Rhodium-catalyzed chemo- and regioselective decarboxylative addition of β-ketoacids to allenes: Efficient construction of tertiary and quaternary carbon centers

Article abstract of DOI:10.1021/ja411397g

A rhodium-catalyzed chemo- and regioselective intermolecular decarboxylative addition of β-ketoacids to terminal allenes is reported. Using a Rh(I)/DPPF system, tertiary and quaternary carbon centers were formed with exclusively branched selectivity under mild conditions. Preliminary mechanism studies support that the carbon-carbon bond formation precedes the decarboxylation and the reaction occurs in an outer-sphere mechanism.

Full text of DOI:10.1021/ja411397g

Communication
pubs.acs.org/JACS
Rhodium-Catalyzed Chemo- and Regioselective Decarboxylative
Addition of β‑Ketoacids to Allenes: Efficient Construction of Tertiary
and Quaternary Carbon Centers
Changkun Li and Bernhard Breit*
Institut fur Organische Chemie, Albert-Ludwigs-Universitat Freiburg, Albertstrasse 21, 79104 Freiburg im Breisgau, Germany
̈
̈
S
* Supporting Information
problem: (a) the carboxylic acid may initiate the reaction
through the formation of the allyl rhodium intermediate; (b)
the nucleophilicity of α-carbon would be enhanced; (c) CO2
can be eliminated spontaneously as a traceless directing group9
(eq 2). Herein, we present a rhodium-catalyzed regioselective
decarboxylative10 addition of β-ketoacids11 to allenes12 as an
efficient method to construct tertiary and quaternary13 carbon
centers under mild and neutral reaction conditions, of which
can be regarded as an alternative to enolate allylation under
basic conditions.14
ABSTRACT: A rhodium-catalyzed chemo- and regiose-
lective intermolecular decarboxylative addition of β-
ketoacids to terminal allenes is reported. Using a Rh(I)/
DPPF system, tertiary and quaternary carbon centers were
formed with exclusively branched selectivity under mild
conditions. Preliminary mechanism studies support that
the carbon−carbon bond formation precedes the decar-
boxylation and the reaction occurs in an outer-sphere
mechanism.
We commenced our studies employing cyclohexylallene 1a
(1.0 equiv) and benzoylacetic acid 2a (1.2 equiv) as model
substrates (Table 1). In the presence of 2.5 mol % of
ransition-metal-catalyzed allylic alkylation represents one
Tof the most powerful methods to construct C−C bonds in
organic synthesis.1 Under certain conditions, transition metal
catalysts can regioselectively produce the branched allylic
compounds, which provides an opportunity to obtain chiral
molecules (Scheme 1, eq 1, right).2−5 Recently, we developed a
Table 1. Optimization of Reaction Conditions
Scheme 1. Proposed Rhodium Catalyzed Decarboxylative
Addition of β-Ketoacids to Terminal Allenes
a
entry
ligand
x/y
z equiv
time (h)
yield (%)
b
1
2
3
4
5
6
7
8
DPEphos
DPPB
2.5/5.0
2.5/5.0
2.5/5.0
2.5/5.0
2.5/−
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.5
16
16
3
46
b
39
b
rac-BINAP
DPPF
24
1
88
0
DPPF
48
24
2
DPPF
−/5.0
1.0/2.0
0
DPPF
84
87
DPPF
1.0/2.0
2
a
b
Isolated yield. Some allene 1a was recovered, and benzoylacetic acid
2a was all consumed.
