Rhodium(III)-catalyzed arene and alkene C-H bond functionalization leading to indoles and pyrroles

Article abstract of DOI:10.1021/ja1082624

Recently, the rhodium(III)-complex [Cp*RhCl₂]₂ 1 has provided exciting opportunities for the efficient synthesis of aromatic heterocycles based on a rhodium-catalyzed C-H bond functionalization event. In the present report, the use of complexes 1 and its dicationic analogue [Cp*Rh(MeCN)₃][SbF₆]₂ 2 have been employed in the formation of indoles via the oxidative annulation of acetanilides with internal alkynes. The optimized reaction conditions allow for molecular oxygen to be used as the terminal oxidant in this process, and the reaction may be carried out under mild temperatures (60 °C). These conditions have resulted in an expanded compatibility of the reaction to include a range of new internal alkynes bearing synthetically useful functional groups in moderate to excellent yields. The applicability of the method is exemplified in an efficient synthesis of paullone 3, a tetracyclic indole derivative with established biological activity. A mechanistic investigation of the reaction, employing deuterium labeling experiments and kinetic analysis, has provided insight into issues of reactivity for both coupling partners as well as aided in the development of conditions for improved regioselectivity with respect to meta-substituted acetanilides. This reaction class has also been extended to include the synthesis of pyrroles. Catalyst 2 efficiently couples substituted enamides with internal alkynes at room temperature to form trisubstituted pyrroles in good to excellent yields. The high functional group compatibility of this reaction enables the elaboration of the pyrrole products into a variety of differentially substituted pyrroles.

Full text of DOI:10.1021/ja1082624

Published on Web 12/06/2010
Rhodium(III)-Catalyzed Arene and Alkene C-H Bond
Functionalization Leading to Indoles and Pyrroles
David R. Stuart,*,† Pamela Alsabeh,‡ Michelle Kuhn, and Keith Fagnou§
Centre for Catalysis Research and InnoVation, Department of Chemistry, UniVersity of Ottawa,
10 Marie Curie, Ottawa, (Canada) K1N 6N5
Received September 13, 2010; E-mail: dstuart@fas.harvard.edu
Abstract: Recently, the rhodium(III)-complex [Cp*RhCl2]2 1 has provided exciting opportunities for the
efficient synthesis of aromatic heterocycles based on a rhodium-catalyzed C-H bond functionalization
event. In the present report, the use of complexes 1 and its dicationic analogue [Cp*Rh(MeCN)3][SbF6]2 2
have been employed in the formation of indoles via the oxidative annulation of acetanilides with internal
alkynes. The optimized reaction conditions allow for molecular oxygen to be used as the terminal oxidant
in this process, and the reaction may be carried out under mild temperatures (60 °C). These conditions
have resulted in an expanded compatibility of the reaction to include a range of new internal alkynes bearing
synthetically useful functional groups in moderate to excellent yields. The applicability of the method is
exemplified in an efficient synthesis of paullone 3, a tetracyclic indole derivative with established biological
activity. A mechanistic investigation of the reaction, employing deuterium labeling experiments and kinetic
analysis, has provided insight into issues of reactivity for both coupling partners as well as aided in the
development of conditions for improved regioselectivity with respect to meta-substituted acetanilides. This
reaction class has also been extended to include the synthesis of pyrroles. Catalyst 2 efficiently couples
substituted enamides with internal alkynes at room temperature to form trisubstituted pyrroles in good to
excellent yields. The high functional group compatibility of this reaction enables the elaboration of the pyrrole
products into a variety of differentially substituted pyrroles.
prominent role.5,6 The Larock indole synthesis is paramount
among the latter, setting the state-of-the-art for transition metal-
Introduction
catalyzed intermolecular approaches toward indoles,5d,g,7 and
analogous reactions have recently been developed for the
synthesis of pyrroles6c and other heterocycles.5a However, these
approaches necessitate substrate preactivation in the form of
aryl and alkenyl halides, adding to the overall time and cost of
a synthetic sequence (Scheme 1, eq 1). More recently, the
synthesis of aromatic heterocycles has shifted to include
approaches predicated upon a C-H bond functionalization event
Indoles and pyrroles represent two of the most abundant
heterocycles in nature. As constituents of essential amino acids,
embedded in structurally complex biologically active natural
products,1 and possessing desirable properties to materials
scientists,2 the means for their construction have captivated
chemists for over a century (Figure 1).3 Additionally, owing to
the prevalence of these heterocycles in known pharmaceuticals,
they have been termed a “privileged motif” in medicinal
chemistry (Figure 1).4 For these reasons, many advances have
been made in the synthesis of aromatic heterocycles in recent
decades, with transition metal-catalyzed reactions taking a
(5) For a review on transition metal-catalyzed indole synthesis, see: (a)
Zeni, G.; Larock, R. C. Chem. ReV. 2006, 106, 4644. For selected
individual accounts, see. (b) Tida, H.; Yuasa, Y.; Kibayashi, C. J.
Org. Chem. 1980, 45, 2938. (c) Sakamoto, T.; Nagano, T.; Kondo,
Y.; Yamanaka, H. Synthesis 1990, 215. (d) Larock, R. C.; Yum, E. K.
J. Am. Chem. Soc. 1991, 113, 6689. (e) Koerber-Ple, K.; Massiot, G.
Synlett 1994, 759. (f) Chen, C. Y.; Lieberman, D. R.; Larsen, R. D.;
Verhoeven, T. R.; Reider, P. J. J. Org. Chem. 1997, 62, 2676. (g)
Larock, R. C.; Yum, E. K.; Refvik, M. D. J. Org. Chem. 1998, 63,
7652. (h) Yamazaki, K.; Nakamura, Y.; Kondo, Y. J. Org. Chem.
2003, 68, 6011. (i) Watanabe, T.; Arai, S.; Nishida, A. Synlett 2004,
907. (j) Nazare, M.; Schneider, C.; Lindenschmidt, A.; Will, D. W.
Angew. Chem., Int. Ed. 2004, 43, 4526. (k) Jia, Y.; Zhu, J. J. Org.
Chem. 2006, 71, 7826. (l) Leogane, O.; Lebel, H. Angew. Chem., Int.
Ed. 2008, 47, 350.
† Current address: Department of Chemistry & Chemical Biology,
Harvard University, Cambridge, MA 02138.
‡ Current address: Department of Chemistry, Dalhousie University,
Halifax, Nova Scotia, (Canada) B3H 4J3.
§ Deceased, November 11, 2009.
(1) (a) A Beilstein search for indoles with biological activity yielded
>45 000 results. (b) Richter, J. M.; Ishihara, Y.; Masuda, T.; Whitefield,
B. W.; Llamas, T.; Pohjakallio, A.; Baran, P. S. J. Am. Chem. Soc.
2008, 130, 17938, and references therein. (c) Fu¨rstner, A. Angew.
Chem., Int. Ed. 2003, 43, 3582.
(2) (a) Yoon, Z. S.; Kwon, J. H.; Yoon, M.-C.; Koh, M. K.; Noh, S. B.;
Sessler, J. L.; Lee, J. T.; Seidel, D.; Aguilar, A.; Shimizu, S.; Suzuki,
M.; Osuka, A.; Kim, D. J. Am. Chem. Soc. 2006, 128, 14128. (b)
Lee, J.-S.; Kang, N.-Y.; Kim, Y. K.; Samanta, A.; Feng, S.; Kim,
H. K.; Vendrell, M.; Park, J. H.; Chang, Y.-T. J. Am. Chem. Soc.
2009, 131, 10077.
(6) For selected examples of transition metal-catalyzed pyrrole synthesis,
see: (a) Dhawan, R.; Arndtsen, B. A. J. Am. Chem. Soc. 2004, 126,
468. (b) Lu, Y.; Arndtsen, B. A. Angew. Chem., Int. Ed. 2008, 47,
5430. (c) Crawley, M. L.; Goljer, I.; Jenkins, D. J.; Mehlmann, J. F.;
Nogle, L.; Dooley, R.; Mahaney, P. E. Org. Lett. 2006, 8, 5837. (d)
Kel’in, A. V.; Sromek, A. W.; Gevorgyan, V. J. Am. Chem. Soc. 2001,
123, 2075. (e) Gabriele, B.; Salerno, G.; Fazio, A. J. Org. Chem. 2003,
68, 7853.
(3) Gilchrist, T. L. J. Chem. Soc., Perkin Trans. 1 1999, 2849.
(4) de Sa´ Alves, F. R.; Barreiro, E. J.; Fraga, C. A. M. Mini-ReV. Med.
Chem. 2009, 9, 782.
9
18326 J. AM. CHEM. SOC. 2010, 132, 18326–18339
10.1021/ja1082624  2010 American Chemical Society

