Ruthenium-catalyzed decarboxylative allylation of nonstabilized ketone enolates

Article abstract of DOI:10.1021/ol049097a

(Matrix Presented) Bipyridyl(pentamethylcyclopentadienyl)ruthenium chloride is an efficient catalyst for the formal [3,3] rearrangement of allyl β-ketoesters. The mechanism of the transformation involves formation of π-allyl ruthenium intermediates, which

Full text of DOI:10.1021/ol049097a

ORGANIC
LETTERS
2004
Vol. 6, No. 15
2603-2605
Ruthenium-Catalyzed Decarboxylative
Allylation of Nonstabilized Ketone
Enolates
Erin C. Burger and Jon A. Tunge*
Department of Chemistry, UniVersity of Kansas, Lawrence, Kansas 66045
tunge@ku.edu
Received May 17, 2004
ABSTRACT
Bipyridyl(pentamethylcyclopentadienyl)ruthenium chloride is an efficient catalyst for the formal [3,3] rearrangement of allyl â-ketoesters. The
mechanism of the transformation involves formation of π-allyl ruthenium intermediates, which are selectively attacked at the more substituted
allyl terminus by freely diffusing enolates. Decarboxylation of â-ketocarboxylates allows generation of enolates under extremely mild conditions.
Efficient catalysis of sigmatropic rearrangements is a long-
standing goal in chemical synthesis. Despite the recognized
synthetic utility of Claisen-type rearrangements, only recently
have successful approaches to catalysis been realized.1
Although both anionic2 and cationic3 acceleration of Claisen
rearrangements are well-known, the majority of attempts at
catalysis have focused on Lewis acid activation. A conceptu-
ally different approach toward catalysis is the oxidative
addition of an allyl vinyl ether to a transition metal,
producing an enolato-metal-allyl that could couple these
fragments to form the desired unsaturated ketone (Scheme
1).
Given the difficulty of metal insertion into ether C-O
bonds, we have directed our initial efforts toward catalysis
of the Carroll (decarboxylative Claisen) rearrangement,4
which utilizes the more reactive and more easily synthesized
allyl-â-ketoesters. Herein we report that bipyridyl(penta-
methylcyclopentadienyl)ruthenium chloride [Cp*Ru(bpy)Cl]
is an efficient and selective catalyst for the rearrangement
of allyl-â-ketoesters.
Saegusa reported the first catalytic rearrangement of allyl-
â-ketoesters using palladium; however, the primary products
were those of [1,3] rearrangement.5 This result was explained
by the intermediacy of palladium π-allyl complexes that are
(1) (a) Yoon, T. P.; Dong, V. M.; MacMillan, D. W. C. J. Am. Chem.
Soc. 1999, 121, 9276-9277. (b) Abraham, L.; Czerwonka, R.; Hiersmann,
M. Angew. Chem., Int. Ed. 2001, 40, 4700-4703. (c) Anderson, C. E.;
Overman, L. E. J. Am. Chem. Soc. 2003, 125, 12412-12413. (d) Ito, H.;
Taguchi, T. Chem. Soc. ReV. 1999, 28, 43-50.
Scheme 1
(2) (a) Denmark, S. E.; Marlin, J. E. J. Org. Chem. 1987, 52, 5742-
5743. (b) Denmark, S. E.; Harmata, M. A. J. Org. Chem. 1983, 48, 3369-
3370. (c) Koreeda, M.; Luengo, J. I. J. Am. Chem. Soc. 1985, 107, 5572-
5573.
(3) (a) Boeckman, R. K.; Flann, C. J.; Poss, K. M. J. Am. Chem. Soc.
1985, 107, 4359-4362. (b) Dhanalekshmi, S.; Venkatachalam, C. S.;
Balasubramanian, K. K. J. Chem. Soc., Chem. Commun. 1994, 511-512.
(4) (a) Carroll, M. F. J. Chem. Soc. 1940, 704-706. (b) Kimel, W.; Cope,
A. C. J. Am. Chem. Soc. 1943, 65, 1992-1998.
(5) (a) Tsuda, T.; Chujo, Y.; Nishi, S.-I.; Tawara, K.; Saegusa, T. J.
Am. Chem. Soc. 1980, 102, 6381-6384. (b) Tsuji, J. Pure Appl. Chem.
1982, 54, 197-206. (c) Tsuji, J.; Yamada, T.; Minami, I.; Yuhara, M.;
Nisar, M. Shimizu, I. J. Org. Chem. 1987, 52, 2988-2995.
10.1021/ol049097a CCC: $27.50 © 2004 American Chemical Society
Published on Web 06/19/2004