rhodium-catalyzed regioselective addition of carboxylic acids
and anilines to alkynes or allenes to furnish branched allylic
esters and amines in an atom-economic manner,6 which avoids
the installation of leaving groups on the substrates and
generation of waste.7 Unfortunately, many efforts to extend
the reactions to ketones (carbon nucleophiles) have failed (eq
1, left). The possible reasons are the relatively weak acidities
(or the ability of oxidative addition with rhodium complexes)
and the difficulty to form the nucleophilic enolate under
nonbasic conditions. Inspired by the decarboxylative enolate
formation during the biosynthesis of polyketides and fatty
acids,8 we thought that the installation of a carboxylic
group6b−d to the α-position of the ketone may address the
[Rh(cod)Cl]2 and 5.0 mol % DPEphos ligand,6c,d the reaction
proceeded smoothly at room temperature in DCE to give 46
mol % of the branched γ,δ-unsaturated ketone 3a as the only
regiomer, along with acetophenone as a byproduct (entry 1).15
Next, different bidentate ligands were tested (entries 2−4)
Received: November 8, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja411397g | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
among which the DPPF derived catalyst showed the highest
reactivity, consuming the allene substrate 1a in 1 h at room
temperature to give 88% isolated yield of 3a, without any trace
of allyl benzoylacetate (entry 4, see eq 2) detected. Control
experiments showed that both the rhodium catalyst and the
ligand are mandatory for the reaction (entries 5 and 6). By
reducing the catalyst loading to 1 mol %, the reaction was still
efficient to give a similar yield within 2 h (entry 7). When 1.5
equiv of benzoylacetic acid 2a was used in the reaction, the
product yield increased (entry 8).
(Table 3). Commercially available 3-methyl-1,2-butadiene
(4a) did react with benzoylacetic acid (2a) under the standard
Table 3. Scope of Rhodium-Catalyzed Regioselective
Synthesis of Quaternary Carbon Centers
With the optimized conditions in hand, we investigated the
scope of the decarboxylative addition reaction (Table 2). The
Table 2. Scope of Rhodium-Catalyzed Regioselective
Synthesis of Tertiary Carbon Centers
a
Reaction was run at 40 °C.
conditions to furnish 5a in 82% yield with concomitant
formation of a quarternary carbon center. The allene with a
benzyl substituent (4b) also gave the desired product 5b in
78% yield, even though this substrate is potentially vulnerable
to undergo isomerization to the corresponding 1,3-diene via β-
hydride elimination.17 Allenes with silyl ether and ester
protecting groups gave the quaternary carbon centers (5c and
5d) smoothly as well. With the allene 4b as the model
substrate, different β-ketoacids were examined in aromatic
systems with an electron-donating group (5e), a halogen (5h),
a naphthyl (5f), and a thiophene heterocycle (5g) that reacted
smoothly. Methyl, n-propyl, isopropyl, and tert-butyl groups can
all be tolerated in the decarboxlative addition reactions (5i to
5l), which is difficult in the allylic substitution reactions with
unsymmetrical ketone enolates under basic conditions. Aryl-
substituted quaternary carbon centers can also be formed (5o
and 5p). Furthermore, for the preparation of quaternary carbon
stereocenters, the allylic substitution reaction requires isomeri-
cally pure trisubstituted alkenes as substrates, which can be
synthetically challenging.13e Conversely, 1,1-disubstitued al-
lenes are prepared in one or two steps from commercially
available starting materials.16
Furthermore, the commercially available acetone-1,3-dicar-
boxylic acid 6 reacted with 2.5 equivalents of cyclohexylallene
1a under the rhodium(I)/DPPF condition to give 55%
symmetrical diallylic product 7, along with 34% monoallylic
product 3l (Scheme 2, eq 3). A subsequent ring closing
metathesis with Grubbs’ second generation catalyst furnished
the 3,6-dicyclohexylcyclohept-4-enone 8 as a separable mixture
(8-meso and 8-rac) (eq 4). This two-step sequence allows the
ready preparation of a cyclohept-4-enone from commercially
available 1a and 6.
terminal allenic substrates used were readily prepared in one or
two steps from commercially available starting materials.16
A
linear alkyl-substituted allene (3b) and an α-branched alkyl-
substituted allene (3c) were both suitable substrates for the
addition reaction. Protected amines (3d and 3e) and alcohols
(3f and 3g) were tolerated and furnished the desired products
smoothly, allowing for further functional group manipulations.