Rhodium(III)-Catalyzed Formation of Indoles and Pyrroles
A R T I C L E S
Figure 1. Representative important indoles and pyrroles.
which represents an inherently more efficient and atom eco-
nomical approach (Scheme 1, eq 2).8 Within these methods,
palladium, copper, and rhodium catalysts have been employed
in the synthesis of isocoumarins,9 indoles,10 indolines,11 pyr-
roles,12 isoquinolines,13 and isoquinolones and pyridones14 among
other heterocycles.15 Rhodium(III)-catalysts have been particularly
useful in these transformations, in which rhodium is often introduced
as the [Cp*RhCl2]2 1 catalyst precursor.9,10b,12-15
Our group has successfully implemented this approach,
employing rhodium(III)-catalysts, in the preparation of a number
of heterocycles,10b,13b,14a as well as in the hydroarylation of
unactivated internal alkynes.16 However, in our previous indole
synthesis (eq 3), and in many other cases within this reaction
Scheme 1. Recent Transition Metal-Catalyzed Approaches to
Aromatic Heterocycles
(7) For representative applications in medicinal chemistry, see: (a) Lanter,
J. C.; Fiordeliso, J. J.; Alford, V. C.; Zhang, X.; Wells, K. M.; Russell,
R. K.; Allan, G. F.; Lai, M.-T.; Linton, O. L.; Lundeen, S.; Sui, Z.
Bioorg. Med. Chem. Lett. 2007, 17, 2545. (b) Watterson, S. H.; Murali
Dhar, T. G.; Ballentine, S. K.; Shen, Z.; Barrish, J. C.; Cheney, D.;
Fleener, C. A.; Rouleau, K. A.; Townsend, R.; Hollenbaugh, D. L.;
Iwanowicz, E. J. Bioorg. Med. Chem. Lett. 2003, 13, 1273. (c) Curtis,
N. R.; Kulagowski, J. J.; Leeson, P. D.; Ridgill, M. P.; Emms, F.;
Freedman, S. B.; Patel, S.; Patel, S. Bioorg. Med. Chem. Lett. 1999,
9, 585. (d) Ujjainwalla, F.; Walsh, T. F. Tetrahedron Lett. 2001, 41,
6441. For recent representative examples of the Larock indole synthesis
in total synthesis, see: (e) Newhouse, T.; Lewis, C. A.; Baran, P. S.
J. Am. Chem. Soc. 2009, 131, 6360. (f) Newhouse, T.; Baran, P. S.
J. Am. Chem. Soc. 2008, 130, 10886. (g) Newhouse, T.; Lewis, C. A.;
Eastman, K. J.; Baran, P. S. J. Am. Chem. Soc. 2010, 132, 7119. (h)
Garfunkle, J.; Kimball, F. S.; Trzupek, J. D.; Takizawa, S.; Shimamura,
H.; Tomishima, M.; Boger, D. L. J. Am. Chem. Soc. 2009, 131, 16036.
(8) For a recent and comprehensive review of rhodium(III)-catalyzed
formation of heterocycles via annulative processes with internal
alkynes, see: Satoh, T.; Miura, M. Chem.sEur. J. 2010, 16, 11212.
(9) (a) Ueura, K.; Satoh, T.; Miura, M. Org. Lett. 2007, 9, 1407. (b) Ueura,
K.; Satoh, T.; Miura, M. J. Org. Chem. 2007, 72, 5362.
(10) (a) Wurtz, S.; Rakshit, S.; Neumann, J. J.; Droge, T.; Glorius, F.
Angew. Chem., Int. Ed. 2008, 47, 7230. (b) Stuart, D. R.; Bertrand-
Laperle, M.; Burgess, K. M. N.; Fagnou, K. J. Am. Chem. Soc. 2008,
130, 16474. (c) Shi, Z.; Zhang, C.; Li, S.; Pan, D.; Ding, S.; Cui, Y.;
Jiao, N. Angew. Chem., Int. Ed. 2009, 48, 4572. (d) Bernini, R.; Fabrizi,
G.; Sferrazza, A.; Cacchi, S. Angew. Chem., Int. Ed. 2009, 48, 8078.
(e) Tan, Y.; Hartwig, J. F. J. Am. Chem. Soc. 2010, 132, 3676. (f)
Chen, J.; Song, G.; Pan, C.-L.; Li, X. Org. Lett. 2010, 12, 5426.
9
J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010 18327

A R T I C L E S
Stuart et al.
Table 1. Optimization of Second-Generation Conditions
entry
precatalyst (mol %)
additive (mol %)
Cu(OAc)2 (mol %)
temp
time (h)
1H NMR yield (6a)c
1
2
3
4
5
6
7
8
9
[Cp*RhCl2]2 (2.5)
[Cp*RhCl2]2 (2.5)
[Cp*RhCl2]2 (2.5)
[Cp*RhCl2]2 (2.5)
AgSbF6 (10)
AgSbF6 (10)
AgSbF6 (10)
AgSbF6 (10)
210b
210b
10
10
20
20
20
20
20
120 °C
60 °C
60 °C
60 °C
60 °C
60 °C
21 °C
21 °C
21 °C
1
6
6
22
22
22
30
96
96
86(79)
92(90)
55
68
81
[Cp *RhCl2]2 (2.5)
AgSbF6 (10)
[Cp*Rh (MeCN)3][SbF6]2 (5)
[Cp*Rh (MeCN)3][SbF6]2 (5)
[Cp*Rh (MeCN)3][SbF6]2 (5)
[Cp*Rh (MeCN)3][SbF6]2 (5)
s
s
s
s
93(90)
9d
12d
95d, e
a Conditions: 4a (1 equiv 0.2 M), 5a (1.1 equiv), precatalyst 1 or 2, additive (if applicable), Cu(OAc)2 ·H2O, O2 balloon (if applicable), t-AmOH,
temperature, specified time. b No balloon of O2 used. c NMR yield vs trimethoxybenzene as internal standard; isolated yields in parentheses. d GCMS
yield vs trimethoxybenzene as internal standard. e Reaction performed in acetone.
class, high temperatures (>100 °C) and stoichiometric amounts
of a metal oxidant17 (2 equiv of Cu(OAc)2) are employed.
Additionally, many of these reactions have so far displayed
limited compatibility with respect to the alkyne coupling
partner.18 Symmetrical and unsymmetrical dialkyl alkynes are
noticeably absent from many recently reported annula-
tions,10f,12,14a,b,d,e,15a,c and, when present, typically provide
diminished yields of the desired product in comparison to other
alkynes.10b,13c,14c,15b Unsymmetrical dialkyl alkynes are also
complicated by low ratios of often inseparable regioisomers.10b,14c
Given the recent developments for the synthesis of heterocycles
catalyzed by 1, a detailed mechanistic understanding of this class
of reaction will only aid in the further expansion of existing
methods and pave the way for future endeavors. In the context
of heterocycle synthesis,13a Jones has carried out an elegant
mechanistic study on the cyclorhodation of benzylamines with
stoichiometric amounts of 1.19 However, a complete mechanistic
study of catalytic oxidative annulation reactions utilizing 1 (or
related analogues) has yet to be reported and prompted our
interest in a broader understanding of the reaction.
(11) Houlden, C. E.; Bailey, C. D.; Ford, J. G.; Gagne´, M. R.; Lloyd-
Jones, G. C.; Booker-Milburn, K. I. J. Am. Chem. Soc. 2008, 130,
10066.
(12) During the preparation of this manuscript, a conceptually interesting
approach employing allylic, and one example of vinylic, C-H bond
functionalization in the formation of pyrroles was reported; see:
Rakshit, S.; Patureau, F. W.; Glorius, F. J. Am. Chem. Soc. 2010,
132, 9585.
In this account, we report our findings in the development of
general and mild conditions for the oxidative annulation of
acetanilides with internal alkynes catalyzed by the rhodium(III)-
complexes 1 and [Cp*Rh(MeCN)3][SbF6]2 2. The mild condi-
tions support the use of molecular oxygen as the terminal oxidant
and are compatible with a more diverse set of internal alkynes
than previously reported (eq 4). The synthetic utility of this
transformation is further highlighted by an expedient synthesis
of paullone 3, the parent compound of a family of compounds
with known biological activity. Additionally, we have under-
taken a mechanistic evaluation of the reaction based on isotope
labeling experiments and kinetic analysis. The resulting infor-
mation has provided answers to some of the issues of regiose-
lectivity for the acetanilide metalation and alkyne insertion steps
and guided our development of conditions for improved
regioselectivity of meta-substituted acetanilides. Finally, the
method is extended to include the room temperature preparation
of pyrroles, and the utility of the method is illustrated by the
(13) For an example of stoichiometric [Cp*RhCl2]2-mediated isoquinoline
salt formation, see: (a) Li, L.; Brennessel, W. W.; Jones, W. D. J. Am.
Chem. Soc. 2008, 130, 12414. For examples of isoquinoline synthesis
employing catalytic 1 and 2, see: (b) Guimond, N.; Fagnou, K. J. Am.
Chem. Soc. 2009, 131, 12050. (c) Fukutani, T.; Umeda, N.; Hirano,
K.; Satoh, T.; Miura, M. Chem. Commun. 2009, 5141.
(14) (a) Guimond, N.; Gouliaras, C.; Fagnou, K. J. Am. Chem. Soc. 2010,
132, 6908. (b) Mochida, S.; Umeda, N.; Hirano, K.; Satoh, T.; Miura,
M. Chem. Lett. 2010, 39, 744. (c) Hyster, T. K.; Rovis, T. J. Am.
Chem. Soc. 2010, 132, 10565. (d) Song, G.; Chen, D.; Pan, C.-L.;
Crabtree, R. H.; Li, X. J. Org. Chem. 2010, 75, 7487. (e) Su, Y.;
Zhao, M.; Han, K.; Song, G.; Li, X. Org. Lett. 2010, 12, 5462.
(15) (a) Umeda, N.; Tsurugi, H.; Satoh, T.; Miura, M. Angew. Chem., Int.
Ed. 2008, 47, 4019. (b) Morimoto, K.; Hirano, K.; Satoh, T.; Miura,
M. Org. Lett. 2010, 12, 2068. (c) Wang, F.; Song, G.; Li, X. Org.
Lett. 2010, 12, 5430.
(16) Schipper, D. J.; Hutchinson, M.; Fagnou, K. J. Am. Chem. Soc. 2010,
132, 6910.
(17) For pioneering work employing catalytic copper and air as the terminal
oxidant, see ref 9.
(18) This aspect of the chemistry was identified in a Synfacts Highlight of
our previously reported indole synthesis, ref 10b. Synfacts 2009, 3,
254.
(19) Li, L.; Brennessel, W. W.; Jones, W. D. Organometallics 2009, 28,
3492.
9
18328 J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010

