Ruthenium-catalyzed stereospecific decarboxylative allylation of non-stabilized ketone enolates

Article abstract of DOI:10.1039/b503568f

The ruthenium-catalyzed stereospecific decarboxylative allylation of ketone enolates provides access to γ,δ-unsaturated ketones with good yields and enantioenrichments. The Royal Society of Chemistry 2005.

Full text of DOI:10.1039/b503568f

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Ruthenium-catalyzed stereospecific decarboxylative allylation of
non-stabilized ketone enolates{
Erin C. Burger and Jon A. Tunge*
Received (in Bloomington, IN, USA) 7th March 2005, Accepted 7th April 2005
First published as an Advance Article on the web 21st April 2005
DOI: 10.1039/b503568f
esterification with diketene.9 We were encouraged to find the
The ruthenium-catalyzed stereospecific decarboxylative allyla-
product (3a) was produced with a 79% ee (83% cee){ upon
treatment of 1a with 2.5 mol% [Cp*RuCl]4 and 10 mol%
bipyridine at room temperature in CH2Cl2. Interestingly, the
regioselectivity of the decarboxylative rearrangement (3:4 5 75:1)
is much higher than that reported for the related ruthenium-
catalyzed allylation of dimethyl malonate (12:1).5 Furthermore, we
have confirmed that the reaction occurs with net retention of
configuration by hydrogenation of 3a to produce (R)-(2)-4-
phenyl-2-hexanone.10 Thus, our allylation of a non-stabilized
ketone enolate is stereochemically analogous to Trost’s ruthenium-
catalyzed allylation of stabilized malonate nucleophiles.5 However,
racemization was not observed in reactions of allylic carbonates
with stabilized nucleophiles, indicating that our reaction may differ
mechanistically.
tion of ketone enolates provides access to c,d-unsaturated
ketones with good yields and enantioenrichments.
The ability to synthesize small molecules that are highly
enantioenriched is an important goal in organic synthesis.
Optically active c,d-unsaturated ketones are particularly attractive
targets due to the complementary reactivity of the electrophilic
carbonyl and the nucleophilic olefin. Moreover, these molecules
can be further elaborated by diastereoselective additions in order
to synthesize more complex targets as single enantiomers.1 During
the course of our studies2 on a catalytic version of the
decarboxylative Claisen (Carroll) rearrangement,3,4 we disclosed
the use of [Cp*RuCl]4 as a catalyst to provide c,d-unsaturated
ketones (3) in high yields via the regioselective decarboxylative
allylation of non-stabilized ketone enolates (Scheme 1).2a We now
wish to report that our catalytic decarboxylative rearrangement is
also highly stereospecific, allowing the synthesis of enantioenriched
c,d-unsaturated ketones. Furthermore, we have identified the
primary mechanism for racemization and used this knowledge to
maximize the stereospecificity.
A variety of other allylic b-keto esters underwent decarbox-
ylative rearrangement to give good yields of the corresponding
unsaturated ketones (3) with good retention of stereochemistry
(Table 1). While aryl substitution on the allyl fragment is required
for a facile, high-yielding reaction, the reaction tolerates a variety
of electron donating and withdrawing substituents on the phenyl
ring. Furthermore, in contrast to results obtained for straight-
chain b-keto esters (2), reaction times for branched b-keto esters
(1) do not show a strong dependence on the electronic nature of
the aryl substituent.2a For example, while an electron withdrawing
p-NO2 substituent does not have a large effect on the rate of
reaction of branched b-keto ester 1e (vs. 1a), the rate of reaction of
NO2-substituted straight-chain b-keto ester 2e is decreased ca. 20-
fold relative to 2a.
Stereospecific allylations of stabilized malonate enolates have
been previously reported with ruthenium,5 rhodium,6 iridium,7 and
palladium.8 Evans recently expanded the scope of Rh-catalyzed
allylic alkylations to include the first examples of a stereospecific
allylation of non-stabilized ketone enolates which proceeds via the
in situ formation of copper enolates.6c,d This prompted us to
examine the stereochemical fidelity of our ruthenium-catalyzed
allylation of non-stabilized enolates.