Additionally, a variety of β-ketoacids were tested with
cyclohexylallene 1a as the model substrate. Electron-donating
(3h) and halogenated (3k) aromatic rings, naphthyl (3i), and
heterocycles such as a thiophene (3j) were compatible with the
reaction conditions. Replacing the aromatic groups with alkyl
groups bearing potentially acidic α-hydrogens showed little to
no complication. Primary (3l and 3m), secondary (3n), and
tertiary (3o) alkyl groups can all be introduced. In particular
3m is an interesting case, because the differentiation of a
methylpropylketone in enolate chemistry toward methyl
allylation is extremely difficult. Furthermore, β-ketoacids with
chloride (3p) and alkenic functions (3r) gave the coupling
products in satisfactory to good yields. Reactions with
phenylallene also gave high yields (3s and 3t). In all cases,
only branched products were observed.
There are two fundamental questions in the mechanism of
this rhodium-catalyzed decarboxylative allylation reaction: (a)
decarboxylation and allylation, and which of these steps
precedes the other; (b) inner-sphere or outer-sphere
We then tested the more challenging 1,1-disubstituted
allenes as substrates for quaternary carbon construction
B
dx.doi.org/10.1021/ja411397g | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
Scheme 2. Rhodium Catalyzed Decarboxylative Diallylation
of Acetone-1,3-dicarboxylic Acid and RCM to Cyclohept-4-
enone
Scheme 4. Proposed Catalytic Cycle
mechanism regarding the nucleophilic attack.10 To address the
first question, because the benzoylacetic acid 2a is not soluble
in DCE, the reaction of 1a and 2a in the presence of the
[Rh(cod)Cl]2/DPPF catalyst system was stopped immediately
when the solid 2a disappeared after approximately 1 h. The
NMR of the crude reaction mixture showed both the α-allyl-β-
ketoacid intermediate 9 and the product 3a after decarbox-
ylation. Intermediate 9 can decompose slowly to product 3a
within 24 h.16 Furthermore, when solid 2a disappeared,
trimethylsilyldiazomethane was added to methylate the
intermediate 9. The β-ketoester 10 can be isolated in 32%
yield with a 7:1 dr, along with 35% of 3a (scheme 3). These
allene 1 and generate the π-allyl-rhodium intermediate B.19
Another molecule of β-ketoacid 2 attacks the allylic carbon in B
through its enol form followed by the release of β-ketoacid 9
and 2. A second allylation of α-allyl-β-keto-acid 9 is for steric
reasons certainly significantly slower.21 Instead, we propose that
9 enters a second catalytic cycle involving a rhodium-catalyzed
decarboxylation via intermediate C generating the final product
3.22
In conclusion, we have developed a highly regioselective
decarboxylative addition of β-ketoacids to terminal allenes to
produce γ,δ-unsaturated ketones. Branched tertiary and
quaternary carbon centers were constructed under very mild
reaction conditions with commercially available [Rh(cod)Cl]2
and DPPF. This reaction releases CO2 as the only byproduct,
which shows high atom economy and provides a mild
alternative to existing enolate allylation under basic conditions.
Preliminary mechanistic studies suggest that the carbon−
carbon bond formation precedes the decarboxylation and the
reaction occurs in an outer-sphere mechanism. Further
investigation on the asymmetric version of this reaction and
the attempt to extend the substrates from allenes to alkynes are
ongoing.