Rhodium(III)-Catalyzed Formation of Indoles and Pyrroles
A R T I C L E S
ability to elaborate the primary reaction products into pyrroles
with a variety of substitution patterns.
respectively, compare entries 1 and 6, Table 1). Additionally, a
multitude of 5-, 6-, and 7-substituted indoles were readily
prepared using this reaction. Para-substituted acetanilides bearing
electron-donating, electron-withdrawing, and halide substituents
afforded the corresponding 5-substituted indoles 6b-f in good
yields (Chart 1, entries 1-5). Meta-substitution was tolerated
in the case of both symmetrical and unsymmetrical acetanilides.
Electron-rich and electron-poor 3,5-substituted acetanilides gave
4,6-disubstituted indoles 6g and 6i in good yields (Chart 1,
entries 6 and 8). However, it was evident that steric effects play
a significant role in reaction efficiency, as 3,5-dimethylaceta-
nilide provided indole 6h in very low isolated yield (7%), even
under extended reaction times (Chart 1, entry 7). When
3-substituted acetanilides were employed, the inherent regiose-
lectivity of the reaction was for the more sterically accessible
C-H bond. Steric effects dominate this process over electronic
effects, as 6k was produced as a single regioisomer in good
isolated yield (Chart 1, entry 10). Other electronically diverse
meta-substituted acetanilides also react at the more sterically
accessible C-H bond to furnish 6-substituted indoles. An
electron-donating methoxy group was compatible, providing
good isolated yield of the 6-substituted indole 6j as the major
regioisomer (Chart 1, entry 9). The slightly diminished yield
under the second-generation conditions likely reflected the
reduced regioselectivity under these conditions (vide infra). An
acetanilide bearing an electron-withdrawing substituent, such
as a methyl ester, reacted with regioselectivity consistent with
other meta-substituted acetanilides, but with much lessened
reactivity, producing indole 6l in 22% isolated yield (Chart 1,
entry 11). Additionally, a benzo-fused acetanilide provided the
corresponding indole in good yield (Chart 1, entry 12). Ortho-
substitution was also tolerated yielding 7-methyl-substituted or
benzo[g]indoles, 6n, and 6o, respectively (Chart 1, entries 13
and 14).
Results and Discussion
Indole Formation: Reaction Development. In the coupling
of acetanilide 4a with 1-phenyl-1-propyne 5a, our previous
optimization identified the requirement for a silver additive to
sequester chloride anions from precatalyst 1 in order to obtain
catalytic activity.10b The secondary investigation of reaction
conditions presented here focused on the use of molecular
oxygen as the terminal oxidant in conjunction with a catalytic
amount of a transition metal oxidant. In order to do so, we first
explored the effect of reaction temperature and have found that
it can be reduced to 60 °C with a concomitant increase in
reaction time, resulting in an improved yield of 6a (Table 1,
entries 1 and 2). Employing 10 mol % copper(II) acetate with
a balloon of oxygen, we were pleased to find that we could
still obtain moderate yields of 6a (Table 1, entry 3). Addition-
ally, the reaction yield could be restored to synthetically useful
levels by further increasing the reaction time to 22 h and slightly
increasing the loading of copper(II) acetate to 20 mol % (Table
1, entries 4 and 5). These studies also identified 2 as a competent
catalyst precursor for this reaction,20 providing 6a in excellent
yield while simplifying the reaction setup (Table 1, entry 6).
The bench-stable powder 2 is easily prepared, may be weighed
out to air, is insensitive to oxygen and moisture, and provides
the advantage of avoiding the use of hygroscopic AgSbF6. When
the reaction was carried out at room temperature (21 °C)
under the otherwise optimized conditions of entry 6 in Table 1,
the yield reached 9% and 12% after 30 and 96 h, respectively
(Table 1, entries 7 and 8). These low yields were likely due to
the insolubility of many of the reaction components in t-AmOH
at room temperature; however, over an extended period of time
(96 h), a GCMS yield of 95% was obtained in acetone as the
solvent at 21 °C (Table 1, entry 9). This result demonstrates
the robust nature of the catalyst such that even after 4 days,
catalytic turnover was observed. In the context of developing
an overnight reaction, the conditions described in entry 6 were
taken as optimal and used in further investigations of the general
utility of the reaction.21 We have also explored the effect of
the aniline-nitrogen protecting group and found that in the
absence of a protecting group no reaction was observed.
Additionally, while more sterically encumbered N-pivaloyla-
niline provides a diminished yield of the desired indole, the
use of substrates with attenuated Lewis basicity at the amide
oxygen, such as N-trifluoroacetylaniline and N-Boc-aniline, were
ineffective in providing any appreciable amount of 6a.
A range of new and functionally diverse alkynes were
compatible under our second-generation conditions (Chart 2).
As previously observed, both aryl/aryl and alkyl/alkyl sym-
metrically substituted internal alkynes produced 2,3-disubstituted
indoles 7b and 7c in moderate to excellent yields, with the
second-generation conditions providing enhanced yield in the
case of the former (Chart 2, entries 1 and 2). However, when
an unsymmetrical alkyl/alkyl substituted internal alkyne was
employed in the reaction, a mixture of C2/C3 indole regioiso-
mers was obtained (Chart 2, entry 3). In this case, a low yield
was observed and while the regioselectivity of alkyne insertion
was low (1.2:1), it was consistent with other rhodium(III)-
catalyzed annulations.10b,13b,14c Given the significant challenge
in the preparation of this class of compound by transition metal-
catalyzed annulations of internal alkynes, continued research
will be necessary to find a synthetically useful solution to the
formation of unsymmetrical 2,3-dialkylindoles. Conversely,
alkyl/aryl internal alkynes reacted with near exclusive regiose-
lectivity (>40:1), placing the aryl substituent proximal to the
indole nitrogen. A number of different functional groups,
including an electron-donating group, 7f, and a halide, 7g, were
compatible on the aryl moiety of this class of alkyne (Chart 2,
entries 5 and 6). Unsaturated heterocycles, such as thiophene,
7h, and indole, 7i, may also serve as the aryl moiety, providing
a heterobiaryl linkage at the C2 position of indole in good yields
(Chart 2, entries 7 and 8). The mild second-generation conditions
allowed for additional functional groups on the alkyl moiety of
alkyl/aryl internal alkynes. A cyclic alkyl group, such as
cyclopropyl, led to 3-cyclopropylindole 7j in moderate yield
Indole Formation: Reaction Scope and Limitations. Having
arrived at mild second-generation conditions, we now report
the full compatibility of the reaction with respect to both the
acetanilide and alkyne coupling partners under both first- and
second-generation conditions. When various acetanilides were
examined, yields were generally comparable or better under the
second-generation conditions (for example 6a is prepared in 79%
and 90% yield under first- and second-generation conditions,
(20) Subsequent to its use in our indole forming reaction, catalyst 2 has
been utilized in the synthesis of isoquinolines (ref 13b) and in the
hydroarylation of unactivated alkynes (ref 16) by our group. This
catalyst (2) will soon be commercially available from Strem Chemicals,
Inc.
(21) It should be noted that the reaction set-up is operationally very simple
and that there is no need for special precautions to exclude atmospheric
air or moisture.
9
J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010 18329

