Allylation of N-Heterocycles with allylic alcohols employing self-assembling palladium phosphane catalysts

Article abstract of DOI:10.1021/ol800073v

The first palladium catalyst system that allows the direct allylation of indoles with allylic alcohols as substrates with water being the only byproduct is presented. The application of self-assembing ligands based on complementary hydrogen bonding was the key to success.

Full text of DOI:10.1021/ol800073v

ORGANIC
LETTERS
2008
Vol. 10, No. 6
1207-1210
Allylation of N-Heterocycles with Allylic
Alcohols Employing Self-Assembling
Palladium Phosphane Catalysts
Ippei Usui, Stefan Schmidt, Manfred Keller, and Bernhard Breit*
Institut fu¨r Organische Chemie und Biochemie, Albert-Ludwigs-UniVersita¨t Freiburg,
Albertstrasse 21, 79104, Freiburg
bernhard.breit@chemie.uni-freiburg.de
Received January 11, 2008
ABSTRACT
The first palladium catalyst system that allows the direct allylation of indoles with allylic alcohols as substrates with water being the only
byproduct is presented. The application of self-assembing ligands based on complementary hydrogen bonding was the key to success.
N-Heterocycles are important building blocks and pharma-
cophores in natural products and synthetic drugs. As such,
their synthesis and functionalization has been the subject of
numerous studies.1 However, catalytic processes which allow
the construction of new C/C bonds between heterocycles and
readily available bifunctional organic substrates which
minimize the formation of a byproduct (salt etc.) are rare.2
In this context, the transition-metal-catalyzed allylic alky-
lation could become an interesting alternative. Most fre-
quently, allyl acetates and carbonates have been used as
electrophiles in these reactions. However, use of these
substrates is associated with the disadvantage of generating
a stoichiometric amount of a couple product (carboxylate or
alcohols/alcoholates and carbon dioxide). Furthermore, these
starting materials have to be prepared in an extra step from
the corresponding allylic alcohol. Thus, ideal substrates
would be the allylic alcohols themselves, with water being
the only byproduct in this case. Only a few studies have
addressed this problem.3 However, in order to employ
N-heterocycles as nucleophiles, an additional activator (BEt3)
was required which results in the formation of boron-
containing couple products.4
Recently, self-assembling ligand systems based on comple-
mentary hydrogen-bonding have been developed in our
laboratory. This approach is intrinsically combinatorial and
has potential for high throughput screening.5 From these
studies excellent catalysts have emerged for regioselective
hydroformylation of terminal olefins,6 asymmetric hydroge-
nation with Rh catalyst,7 and Ru-catalyzed hydration of
alkynes and nitriles.8,9 The latter studies suggested that the
hydrogen-bonding network of the self-assembling ligands
might have the propensity for activation of protic nucleo-
(3) (a) Ozawa, F.; Okamoto, H.; Kawagishi, S.; Yamamoto, S.; Minami,
T.; Yoshifuji, M. J. Am. Chem. Soc. 2002, 124, 10968. (b) Ozawa, F.;
Ishiyama, T.; Yamamoto, S.; Kawagishi, S.; Murakami, H.; Yoshifuji, M.
Organometallics 2004, 23, 1698. (c) Kayaki, Y.; Koda, T.; Ikariya, T. J.
Org. Chem. 2004, 69, 2595. (d) Kinoshita, H.; Shinokubo, H.; Oshima, K.
Org. Lett. 2004, 6, 4085. (e) Utsunomiya, M.; Miyamoto, Y.; Ipposhi, J.;
Ohshima, T.; Mashima, K. Org. Lett. 2007, 9, 3371.
(4) (a) Kimura, K.; Furumata, M.; Mukai, R.; Tamaru, Y. J. Am. Chem.
Soc. 2005, 127, 4592. (b) Trost, B. M.; Quancard, J. J. Am. Chem. Soc.
2006, 128, 6314.
(1) Joule, J. A.; Mills, K. Heterocyclic Chemistry, 4th ed.; Blackwell
Science: Oxford, U.K., 2000.
(2) (a) Kakiuchi, F.; Murai, S. Acc. Chem. Res. 2002, 35, 826. (b)
Kakiuchi, F.; Chatani, N. AdV. Synth. Catal. 2003, 345, 1077. (c) Dyker,
G. Handbook of C-H actiVation Transformation; Wiley-VCH: Weinheim,
Germany, 2005; Vols. 1 and 2. (d) Ritleng, V.; Stirlin, C.; Pfeffer, M. Chem.
ReV. 2002, 102, 1731. (e) Campeau, L. C.; Fagnou, K. Chem. Commun.
2006, 1253. (f) Alberico, D.; Scott, M. E.; Lautens, M. (g) Miura, M.;
Nomura, M. Top. Curr. Chem. 2002, 219, 211.
10.1021/ol800073v CCC: $40.75
© 2008 American Chemical Society
Published on Web 02/27/2008