Scheme 3. Detection, Capture of Reaction Intermediate, and
Crossover Experiment
ASSOCIATED CONTENT
■
S
* Supporting Information
Experimental procedures and analytic data for synthesized
1
compounds, including H and 13C NMR spectra. This material
is available free of charge via the Internet at http://pubs.acs.org.
results suggest that allylation precedes decarboxylation in this
rhodium-catalyzed reaction. With regard to the second
question, a crossover reaction was conducted. While
benzoylacetate 11a was found to be completely unreactive
under standard catalysis conditions, when acetoacetic acid 2f
was added, the crossover product 3l could be observed and
isolated in 24% yield. Another 12% of 3a was formed most
likely from the released benzoylacetic acid 2a. 36% of the
benzoylacetate 11a was recovered.18 These experiments suggest
the following conclusions: first, a π-allyl rhodium intermedi-
ate19 (vide infra), generated through ionization of 11a, seems to
be involved; second, the π-allyl rhodium intermediate reacts
with another β-ketoacid; and third, the reaction proceeds more
likely via an outer-sphere mechanism,20 although other
mechanistic alternatives cannot be ruled out at this point.
Based on the experiments above, we propose the following
reaction mechanism (Scheme 4). Carboxylic acid 2 reacts with
the rhodium catalyst to give intermediate A. Two pathways are
possible from A. One is to release carbon dioxide and give the
byproduct methyl ketone.15 The other one is to insert into
AUTHOR INFORMATION
Corresponding Author
bernhard.breit@chemie.uni-freiburg.de
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by the DFG, the International
Research Training Group “Catalysts and Catalytic Reactions for
Organic Synthesis” (IRTG 1038), and the Krupp Foundation.
We thank the Alexander von Humboldt Foundation (C.L.) for
a postdoctoral fellowship.
REFERENCES
■
(1) For reviews on transition-metal-catalyzed allylic substitution, see:
(a) Lu, Z.; Ma, S. Angew. Chem., Int. Ed. 2008, 47, 258. (b) Trost, B.
M.; Crawley, M. L. Chem. Rev. 2003, 103, 2921. (c) Trost, B. M.; Van
Vranken, D. L. Chem. Rev. 1996, 96, 395.
C
dx.doi.org/10.1021/ja411397g | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
(2) For selected recent examples on palladium-catalyzed allylic
alkylation to give selectively the branched products, see: (a) Chen, J.-
P.; Peng, Q.; Lei, B.-L.; Hou, X.-L.; Wu, Y.-D. J. Am. Chem. Soc. 2011,
133, 14180. (b) Trost, B. M.; Malhotra, S.; Chan, W. H. J. Am. Chem.
Soc. 2011, 133, 7328. (c) Chen, J.-P.; Ding, C.-H.; Liu, W.; Hou, X.-L.;
Dai, L.-X. J. Am. Chem. Soc. 2010, 132, 15493. (d) Zhang, P.; Brozek,
L. A.; Morken, J. P. J. Am. Chem. Soc. 2010, 132, 10686.
(3) For selected examples on rhodium-catalyzed allylic alkylation to
give selectively the branched products, see: (a) Evans, P. A.; Oliver, S.;
Chae, J. J. Am. Chem. Soc. 2012, 134, 19314. (b) Kazmaier, U.; Stolz,
D. Angew. Chem., Int. Ed. 2006, 45, 3072. (c) Evans, P. A.; Lawler, M. J.
J. Am. Chem. Soc. 2004, 126, 8642. (d) Ashfeld, B. L.; Miller, K. A.;
Martin, S. F. Org. Lett. 2004, 6, 1321. (e) Hayashi, T.; Okada, A.;
Suzuka, T.; Kawatsura, M. Org. Lett. 2003, 5, 1713. (f) Evans, P. A.;
Uraguchi, D. J. Am. Chem. Soc. 2003, 125, 7158. (g) Evans, P. A.;
Nelson, J. D. J. Am. Chem. Soc. 1998, 120, 5581.
(4) For selected examples on iridium-catalyzed regioselective allylic
alkylation, see: (a) Liu, W.-B.; Reeves, C. M.; Virgil, S. C.; Stoltz, B. M.
J. Am. Chem. Soc. 2013, 135, 10626. (b) Krautwald, S.; Sarlah, D.;
Schafroth, M. A.; Carreira, E. M. Science 2013, 340, 1065. (c) Hamilton,
J. Y.; Sarlah, D.; Carreira, E. M. Angew. Chem., Int. Ed. 2013, 52, 7532.
(d) Chen, W.; Hartwig, J. F. J. Am. Chem. Soc. 2013, 135, 2068.