A R T I C L E S
Stuart et al.
Chart 1. Acetanilide Compatibility under Both First- and Second-Generation Conditions
a Conditions: First generation (red): Acetanilide (1 equiv 0.2 M), 5a (1.1 equiv), 1 (2.5 mol %), AgSbF6 (10 mol %), Cu(OAc)2 ·H2O (2.1 equiv), t-AmOH,
120 °C, 1-3 h. Second generation (blue): Acetanilide (1 equiv 0.2 M), 5a (1.1 equiv), 2 (5 mol %), Cu(OAc)2 ·H2O (20 mol %), O2 balloon, t-AmOH, 60
°C, 20 h. b Isolated yields are reported above. c Regioselectivity based on isolated yield of regioisomers. d Regioselectivity based on crude 1H NMR spectrum.
(Chart 2, entry 9). Ethyl propiolate esters, 3-alkynoates, and a
diethylmalonate moiety at the propargylic position of the alkyne
were all tolerated, providing functionalized indoles in moderate
to good yields (Chart 2, entries 10-12). Unfortunately, under
both first- and second-generation conditions, terminal alkynes
were incompatible, largely producing alkyne homocoupling
products. However, while the reaction yield was low, TMS-
substituted 1-octyne 5o reacted with acetanilide 4a under
modified second-generation conditions to yield the indole
product, 7o, with the TMS group proximal to the indole nitrogen.
This provides the opportunity for TMS-group cleavage from
the indole product and the ability to access C2 unsubstituted
indoles.
occurring, producing C3-benzyl ether 8.22 In t-AmOH this
process was largely dependent on the nature of the protecting
group and a screen of common alcohol protecting groups proved
most fruitful. Due to its good leaving group ability, the use of
acetate as a protecting group resulted in the significant produc-
tion of 8 (Table 2, entry 2). Although the use of benzyl and
silyl groups equally reduced the production of the solvolysis
product, tert-butyldimethylsilyl-protected alcohol provided the
desired product in the highest yield (Table 2, entries 3-5).
Furthermore, control experiments revealed that 8 was observed
when indole products 7q-t were heated alone in tert-amyl
alcohol.23 The use of a TBS-protecting group could also be
extended to include protected homopropargyl alcohols as
coupling partners in good yield (Chart 2, entry 13).
Unprotected propargyl alcohols were unreactive when sub-
jected to either set of reaction conditions. Thus, an investigation
of common protecting groups was conducted to allow for the
use of this important class of compound (Table 2). These studies
found that a competitive solvolysis of the indole product was
(22) Other solvents were tested in order to avoid solvolysis of the product
with t-AmOH. However, slow reactions and lower yields of the desired
product were observed.
(23) See Supporting Information for further details.
9
18330 J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010

Rhodium(III)-Catalyzed Formation of Indoles and Pyrroles
A R T I C L E S
Chart 2. Alkyne Compatibility under Both First- and Second-Generation Conditions
a Conditions: First generation (red): 4a (1 equiv 0.2 M), alkyne (1.1 equiv), 1 (2.5 mol %), AgSbF6 (10 mol %), Cu(OAc)2 ·H2O (2.1 equiv), t-AmOH,
120 °C, 1-3 h. Second generation (blue): 4a (1 equiv, 0.2 M), alkyne (1.1 equiv), 2 (5 mol %), Cu(OAc)2 ·H2O (20 mol %), O2 balloon, t-AmOH, 60 °C,
20 h. b Isolated yields are reported above. c Isolated yield of a mixture of regioisomers (major isomer shown). d Yield of mixture of regioisomers and ratio
based on crude H NMR spectrum vs trimethoxybenzene as internal standard. e i-PrOAc used as solvent.
1
Synthetic Applications: Synthesis of Paullone. Having estab-
lished a method to efficiently construct indoles with high
Table 2. Screen of Potential Propargyl Alcohol Protecting Groups
functional group compatibility, we investigated the feasibility
of applying our method to the synthesis of paullone 3. The
paullone family of compounds was originally discovered by
Sausville during studies on a semisynthetic flavinoid, flavopiri-
dol, known to inhibit cyclin-dependent kinases (CDKs) 1, 2,
and 4.24 Over the past decade, numerous derivatives of paullone
entry
alkyne
R group
yieldb of 7p-t
yieldb of 8
have been found to possess biological activity similar to that
of CDK-1, -2, and -5 inhibitors,25 to have in vitro antitumor
1
2
3
4
5
5p
5q
5r
5s
5t
H
Ac
Bn
TBS
TIPS
7p: s
s
7q: 16%
7r: 35%
7s: 67%
7t: 35%
59%
11%
11%
17%
(24) Sausville, E. A.; Zaharevitz, D.; Gussio, R.; Meijer, L.; Louarn-Leost,
M.; Kunick, C.; Schultz, R.; Lahusen, T.; Headlee, D.; Stinson, S.;
Arbuck, S. G.; Senderowicz, A. Pharmacol. Ther. 1999, 82, 285.
(25) Zaharevitz, D.; Gussio, R.; Leost, M.; Senderowicz, A.; Lahusen, T.;
Kunick, C.; Meijer, L.; Sausville, E. A. Cancer Res. 1999, 59, 2566.
a Conditions: 4a (1 equiv, 0.2 M), alkyne (1.1 equiv), 2 (5 mol %),
Cu(OAc)2 ·H2O (20 mol %), O2 balloon, t-AmOH, 60 °C, 20 h.
b Isolated yields.
9
J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010 18331