known to undergo nucleophilic attack at the least substituted
allyl terminus.6
appears that the role of the ligand is to form a monomeric
ruthenium species from the tetrameric precatalyst; however,
strongly coordinating solvents such as CH3CN are deleterious
to the rates. Other ruthenium(II) sources such as
(CH2CMeCH2)2Ru(COD) (4) and CpRu(NCCH3)3PF6 (5)
were not as effective.
A brief study of substituent effects shows that electron-
donating groups on the aryl substituent accelerate the reaction
and the reactions are more sluggish with electron-withdraw-
ing substituents (Table 2). Importantly, substituted nucleo-
In the interest of discovering catalysts that would provide
primarily the products of [3,3] rearrangement, we chose to
investigate various ruthenium catalysts. Ruthenium carboxy-
lates are known to readily decarboxylate7 much like the
carboxylates of palladium, and the chemistry of ruthenium
allyls is well-studied.8 For example, Trost has shown that
Cp*Ru(NCCH3)3PF6 catalyzes the allylation of stabilized
nucleophiles with high regioselectivity for addition at the
more substituted allyl terminus.8b
The rearrangement of cinnamyl â-ketoester (R ) Ph, 1a)
was used to investigate the activity of various ruthenium
catalysts and conditions (Scheme 2). Initially, it was found
Table 2. Ruthenium-Catalyzed Carroll Rearrangementsa
substrate
R1
R2
time
yield
1a
1b
1c
1d
1e
1f
Ph
p-tolyl
o-tolyl
p-C6H4OMe
o-C6H4OMe
H
H
H
H
H
H
1.5 h
2 h
5 d
15 min
15 min
30 min
94
96
81
91
93
91
Scheme 2
1g
1h
1i
p-C6H4Cl
p-C6H4CF3
H
Ph
H
H
Ph
Ph
4 h
96
90
93
67
40 h
1.5 h
1 h
that 2.5 mol % [Cp*RuCl]4 in CH2Cl2 led to complete
rearrangement, but the reaction was quite slow and the
branched ([3,3]) to linear ([1,3]) ratio was not that high
(Table 1). Addition of 10 mol % bipyridine accelerated the
1j
a Reaction times and isolated yields for 0.2 M substrate, 2.5 mol %
[Cp*RuCl]4, and 10 mol % bipyridine in CH2Cl2 at 25 °C.
Table 1. Activity and Selectivity of Carroll Rearrangement
Catalysts toward Reaction with 1a (R1 ) Ph, R2 ) H)a
philes react smoothly but require longer reaction times and
the diastereoselectivities are low (Scheme 3).
catalyst
ligand time (h) solvent (temp) conversion (2:3)a
[Cp*RuCl]4
1.5
1.5
1.5
4
4
4
1.5
16
16
16
16
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
THF
THF
CH3CN (50 °C)
THF (65 °C)
CH2Cl2 (40 °C)
CH2Cl2 (40 °C)
CH2Cl2
9 (6.2)
[Cp*RuCl]4 2 py
[Cp*RuCl]4 bpy
[Cp*RuCl]4 TMEDA
[Cp*RuCl]4
[Cp*RuCl]4 bpy
[Cp*RuCl]4
4
4
4
5
19 (2.5)
100 (>19)b
100 (9.1)
100 (14) c
100 (10.7)
66 (3)
100 (0.5)
nr
nr
Scheme 3
bpy
bpy
bpy
Next, the rearrangement of 1i (the regioisomer of 1a) was
shown to give 2a, the product of formal [1,3] rearrangement
(Scheme 4). This experiment demonstrates that the ruthenium-
49 (10)
a Ratios of branched (2) to linear (3) product were determined at complete
conversion. b Minor isomer was not detected by 1H NMR spectroscopy.
c 1.25 mol % catalyst was used.
Scheme 4
reaction and resulted in quantitative conversion to the [3,3]
rearrangement product in 1.5 h at room temperature. Addition
of 10 or 20 mol % pyridine led to smaller rate increases. It
(6) Trost, B. M.; Van Vranken, D. L. Chem. ReV. 1996, 96, 395-422.
(7) (a) Roper, W. R.; Wright, L. J. J. Organomet. Chem. 1982, 234,
C5-C8. (b) Casey, C. P.; Singer, S. W.; Powell, D. R. Can. J. Chem. 2001,
79, 1002-1011.
(8) (a) Matsushima, Y.; Onitsuka, K.; Kondo, T.; Mitsudo, T.; Takahashi,
S. J. Am. Chem. Soc. 2001, 123, 10405-10406. (b) Trost, B. M.; Fraisse,
P. L.; Ball, Z. T. Angew. Chem., Int. Ed. 2002, 41, 1059-1061. (c) Renaud,
J. L.; Bruneau, C.; Demerseman, B. Synlett 2003, 3, 408-410. (d) Kondo,
H.; Kageyama, A.; Yamaguchi, Y.; Haga, M.; Kirchner, K.; Nagashima,
H. Bull. Chem. Soc. Jpn. 2001, 74, 1927-1937.
catalyzed rearrangement is highly regioselective but not
regiospecific like the uncatalyzed Carroll rearrangement. This
observation is most simply interpreted in terms of an
intermediate ruthenium π-allyl complex, which exhibits a
2604
Org. Lett., Vol. 6, No. 15, 2004