(e) Liu, W.-B.; Zheng, C.; Zhuo, C.-X.; Dai, L.-X.; You, S.-L. J. Am.
Chem. Soc. 2012, 134, 4812. (f) Ye, K.-Y.; He, H.; Liu, W.-B.; Dai, L.-
X.; Helmchen, G.; You, S.-L. J. Am. Chem. Soc. 2011, 133, 19006.
(g) Lipowsky, G.; Miller, N.; Helmchen, G. Angew. Chem., Int. Ed.
2004, 43, 4595.
(5) For selected examples on other transition metal catalyzed branch-
selective allylic alkylations, for Mo, see: (a) Trost, B. M.; Miller, J. R.;
Hoffman, C. M., Jr. J. Am. Chem. Soc. 2011, 133, 8165. For Fe, see:
(b) Plietker, B. Angew. Chem., Int. Ed. 2006, 45, 1469. For Ru, see:
(c) Sundararaju, B.; Achard, M.; Demerseman, B.; Toupet, L.; Sharma,
G. V. M; Bruneau, C. Angew. Chem., Int. Ed. 2010, 49, 2782.
(6) (a) Cooke, M. L.; Xu, K.; Breit, B. Angew. Chem., Int. Ed. 2012,
51, 10876. (b) Lumbroso, A.; Abermil, N.; Breit, B. Chem. Sci. 2012, 3,
789. (c) Koscher, P.; Lumbroso, A.; Breit, B. J. Am. Chem. Soc. 2011,
133, 20746. (d) Lumbroso, A.; Koschker, P.; Vautravers, N. R.; Breit,
B. J. Am. Chem. Soc. 2011, 133, 2386.
J. Org. Lett. 2013, 15, 3790. (b) Nagao, K.; Yokobori, U.; Makida, Y.;
Ohmiya, H.; Sawamura, M. J. Am. Chem. Soc. 2012, 134, 8982.
(c) Jung, B.; Hoveyda, A. H. J. Am. Chem. Soc. 2012, 134, 1490.
(d) Lin, M.; Kang, G.-Y.; Guo, Y.-A.; Yu, Z.-X. J. Am. Chem. Soc. 2012,
134, 398. (e) Zhang, P.; Le, H.; Kyne, R. E.; Morken, J. P. J. Am. Chem.
Soc. 2011, 133, 9716. (f) Dabrowski, J. A.; Gao, F.; Hoveyda, A. H. J.
Am. Chem. Soc. 2011, 133, 4778. (g) Moran, J.; Preetz, A.; Mesch, R.
A.; Krische, M. J. Nat. Chem. 2011, 3, 287.
(14) For examples, see: (a) Weix, D. J.; Hartwig, J. F. J. Am. Chem.
Soc. 2007, 129, 7720. (b) Graening, T.; Hartwig, J. F. J. Am. Chem. Soc.
2005, 127, 17192. (c) Evans, P. A.; Leahy, D. K. J. Am. Chem. Soc.
2003, 125, 8974.
(15) Benzoylacetic acid 2a was consumed within 3 h in the presence
of the Rh(I)/dppf catalyst, while benzoylacetic acid 2a decomposes in
48 h without the rhodium complex.
(16) For the details, see Supporting Information
(17) Al-Masum, M.; Yamamoto, Y. J. Am. Chem. Soc. 1998, 120,
3809.
(18) For palladium-catalyzed intermolecular coupling of β-ketoacids
with allylic acetates to give linear products, see: Tsuda, T.; Okada, M.;
Nishi, S.; Saegusa, T. J. Org. Chem. 1986, 51, 421. The rhodium-
catalyzed reaction gave the branched product in 26% yield. For detail,
see Supporting Information.