A R T I C L E S
Stuart et al.
Scheme 2. Retrosynthesis and Forward Synthetic Route to
Paullone 3a
Scheme 3. Proposed Catalytic Cycle under Our First-Generation
Conditions
theses30a,c,f that the azepine ring may be closed by intramolecular
amide bond formation from the corresponding aniline and
benzylic ester (or equivalent carbonyl). Thus, our retrosynthesis
was reduced to the formation of the appropriately substituted
indole, which could arrive from the coupling of 4a and an
appropriately functionalized alkyne 10 (Scheme 2).
Alkyne 10 was prepared in four steps in 50% overall yield
with chromatographic purification necessary only after the final
step (Scheme 2).23 The alkyne was then set for oxidative
coupling with 4a under the second-generation conditions,
yielding indole 11 in 71% isolated yield. As previously observed
in a screen of compatible nitrogen protecting groups, exclusive
chemoselectivity was observed for reaction with rhodium at the
acetanilide in the presence of the N-Boc-aniline, two functional
groups with significant structural similarity. Following chro-
matographic purification, and with the indole core set, the final
sequence of reactions was carried out in three steps. Treatment
of 11 with K2CO3 in MeOH/DCM followed by acidic (TFA)
cleavage of the Boc group affords the unprotected indole and
aniline. Exposure of the crude product to basic (DBU) conditions
effects an intramolecular amide bond formation, affording
paullone 3 in 33% yield from 11 and 12% overall yield from
commercially available compounds in eight steps using only
three chromatographic purifications (Scheme 2).
Mechanistic Investigation. A general catalytic cycle, under
our first-generation conditions, is shown in Scheme 3 based on
our previous preliminary mechanistic investigation10b and the
proposals put forth by Larock5g as well as Satoh and Miura for
other annulative processes with internal alkynes.9 Jones has
revealed that the active metalating agent in the stoichiometric
cyclometalation of benzylimines with [Cp*RhCl2]2 1 was the
cationic complex I (Scheme 3).19 Given the similarity of our
reaction and the large excess of acetate in our first-generation
conditions, we propose a similar scenario for this reaction.
Complex I is presumably generated from the combination of
catalyst precursor 1, AgSbF6, and acetate (from Cu(OAc)2) and,
based on our previous work, reversibly metalates acetanilide to
form a cyclometalated intermediate. Alkyne coordination to
rhodium and 1,2-migration of the rhodium-carbon bond across
the alkyne results in the formation of the requisite carbon-carbon
bond and a six-membered metallacycle. The carbon-nitrogen
bond of the indole product is formed after reductive elimination,
at which time the metal catalyst is reduced to Rh(I). Reoxidation
of the reduced catalyst with the copper(II) oxidant restores the
catalytically active rhodium(III)-complex (Scheme 3).
a Conditions: (a) 9 (1 equiv), ethyl diazoacetate (1.05 equiv, 0.8 M),
CuI (10 mol %), MeCN, r.t., 12 h; 70%. (b) 4a (1 equiv, 0.2 M), 10 (1.1
equiv), 2 (5 mol %), Cu(OAc)2 ·H2O (20 mol %), O2 (atm), t-AmOH, 60
°C, 20 h; 71%. (c) 11 (1 equiv), K2CO3 (10 equiv), MeOH: DCM (1: 1),
r.t., 12 h; 97%. (d) TFA (0.1 M), r.t., 30 min; 94%. (e) DBU (1.2 equiv),
DMF (0.1 M), 150 °C, 19 h; 36% (33% over three steps).
activity,26 in vitro antiproliferative activity,27 and antileishmanial
activity,28 and to inhibit glycogen synthase kinase-3 (GSK-3),
which may find use in the treatment of diabetes.29 As such,
there has been a significant effort directed toward efficient
syntheses of this class of compounds.30
The majority of previous routes toward paullone (or its
derivatives) are initiated with a preformed indole which is
therein functionalized during the synthesis.30a,b,d,e Our group’s
interest in the synthesis of heterocycles via reactions that take
place at C-H bonds and the lack of synthetic routes for the
construction of the indole from simple starting materials
prompted us to develop an efficient route to paullone 3,
employing the oxidative coupling of acetanilide 4a and an
appropriately functionalized alkyne (Scheme 2). There was
sufficient literature precedent from previous paullone syn-
(26) Schultz, C.; Link, A.; Leost, M.; Zaharevitz, D.; Gussio, R.; Sausville,
E. A.; Meijer, L.; Kunick, C. J. Med. Chem. 1999, 42, 2909.
(27) Kunick, C.; Schultz, C.; Lemcke, T.; Zaharevitz, D.; Gussio, R.; Jalluri,
R. K.; Sausville, E. A.; Leost, M.; Meijer, L. Bioorg. Med. Chem.
Lett. 2000, 10, 567.
(28) Reichwald, C.; Shimony, O.; Dunkel, U.; Sacerdoti-Sierra, N.; Jaffe,
C. L.; Kunick, C. J. Med. Chem. 2008, 51, 659.
(29) Stukenbrock, H.; Mussmann, R.; Geese, M.; Ferandin, Y.; Lozach,
O.; Lemcke, T.; Kegel, S.; Lomow, A.; Burk, U.; Dohrmann, C.;
Meijer, L.; Austen, M.; Kunick, C. J. Med. Chem. 2008, 51, 2196.
(30) For previous syntheses of paullone and its derivatives, see: (a) Baudoin,
O.; Cesario, M.; Gue´nard, D.; Gue´ritte, F. J. Org. Chem. 2002, 67,
1199. (b) Bremner, J. B.; Sengpracha, W. Tetrahedron 2005, 61, 5489.
(c) Henry, N.; Blu, J.; Be´ne´teau, V.; Me´rour, J.-Y. Synthesis 2006,
3895. (d) Avila-Za´rraga, J. G.; Lujan-Montelongo, A.; Covarrubias-
Zu´n˜iga, A.; Romero-Ortega, M. Tetrahedron Lett. 2006, 47, 7987.
(e) Joucla, L.; Popowycz, F.; Lozach, O.; Meijer, L.; Joseph, B. HelV.
Chem. Act. 2007, 90, 753. (f) Tobisu, M.; Fujihara, H.; Koh, K.;
Chatani, N. J. Org. Chem. 2010, 75, 4841.
9
18332 J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010

Rhodium(III)-Catalyzed Formation of Indoles and Pyrroles
A R T I C L E S
Table 3. Temperature-Dependence of the Reversibility of the
Cyclorhodation
%H incorporationb
entry
conditions
temp, °C
4a-dn
6a-dn
1
2
A
B
B
B
120
90
60
75
20
0
27
4
0
3
4c
60
>90
s
Figure 2. Kinetic plot for rate dependence on the concentration of 2.
Conditions: 4a (1 equiv, 0.15 M), 5a (1.1 equiv), 2 (0-0.018 M),
Cu(OAc)2 ·H2O (2.1 equiv), trimethoxybenzene (1 equiv) as internal
standard, t-AmOH, 60 °C, 80 s.
a Conditions: A: 4a-d5 (1 equiv, 0.2 M), 5a (1.1 equiv), 1 (2.5 mol
%), AgSbF6 (10 mol %), Cu(OAc)2 ·H2O (2.1 equiv), t-AmOH, 120 °C,
4 min. B: 4a-d5 (1 equiv, 0.2 M), 5a (1.1 equiv), 2 (5 mol %),
Cu(OAc)2 ·H2O (20 mol %), O2 balloon, t-AmOH, 60 or 90 °C, 5 h.
b Reactions were stopped at ∼50% conversion, and % H incorporation
was determined by 1H NMR spectroscopy on pure isolated material.
c No alkyne 5a added to reaction mixture.
conjunction with other experimental data obtained,31 is in accord
with previous experimental19,32 and computational findings33
for a CMD mechanism in other cyclometalation reactions.34
We have identified the current limitations of the more mild
second-generation conditions to include the poor reactivity of
alkyl/alkyl-substituted internal alkynes, very low regioselectivity
for unsymmetrical alkyl/alkyl-substituted internal alkynes, and
reduced regioselectivity of meta-substituted acetanilides relative
to the first-generation conditions. Subsequent to the development
of our second-generation conditions, the mechanism of the
reaction was explored in order to verify our initial proposal and
address these limitations. We primarily employed kinetic
analysis of the reagents and isotope labeling experiments to gain
insight into these aspects of the chemistry and, where applicable,
to aid in the development of improved reaction conditions for
these processes.
Determination of the Order of Reagents. Our kinetic analysis
commenced with the determination of the rate order of each of
the primary reaction components (catalyst, acetanilide, and
alkyne) under the second-generation conditions.35 The reaction
conditions used in the kinetic analysis were a slight modification
of the second-generation conditions such that 2 equiv of the
Cu(OAc)2 ·H2O oxidant was employed in place of 20 mol %
Cu(OAc)2 ·H2O and an oxygen atmosphere.36 First-order de-
pendence of the reaction rate on rhodium concentration was
obtained from the linear plot of initial rate (∆[6a]/∆t) vs initial
rhodium concentration ([2]) (Figure 2). Saturation kinetics were
observed with respect to acetanilide concentration when similar
experiments were carried out by measuring the initial rate over
a range of 4a concentrations (Figure 3). These combined data
advocate that the acetanilide reversibly binds to rhodium, most
likely via its Lewis basic amide oxygen, before the rate-
determining step, and are consistent with the catalyst and
acetanilide being saturated as a coordination complex prior to
irreversible and rate-determining C-H bond cleavage.
Reversibility of Acetanilide Cyclorhodation and the Pre-
sence of a Primary DKIE. Our previous study10b found that,
under the first-generation conditions, the acetanilide cyclorho-
dation was a reversible process, resulting in 75% loss of
deuterium exclusively ortho to the amide using 4a-d5 (Table 3,
entry 1). These results suggested that the metalation is directed
by the Lewis basic amide group and are not in line with
traditional Friedel-Crafts chemistry. Further experiments were
carried out to determine if this was also true of the more mild
second-generation conditions. Indeed, it was found that the
cyclorhodation was intrinsically reversible in the absence of
alkyne under the second-generation conditions (Table 3, entry
4). However, with all reaction components present, the meta-
lation of acetanilide was found to be irreversible, with no loss
in deuterium from 4a-d5 (Table 3, entry 3). Interestingly, a trend
of decreasing reversibility of metalation (H incorporation) with
decreasing temperature was observed (Table 3, entries 1-3).
The irreversibility of the C-H bond functionalization event
under the more mild conditions prompted us to investigate the
presence of a deuterium kinetic isotope effect (DKIE). The
average initial rate of the reaction was 3.1 ((0.5) × 10-4 M/s
using acetanilide 4a and 0.74 ((0.09) × 10-4 M/s using
acetanilide 4a-d5. A comparison of these initial rates revealed
a primary DKIE of 4.2 ( 0.9, suggesting that the rate-
determining step involved rhodium-catalyzed C-H bond cleav-
age (eq 5). Additionally, a DKIE of this magnitude is consistent
with a concerted metalation-deprotonation (CMD) mechanism
being operative in this process. Furthermore, this data, in
(31) A Hammett correlation and an Arrehnius plot are also provided in
the Supporting Information.
(32) Ryabov, A. D.; Sakodinskaya, I. K.; Yatsimirsky, A. K. J. Chem. Soc.,
Dalton Trans. 1985, 2629.
(33) (a) Davies, D. L.; Donald, S. M. A.; Macgregor, S. A. J. Am. Chem.
Soc. 2005, 127, 13754. (b) Davies, D. L.; Donald, S. M. A.; Al-Duaij,
O.; Macgregor, S. A.; Po¨lleth, M. J. Am. Chem. Soc. 2006, 128, 4210.
(34) For a historical perspective on the evolution of the CMD mechanism
in the chemical literature, including examples involving cyclometa-
lation reactions, see: Lapointe, D.; Fagnou, K. Chem. Lett. 2010, 39,
1118.
(35) See the Supporting Information for exact details of data collection
for the kinetic experiments.
(36) This modification to the second-generation reaction conditions was
done for practical purposes of data collection and determined to be
insignificant to initial rate determinations because under a relevant
concentration range (0.2-2.1 equiv) of Cu(OAc)2 ·H2O, the reaction
rate exhibited zero-order dependence on [Cu(OAc2)]; see Supporting
Information for more details.
9
J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010 18333