To begin, model substrate (S)-1a (R1 5 Me; R2 5 H; R3 5 Ph),
was prepared with high enantiopurity (95% ee) via enzymatic
resolution of the corresponding allylic alcohol followed by
In addition to variation of the aryl group, a variety of enolates
(1f–1h) are accessible through decarboxylation. The results of these
studies indicate that enolate generation is regiospecific, with the
regiochemistry determined solely by the position that carries the
carboxylate group. It is particularly noteworthy that we are able to
access the terminal enolate of benzyl acetone (from 1g), which
cannot be accomplished with standard base-induced enolate
formation.11
The effects of nitrogenous bidentate ligands other than
bipyridine were also briefly explored. It was found that
TMEDA (tetramethyethylenediamine) leads to decreased retention
of stereochemistry for most substrates; the exception to this is 1a,
which produced 3a with slightly elevated conservation of
configuration using TMEDA as a ligand. Since there was no
clear trend, the origin of these effects is not completely understood.
Having demonstrated that the rearrangement is enantiospecific,
we turned our attention to identifying the mechanism(s) respon-
sible for imperfect retention of configuration. Racemizations of
Scheme 1 Decarboxylative allylation.
{ Electronic supplementary information (ESI) available: experimental and
characterization of new compounds. See http://www.rsc.org/suppdata/cc/
b5/b503568f/
*tunge@ku.edu
This journal is ß The Royal Society of Chemistry 2005
Chem. Commun., 2005, 2835–2837 | 2835

View Article Online
Table 1 Reaction with various aryl substratesa
Substrate
R1
R2
R3
Time
3:4b
75:1
—
101:1
19:1
.19:1
—
89:1
.19:1
.19:1
59:1
20:1
—
ceec (%)
Yield (%)
1a
1a
1b
1c
1c
1d
1d
1e
1e
1f
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Bn
H
H
H
H
H
H
H
H
H
Me
H
H
H
Ph
Ph
1.5 h
4 h
2 h
15 min
—
4 h
30 min
3 h
2 h
2 h
3 h
83
86
N/A
81
83
N/A
70
56
49
N/A
76
71
49
68
87d
87
p-MeC6H4
p-(MeO)C6H4
p-(MeO)C6H4
p-ClC6H4
p-ClC6H4
p-(NO2)C6H4
p-(NO2)C6H4
Ph
93
52d
86
94
98
45d
93e
72
1g
1g
1h
a
Ph
Ph
Ph
Bn
i-Pr
1 h
3 h
79
58
38:1
Reactions run at 0.1 M in CH2Cl2 at room temperature with 2.5 mol% [Cp*RuCl]4 and 10 mol% bipyridine. Determined by GC or 1H
b
NMR spectroscopy of the crude reaction mixture; . 19:1 indicates that the minor regioisomer was not observed by NMR spectroscopy.
Determined on a Diacel Chiralpak AD or OD-H HPLC column. With 10 mol% TMEDA ligand. cee of major diastereomer; dr 5 1:1.5.
c
d
e
related palladium p-allyl complexes have been shown to occur by
p–s–p interconversion of the enantiotopic allyl faces and/or by a
bimetallic mechanism involving degenerate substitution of a
palladium p-allyl by Pd0.12 It has also been suggested,8a and later
refuted,12 that nonselective attack of the acetate leaving group on
the p-allyl is responsible for reduced stereospecificity in alkylations
of allylic acetates. While much less is known about the mechanism
of racemization of ruthenium p-allyl complexes, Trost has shown
that racemization by p–s–p allyl interconversion is slow at
ambient temperature.5
To begin, a brief survey of reaction conditions demonstrated
that the concentration of the reaction mixture impacts the
stereospecificity of decarboxylative allylation. For example, when
various concentrations of 1c (0.32 M to 0.04 M) were allowed to
undergo catalytic rearrangement with 2.5 mol% [Cp*RuCl]4 and
10 mol% bpy, it was found that as the concentration of 1c
decreased, the retention of stereochemistry in product 3c increased
from 86% to 94% cee. While this difference is significant, the
magnitude is much smaller than would be predicted based on a
bimetallic racemization mechanism;12 an 8-fold increase in catalyst
concentration would be expected to increase the rate of
racemization by a factor of 64.