philes through hydrogen bonding.10 This let us speculate,
that the corresponding palladium catalysts of our self-
assembling ligand systems might also allow for activation
of unreactive allylic alcohols as substrates in the course of
the palladium-catalyzed allylic substitution. Herein, we report
on the application of self-assembling ligand-derived pal-
ladium catalysts for the allylic alkylation of indoles and
pyrroles with allylic alcohols omitting the need for additives.
In our first experiments, we explored whether the principle
of self-assembling of monodentate ligands through comple-
mentary hydrogen bonding is transferable to palladium
catalysts. Thus, [Pd(cod)Cl2] was reacted with 3-DPICon
(1)11 and 6-DPPAP (2)11 in CH2Cl2 containing residual water
to furnish the yellow crystalline palladium complex 6
(Scheme 1). From the X-ray plot of 6 depicted in Figure 1,
Figure 1. PLATON plot of [(3-DPICon‚H2O)(6-DPPAP)PdCl2]
(6) in the solid state. Selected interatomic distances (Å) and angles
(deg): Pd-P1 2.2742(6), Pd-P2 2.2715(6), O1-N22 2.863(3),
N1-O3 2.805(3), P1-Pd-P2 98.72(2), O1-H-N22 167(3), N1-
H-O3 161(2). Pd ) green, P ) orange, Cl ) yellow, O ) red, N
) blue. H atoms bound to C atoms are omitted for clarity.
Scheme 1. Self-Assembling Ligands Employed in This Study
speculated that water might be replaced in the course of a
catalytic cycle by an allylic alcohol (see Scheme 2). This
may allow for activation of the hydroxyl function to become
a better leaving group in the course of an allylic substitution
reaction.
Scheme 2. Proposed Reaction Mechanism
it is obvious that cis-coordinated 3-DPICon (1)/6-DPPAP
(2) ligands form the expected hydrogen-bonded hetero dimer
and act as a chelating ligand for the palladium metal center.
Interestingly, an additional molecule of water was incorpo-
rated in the NH-N hydrogen bond system. Thus, we
(5) For alternative approaches to self-assembling ligand/catalyst systems,
see: (a) Breit, B. Angew. Chem. 2005, 117, 6976; Angew. Chem., Int. Ed.
2005, 44, 6816. (b) Takacs, J. M.; Reddy, D. S.; Moteki, S. A.; Wu, D.;
Palencia, H. J. Am. Chem. Soc. 2004, 126, 4494. (c) Slagt, V. F.; Roeder,
M.; Kamer, P. C.; van Leeuwen, P. W. N. M.; Reek, J. N. H. J. Am. Chem.
Soc. 2004, 126, 4056. (d) Slagt, V. F.; van Leeuwen, P. W. N. M. Angew.
Chem. 2003, 115, 5777; Angew. Chem., Int. Ed. 2003, 42, 5619. (e) Slagt,
V. F.; van Leeuwen, P. W. N. M. Chem. Commun. 2003, 2474. (f) Liu, Y.;
Sandoval, C. A.; Yamaguchi, Y.; Zhang, X.; Wang, Z.; Kato, K.; Ding, K.
J. Am. Chem. Soc. 2006, 128, 14212. (g) Sandee, A. J.; van der Burg, A.
M.; Reek, J. N. H. Chem. Commun. 2007, 864. (h) Rivillo, D.; Gulyas, H.;
Benet-Buchholz, J.; Escudero-Adan, E. C.; Freixa, Z.; van Leeuwen, P. W.
N. M. Angew. Chem. 2007, 119, 7385; Angew. Chem., Int. Ed. 2007, 46,
7247. (i) Hattori, G.; Hori, T.; Miyake, Y.; Nishibayashi, Y. J. Am. Chem.
Soc. 2007, 129, 12930.
(6) Breit, B.; Seiche, W. J. Am. Chem. Soc. 2003, 125, 6608.
(7) (a) Weis, M.; Waloch, C.; Seiche, W.; Breit, B. Angew. Chem. 2007,
119, 3097. Angew. Chem., Int. Ed. 2007, 46, 3037.
(8) Chevallier, F.; Breit, B. Angew. Chem. 2006, 118, 1629; Angew.
Chem., Int. Ed. 2006, 45, 1599.
(9) Smejkal, T.; Breit, B. Organometallics 2007, 26, 2461.
(10) Oshiki, T.; Yamashita, H.; Sawada, K.; Utsunomiya, M.; Takahashi,
K.; Takai, K. Organometallics, 2005, 24, 6287.
Unfortunately, complex 6 was not effective as a catalyst
precursor to allylic substitution reaction (Table 1, entry 1).
Also, when using π-allyl palladium chloride dimer as the
catalyst precursor, although traces of the desired 3-allylation
product 9a could be observed, the result was unsatisfactory
(entry 2). However, promising observations were made with
[(η3-allyl)Pd(cod)]BF4 as the catalyst precursor in the pres-
ence of 5 mol % of each of the self-assembling ligands 1
and 2 (entry 3). After some experimentation the conditions
(11) 3-DPICon (1): 3-Diphenylphosphanylisoquinolone; 6-DPPAP (2):
N-Pivaloyl-6-diphenylphosphanylaminopyridine; 6-DPAIND (3): 6-Diphe-
nylphosphanyl-1H-7-azaindole; 2-DPPAT (4): N-Pivaloyl-2-diphenylphos-
phanylaminothiazole; 6-DPPon (5): 6-Diphenylphos-phanylpyridone.
1208
Org. Lett., Vol. 10, No. 6, 2008