A R T I C L E S
Stuart et al.
Table 4. DKIEs for a Range of Alkynes and Alkyne Concentrations
entry
alkyne
[alkyne ]
DKIE
product
1
2
3
4
5
5a
5b
5c
5a
5a
0.17 M
0.17 M
0.17 M
0.05 M
0.80 M
4.2 ( 0.9
3.6
3.5
3.5 ( 0.2
3.0 ( 0.5
6a
7b
7c
6a
6a
a Conditions: 4a or 4a-d5 (1 equiv, 0.15 M), alkyne (see above for
concentration), 2 (5 mol %), Cu(OAc)2 ·H2O (2.1 equiv), t-AmOH, TMB
) trimethoxybenzene (1 equiv) as internal standard, 60 °C, 80 s.
Figure 3. Kinetic plot for rate dependence on the concentration of 4a.
Conditions: 4a (0-0.55 M), 5a (1 equiv, 0.17 M), 2 (5 mol %),
Cu(OAc)2 ·H2O (2.1 equiv), trimethoxybenzene (1 equiv) as internal
standard, t-AmOH, 60 °C, 80 s.
served for alkynes 5a and 5b, with alkyne 5a inhibiting the
reaction rate to a slightly greater extent at high alkyne
concentration (0.8 M). However, alkyne 5c, while also having
a sharp increase in rate at very low alkyne concentrations (<0.04
M), displayed a more drastic rate inhibition at higher alkyne
concentrations (>0.08 M, Figure 5). The presence of a primary
DKIE (4.2 ( 0.9 at 0.15 M acetanilide, eq 5) was used to probe
the unusual observed rate dependence on alkyne concentration
and the fidelity of the rate-determining step over alkynes 5a,
5b, and 5c, under relevant concentrations (Table 4). These
experiments demonstrate that the rate-determining step is
invariable for the reactions with these alkynes due to a consistent
and large primary DKIE of ∼3.6 (Table 4, entries 1-3).
Additionally, over the alkyne concentrations used in our kinetic
study, there is little variation in the DKIE, which remains
significantly large (Table 4, entries 1, 4, and 5). This suggests
that the nature and mechanism of the rate-determining step of
the reaction is unchanged over these alkyne concentrations and
remains the C-H bond cleavage step.
Figure 4. Kinetic plots for rate dependence on the concentration of 5a.
Conditions: 4a (1 equiv, 0.15 M), 5a (0-0.8 M), 2 (5 mol %),
Cu(OAc)2 ·H2O (2.1 equiv), trimethoxybenzene (1 equiv) as internal
standard, t-AmOH, 60 °C, 80 s.
Within the catalytic cycle, 1,2-migratory insertion of the
rhodium-carbon arene bond across the alkyne takes place after
the rate-determining C-H bond cleavage step, and therefore,
in this case, zero-order dependence of the reaction rate on alkyne
concentration would be expected (Scheme 3). However, the
observed kinetic data with respect to alkyne concentration was
certainly not zero order and indicated a noninteger dependence
of the reaction rate on alkyne concentration suggesting an
additional role for this component prior to the rate-determining
step. Kinetic profiles of this type have been observed in other
transition metal and organocatalyzed processes38 and have been
attributed to an inhibitory binding event of the reagent in
question (alkyne in this case) to the active catalyst prior to the
rate-determining step. Cognizant of these precedents, we propose
a similar scenario for this reaction in which the alkyne reversibly
coordinates the active catalyst I, effectively occupying the free
coordination site necessary for C-H bond cleavage (Scheme
4). The shape of the curves in Figure 5 also indicate that the
stoichiometry of the rhodium-alkyne complexes formed are
not limited to a 1:1 complex and may differ for the three alkynes
examined.
Figure 5. Normalized kinetic plots for the rate dependence on the
concentration of alkynes 5a-c.
A sharp increase in reaction rate followed by a region of rate
inhibition at high alkyne concentrations was observed when a
similar analysis was carried out over a range of alkyne 5a
concentrations (0-0.80 M, Figure 4). Given this unanticipated
result, alkynes 5b (diphenylacetylene) and 5c (4-octyne) were
subjected to analogous inspection in order to determine if this
effect was a general phenomenon common to other alkynes over
the same concentration range. Figure 5 shows a plot of
normalized initial rate vs alkyne concentration for alkynes 5a,
5b, and 5c.37 While the general behavior of a sharp rate increase
followed by a region of rate inhibition is consistent across the
three alkynes, the magnitude of rate inhibition and shape of the
curves varies significantly. Similar kinetic behavior was ob-
Interestingly a one-pot competition experiment between the
two alkynes at either end of the spectrum of rate inhibition (5b
(37) The lines drawn in Figure 5 are done in order to guide the reader’s
eye and do not represent a mathematical correlation.
(38) (a) Steinhoff, B. A.; Guzei, I. A.; Stahl, S. S. J. Am. Chem. Soc. 2004,
126, 11268. (b) Schultz, M. J.; Adler, R. S.; Zierkiewicz, W.; Privalov,
T.; Sigman, M. S. J. Am. Chem. Soc. 2005, 127, 8499. (c) Zuend,
S. J.; Jacobsen, E. N. J. Am. Chem. Soc. 2007, 129, 15872.
9
18334 J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010

Rhodium(III)-Catalyzed Formation of Indoles and Pyrroles
A R T I C L E S
Scheme 4. Proposed Source of Rate Inhibition by Alkynes
ally, under the optimized reaction conditions, the formation of
12 was typically less than 2%.
Revised Catalytic Cycle. The kinetic data obtained for this
system is most consistent with the catalytic cycle depicted in
Scheme 5. As previously postulated, complex 2 reacted with 1
equiv of acetate to form the active metalating agent I. However,
I may react with alkyne 5a in an inhibitory manner to form
nonproductive coordination complexes Ia and Ib. The active
catalyst I may also be coordinated by 4a via the Lewis basic
amide oxygen to form the coordination complex II which
undergoes irreversible and rate-determining C-H bond cleavage
producing III. Alkyne 5a may then coordinate III, followed
by carbo-rhodation to yield V. In the presence of adventitious
acid, V may react to form the hydroarylation product 12;
however, this is a minor reaction pathway under our optimized
conditions. C-N bond reductive elimination extrudes the desired
indole product 6a along with Cp*Rh(I). The reduced catalyst
is then oxidized back to the active species via catalytic copper(II)
acetate which is then reoxidized by the combination of molecular
oxygen and acetic acid (generated in the reaction). Having
established a viable catalytic cycle consistent with our kinetic
data, we set out to investigate specific aspects of the reaction
that we deemed were limitations of the chemistry, namely the
regioselectivity of metalation of meta-substituted acetanilides
andtheregioselectivityofalkyneinsertionintotherhodium-carbon
bond of complex III.
and 5c) revealed a preference for migratory insertion across the
diaryl alkyne 5b over the dialkyl alkyne 5c (eq 6). This
competition experiment removes any ambiguity due to rate
inhibition and directly demonstrates the preference for migratory
insertion across the diaryl alkyne despite the stronger interaction
between rhodium and the dialkyl alkyne implicit from Figure
5. This is consistent, though to a lesser degree, with a previous
observation by Rovis for a related Rh(III)-catalyzed annulation
of internal alkynes in which exclusive reaction with dipheny-
lacetylene was observed in the presence of an equimolar quantity
of 5-decyne.14c This observation corroborates the empirically
observed lower yields for dialkyl alkynes in these transforma-
tions due to a competitive off-cycle inhibitory pathway that is
much more pronounced for this class of alkyne in comparison
to diaryl alkynes.
Studies on the Regioselectivity of C-H Bond Func-
tionalization Leading to Improved Conditions for Meta-
Substituted Acetanilides. The first rhodium-carbon bond for-
mation, resulting in III, was subject to issues of regioselectivity
when meta-substituted acetanilides were used. This step con-
sistently takes place at the more sterically accessible C-H bond
of meta-substituted acetanilides regardless of the electronic
character of the aromatic ring (Chart 1, entries 9-12). During
our studies on the compatibility of the reaction conditions with
various acetanilides, 3,5-dimethylacetanilide was found to be a
poor substrate, resulting in only 7% isolated yield of the
corresponding indole. Additionally, 3-methylacetanilide, 4k,
provided 6-methylindole, 6k, in good yield as a single regioi-
somer, therefore implying that the acetanilide metalation step
was extremely sensitive to steric effects. In a comparison of
the first- and second-generation conditions with respect to
regioselectivity at the acetanilide, indole 6j was formed as a
9:1 mixture of regioisomers under the first-generation conditions,
whereas the regioselectivity was reduced to 6:1 under the
second-generation conditions. This result was consistent with
the observed lack of reversibility of acetanilide metalation under
the more mild temperatures of the second-generation conditions
(see Table 3). In an effort to improve the site selectivity of
acetanilide metalation under these mild conditions, we formally
investigated the influence of temperature (Table 5) and alkyne
concentration (Table 6) on regioselectivity. In line with our
previous observations we found a small but consistent increase
Isolation and Characterization of a Reaction Side Product.
A significant amount of a side product was obtained when
chlorinated solvents39 were used during the optimization of our
second-generation conditions. Isolation and structural charac-
terization of the compound revealed it to be the hydroarylation
product 12. Intrigued by the fact that this trisubstituted alkene
is produced as a single regio- and stereoisomer, we found that
by removal of the oxidant and addition of a carboxylic acid40
(5 equiv), modest yield of the hydroarylation product could be
obtained (eq 7). This information assisted in the development
of general conditions for the hydroarylation of unactivated
alkynes with arenes and heteroarenes catalyzed by 2.16 While
the observation of 12 is further evidence that C-C bond
formation precedes C-N bond formation, consistent with our
proposed catalytic cycle, when 12 was subjected to our second-
generation reaction conditions, none of the desired product 6a
was obtained. This result indicated that 12 is a nonproductive
side product and not an intermediate en route to the desired
indole, also consistent with our mechanistic proposal. Addition-
(39) The greater formation of the side product in chlorinated solvents is
attributed to the presence of small amounts of HCl due to decomposi-
tion of the solvent.
(40) The pKa of the carboxylic acid was found to be critical to formation
of 12. Acids with pKa < 1 resulted in very low conversion and a
significant amount of 4a. Acids with pKa > 1 provided 12 in similar
yield with pivalic acid giving optimal results; see the Supporting
Information for further details.
9
J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010 18335