(19) For alternative (σ+π) rhodium intermediate, see ref 3g.
(20) Coupling of carboxylic acid and allene with [Rh(cod)Cl]2 and
(−)-DIOP gave above 90% ee, while the same ligand lead to 0% ee in
this decarboxylative C−C bond formation reaction, which suggests the
reactions may occur by different mechanisms.
(21) The reaction of 1-oxocyclohexane-2-carboxylic acid and
cyclohexylallene 1a gave only a trace amount of the desired product
under the standard conditions.
(22) A linear selective addition of β-ketoacid to the allene followed
by a Carroll rearrangement can be ruled out. For details, see
Supporting Information.
(7) Trost, B. M. Science 1991, 254, 1471.
(8) For a review, see: Hill, A. M. Nat. Prod. Rep. 2006, 23, 256.
(9) Rousseau, G.; Breit, B. Angew. Chem., Int. Ed. 2011, 50, 2450.
(10) (a) For a review on decarboxylative allylations, see: Weaver, J.
D.; Recio, A., III; Grenning, A. J.; Tunge, J. A. Chem. Rev. 2011, 111,
1846. For selected examples about transition-metal-catalyzed
decarboxylative allylic alkylation, with Pd:, see: (b) Trost, B. M.; Xu,
J. J. Am. Chem. Soc. 2005, 127, 17180. (c) Behenna, D. C.; Stoltz, B. M.
J. Am. Chem. Soc. 2004, 126, 15044. (d) Burger, E. C.; Tunge, J. A. Org.
Lett. 2004, 6, 4113. (e) Shimizu, I.; Yamada, T.; Tsuji, J. Tetrahedron
Lett. 1980, 21, 3199. (f) Tsuda, T.; Chujo, Y.; Nishi, S.; Tawara, K.;
Saegusa, T. J. Am. Chem. Soc. 1980, 102, 6381. For Ir: (g) He, H.;
Zheng, X.-J.; Li, Y.; Dai, L.-X.; You, S.-L. Org. Lett. 2007, 9, 4339. For
Ru: (h) Burger, E. C.; Tunge, J. A. Org. Lett. 2004, 6, 2603.
(11) For selected recent examples, see: (a) Yuan, H.-N.; Wang, S.;
Nie, J.; Meng, W.; Yao, Q.; Ma, J.-A. Angew. Chem., Int. Ed. 2013, 52,
3869. (b) Zhong, F.; Yao, W.; Dou, X.; Lu, Y. Org. Lett. 2012, 14,
4018. (c) Yang, C.-F.; Wang, J.-Y.; Tian, S.-K. Chem. Commun. 2011,
47, 8343. (d) Evans, D. A.; Mito, S.; Seidel, D. J. Am. Chem. Soc. 2007,
129, 11583. (e) Lalic, G.; Aloise, A. D.; Shair, M. D. J. Am. Chem. Soc.
2003, 125, 2852.
(12) For the atom-economic addition of carbon nucleophiles to
alkynes and allenes, see: (a) Trost, B. M.; Xie, J.; Sieber, J. D. J. Am.
Chem. Soc. 2011, 133, 20611. (b) Liu, C.; Widenhoefer, R. A. Org. Lett.
2007, 9, 1935. (c) Patil, N. T.; Yamamoto, Y. J. Org. Chem. 2004, 69,
6478. (d) Trost, B. M.; Jakel, C.; Plietker, B. J. Am. Chem. Soc. 2003,
̈
125, 4438. (e) Kadota, I.; Shibuya, A.; Gyoung, Y. S.; Yamamoto, Y. J.
Am. Chem. Soc. 1998, 120, 10262. (f) Trost, B. M.; Gerusz, V. J. J. Am.
Chem. Soc. 1995, 117, 5156.
(13) For selected recent examples on allylic quaternary carbon center
formation reactions, see: (a) Sam, B.; Montgomery, T. P.; Krische, M.
D
dx.doi.org/10.1021/ja411397g | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX