Catalytic C-H bond functionalization with palladium(II): Aerobic oxidative annulations of indoles

Article abstract of DOI:10.1021/ja035054y

A palladium-catalyzed aerobic oxidative annulation of indoles is described. We have demonstrated that a variety of factors influence these cyclizations, and in particular the electronic nature of the pyridine ligand is crucial. It is also remarkable that these oxidative cyclizations can proceed in good yield despite background oxidative decomposition pathways, testament to the facile nature with which molecular oxygen can serve as the direct oxidant for Pd(0). We have also shown that the mechanism most likely involves initial indole palladation (formal C-H bond activation) followed by migratory insertion and β-hydrogen elimination. Copyright

Full text of DOI:10.1021/ja035054y

Published on Web 07/22/2003
Catalytic C-H Bond Functionalization with Palladium(II): Aerobic Oxidative
Annulations of Indoles
Eric M. Ferreira and Brian M. Stoltz*
DiVision of Chemistry and Chemical Engineering, California Institute of Technology, Pasadena, California 91125
Received March 7, 2003; E-mail: stoltz@caltech.edu
Table 1. The Effect of Pyridine Substitution on the Aerobic Indole
The challenge of direct C-H bond functionalization continues
Annulation
to be an area of intense interest among academic and industrial
chemists.1 Catalytic transformations of this type involving the indole
nucleus would be of particular value to the synthetic chemist.
Palladium-mediated indole annulations have been used as the key
step in several total syntheses; however, these reactions have
exclusively required stoichiometric quantities of palladium.2 Ad-
+
entry
pyridine ligand
pK (pyrH )
conversion (%)a
a
ditionally, both Fujiwara and Itahara have studied related alken-
ylations and arylations of a variety of arenes including indoles.3
Despite the obvious utility of these indole transformations, variants
employing substoichiometric quantities of palladium have been
highly limited.4 An ideal oxidation catalyst system would employ
an inexpensive, readily abundant terminal oxidant such as molecular
oxygen. Although catalytic Pd(II) oxidations using O2 as the sole
stoichiometric oxidant are known, they are rare. In one example,
Uemura found that pyridine-ligated palladium catalysts are active
for alcohol oxidations under O2.5 Since that key discovery, the Pd-
pyridine system has been utilized for several reactions, including
oxidative ring-cleavage,6a oxidative amination,6b and Wacker-type
cyclizations.6c Additionally, reports by our group and Sigman
demonstrate that this system can be easily modified with chiral
ligands to catalyze highly enantioselective processes.6c,7 The Pd-
pyridine system, however, has never been utilized in oxidative C-C
bond-forming reactions. Intrigued by the potential of this oxidation
system as an entry into indole annulation, we chose to investigate
the aerobic cyclization of indole 1. Herein, we report indole C-H
bond functionalizations under mild, direct dioxygen-coupled ca-
talysis. Moreover, this system has provided access to annulated
indoles in high yields, previously unattainable via precedented
palladium-catalyzed oxidation methods.
To our delight, treatment of indole 1 with catalytic Pd(OAc)2
(10 mol %) and pyridine (40 mol %) under an atmosphere of oxygen
in PhCH3 resulted in oxidative cyclization to 2 (Table 1, entry 3).
The conversion (23%), however, was noticeably low relative to
that from the other oxidative processes reported under the Pd-
pyridine system. We hypothesized that the active palladium species
was not electronically suited to catalyze this transformation with
high efficiency. Consequently, we surveyed a range of electronically
differentiated pyridines with Pd(OAc)2 (Table 1), and observed a
correlation between the electronic nature of the ligand (i.e., pKa)8
and its ability to facilitate the cyclization.9 Presumably, more
electron-poor ligands resulted in a more electrophilic, and therefore
more reactive, palladium catalyst. Ligands that were too electron-
deficient, however, were unable to sufficiently ligate palladium,10
likely hampering both reactivity and Pd(0) reoxidation. Ethyl
nicotinate proved to be the maximally effective ligand for these
indole annulations (Table 1, entry 5), proving that electronic
modulation of the Pd-pyridine system can lead to altered and
unique reactivity.11
1
2
3
4
4-MeO
6.47
5.99
5.25
3.45
3
1
23
52
4-t-Bu
unsubstituted
4-CO2Et
5
3-CO2Et
3.35
76
6
7
8
9
3-COCH3
3-F
3-CN
3,5-di-Cl
3.18
2.97
1.39
0.90
58
64
55
22
a Measured using GC by consumption of 1 relative to an internal standard
(tridecane).
Table 2. Solvent Effects in the Reaction of 1 to 2
entry
solvent
% conversion
% yield (GC)
1
2
3
4
5
6
7.
8
toluene
88
85
95
94
95
86
91
99
33
40
49
53
58
25
76
82b
chlorobenzene
butyl acetate
tert-amyl alcohol
pinacolone
AcOH
pinacolone/AcOH (4:1)
tert-amyl alcohol/AcOH (4:1)
a 10 mol % Pd(OAc)2, 40 mol % ethyl nicotinate, 1 atm O2, 80 °C, 24
h, 0.1 M substrate 1 in solvent. b Isolated yield.
as we moved from nonpolar aromatic solvents toward more polar
solvents, suggesting the potential stabilization of charged intermedi-
ates in the catalytic cycle. Although the reactions proceeded to high
conversions in high-boiling polar solvents such as tert-amyl alcohol
and pinacolone, the disparity between the consumption of 1 and
the yield of 2 was still troubling.12 We reasoned that upon prolonged
exposure to O2, the indole products were oxidatively decompos-
ing.13,14 Interestingly, this problem could be minimized by the
addition of AcOH as cosolvent. Since oxidative decomposition of
indoles is known to proceed through oxygen addition at C(3), it is
possible that under our conditions, AcOH acts as a competitive
inhibitor via protonation at the C(3) position in these products.15
With this optimized system, good to excellent yields of cyclized
product were obtained (entry 8).
As shown in Table 3, a range of substituted indoles participate
in this oxidative cyclization reaction. In addition to the C(3) to
C(2) cyclizations (entries 1-10), annulation can proceed from the
C(2) to C(3) positions (entries 11 and 12) as well as from N(1) to
C(2) (entry 13). Furthermore, ring sizes of 5 and 6 (entry 10) are
available from the oxidative cyclization.
With this Pd catalyst in hand, we turned our attention to solvent
effects (Table 2). An increase in reaction efficiency was observed
Particularly interesting was the rate difference observed in the
cyclization of substrates possessing stereocenters in the tether
9
9578
J. AM. CHEM. SOC. 2003, 125, 9578-9579
10.1021/ja035054y CCC: $25.00 © 2003 American Chemical Society

C O M M U N I C A T I O N S
Table 3. Pd-Catalyzed Oxidative Indole Cyclizationa
Scheme 1
activation and uses molecular oxygen as the sole stoichiometric
oxidant. To our knowledge, this represents the first example where
the electronics of the Pd-pyridine catalytic system have been
rationally evaluated, resulting in the discovery of the unique
observed reactivity. Furthermore, this is the first report of oxidative
indole annulations carried out using catalytic palladium. It is also
remarkable that these oxidative cyclizations can proceed in good
yield despite background oxidative decomposition pathways, testa-
ment to the facile nature with which molecular oxygen can serve
as the direct oxidant for Pd(0). The discovery of specific electronic
and medium effects in this system should be applicable to numerous
other aromatic and olefinic systems, as well as asymmetric catalysis.
Studies in these areas are currently ongoing.
Acknowledgment. We thank the NIH-NIGMS (GM65961-01)
and the NSF (predoctoral fellowship to E.M.F.) for generous
financial support, and Professors Andra´s Kotschy and Greg Fu for
stimulating discussions.
Supporting Information Available: Experimental details and
characterization data for all new compounds (PDF). This material is
available free of charge via the Internet at http://pubs.acs.org.
References
(1) For a recent review, see; Ritleng, V.; Sirlin, C.; Pfeffer, M. Chem. ReV.
2002, 102, 1731-1769.
(2) (a) Trost, B. M.; Godleski, S. A.; Geneˆt, J. P. J. Am. Chem. Soc. 1978,
100, 3930-3931. (b) Cushing, T. D.; Sanz-Cervera, J. F.; Williams, R.
M. J. Am. Chem. Soc. 1993, 115, 9323-9324. (c) Baran, P. S.; Corey, E.
J. J. Am. Chem. Soc. 2002, 124, 7904-7905. (d) Baran, P. S.; Guerrero,
C. A.; Corey, E. J. J. Am. Chem. Soc. 2003, 125, 5628-5629.
(3) (a) Jia, C.; Kitamura, T.; Fujiwara, Y. Acc. Chem. Res. 2001, 34, 633-
639. (b) Itahara, T.; Ikeda, M.; Sakakibara, T. J. Chem. Soc., Perkin Trans.
1 1983, 1361-1363 and references therein.
a 10 mol % Pd(OAc)2, 40 mol % ethyl nicotinate, 1 atm O2, 80 °C, 0.1
M substrate in tert-amyl alcohol:AcOH (4:1). b Typically used as a mixture
of olefin isomers, see Supporting Information. c Isolated yields. d Product
isolated as a 58:42 mixture of E and Z isomers. e 0.1 M in pinacolone.
f 0.1 M in tert-amyl alcohol.
(4) Hegedus, L. S. Angew. Chem., Int. Ed. Engl. 1988, 27, 1113-1126.
(5) Nishimura, T.; Onoue, T.; Ohe, K.; Uemura, S. J. Org. Chem. 1999, 64,
6750-6755.
(6) (a) Nishimura, T.; Ohe, K.; Uemura, S. J. Am. Chem. Soc. 1999, 121,
2645-2646. (b) Fix, S. R.; Brice, J. L.; Stahl, S. S. Angew. Chem., Int.
Ed. 2002, 41, 164-166. (c) Trend, R. M.; Ramtohul, Y. K.; Ferreira, E.
M.; Stoltz, B. M. Angew. Chem., Int. Ed. 2003, 42, 2892-2895.
(7) (a) Ferreira, E. M.; Stoltz, B. M. J. Am. Chem. Soc. 2001, 123, 7725-
7726. (b) Bagdanoff, J. T.; Ferreira, E. M.; Stoltz, B. M. Org. Lett. 2003,
5, 835-837. (c) Jensen, D. R.; Pugsley, J. S.; Sigman, M. S. J. Am. Chem.
Soc. 2001, 123, 7475-7476.
(entries 8 and 9), coupled to the sense of diastereoinduction
observed in entry 8 and the lack of selectivity for entry 9. These
results pointed toward a mechanism wherein initial palladation at
C(2)2a is followed by olefin insertion and â-hydrogen elimination
(C-H bond functionalization). Palladation would be slowed by the
branched 3R site (e.g., entry 9), and the insertion facial bias would
be guided by the principles of acyclic stereocontrol (e.g., A(1,3)
interactions in entry 8). An alternative mechanism would involve
Pd-mediated olefin activation and subsequent attack by the indole
nucleus. To distinguish between these two possibilities, substrate
3 was exposed to our standard annulation conditions and, as
anticipated, afforded tetracycle 4 as a single diastereomer (Scheme
1).16 The observed stereochemistry of diastereomer 4 supports a
mechanism involving initial palladation, if the requirements for syn
migratory insertion and syn â-hydrogen elimination are operative.
This mechanism is in full agreement with those previously proposed
for related reactions.2,3b
(8) See Supporting Information for a full list of pKa references.
(9) Although the use of pKa values to determine ligand nature may be an
oversimplification, it has served as a useful gauge in our study.
(10) Pd-black was observed in Table 1, entries 7-9.
(11) (a) The Pd source was also found to be critical. Pd(OAc)2 was much more
effective than PdCl2 or Pd(TFA)2 for these annulation reactions. (b)
Attempts to convert 1f2 using a variety of other Pd(II) oxidative protocols
have met with minimum success, see Supporting Information.
(12) (a) GC conversion was measured by consumption of 1 relative to an
internal standard (tridecane). (b) GC yield was measured by the amount
of 2 relative to an internal standard (tridecane).
(13) There are numerous reports dealing with the oxidation of indole by O2.
For a list of key references, see Supporting Information.
(14) For control experiments, see Supporting Information.
(15) AcOH is often used as a solvent for aryl palladations.3 Thus, other possible
effects cannot be ruled out, such as favorable interactions with catalytic
intermediates. Studies elucidating the exact effect are underway.
(16) Relative configuration determined by NOE analysis, see Supporting
Information.
In conclusion, we have developed a remarkably mild oxidative
C-C bond-forming reaction that relies on formal C-H bond
JA035054Y
9
J. AM. CHEM. SOC. VOL. 125, NO. 32, 2003 9579