A R T I C L E S
Stuart et al.
Scheme 5. Revised Catalytic Cycle
Table 5. Influence of Temperature on Regioselectivity of Indole
Formation with 4j
Table 6. Influence of Alkyne Concentration on Regioselectivity of
Indole Formation with 4j
entry
temp (°C)
1H NMR yield (%)b
6j (6-OMe:4-OMe)b
entry
initial [5a] (M)
1H NMR yield (%)b, c
6j (6-OMe:4-OMe)b
1
2
3
4
5
40
50
60
70
80
91
97
99
94
89
5.0:1
5.5:1
6.1:1
6.5:1
7.2:1
1
2
3
4
5
0.04
0.10
0.23
1.00
2.00
17(85)
46(93)
89
84
84
8.7:1
7.4:1
7.2:1
3.0:1
2.4:1
a Conditions: 4j (0.1 mmol, 0.2 M, 1 equiv), 5a (0.11 mmol, 1.1
equiv), 2 (5 mol %), Cu(OAc)2 ·H2O (2.1 equiv), t-AmOH, temperature
(see above). b Yield and ratio of mixture of 6-OMe and 4-OMe
regioisomers determined by 1H NMR spectroscopy versus
1,3,5-trimethoxybenzene as an internal standard.
a Conditions: 4j (0.1 mmol, 0.2 M, 1 equiv), 5a (see above for initial
concentration), 2 (5 mol %), Cu(OAc)2 ·H2O (2.1 equiv), t-AmOH, 80
°C. b Yield and ratio of mixture of 6-OMe and 4-OMe regioisomers
1
determined by H NMR spectroscopy versus 1,3,5-trimethoxybenzene as
an internal standard. c Yield in parentheses determined by 1H NMR
spectroscopy and based on the alkyne as the limiting reagent.
in regioselectivity over a range of increasing temperatures.41
Consequently, the regioselectivity was increased from 5:1 at
40 °C to ∼7:1 at 80 °C (Table 5, compare entries 1 and 5). A
much larger correlation between regioselectivity and initial
alkyne concentration was observed when the reaction was
carried out at 80 °C (Table 6). At low alkyne concentration,
the regioselectivity is further increased to ∼9:1, while, at high
initial alkyne concentration (2M, 10 equiv of alkyne relative to
acetanilide), the regioselectivity was dramatically reduced to
2.4:1. With this knowledge in hand, when the reaction was
carried out at 80 °C with a portion-wise (0.2 equiv/∼15 min)
addition of the alkyne, an ∼10:1 ratio and an excellent combined
Studies on the Regioselectivity of the 1,2-Alkyne Migra-
tion Event. In contrast to the regioselectivity at the acetanilide,
the effect of sterics appeared to have only a modest influence
on the regioselectivity of the alkyne insertion (Chart 2, entry
3). Interestingly, the regioselectivity of this process was
opposite to that previously reported by Larock.5g In his
account, Larock proposed that an unfavorable steric interac-
tion between the larger alkyne substituent (RL) and the aryl
1
yield of 96% (by H NMR) for the mixture of regioisomers
was obtained (eq 8).42
(41) This is opposite to what was previously observed by Jones for the
stoichiometric cyclorhodation of benzylimines; see ref 19.
9
18336 J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010

Rhodium(III)-Catalyzed Formation of Indoles and Pyrroles
A R T I C L E S
Scheme 6. Rationale for Regioselectivity of Alkyne Insertion Based
on Steric Effects
Pyrrole Formation: Reaction Development. Despite the wide
applicability of pyrroles in pharmaceuticals4,44 and biologically
active natural products1c and materials,2 methods for their syntheses
have not enjoyed the same level of ingenuity as those for their
heterocyclic counterpart, the indole. Our development of mild
conditions for the efficient synthesis of indoles and the broad
substrate scope displayed by this reaction prompted us to explore
the extension of this method to vinylic C-H bond functionalization
in the context of a pyrrole synthesis.12 A screen of a number of
different enamides identified 16a as a competent substrate for
reaction development. When coupled with 1-phenyl-1-propyne 5a
under the reaction conditions described in entry 6 of Table 1, the
desired pyrrole 17a was obtained in 93% GCMS yield (eq 10).
Encouraged by this result, the pyrrole formation was tested under
the room temperature reaction conditions (21 °C in acetone as
solvent; Table 1, entry 9), and, in this case, an 87% GCMS yield
and 81% isolated yield was obtained after 16 h (eq 10). Further-
more, this reaction may be performed on gram-scale (1 g of 16a),
under reduced catalyst loading (3 mol % of 2) with no loss in yield
(80% isolated yield of 17a).23 Therefore, these two sets of
conditions were used to test the generality of the reaction, primarily
employing the room temperature conditions.
ring forced this substituent proximal to the indole nitrogen
(Scheme 6). Whereas, in the reaction described in this report,
we propose that the bulky pentamethylcyclopentadienyl (Cp*)
ligand on rhodium forces the larger group to be distal to the
indole nitrogen (Scheme 6). This result is consistent with
the observed regioselectivity when unsymmetrical alkynes
have been employed in other rhodium-catalyzed annulations
with Cp*Rh(III)-catalysts.13b,14a,c
The electronic effects of the alkyne substituents appear to
have more influence over the regioselectivity of alkyne
insertion than do the steric effects. This is particularly true
when these two substituents have disparate electronic proper-
ties. Alkyne 13, having two electronically distinct aryl groups,
was prepared and subjected to the second-generation condi-
tions in order to demonstrate the electronic bias for alkyne
insertion (eq 9). The resulting mixture of indole regioisomers
could not be separated by conventional chromatography;
however, benzene-d6 provided sufficient resolution in the 1H
NMR spectrum to allow for characterization of the mixture
by 1D and 2D NMR methods.23 The mixture of isomers was
formed in a 2.7:1 ratio in favor of the electron-deficient aryl
group proximal to the indole nitrogen.43 In this specific case
of alkyne 13, adorned with two electronically dissimilar aryl
substituents, the more electron-deficient substituent is deliv-
ered proximal to the indole nitrogen. However, this trend
does not extend to the coupling of ethyl 3-phenylpropiolate
5k in which the more electron-withdrawing alkyne substituent
(CO2Et) is delivered distal to the indole nitrogen. Other
interactions of the alkyne substituents with the catalyst may
be operative in the case of 5k (and related alkynes),
overriding any inherent electronic bias of the alkyne insertion
event.
Pyrrole Formation: Reaction Scope and Limitations. While
other processes of this nature have displayed a relatively limited
substrate scope with respect to the alkyne,12 the very mild reaction
conditions developed here allow a wide variety of unactivated
internal alkynes to be employed. Symmetrical diaryl and dialkyl
internal alkynes reacted smoothly to afford pyrroles in moderate
to excellent yields (Chart 3, entries 2 and 3). While a moderate
56% yield of 2,3-dialkyl-substituted pyrrole 17c was obtained at
room temperature, the yield was increased to a more synthetically
useful 83% by employing conditions at a slightly higher temper-
ature (60 °C in t-AmOH). In addition to 1-phenyl-1-propyne 5a
other alkyl/aryl unsymmetrical internal alkynes reacted well to give
(42) A method involving syringe pump addition of the alkyne was explored
but was complicated by the small scale of the reaction. Therefore,
syringe pump addition was abandoned for the sequential addition
protocol.
(43) In the pyrrole forming reaction, a mixture of regioisomers in a 3.5:1
ratio is observed with the same selectivity as that observed in the
indole-forming reaction.
(44) Fensome, A.; et al. J. Med. Chem. 2005, 48, 5092.
9
J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010 18337

A R T I C L E S
Stuart et al.
Chart 3. Enamide and Alkyne Compatibility in the Formation of Pyrroles
a Conditions (at 60 °C; red): Enamide (1 equiv 0.2 M), alkyne (1.1 equiv), 2 (5 mol %), Cu(OAc)2 ·H2O (20 mol %), O2 balloon, t-AmOH, at 60 °C, 4 h.
b Conditions (at 21 °C, blue): as above, in acetone, 21 °C, 16 h. c Isolated yields are reported above. d See Supporting Information for details of TMS
cleavage. e 2 equiv of enamide used to 1 equiv of alkyne.
trisubstituted pyrroles with excellent levels of regioselectivity for
the aryl group proximal to the pyrrole nitrogen (Chart 3, entries
4-10). Notable among these examples was the variety of functional
groups that were tolerated on the alkyl substituent of the alkyne.
Pyrroles with a C3 ester (Chart 3, entries 6 and 7), C3 benzyl ester
(Chart 3, entry 8), benzyl silyl ether (Chart 3, entries 5 and 9), and
benzyl amine protected as a phthalimide (Chart 3, entry 10) were
obtained in moderate to good yields under the reaction conditions.
While the enamide scope was limited at this stage, initial examples
demonstrate that aryl and heteroaryl enamides were also compatible
resulting in biaryl and heterobiaryls (Chart 3, entries 13-15);
additionally, the use of N-vinylacetamide provided direct access
to 2,3-substituted pyrroles, albeit in low yield (Chart 3, entry 16).
The synthetic utility of the reaction was further qualified by the
use of TMS-substituted alkynes, which allowed for the incorpora-
tion of a silyl group onto the pyrrole (Chart 3, entry 11 and 12).
Manipulation of these groups, as well as the methyl ester (in
compounds 17a-l), ultimately provided access to a number of other
pyrrole substitution patterns (vide infra).23
Synthetic Applications: Strategies To Access Multiple
Pyrrole Substitution Patterns. The high functional group compat-
ibility of the pyrrole forming reaction from enamides and alkynes
directly lead to the ability to access a variety of pyrrole substitution
patterns. Namely, the incorporation of trimethylsilyl and ester func-
tionality into the alkyne and enamide, respectively, allowed for facile
and high yielding removal of these groups after pyrrole formation
occurred. Scheme 7 provides conditions for the removal of these
groups as well as representative examples of the possible pyrrole
substitution patterns that may be obtained.
Removal of the methyl ester in 17a may be accomplished in
high yield (86%) over two steps involving saponification of the
ester using LiOH followed by TFA-mediated decarboxylation. This
9
18338 J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010

Rhodium(III)-Catalyzed Formation of Indoles and Pyrroles
A R T I C L E S
Scheme 7. Strategies To Access Multiple Pyrrole Substitution Patternsa
a Conditions: (a) 2 (5 mol %), Cu(OAc)2 ·H2O (20 mol %), O2 balloon, acetone, 21 °C, 16 h. (b) 1 N HCl(aq), THF 60 °C, 17 h. (c) LiOH ·H2O (10 equiv,
THF:H2O (1:1), 60 °C, 17 h. (d) TFA (0.1 M), 0-21 °C, 40 min to overnight. (e) TBAF (1.2 equiv), THF, 21 °C, 18 h. (f) 2 N HCl (in Et2O), 21 °C, 18 h.
process was used in the formation of pyrroles 18b, 18d, and 18e,
and, in all cases, chromatographic purification is only required on
the final product. It should also be noted that in these cases under
the basic conditions needed for ester saponification, the N-acetyl
group was also cleaved. The C2-TMS group may be cleaved from
the crude pyrrole product by exposure to TBAF (1 M in THF) to
yield the C2-unsubstituted pyrrole 17k. The two-step protocol of
oxidative coupling/TMS-cleavage was performed in 55% overall
yield (one chromatographic purification) (Scheme 7). These condi-
tions were, however, unsuitable for cleavage of the C3-TMS group
of 17l. In this case, anhydrous HCl (2 N in Et2O) is necessary and
the TMS group was cleaved in near quantitative yield without the
need for chromatographic purification, providing pyrrole 18c
(Scheme 7). It should be noted that while exposure to TBAF or
aqueous HCl will cleave the N-acetyl moiety, anhydrous HCl leaves
this group intact, providing orthogonal methods for TMS group
cleavage.
Given the operationally simple and high yielding methods
outlined above, six of the nine possible pyrrole substitution patterns
may be accessed as 2,3,5-, 2,3-, 2,4-, 2,5-, 2-, and 3-substituted
pyrroles in good overall yields. Compound 18e is an interesting
example demonstrating the power of this method. Its synthesis has
only been reported on one other occasion in the literature and was
achieved in five steps in 25% overall yield from a preformed
pyrrole.45 The inherent challenge in the preparation of this class
of compound directly from pyrrole derives from the need to alkylate
the least nucleophilic position of the heterocycle. Additionally, the
use of primary alkyl halides in Friedel-Crafts chemistry typically
suffer from carbocation rearrangements. Thus, this method allows,
among other things, the construction of C3-alkylated pyrroles from
simple and readily available starting materials in good yield and
in an expedient manner.
ium(III)-catalyzed C-H bond functionalization event of arenes and
alkenes. The reactions feature good substrate scope with respect
to both coupling partners and employ molecular oxygen as the
stoichiometric terminal oxidant. A mechanistic investigation of the
indole-forming reaction established a viable catalytic cycle that was
consistent with the kinetic data obtained and revealed an inhibitory
effect at high alkyne concentrations that was largest for alkyl/alkyl-
substituted alkynes. Additionally, the regioselectivity of the C-H
bond cleavage was found to be dependent on both temperature
and alkyne concentration and thus enabled the development of
improved conditions for enhanced regioselectivity of meta-
substituted acetanilides. The mechanism of the C-H bond cleavage
was investigated, and the evidence gathered points toward an
irreversible and rate-determining CMD transition state. Finally, the
utility of these protocols were demonstrated by the facile and
expedient synthesis of paullone, a compound with known bioac-
tivity, and the ease with which a number of pyrrole substitution
patterns may be accessed from the cleavage of common functional
groups. Given these points, the reactions developed here should
find widespread use in synthetic organic chemistry, and the
mechanistic insights garnered should aid in the future development
of rhodium(III)-catalyzed heterocycle syntheses.
Acknowledgment. We thank NSERC, the Research Corporation,
the Sloan Foundation, the ACS (PRF AC), Merck Frosst, Merck Inc.,
Amgen, Eli Lilly, Astra Zeneca, and Boehringer Ingelheim for financial
support. D.R.S. thanks NSERC for a postgraduate scholarship (PGS-
D). Professors Tom Rovis, Laurel Schafer, and Andre´ Beauchemin
are acknowledged for insightful comments and assistance during the
preparation of this manuscript. Dr. Benoˆit Lie´gault and Sophie
Rousseaux are acknowledged for critical reading of the manuscript.
Conclusion
Supporting Information Available: Detailed experimental
procedures, characterization data for all new compounds, and
complete ref 44. This material is available free of charge via
the Internet at http://pubs.acs.org.
We have developed a novel and mild method for the construction
of highly functionalized indoles and pyrroles based on a rhod-
(45) Reaxys search of compound 18e provided only one hit: Garrido,
D. O. A.; Buldain, G.; Ojea, M. I.; Frydman, B. J. Org. Chem. 1988,
53, 403.
JA1082624
9
J. AM. CHEM. SOC. VOL. 132, NO. 51, 2010 18339