Scheme 2 Mechanism of racemization.
fact, we have evidence that both of these routes contribute to
production of racemic 3.
Another important observation was that enantioenrichment of 3
is dependent on conversion. For example, the ee of 3a, produced
from the rearrangement of 1a (93.5% ee), gradually decreased from
93% to 76% over a 20 minute reaction period.
In the case of 1c, degradation of enantiopurity of the starting
material was directly observed. It was found that the ee of 1c
decreased from 94% to 78% after five minutes of catalysis (ca. 50%
conv.). Thus, the racemization of 1 contributes to imperfect
retention of stereochemistry in product 3. While ionization to
p-allyl complex 5 followed by p–s–p isomerization could explain
this result, we do not believe that this is the case for the following
reasons: 1) Trost has shown that p–s–p isomerization in related
ruthenium-p-allyl complexes is slow and 2) we directly observed
substantial quantities (20%) of isomerization product 2c in the
reaction mixture, implicating equilibration of 1 and 2 as a simple
mechanism for racemization of 1 (mechanism A).
Further investigation of the course of the reaction via 1H NMR
spectroscopy proved to be particularly revealing. After five
minutes of catalysis (ca. 50% conv.) under standard reaction
conditions we observed partial isomerization of the branched,
chiral b-keto ester 1c to linear, achiral b-keto ester 2c (4:1 ratio of
1c:2c). The isomerization of 1 to 2 was subsequently observed for
all substrates (Scheme 2). This isomerization can lead to loss of
stereochemistry via two related routes: A) the racemization of 1 via
equilibration between 1 and 2 and B) the production of racemic 3
from the decarboxylative rearrangement of achiral substrate 2. In
A further observation from the 1H NMR spectroscopic analysis
is that branched isomer 1 undergoes rearrangement to 3 faster than
2836 | Chem. Commun., 2005, 2835–2837
This journal is ß The Royal Society of Chemistry 2005

View Article Online
the corresponding linear isomer 2. This can be in part explained
by a more facile pre-coordination of the Ru catalyst to the
monosubstituted allyl fragment found in 1 compared to the
conjugated, disubstituted allyl fragment present in 2. Based on this
result, we surmised that the difference in reactivity between
substrates 1 and 2 could be exploited in order to generate products
with increased retention of stereochemistry by avoiding the
production of racemic 3 through mechanism B. For example,
when substrate 1d (96% ee) was allowed to react for 4 hours in
order to maximize conversion to product, 3d was isolated in 70%
yield and 83% ee (86% cee, Table 1). However, 1H NMR
spectroscopy indicated that after 30 minutes only linear, achiral
b-keto ester 2 remained. Thus, allowing the reaction to proceed for
only 30 minutes allowed us to isolate 3d in 90% ee (94% cee), albeit
in lower chemical yield (56%).
allyl b-keto esters is necessary in order to maximize the
stereospecificity of decarboxylative allylations.
In summary, we have shown that the decarboxylative allylation
of ketone enolates using a [Cp*RuCl]4/bipyridine catalyst proceeds
in a stereospecific manner. Furthermore, imperfect stereospecificity
has been attributed to a ruthenium-catalyzed isomerization of
starting material through reversible formation of p-allyl ruthenium
intermediates. This mechanistic insight allowed us to develop a
highly stereospecific decarboxylative allylation of non-stabilized
ketone enolates.
Erin C. Burger and Jon A. Tunge*
Department of Chemistry, University of Kansas, Lawrence, Kansas.
