
in determining the regiochemical outcome. For this purpose,
N-methylindole was selected as a model substrate.
Results and Discussion
The Reaction Order of Iodobenzene, Indole, and the
Catalyst. According to Scheme 2, the first step of the catalytic
cycle involves formation of an aryl-palladium(II) intermediate
via the oxidative addition of iodobenzene to a Pd(0) species.
The reaction rate for the phenylation of 1-methylindole was
measured over a range of iodobenzene concentrations (Sup-
porting Information). These kinetic experiments determined that
the reaction is zero order in iodobenzene, indicating that the
rate-determining step occurs after the oxidative addition, most
likely within the indole-functionalization sequence. Indeed, the
reaction was first order in both indole and catalyst (see
Supporting Information). This represents a favorable scenario
as kinetic isotope experiments may shed more light on the
mechanism of the C-H bond functionalization.
Alternative Mechanistic Pathways. There are three reaction
mechanisms that may rationalize the strong preference for C-2
arylation of indole: (1) the electrophilic metalation-migration,
(2) nonelectrophilic metalation of the 2-position, and (3) carbo-
metalation, that is, Heck-type reaction (Scheme 3). The elec-
trophilic substitution of the indole ring is well established, and
a strong preference for the 3-position is known.9 Thus, a C3 f
C2 migration (1,2-migration) of palladium has to take place if
this mechanism is operative. Alternatively, the direct C2-
palladation of indole via a nonelectrophilic pathway (e.g., via
σ-bond metathesis) would also explain the observed regiose-
lectivity. C-2 palladation of indole has been reported; however,
the presence of a strong directing group was required [cf.,
palladation of 1-(2′-pyridyl)-indole].10 The carbo-metalation or
the Heck-type reaction is also feasible; however, this pathway
requires anti-dehydropalladation (or anti-â-hydride elimina-
tion).11 To gain deeper insight into the mechanism of this
This new 1-methylindole substrate required a full-scale
optimization effort, which led to addressing two key issues: (1)
the choice of base, and (2) formation of the biphenyl side
product.8 The first issue was solved when we found CsOAc to
be an effective base, while MgO was ineffective, lending further
support for its role in the free indole arylation. The second
problem, the formation of biphenyl, was marginalized by
decreasing the palladium catalyst loading. The entire reaction
system can be described by a qualitative kinetic model shown
in Scheme 2. The first step, oxidative addition between
palladium(0) and the aryl-halide, proceeds to an aryl-palladium-
(II) intermediate which is partitioned between two major
competing pathways: (1) cross-coupling with the substrate to
furnish the desired product or (2) formation of byproduct
biphenyl. Despite its simplicity, this model suggested that the
formation of the major side product, biphenyl, may be dimin-
ished by simply decreasing the amount of catalyst.8 This
favorable adjustment was inspired by an informed assumption
that the biphenyl formation proceeds via a bimolecular trans-
metalation of the aryl-palladium species (Scheme 2).
Importantly, both the indole magnesium salt (Scheme 1) and
the 1-alkylindole (Table 1) showed a strong preference for C-2
arylation, validating the latter as a suitable model substrate for
the subsequent mechanistic studies. The resulting robust protocol
required low catalyst loading (<0.5 mol % of Pd) and was
(6) Palladium-catalyzed arylation of other electron-sufficient heteroarenes
afforded the “electrophilic regioselectivity”. (a) Ohta, A.; Akita, A.;
Ohkuwa, T.; Chiba, M.; Funkunaga, R.; Miyafuji, A.; Nakata, T.; Tani,
N.; Aoyagi, Y. Heterocycles 1990, 31, 1951-1958. (b) Sommai, P.-A.;
Tetsuya, S.; Yoshiki, K.; Masahiro, M.; Masakatsu, N. Bull. Chem. Soc.
Jpn. 1998, 71, 467-473. (c) Park, C.-H.; Ryabova, V.; Seregin, I. V.;
Sromek, A. W.; Gevorgyan, V. Org. Lett. 2004, 6, 1159-1162. (d) Li,
W.; Nelson, D. P.; Jensen, M. S.; Hoerrner, R. S.; Javadi, G. J.; Cai, D.;
Larsen, R. D. Org. Lett. 2003, 5, 4835-4837.
(9) Jackson, A. H.; Lynch, P. P. J. Chem. Soc., Perkin Trans. 2 1987, 1215-
1219.
(10) (a) Tollari, S.; Demartin, F.; Cenini, S.; Palmisano, G.; Raimondi, P. J.
Organomet. Chem. 1997, 527, 93-102. (b) Nonoyama, M.; Nakajima, K.
Polyhedron 1998, 18, 533-543.
(7) Cobalt-catalyzed arylation of thiazole also furnished products consistent
with the “electrophilic regiochemistry”: Sezen, B.; Sames, D. Org. Lett.
2003, 5, 3607-3610.
(8) Lane, B. S.; Sames, D. Org. Lett. 2004, 6, 2897-2900.
9
J. AM. CHEM. SOC. VOL. 127, NO. 22, 2005 8051

A R T I C L E S
Lane et al.
Scheme 3. Possible Reaction Mechanisms in the Arylation of N-Alkylindoles
Scheme 4. Kinetic Isotope Effect at the C-2 and C-3 Positions of
Table 2. Confirmation of the KIE by Competitive Substrate
Studiesa
1-Methylindolea
a Values determined from three kinetic trials; time points were averaged,
and the rate constant was derived from a first-order plot (see Supporting
Information).
transformation, the kinetic isotope effect was measured for the
2- and 3-positions of indole.
The Kinetic Isotope Effect at the 2- and 3-Positions of
Indole: Kinetic Studies. Both 3- and 2-deutero-1-methylindole
were synthesized according to literature procedures,12 and the
kinetic constants derived for these substrates were the average
of three experiments conducted under pseudo-first-order condi-
tions in substrate (Supporting Information). As usual, these
values were compared to those for nondeuterated materials and
were used to calculate the kinetic isotope effect (Scheme 4).
The KIE for 2-deutero-1-methylindole was 1.2, a value too small
for the cleavage of this bond to be involved in the rate-limiting
step. As a confirmation, the loss of deuterium in the starting
material was examined and found significantly slower in
comparison to the arylation reaction.
a The product ratios were determined by HPLC versus an internal
standard. Values reported are the average of three trials. The KIE value
used to calculate the expected product ratios was taken from Scheme 4.
Similar results were obtained with 1-methylindole and 1-n-octylindole, as
well as 1-methylindole and 1-n-propylindole (see Supporting Information).
Scheme 5. Generic Electrophilic Substitution Mechanism
addition, no deuterium scrambling was detected via NMR.
Interestingly, the larger KIE value was obtained for the
3-position where the substitution does not occur and thereby
represents the secondary KIE.
The Kinetic Isotope Effect at the 3-Position of Indole:
Confirmation by Competitive Reaction Studies. In light of
the unexpected results described above, we decided to verify
the KIE value at position C-3 by an alternative means,
employing the method of competing reactions. According to
this technique, two closely related substrate derivatives are
submitted to the reaction conditions of interest, in the same flask,
and the ratio of products is determined. The KIE(H/D) is then
derived from the effect of deuterium labeling, in one of these
Perhaps more surprising was the KIE of 1.6 obtained for
3-deutero-1-methylindole (Scheme 4). Importantly, we con-
firmed that the deuterium loss in the starting material was
significantly slower in comparison to the arylation reaction;
1-methyl-2-phenyl-3-deutero-indole was the main product. In
(11) (a) Glover, B.; Harvey, K. A.; Sharp, M. J.; Tymoschenko, M. F. Org.
Lett. 2003, 5, 301-304. (b) Lautens, M.; Fang, Y.-Q. Org. Lett. 2003, 5,
3679-3682. (c) Ikeda, M.; El Bialy, S. A. A.; Yakura, T. Heterocycles
1999, 51, 1957-1970.
(12) (a) Daunis, J.; Soufiaoui, M.; Laude, B. Org. Mass Spectom. 1979, 14,
121-125. (b) Sundberg, R. J.; Russell, H. F. J. Org. Chem. 1973, 38, 3324-
3330.