replacing cinnamyl alcohol 8a by the less reactive parent
allyl alcohol 8b, the 6-DPPon derived catalyst performed
best (compare entries 5 and 6) although the overall yields
were still unsatisfactory. Hence, in a second stage we decided
Table 1. Optimizing Reaction Conditions
Table 3. Scope of Allylic Alcohol
yieldb
solvent T (°C) t (h) (%)
entrya
catalyst
1
2
6
PhCl
PhCl
rt
rt
65
0
[(η3-allyl)PdCl]2
65 trace
65 32 (0)
20 84 (3)
3-DPICon(1)/6-DPPAP(2)
[(η3-allyl)Pd(cod)]BF4
3-DPICon(1)/6-DPPAP(2)
[(η3-allyl)Pd(cod)]BF4
3-DPICon(1)/6-DPPAP(2)
3
PhCl
rt
4c
toluene 50
a Pd/P (25/50 µmol), indole (0.50 mmol), and cinnamyl alcohol (0.55
mmol) were reacted in toluene (0.25 mL). b Determined from 1H NMR
employing t-Bu2biphenyl (0.056 mmol ) 2/18 mmol) as internal standard;
yield of 10a in parentheses. c Utilizing 3 mol % of catalyst and 0.50 mmol
of cinnamyl alcohol.
shown in entry 4 were identified as best. Thus, employing 3
mol % of the self-assembling catalyst in toluene at 50 °C,
84% of the desired 3-allylated indole 9a with only 3% of
the easily separable bis-allylation product 10a were obtained.
The inherent advantage of the self-assembly approach is
its implemented combinatorial flexibility for optimization.
Thus, next we decided to identify the best self-assembly
platform in order to achieve optimal results in the title
transformation. Thus, replacing aminopyridine 2 by the
thiazole 4, which has recently been shown to be an ideal
system for rhodium-catalyzed hydroformylation,6b did not
lead to a better catalyst (Table 2, entry 2), neither did the
Table 2. Identifying the Best Self-assembling Ligand System
a Isolated yield. b 3 equiv of allylic alcohol (8b) was utilized. c Employed
3 mol % of 1, 2 as ligands. d Utilizing N-methylindole (7d) instead of
N-unsubstituted indole (7a).
entry
L1/L2
yielda (%)
entry
L1/L2
yielda (%)
to optimize the most simple 6-DPPon system by variation
of the electronic properties of the phosphine donors. It has
previously been shown that the electrophilicity and hence
reactivity of a palladium π-allyl intermediate toward nu-
cleophiles can be increased on increasing the π-acceptor
strength of the ligands attached to palladium.12 Accordingly,
we observed increased catalyst activity when increasing the
π-acceptor strength of the phosphine donors attached to the
pyridone platform (entries 7-11) with the bis-m-trifluorom-
ethyl-substituted phosphine 5d furnishing the best catalyst
(entries 7-10). Thus, good yield and excellent chemoselec-
tivity for the C3-allylation product were obtained (entry 11).
1
2
3
4
5b
6b
1/2
1/4
3/2
5a/5a
1/2
84 (3)
76 (4)
0
79 (6)
42 (0)
60 (0)
7c
5a/5a
5b/5b
5c/5c
5d/5d
5d/5d
10 (0)
22 (0)
41 (3)
40 (0)
80
8c
9c
10c
11b
5a/5a
a Estimated from 1H NMR with use of t-Bu2biphenyl (0.056 mmol )
2/18 mmol) as internal standard; yield of 10a in parentheses. b Utilizing
allyl alcohol (8b) instead of cinnamyl alcohol (8a). c Utilizing allyl alcohol
(8b) instead of cinnammyl alcohol (8a) at rt.
replacement of the isoquinolone 1 by the azaindole ligand
3. However, the simplest self-complementary 6-DPPon
ligand 5a provided a rather active catalyst (entry 4). Notably,
(12) Kuhn, O.; Meyr, H. Angew. Chem. 1999, 111, 356; Angew. Chem.,
Int. Ed. 1999, 38, 343.
Org. Lett., Vol. 10, No. 6, 2008
1209