E-mail: tunge@ku.edu
Notes and references
This strategy was useful for improving cee’s of substrates for
which isomeric forms 1 and 2 have substantially different
rearrangement kinetics. Substrates 1e and 2e provide the most
striking example of the differential reactivity between the two
isomers of starting material. In this case the reaction of 2e is
extremely slow (vide supra) under the conditions of catalysis,
virtually eliminating the reaction pathway leading from 2 to
racemic product. Presumably the difference in rates is due to the
presence of the strongly withdrawing nitro group. Bruneau, et al.
have suggested that, prior to oxidative addition and p-allyl
formation, Ru pre-coordinates to the olefin.13 Backbonding into
the p* orbital of the alkene is expected to be greatest for electron
deficient aryl substituted alkenes (2). This in turn decreases the
nucleophilicity of Ru, which will raise the barrier for oxidative
addition to form the reactive p-allyl ruthenium species (5). Thus,
with electron withdrawing aryl groups, 2 is essentially unreactive,
preventing the formation of racemic 3. However, with electron
donating aryl groups (p-OMe) the barrier for racemization is low
enough that we have observed the partial racemization of 1c.
The addition of an a-methyl group also leads to a large disparity
between the reaction rates of 1f and isomeric 2f, allowing the
isolation of 3f in a 90% ee from 97% ee starting material (93% cee).
This is compared to 3a (non-methylated), which was isolated in a
79% ee from 95% ee starting material (83% cee). Clearly,
understanding and utilizing the differing reaction rates of isomeric
{ Conservation of enantiomeric excess (cee) 5 [product ee/reactant ee] 6
100.
1 A. M. M. Castro, Chem. Rev., 2004, 104, 2939.
2 (a) E. Burger and J. Tunge, Org. Lett., 2004, 6, 2603; (b) E. Burger and
J. Tunge, Org. Lett., 2004, 6, 4113; (c) E. Burger and J. Tunge, Eur. J.
Org. Chem., 2005, DOI: 10.1002/ejoc.200400865.
3 M. Carroll, J. Chem. Soc., 1940, 704.
4 (a) T. Tsuda, Y. Chujo, S.-I. Nishi, K. Tawara and T. Saegusa, J. Am.
Chem. Soc., 1980, 102, 6381–6384; (b) I. Shimizu, T. Yamada and
J. Tsuji, Tetrahedron Lett., 1980, 21, 3199–3202.
5 B. Trost, P. Fraisse and Z. Ball, Angew. Chem. Int. Ed., 2002, 41, 1059.
6 (a) P. A. Evans and J. Nelson, J. Am. Chem. Soc., 1998, 120, 5581; (b)
P. A. Evans and L. Kennedy, Org. Lett., 2000, 2, 2213; (c) P. A. Evans
and D. Leahy, J. Am. Chem. Soc., 2003, 125, 8974; (d) P. A. Evans and
M. Lawler, J. Am. Chem. Soc., 2004, 126, 8642.
7 (a) B. Bartels and G. Helmchen, Chem. Commun., 1999, 741.
8 (a) B. Trost and T. Verhoeven, J. Am. Chem. Soc., 1980, 102, 4730; (b)
T. Hayashi, T. Hagihara, M. Konishi and M. Kumada, J. Am. Chem.
Soc., 1983, 105, 7767; (c) T. Konno, K. Nagata, T. Ishihara and
H. Yamanaka, J. Org. Chem., 2002, 67, 1768.
9 (a) K. Burgess and L. Jennings, J. Am. Chem. Soc., 1990, 112, 7434; (b)
K. Burgess and L. Jennings, J. Am. Chem. Soc., 1991, 113, 6129.
10 K. Soai, S. Yokoyama, T. Hayasaka and K. Ebihara, J. Org. Chem.,
1988, 53, 4148.
11 E. J. Corey and A. W. Gross, Tetrahedron Lett., 1984, 25, 495.
12 P. B. Mackenzie, J. Whelan and B. Bosnich, J. Am. Chem. Soc., 1985,
107, 2046.
13 M. Mbaye, B. Demerseman, J. Renaud, L. Toupet and C. Bruneau,
Angew. Chem. Int. Ed., 2003, 42, 5066.
This journal is ß The Royal Society of Chemistry 2005
Chem. Commun., 2005, 2835–2837 | 2837