large preference for addition of the enolate to the more
substituted allyl terminus.8
Scheme 7
To further probe the mechanism of this transformation,
we tested whether the reaction proceeds intra- or intermo-
lecularly. Thus, a 1:1 mixture of allyl-â-ketoesters 8a and
1g were treated under the standard reaction conditions
(Scheme 5). Gas chromatographic and mass spectrometric
Scheme 5
sible for the catalysis (Scheme 7). This 18 e- complex likely
ionizes to catalytically active 16 e- Cp*Ru(bpy)+, consistent
with the previously reported rapid ionization of Cp*Ru-
(TMEDA)Cl [TMEDA ) tetramethylethylenediamine].12
Activation by ionization is also consistent with the poor
reactivity of these catalysts in acetonitrile, which will disfavor
formation of a coordinatively unsaturated cationic ruthenium
complex. This ruthenium complex reacts with the allyl-â-
ketoester resulting in ruthenium allyl formation. The decar-
boxylation of the resulting acetoacetate requires further
mechanistic investigation; however, ruthenium is likely
involved in this process because acetoacetate anion does not
spontaneously decarboxylate at ambient temperature.5a,13
In conclusion, Cp*Ru(bpy)Cl is a selective catalyst for
the decarboxylative rearrangement of allyl-â-ketoesters to
γ,δ-unsaturated ketones. The mechanism of catalysis likely
involves ruthenium-π-allyl intermediates that are selectively
attacked at the more substituted allyl terminus. In this regard,
the reaction is equivalent to the allylation of nonstabilized
ketone enolates, an important goal of allylation chemistry.6,14
Currently, use of boron or tin enolates has allowed use of
ketone enolates for palladium-catalyzed allylic alkylation of
allyl acetates; however these reactions result in substitution
at the least hindered allyl terminus. Therefore, the reaction
described here broadens the scope of allylic alkylations and
eliminates the need for boron or tin additives. Extension of
this methodology to the asymmetric rearrangement is cur-
rently underway.
analysis of the resulting mixture shows approximately equal
quantities of products 2g, 9a, and crossover products 2a and
9g. This result indicates that freely diffusing enolates are
formed, which then add to ruthenium π-allyl complexes.9
The observation that 9a and 9g10 are the only products
derived from the propionyl acetate 8a demonstrates that the
regiochemistry of enolate generation is preserved. In fact,
treatment of 8a under the standard reaction conditions
produced only 9a, confirming that enolate formation is under
kinetic control.
Next, the reaction was run in the presence of dimethyl-
malonate in order to probe the lifetime of the enolate. If the
enolate is long-lived, it should be protonated by the more
acidic malonate, ultimately providing 10 as the product
(Scheme 6). The fact that the reaction is unhampered by the
Scheme 6
Acknowledgment. This research has been supported by
the University of Kansas. E.C.B. thanks the Madison and
Lila Self-Fellowship for support.
addition of 1 equiv of dimethylmalonate shows that the
addition of enolate to the allyl ligand is much faster than
deprotonation of the acidic malonate.
Supporting Information Available: Experimental pro-
cedures and spectroscopic data of all compounds. This
material is available free of charge via the Internet at
http://pubs.acs.org.
1
Investigation of the ruthenium speciation by H NMR
spectroscopy showed that under the conditions of catalysis
ruthenium exists as a 1:1 complex with bipyridine. On the
basis of this evidence and the known coordination chemistry
of [Cp*RuCl]4,11 we suggest that Cp*Ru(bpy)Cl is respon-
OL049097A
(11) Koelle, U. Chem. ReV. 1998, 98, 1313-1334.
(12) Gemel, C.; Mereiter, K.; Schmid, R.; Kirchner, K. Organometallics
1997, 16, 5601-5603.
(13) We are grateful to a reviewer for pointing out this fact.
(14) (a) Braun, M.; Laicher, F.; Meier, T. Angew. Chem., Int, Ed. 2000,
39, 3494-3497. (b) Trost, B. M.; Schroeder, G. M. J. Am. Chem. Soc.
1999, 121, 6759-6760.
(9) Bimolecular addition of a ruthenium enolate to a ruthenium allyl
remains a possibility.
(10) The structure of 9g is tentatively assigned as shown based only on
the observation of its molecular ion. The regiochemistry is assumed to be
the same as all other substrates investigated.
Org. Lett., Vol. 6, No. 15, 2004
2605