More substituted allylic alcohols 8c-h can be applied as
substrates as well, and the corresponding 3-indole monoal-
lylation products 9 were obtained selectively and in good
yields (Table 3, entries 1-7). The results obtained for the
unsymmetrical π-allyl systems (entries 2-7) suggest the
presence of a π-allylpalladium intermediate. For the meth-
allylic alcohol substrate the bismethallylation product at C3
and N was isolated. On the other hand, when the N-
methylindole was employed the C3-allylation product 9g
could be isolated in 94% yield.
Table 5. Nitrogen Nucleophiles
Furthermore, several substituted indole derivatives 7b-f
as well as pyrrole (11) were successfully allylated employing
the new self-assembly catalyst to give the corresponding
products in good to excellent yields (Table 4). We also briefly
Table 4. Scope of N-Heterocycles
a Isolated yield.
II. Nucleophile attack and proton transfer lead to product
alkene complex III which undergoes ligand exchange with
the allylic alcohol substrate liberating the allylation product
and water as the only byproduct.
In conclusion, we have developed the first palladium
catalyst system that allows the direct allylation of indoles
with allylic alcohols as substrates with water being the only
byproduct. Key to success was the application of self-
assembling ligands based on complementary hydrogen-
bonding. We propose that the ligand hydrogen-bonding
network goes beyond its role in ligand structure determining
and suggest that the hydrogen-bonding system assists the
hydroxy group to become a better leaving group in the course
of this allylic substitution process. Hence, this is an example
of a dual transition metal/organocatalysis, which opens up
numerous possibilities for the application of these self-
assembling ligands to homogeneous catalysis including
asymmetric versions.
a Isolated yield. b 36 h. c No reaction. d 85% conversion.
looked into nitrogen nucleophiles such as aniline derivatives,
which gave excellent yields of the monoallylation products
as well (Table 5).
Acknowledgment. This work was supported by the Fonds
der Chemischen Industrie, the DFG via the International
Research Training Group “Catalysts and Catalytic Reactions
for Organic Synthesis” (GRK 1038), and the Alfried Krupp
Award for young university teachers of the Krupp foundation
(to B.B.).
Although the exact details of the reaction mechanism have
not been clarified yet, we propose the following rational
(Scheme 2). In agreement with the experimental observation
that water can be incorporated into the ligand hydrogen-
bonding network (Figure 1), we suggest that the hydroxy
function of an allyl alcohols may be bound in a similar
manner to give alkene complex I. As a result, the hydroxy
substituent may become a better leaving group and thus
facilitates the formation of the π-allylpalladium intermediate
Supporting Information Available: Expermental pro-
cedures and spectroscopic data. This material is available
free of charge via the Internet at http://pubs.acs.org.
OL800073V
1210
Org. Lett., Vol. 10, No. 6, 2008