9
8052 J. AM. CHEM. SOC. VOL. 127, NO. 22, 2005

C-2 and C-3 Arylation of Indoles
A R T I C L E S
Scheme 6. Electrophilic Pathway for Arylation of Indole
substrates, on the observed ratio of products. The key advantage
of this method is that it does not require quantification of rate
constants and thus provides an independent confirmation of the
KIE. Furthermore, the reactions are conducted under optimized
conditions, that is, those used in a typical preparative reaction,
as opposed to pseudo-first-order conditions.
carbopalladation mechanism or the Heck-type mechanism
followed by anti-dehydropalladation.15 Carbopalladation of
indole, followed by isomerization and syn-â-hydride elimination,
should also be considered (not shown). A plausible mechanism
for the isomerization step is a reversible R-hydride elimination,
proceeding via a palladium hydride-carbene intermediate.
Although this pathway is known for other transition metals (i.e.,
Pt, Ru),16 it has not been observed with palladium; moreover,
a large kinetic isotope effect is usually associated with this
process (KIE > 4).17 The most likely candidate, however, seems
to be the electrophilic palladation pathway, supported by the
kinetic studies and the Hammett plot (see below). The substrate
specificity also provides circumstantial evidence for the elec-
trophilic mechanism (e.g., 1-acetylindole was completely inert
under the reaction conditions, demonstrating the need for the
sufficient electron density in the indole ring, Table 1, entry 6).
The electrophilic pathway may in fact be reconciled with the
obtained data if a migration of palladium (C3 f C2) is
considered (Schemes 3 and 6). The electrophilic substitution is
often considered to be a two-step process consisting of the
reversible electrophile-arene complex formation and subsequent
deprotonation (Scheme 5). The occurrence and magnitude of
the KIE is related to the relative ratio of the individual rate
constants (k1, k-1, and k2) and varies widely (0-6).14,18 In most
cases, the deprotonation is much faster that the formation of
the electrophile-arene complex (k2 > k1), and, as a conse-
quence, in many electrophilic substitution reactions, like nitra-
tion, no KIE is observed.19 However, if k-1 is large as compared
to k2, a steady-state approximation for the concentration of the
electrophile-arene complex can be made; that is, its concentra-
In this case, two indole derivatives, 1-methylindole and
1-ethylindole, were chosen, and the inherent reactivity difference
between these substrates was determined under the arylation
conditions (1.30 ( 0.004 for 3/4 ratio, Table 2). Subsequently,
two sets of experiments were performed wherein one of the
two substrates was replaced with a 3-deutero-analogue. For
example, in experiment 2 (Table 2), the two substrates were
1-methyl-3-deutero-indole and 1-ethylindole. Considering both
the intrinsic reactivity difference between 1-methyl- and 1-ethyl-
indole (exp. 1, Table 2) and the KIE determined previously,
the calculated product ratio 3/4 was 0.81. In simple terms, the
higher intrinsic reactivity of 1-methylindole should be overrid-
den by the KIE, and consequently 1-ethylindole should be more
reactive. This proved to be the case, and the observed ratio 0.92
( 0.2 was in a good agreement with the calculated value. A
good agreement was also found with the reversed labeling
pattern (exp. 3, Table 2). Finally, as a control, the intrinsic
reactivity ratio was restored when both substrates were labeled.
Although this method does not measure the KIE as accurately
as the kinetic experiments, it does confirm the presence and
approximate magnitude of the KIE. The experiments described
above substantiate the large secondary KIE at the 3-position of
indole (1.6) in the palladium-catalyzed arylation reaction. It is
important to reiterate that the competition experiments are
conducted under reaction conditions used in a normal preparative
reaction, and not the pseudo-first-order ones used in the kinetic
measurements.
(13) (a) Jones, W. D. Acc. Chem. Res. 2003, 36, 140-146. (b) Lloyd-Jones, G.
C.; Slatford, P. A. J. Am. Chem. Soc. 2004, 126, 2690-2691.
(14) Zollinger, H. AdV. Phys. Org. Chem. 1964, 2, 163-200.
(15) Large KIE have been observed for the anti-dehydropalladation. (a) Takacs,
J. M.; Lawson, E. C.; Clement, F. J. Am. Chem. Soc. 1997, 119, 5956-
5957. See also: (b) Maeda, K.; Farrington, E. J.; Galardon, E.; John, B.
D.; Brown, J. M. AdV. Synth. Catal. 2002, 344, 104-109. (c) Farina, V.;
Hossain, A. Tetrahedron Lett. 1996, 37, 6997-7000.
(16) (a) Holtcamp, M. W.; Henling, L. M.; Day, M. W.; Labinger, J. A.; Bercaw,
J. E. Inorg. Chim. Acta 1998, 270, 467-478. (b) Ho, V. M.; Watson, L.
A.; Huffman, J. C.; Caulton, K. G. New J. Chem. 2003, 27, 1446-1450.
(17) Schrock, R. R.; Seidel, S. W.; Moesch-Zanetti, N. C.; Shih, K.-Y.;
O’Donoghue, M. B.; Davis, W. M.; Reiff, W. M. J. Am. Chem. Soc. 1997,
119, 11876-11893.
(18) Lowry, T. H.; Richardson, K. S. Mechanism and Theory in Organic
Chemistry, 3rd ed.; 1987; p 1090.
(19) Bonner, T. G.; Bowyer, F.; Williams, G. J. Chem. Soc. 1953, 2650-2652.
A Mechanistic Interpretation: Electrophilic Arylation. We
are aware that the observed KIE may not be directly related to
a single elementary reaction step but instead may reflect a
combination of isotope effects stemming from multiple reactions
in a complex system.13,14 Nonetheless, these results do provide
some important insights. For instance, they cast serious doubts
about the mechanism wherein direct metalation (e.g., via σ-bond
metathesis) at position C-2 takes place; this is primarily due to
the large KIE at C-3. An alternative mechanism involves the
9
J. AM. CHEM. SOC. VOL. 127, NO. 22, 2005 8053

A R T I C L E S
Lane et al.
Figure 2. Hammett plot derived for the palladium-catalyzed, selective C-2 arylation of 6-substituted-1-methylindoles. 1-Methyl-6-nitroindole, shown in
red, is not included in linear regression.
Scheme 7. Electrophilic Cyclization of 3-(3-Hydroxypropyl)-indole
to that for C-2 (apparent primary KIE), may be explained by
this mechanistic interpretation. To the best of our knowledge,
1.6 represents the highest value reported in the literature for a
secondary KIE exerted by a single C-D bond (previous values
are <1.3, solvolysis reactions).23 Once again, we stress that the
observed value may be a composite of several rate-contributing
steps and thus drawing detailed mechanistic conclusions regard-
ing a single elementary step would be erroneous. At the same
time, the KIE observed at the 3-position supports the interpreta-
tion is taken to be very small and constant. The deprotonation
then becomes kinetically significant, and as a consequence larger
isotope effects are observed, as in the sulfonation of nitrobenzene
(kH/kD ) 1.59-1.69).20 Moreover, very large primary isotope
effects have been observed in the mercuration (kH/kD ) 6.0)
and palladation (kH/kD ) 5.0) of arenes in acidic media.21 In
these electrophilic metalations, the deprotonation (k2) is rate-
limiting, which is reflected in a large primary KIE. More
recently, a large primary KIE has also been measured in the
intramolecular palladium-catalyzed cyclization reactions; these
latter results are consistent with the electrophilic metalation
mechanism.22
The electrophilic arylation of indole is, however, more
complex than the simple cases mentioned above. Let us consider
Scheme 6; for simplicity, the initial oxidative addition of
iodobenzene to palladium(0) was omitted as this step does not
contribute to the overall rate.
tion provided herein, involving the electrophilic attack at C-3,
followed by palladium migration.
The driving force for this migration is related to stabilization
of the carbon-palladium bond by the adjacent nitrogen atom
(fast migration of 3-lithio-indole to 2-lithio-indole has also been
reported).24,25 The metal migration around the arene nucleus has
long been studied, and the higher stability of organometallic
intermediates wherein the metal is attached to an electron-
deficient carbon is well recognized.26 Organopalladium inter-
mediates are known to undergo migrations, as was recently
demonstrated by a 1,4-migration in the system derived from
ortho-iodobiaryls.27
Similar 1,2-migration of alkyl groups has been demonstrated
in the electrophilic substitution of 3-substituted indoles.28 For
example, formation of intermediate 12 has been supported by
(23) The large secondary KIE (1.6) has been reported for solvolysis of acetyl
chloride. However, this value reflects an accumulative effect of three C-D
bonds versus three C-H bonds. (a) Karabatsos, G. J.; Sonnichsen, G. C.;
Papaioannou, C. G.; Scheppele, S. E.; Shone, R. L. J. Am. Chem. Soc.
1967, 89, 463-465. (b) Bentley, T. W. J. Org. Chem. 2004, 69, 1756-
1759.
(24) Saulnier, M. G.; Gribble, G. W. J. Org. Chem. 1982, 47, 757-761.
(25) C-2 selectivity has also been observed in the formation of a zwitterionic
intermediate in the reaction of 1-methylindole and B(C6F5)3. An analogous
C3 f C2 migration was proposed and supported by theoretical calculations,
which showed that the C-2 adduct was more stable than the C-3 adduct by
8.4 kcal/mol. Although no experimental evidence has been presented in
support of this mechanism, the indole-borane adducts represent relevant
models for our system. Focante, F.; Camurati, I.; Nanni, D.; Leardini, R.;
Resconi, L. Organometallics 2004, 23, 5135-5141.
(26) (a) Yudin, L. G.; Pavlyuchenko, A. I.; Kost, A. N. Zh. Obshch. Khim. 1969,
39, 2784-2784. (b) Taylor, E. C.; Kienzle, F.; Robey, R. L.; McKillop,
A.; Hunt, J. D. J. Am. Chem. Soc. 1971, 93, 4845-4850.
(27) Larock, R. C. J. Am. Chem. Soc. 2002, 124, 14326-14327.
(28) (a) Jackson, A. H.; Smith, P. J. Chem. Soc., Chem. Commun. 1967, 264-
266. (b) Ganesan, A.; Heathcock, C. H. Tetrahedron Lett. 1993, 34, 439-
440.
The first step involves the formation of intermediate 6 by
the electrophilic addition of an aryl-palladium(II) species to the
3-position of indole. If the reverse step (k-1) is fast as compared
to both the forward step (k1) and the migration (k2), then the
formation of 6 and (or) the migration may become kinetically
relevant. In such instance, k3 must be faster than k2, and k4 must
be slower than k2. The unusual features of this system, the large
value of the KIE at position C-3 (â-isotope effect) in comparison
(20) Brand, J. C. D.; Jarvie, A. W. P.; Horning, W. C. J. Chem. Soc. 1959,
3844-3853.
(21) (a) Davidson, J. M.; Triggs, C. J. Chem. Soc. A 1968, 1324-1330. (b)
Fung, C. W.; Khorramdel-Vahed, M.; Ranson, R. J.; Roberts, R. M. G. J.
Chem. Soc., Perkin Trans. 2 1980, 267-272.
(22) (a) Hennessy, E. J.; Buchwald, S. L. J. Am. Chem. Soc. 2003, 125, 12084-
12085. (b) Campeau, L.-C.; Parisien, M.; Leblanc, M.; Fagnou, K. J. Am.
Chem. Soc. 2004, 126, 9186-9187.
9
8054 J. AM. CHEM. SOC. VOL. 127, NO. 22, 2005

C-2 and C-3 Arylation of Indoles
A R T I C L E S
Scheme 8. Indole Arylation with Sterically Bulky Iodoarenesa
Table 3. Substrate Scope: Substitution on Indolea
a Conditions: (a) 0.5 mol % Pd(OAc)2, 2 mol % PPh3, 1.2 equiv of ArI,
2 equiv of CsOAc, DMA, 150 °C, 24 h. (b) 5 mol % Pd(OAc)2, 20 mol %
PPh3, 2.5 equiv of ArI, 1.2 equiv of MgO, dioxane/DMF (1:2), 150 °C, 18
h.
2). The σ°R values were developed for instances where there is
minor perturbation of the substituted benzene-ring system,
similar to the instance here.29 A reasonable relationship between
the σ°R values and the rate constants was observed for the
substrates tested, with the exclusion of 1-methyl-6-nitroindole.
The negative F-value derived from the plot is consistent with
the electrophilic palladation/migration mechanism.
A similar trend between reactivity and electron density of
substituted indoles was also observed in terms of isolated yields
obtained in analogous preparative reactions, Table 3. Substrates
containing strongly electron-withdrawing substituents exhibited
lower yields than their more electron-rich counterparts. For
example, the yields in entries 3, 4, and 9 were considerably
depressed in contrast to entries 1, 7, and 8 (Table 3). The only
notable exemption was entry 5, where significant decomposition
of both the starting material and the product was observed. The
main consequence associated with the low reactivity of the
electron-deficient indoles was the greater production of bi-
phenyl byproducts (this competing pathway is outlined in
Scheme 2).
A Mechanistic Rationale for the Observed Regioselectiv-
ity. Although the high C-2 selectivity was observed for both
1-alkylindole and an indole magnesium salt (C-2 arylation),
haloarene donors substituted in the ortho-position afforded a
mixture of C-2 and C-3 arylation products (Scheme 8). These
results may be rationalized by the electrophilic metalation
mechanism in the following terms. In the case of bulky
haloarenes, the palladium migration is slower, and as a
consequence deprotonation of intermediate 6 (and thus C-3
arylation, Scheme 6) becomes competitive, affording a mixture
of regioisomers.
The Development of New C-3 Arylation Protocols. The
mechanistic interpretation offered herein serves not only as a
rationale for the observed regioselectivity, but also as a guiding
hypothesis for the development of new, improved arylation
methods. The examples shown in Scheme 8 demonstrate how
the selectivity can be altered if the steric bulk of the aryl-
palladium(II) intermediate in increased. Consequently, consider-
ing intermediate 14 (Figure 3), palladium migration and
ultimately the regio-course of the arylation may be influenced
by several factors. In addition to the steric bulk of the aryl ring
a All values based on isolated yields; conditions are further described in
the Supporting Information. (a) Performed with 0.1 mol % Pd(OAc)2, 0.4
mol % PPh3; reaction time was 48 h.
labeling experiments, and it has been speculated that the 1,2-
migration is rate-limiting in this transformation (Scheme 7).
Hammett Plot of 6-Substituted 1-Methylindoles. The first-
order rate dependence on indole concentration allowed for
investigating the substitution effect on the reaction rate (Ham-
mett plot) and thus provided strong experimental support for
the electrophilic palladation mechanism; that is, the build-up
of positive charge at the 3-position gave rise to a negative
F-value. The reaction rates of a variety of indoles substituted
with both electron-releasing and electron-withdrawing groups
in the 6-postion, para to the reactive center (3-position), were
determined. The kinetic data were then used to generate a
Hammett plot; as a consequence of the relationship between
the 3- and 6-positons of indole, σ°R values were used and found
to provide the best correlation with the rate constants (Figure
(29) Bromilow, J.; Brownlee, R. T. C.; Lopez, V. O.; Taft, R. W. J. Org. Chem.
1979, 26, 4766-4770.
9
J. AM. CHEM. SOC. VOL. 127, NO. 22, 2005 8055

A R T I C L E S
Lane et al.
Figure 3. The factors that affect palladium migration and ultimately regioselectivity of indole arylation.
Scheme 9. New Methods for Direct C-3 Arylation of Indole; Regioselectivity Can Be Controlled by the Choice of Magnesium Salta
a Conditions: Pd(OAc)2 (2.5 mol %), Ph3P (10 mol %), IMes (2.5 mol %), 24 h. (a) Reaction yield based on indole. (b) Reaction yield based on PhI. See
Supporting Information for detailed procedures. Imes ) 1,3-bis-mesitylimidazolyl carbene.
on palladium, as already discussed, these include the ligand on
palladium, and the ligand/solvation around the magnesium
metal. This hypothesis led us to explore the possibility of direct
phenylation of the 3-position, controlled by the choice of the
magnesium salt.
and Mg(HMDS)2, which led to even higher selectivity (26:1)
and allowed indole be used as the limiting agent.
It is important to note that the magnesium indole salts are
most likely more complex in structure than depicted in Scheme
9. Although the indole Grignard reagents have been studied for
many years, and used primarily in the context of indole
C-alkylation reactions,31 the solution structure of these salts with
respect to the magnesium coordination sphere has not been
elucidated. These issues will be addressed in the future
optimization studies.
As an important hint, we observed that phenylation of the
indole Grignard salt prepared from indole and MeMgCl gave a
mixture of 3-phenyl- and 2-phenylindole in a 7:1 ratio. This
stands in start contrast to the method employing MgO, which
afforded 2-phenylindole exclusively (Scheme 9). Clearly, the
nature of the magnesium coordination sphere influences the
regioselectivity of this reaction. Subsequently, we discovered
that the addition of TMEDA (tetramethylethylenediamine) led
to formation of 3-phenylindole with high selectivity (14:1),
presumably via formation of a sterically demanding magnesium
complex (Scheme 9).30 Although highly selective, the latter
methods required a 2-fold excess of the azole magnesium salt
as one equivalent serves to neutralize hydrogen halide formed
in the arylation reaction. However, this shortcoming was
eliminated via formation of a new magnesium salt from indole
Finally, the choice of the palladium ligand also proved
important as suggested by our hypothesis. Thus, the use of IMes
ligand (1,3-bis-mesitylimidazolyl carbene) in place of Ph3P led
to an improvement in both the yield and the selectivity;
furthermore, bromobenzene gave better results in comparison
to iodobenzene (Scheme 9).
Thus, a highly selective method for C-3 arylation of indole
was developed. A switch from C-2 to C-3 arylation was
achieved by the choice of magnesium base. These new condi-
(31) (a) Powers, J. C.; Meyer, W. P.; Parsons, T. G. J. Am. Chem. Soc. 1967,
89, 5812-5820. (b) Heacock, R. A.; Kasparek, S. AdV. Heterocycl. Chem.
1969, 10, 43-112. (c) Reinecke, M. G.; Sebastian, J. F.; Johnson, H. W.;
Pyun, C. J. Org. Chem. 1971, 36, 3091-3095. (d) Powers, J. C. Chem.
Heterocycl. Compd. 1972, 25, 127-178.
(30) (a) Magnuson, J. A.; Roberts, J. D. J. Org. Chem. 1972, 37, 133-135. (b)
Marsch, M.; Harms, K.; Massa, W.; Boche, G. Angew. Chem., Int. Ed.
Engl. 1987, 26, 696-697. (c) Swiss, K. A.; Liotta, D. C. J. Am. Chem.
Soc. 1990, 112, 9393-9394.
9
8056 J. AM. CHEM. SOC. VOL. 127, NO. 22, 2005

C-2 and C-3 Arylation of Indoles
A R T I C L E S
tions represent an exciting lead for the development of efficient
and practical arylation methods.
methods. It was shown that the choice of magnesium salt affects
the regioselectivity of the arylation reaction, presumably via
the steric demand of the magnesium coordination sphere.
Inspired by these findings, a new catalytic method was designed
for selective C-3 arylation of (NH)-indole, complementing the
C-2 selective protocol, as well as N-arylation methods. The
insight generated in this work will provide the rational basis
for the future studies aimed at optimizing the reaction conditions,
and expanding the substrate scope and functional group toler-
ance.
Conclusions
This paper describes mechanistic investigations aimed at
providing a rational explanation for the high C-2 selectivity in
the palladium-catalyzed arylation of indoles. It was demonstrated
that the reaction was first-order in both the substrate and the
catalyst. Moreover, a Hammett plot revealed the negative
F-value, indicating that a positive charge is accumulated at the
3-position of indole and affording a strong support for the
electrophilic palladation mechanism.
The KIE was determined for both C-2 and C-3 positions of
indole, yielding the values 1.2 and 1.6, respectively. Interest-
ingly, the larger KIE value was obtained for the 3-position where
the substitution does not occur and thereby represents the
secondary KIE. These results were confirmed by an independent
study, employing the method of competitive reactions. These
data support an electrophilic substitution mechanism that
features a 1,2-migration of palladium. The electrophilic attack
of the arylpalladium species on indole and (or) the migration
of palladium represent the slow step(s) of the catalytic cycle.
This mechanism has a close parallel in classical electrophilic
substitution reactions of indoles substituted in the 3-position.
The mechanistic interpretation offered herein served not only
as a rationale for the observed regioselectivity, but also as a
guiding hypothesis for the development of new arylation
Acknowledgment. We would like to thank Professor Thomas
Katz and Ms. Bengu¨ Sezen for valuable discussions. Funding
was provided by the National Science Foundation, GlaxoSmith-
Kline, Johnson & Johnson, and Merck Research Laboratories.
D.S. is a recipient of the Unrestricted Grant in Synthetic
Chemistry from the Bristol-Myers Squibb, AstraZeneca Excel-
lence in Chemistry Award, and the Pfizer Award for Creativity
in Organic Synthesis. We would like to thank Dr. J. B. Schwarz
(editorial assistance) and Vitas Votier Chmelar (intellectual
support).
Supporting Information Available: Experimental procedures,
spectral data, and base optimization data. This material is
available free of charge via the Internet at http://pubs.acs.org.
JA043273T
9
J. AM. CHEM. SOC. VOL. 127, NO. 22, 2005 8057