Highly efficient nickel-catalyzed cross-coupling of succinic and glutaric anhydrides with organozinc reagents

Article abstract of DOI:10.1021/ja044588b

A nickel-catalyzed alkylation of succinic and glutaric anhydrides with alkyl- and arylzinc reagents has been developed. A dramatic olefin effect has been investigated resulting in the identification of several styrene-based promoters which show pronounced enhancements in reaction rate. The substrate scope with respect to electrophilic and nucleophilic coupling partners has been examined and found to be remarkably broad, allowing for rapid introduction of molecular complexity through the use of functionalized coupling partners. Regioselective alkylation of an unsymmetrical succinic anhydride and a profound effect of pendent coordinating olefins on reaction rate suggest a mechanism involving discrete oxidative addition of the nickel complex into the cyclic anhydride followed by a transmetalation event.

Full text of DOI:10.1021/ja044588b

Published on Web 12/10/2004
Highly Efficient Nickel-Catalyzed Cross-Coupling of Succinic
and Glutaric Anhydrides with Organozinc Reagents
Eric A. Bercot and Tomislav Rovis*
Contribution from the Department of Chemistry, Colorado State UniVersity,
Fort Collins, Colorado 80523
Received September 7, 2004; E-mail: rovis@lamar.colostate.edu
Abstract: A nickel-catalyzed alkylation of succinic and glutaric anhydrides with alkyl- and arylzinc reagents
has been developed. A dramatic olefin effect has been investigated resulting in the identification of several
styrene-based promoters which show pronounced enhancements in reaction rate. The substrate scope
with respect to electrophilic and nucleophilic coupling partners has been examined and found to be
remarkably broad, allowing for rapid introduction of molecular complexity through the use of functionalized
coupling partners. Regioselective alkylation of an unsymmetrical succinic anhydride and a profound effect
of pendent coordinating olefins on reaction rate suggest a mechanism involving discrete oxidative addition
of the nickel complex into the cyclic anhydride followed by a transmetalation event.
Introduction
Anhydrides have only recently garnered attention as competent
acylating agents in metal-mediated reaction manifolds.10
A
Transition metal-catalyzed carbon-carbon bond forming
reactions have become a lynchpin in modern synthetic organic
chemistry.1 Due to the mild nature of reaction conditions
employed, a variety of functional groups are well-tolerated,
making transition metal-catalyzed reactions invaluable in the
context of complex molecule synthesis. In light of the ever-
expanding application of these reaction manifolds, the exploita-
tion of new electrophilic and nucleophilic coupling partners
continues to be of interest.2 New methods for the synthesis of
ketones through the exploitation of novel electrophilic coupling
partners have been of particular interest,3 providing the desired
products without the use of harsh reaction conditions.4
A variety of activated acyl species have seen extensive
application in transition metal-catalyzed synthesis of ketones.
The palladium catalyzed acylation of organostannanes using acid
chlorides,5 first reported by Migita6 and later extensively
developed by Stille,7 represented the first example of the
utilization of acid halides as electrophilic coupling partners.
Following Migita and Stille’s pioneering work, a multitude of
metal-catalyzed methods have been developed employing
activated esters including thioesters8 and aryl trifluoroacetates.9
particularly convenient method for the palladium catalyzed
cross-coupling of boronic acids and mixed acyclic anhydrides,
generated in situ from the parent carboxylic acids, was reported
by Goossen11 and Yamamoto12 independently, providing the
product ketones in good yields. Although the aforementioned
methods represent powerful approaches to the ketone function-
ality, they inherently lack the ability to incorporate stereochem-
ical information during the carbon-carbon bond forming event.
The use of succinic or glutaric anhydride derivatives as
electrophilic coupling partners would provide access to γ- and
δ-keto acid derivatives possessing stereochemically defined
backbones. The realization of this approach would represent a
very efficient entry into 1,4- and 1,5-dicarbonyl systems,
intriguing synthons with a demonstrated importance as inter-
mediates in the synthesis of a variety of heterocyclic systems.13
The desymmetrization of cyclic meso anhydrides by the
addition of heteroatom nucleophiles has been extensively
investigated.14 In contrast, the addition of carbon based nucleo-
philes is less developed. Addition of aromatic nucleophiles to
(10) (a) Jabri, N.; Alexakis, A.; Normant, J. F. Tetrahedron 1986, 42, 1369-
1380. (b) Frost, C. G.; Wadsworth, K. J. Chem. Commun. 2001, 2316-
2317. (c) Cacchi, S.; Fabrizi, G.; Gavazza, F.; Goggiamani, A. Org. Lett.
2003, 5, 289-291. (d) Wang, D.; Zhang, Z. Org. Lett. 2003, 5, 4645-
4648. (e) Yamane, M.; Uera, K.; Narasaka, K. Chem. Lett. 2004, 33, 424-
425. (f) Kazmierski, I.; Bastienne, M.; Gosmini, C.; Paris, J.-M.; Perichon,
J. J. Org. Chem. 2004, 69, 936-942.
(1) Diederich, F.; Stang, P. J. Metal-Catalyzed Cross-coupling Reactions;
Wiley-VCH: Weinheim, 1998.
(2) (a) Luh, T.-Y.; Leung, M.-K.; Wong, K.-T. Chem. ReV. 2000, 100, 3187-
3204. (b) Littke, A. F.; Fu, G. C. Angew. Chem., Int. Ed. 2002, 41, 4176-
4211.
(3) For a review see: Dieter, R. K. Tetrahedron 1999, 55, 4177-4236.
(4) Nahm, S.; Weinreb, S. M. Tetrahedron Lett. 1981, 22, 3815-3818.
(5) The direct addition of main-group organometallics to acid halides is well-
known, see: Shirley, D. A. Org. React. 1954, 8, 28-58.
(6) Kosugi, M.; Shimizu, Y.; Migita, T. Chem. Lett. 1977, 1423-1424.
(7) (a) Milstein, D.; Stille, J. K. J. Am. Chem. Soc. 1978, 100, 3636-3638.
(b) Milstein, D.; Stille, J. K. J. Org. Chem. 1979, 44, 1613-1618.
(8) (a) Tokuyama, H.; Yokoshima, S.; Yamashita, T.; Fukuyama, T.Tetrahedron
Lett. 1998, 39, 3189-3192. (b) Liebeskind, L. S.; Srogl, J. J. Am. Chem.
Soc. 2000, 122, 11260-11261. (c) Wittenberg, R.; Srogl, J.; Egi, M.;
Liebeskind, L. S.; Org. Lett. 2003, 5, 3033-3035.
(11) (a) Goossen, L. J.; Ghosh, K. Angew. Chem., Int. Ed. 2001, 40, 3458-
3460. (b) Goossen, L. J.; Ghosh, K. Eur. J. Org. Chem. 2002, 19, 3254-
3267.
(12) (a) Kakino, R.; Narahashi, H.; Shimizu, I.; Yamamoto, A. Bull. Chem. Soc.
Jpn. 2002, 75, 1333-1345. (b) Kakino, R.; Yasumi, S.; Shimizu, I.;
Yamamoto, A. Bull. Chem. Soc. Jpn. 2002, 75, 137-148. (c) Yamamoto,
A. J. Organomet. Chem. 2002, 653, 5-10.
(13) For an excellent review see: Csende, F.; Sta´jer, G. Heterocycles 2000, 53,
1379-1419.
(14) (a) Spivey, A. C.; Andrews, B. I. Angew. Chem., Int. Ed. 2001, 40, 3131-
3134. (b) For a review on alcoholysis of meso anhydrides, see: Chen, Y.;
McDaid, P.; Deng. L. Chem. ReV. 2003, 103, 2965-2983.
(9) Kakino, R.; Shimizu, I.; Yamamoto, A. Bull. Chem. Soc. Jpn. 2001, 74,
371-376.
9
10.1021/ja044588b CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 247-254
247

A R T I C L E S
Bercot and Rovis
Table 1. Initial Ligand Survey
cyclic anhydrides under Friedel-Crafts acylation conditions
provides the product aromatic ketones in moderate to good
yields.15 However, this manifold is inherently limited to aromatic
nucleophiles bearing the proper substitution patterns. The
corresponding alkylation of meso cyclic anhydrides has received
little attention from the synthetic community. Addition of
strongly nucleophilic species such as Grignard reagents to meso
cyclic anhydrides may lead to mixtures of products including
Meerwein-Ponndorf-Verley reduction, over alkylation, and
epimerization.16 Fu and Shintani reported the alkylation of meso-
glutaric anhydride derivatives by aryl Grignard reagents in the
presence of stoichiometric amounts of (-)-sparteine to provide
the corresponding enantioenriched keto acids in good yields and
selectivities.17 The combination of a transition metal-catalyzed
carbon-carbon bond forming process with concurrent desym-
metrization of a meso compound would represent a powerful
method for the synthesis of keto acid derivatives.
prepared and functional group tolerant,21 and control experi-
ments revealed that diethylzinc does not add to cyclohexanedi-
carboxylic anhydride (THF, 23 °C, 6 h).22 Cyclohexanedicar-
boxylic anhydride 1 was subjected to a catalytic amount of a
variety of nickel-ligand complexes in the presence of diethylzinc
in THF at 0 °C. For instances in which the reaction was slow
as indicated by TLC, it was allowed to warm to ambient
temperature. Catalytic amounts of electron-deficient styrene 3
were added to accelerate reductive elimination in lieu of
potential â-hydride elimination from the presumed acyl-alkyl
nickel intermediate (vide infra), following the precedent of
Knochel.23
We recently reported an alkylative anhydride desymmetri-
zation of cyclic anhydrides in the presence of catalytic amounts
of a nickel catalyst providing the product keto acids in good
yields under mild conditions (eq 1).18 Herein, we detail a full
account of the nickel catalyzed anhydride alkylation including
expansion of the substrate scope with respect to both nucleo-
philic and electrophilic coupling partners, the further refinement
of a nickel (II) precatalyst obviating the need for the use of
air-sensitive nickel(0) compounds, and the results of mechanistic
investigations which shed light on some steps of the catalytic
cycle.
The efficiency of the reaction is highly dependent on the
ligand employed (Table 1). Trialkyl- and triarylmonodentate
phosphines proved rather ineffective under the prescribed
reaction conditions (entries 1-4). Bidentate phosphine com-
plexes are more efficient than adducts derived from monodentate
phosphine ligands and also exhibit a dependence on bite angle.
Bis(diphenylphosphino)ethane (DPPE) furnishes the desired
product in good yield while bis(diphenylphophino)butane
(DPPB) is less effective (entries 5 and 6). The most competent
ligands are 2,2′-bipyridyl (bpy) and (2-diphenylphosphino)-
ethylpyridine (pyphos)24 each efficiently supplying the desired
addition product 2 in less than 3 h at 0 °C (entries 7 and 8).
Each of these ligand-metal complexes deliver analytically pure
material upon acid-base workup.
Promoter Effects. Our working hypothesis regarding the
reaction mechanism involved the initial oxidative addition of
the low-valent nickel catalyst to the electron-deficient C-O
bond of the anhydride followed by a transmetalation event
providing an acyl-alkyl nickel intermediate which upon reductive
elimination provides the desired alkylation product (vide infra,
Scheme 2, Pathway A). Electron-deficient olefins have been
Initial Ligand Screen. Our initial survey began with the
known insertion of low-valent nickel complexes into cyclic
anhydrides.19 We envisioned that in the presence of a suitable
nucleophilic coupling partner the initial oxidative addition adduct
could be intercepted, providing the desired keto acid under mild
conditions. Organozinc reagents were chosen for several
reasons: alkyl zinc reagents are known to transmetalate cyclic
nickel alkoxide intermediates,20 alkyl zinc reagents are easily
(15) (a) Fieser, L. F.; Seligman, A. M. J. Am. Chem. Soc. 1938, 60, 170-176.
(b) Ranu, B. C.; Jana, U. J. Org. Chem. 1999, 64, 6380-6386. (c) Seed,
A. J.; Sonpatki, V.; Herbert, M. R. Org. Synth. 2002, 79, 204-208.
(16) (a) Canonne, P.; Akssira, M. Tetrahedron 1985, 41, 3695-3704. (b) For
the diastereoselective addition of aryl Grignard reagents to cyclic anhy-
drides, see: Real, S. D.; Kronenthal, D. R., Wu, H. Y. Tetrahedron Lett.
1993, 34, 8063-8066. (c) For the addition of alkyl aluminum halides to
anhydrides, see: Cardwell, K.; Hewitt, B.; Ladlow, M.; Magnus, P. J. Am.
Chem. Soc. 1988, 110, 2242-2248.
(17) Shintani, R.; Fu. G. C. Angew. Chem., Int. Ed. 2002, 41, 1057-1059.
(18) (a) Bercot, E. A.; Rovis, T. J. Am Chem. Soc. 2002, 124, 174-175. (b) It
was subsequently found that palladium catalysts participate in this type of
reaction manifold, culminating in the discovery of a catalytic enantiose-
lective anhydride alkylation proceeding with uniformly high selectivities:
Bercot, E. A.; Rovis, T. J. Am. Chem. Soc. 2004, 126, 10248-10249.
(19) (a) Sano, K.; Yamamoto, T.; Yamamoto, A. Chem. Lett. 1983, 115-118.
(b) Sano, K.; Yamamoto, T.; Yamamoto, A. Bull. Chem. Soc. Jpn. 1984,
57, 7, 2741-2747. (c) Yamamoto, T.; Sano, K.; Yamamoto, A. J. Am.
Chem. Soc. 1987, 109, 1092-1100. (d) Castan˜o, A. M.; Echavarren, A.
M. Tetrahedron Lett. 1990, 31, 4783-4786. (e) Castan˜o, A. M.; Echavarren,
A. M. Tetrahedron Lett. 1993, 34, 4361-4362.
(21) Knochel, P.; Singer, R. D. Chem. ReV. 1993, 93, 2117-2188.
(22) The addition of arylzinc halides to maleic anhydride occurs at slightly
elevated temperatures, see: Tarbell, D. S. J. Am. Chem. Soc. 1938, 60,
215-216.
(20) (a) Montgomery, J. Acc. Chem. Res. 2000, 33, 467-473. (b) Amarasinghe,
K. K. D.; Chowdhury, S. K.; Heeg, M. J.; Montgomery, J. Organometallics
2001, 20, 370-372. (c) Montgomery, J. Angew. Chem., Int. Ed. 2004, 43,
3890-3908.
(23) Giovannini, R.; Studemann, T.; Dussin, G.; Knochel, P. Angew. Chem.,
Int. Ed. Engl. 1998, 37, 2387-2390.
(24) Toto, S. D.; Doi, J. T. J. Org. Chem. 1987, 52, 4999-5003.
9
248 J. AM. CHEM. SOC. VOL. 127, NO. 1, 2005

Ni-Catalyzed Cross-Coupling of Anhydrides
A R T I C L E S
contaminated with byproducts29 even after extended reaction
times (entries 5 and 6).30
Table 2. Promoter Survey
Anhydride Substrate Scope. A variety of succinic anhydride
derivatives undergo alkylation (Table 3). Both cis- and trans-
cyclohexanedicarboxylic anhydride provide the desired products
2 and 7 with no loss of stereochemical integrity (entries 1 and
2). Cyclohexenedicarboxylic anhydride derivatives 8, 10 and
12 undergo smooth alkylation providing keto acids 9, 11, and
13 in good yield (entries 3-5). Substituents at the â-position
of the anhydride are well tolerated, as typified by anhydride
14, the Diels-Alder adduct of 2,4-hexadiene and maleic
anhydride, which upon alkylation provides acid 15 containing
4 contiguous stereocenters (entry 6). Tricyclic anhydride adducts
efficiently participate in this reaction manifold providing ready
access to endo- and exo-bicyclic keto acids (entries 7-12).
Anhydride 28, readily available in one step from cyclohep-
tatriene and maleic anhydride,31 provides the desymmetrized
product 29, giving access in two steps to a stereochemically
defined hexasubstituted cyclohexane core (entry 13). Fused
cyclopentane and cyclobutane anhydrides 30 and 32 offer
products 31 and 33 in 71 and 61% yield respectively with no
evidence of epimerization (entries 14 and 15). Monocyclic
succinic anhydrides also take part in the chemistry. Tartaric acid
derived anhydrides 34 and 36 afford acylic keto acids 35 and
37 in good yield with no evidence of elimination of the
potentially labile acetates (entries 16 and 17). meso-Dimethyl-
succinic anhydride 38 as well as the homologated ester-
containing derivative 40 each furnish the corresponding 1,4-
dicarbonyl compounds 39 and 41 with no evidence of
epimerization or ester hydrolysis (entries 18 and 19).
a All reactions run in the presence of Ni(COD)2 (5 mol %), bpy (6 mol
%), and Et2Zn (1.2 eq.) at 0 °C for indicated time.
shown to facilitate the reductive elimination of dialkyl nickel
intermediates.25 The acceleration of reductive elimination is
thought to proceed through initial coordination of the olefin to
the metal center resulting in a π-back-bonding interaction
leading to weakening of the carbon-metal σ-bonds.26 Knochel
and co-workers have exploited the use of catalytic amounts of
electron-deficient olefins in the context of Csp3-Csp3 cross-
coupling, resulting in dramatic enhancements of reaction rate
and efficiency.27 We had noted the benefit of these olefinic
additives in our reaction manifold and sought to quantify this
effect. Therefore, a systematic study of the reductive elimination
promoter was undertaken to probe its effect on reaction rate
and efficiency.
With the catalytic alkylation applied to succinic anhydride
derivatives we sought to further expand the scope of the current
method to glutaric anhydrides. Although the bpy-derived nickel
catalyst was quite effective in the context of succinic anhydride
electrophiles, glutaric anhydrides failed to provide product under
identical conditions. However, the use of the pyphos-derived
nickel catalyst provides for an effective alkylation of glutaric
anhydrides (Table 4).32 Glutaric anhydrides bearing a variety
of substitution at the 4-position participate in this transformation.
Methyl glutaric anhydride 42 as well phenyl glutaric anhydride
44 each afford good yields of the corresponding ethyl ketones
43 and 45 (entries 1 and 2). Heteroatom substituents at the
4-position are also well tolerated providing â-benzyloxy acid
47 and N-protected â-amino acid derivative 49 in modest yields
(entries 3 and 4). The HMGA derived anhydride 50, bearing a
potentially labile â-acetate group, efficiently provides keto acid
51 in 75% yield (entry 5). Substitution at the R-position of the
anhydride does not adversely affect the reaction with anhydride
52 smoothly providing acylic keto acid 53 containing a skipped
1,3-syn-dimethyl relationship in the backbone (entry 6). Bridged
bicyclic glutaric anhydrides 54 and 56 afford 1,3-syn-cyclo-
pentane 55 and 1,3-syn-cyclohexane 57 in 90 and 88% yield
The nature of the promoter has a dramatic effect on reaction
rate and conversion (Table 2).28 The reaction proceeds smoothly
in the absence of promoter at 0 °C to provide a good yield of
addition product 2 although the reaction requires 21 h to proceed
to completion (entry 1). The addition of 10 mol % (2 equiv to
catalyst) of 4-trifluoromethyl styrene (3) results in a dramatic
rate acceleration, providing the product acid 2 in 80% isolated
yield in less than 5 min (entry 2). Variation of the electronic
properties of the styrene-based promoter has little effect on the
efficiency of the transformation, but a small effect on the
reaction rate was observed. Styrene (4) provides the product
keto acid in 68% yield in 25 min. while the more electron
deficient 4-fluorostyrene (5) provides 2 in 80% yield in 30 min
(entries 4 and 5). The use of trifluorotoluene or acrylonitrile
seems to suppress reactivity providing poor yields of 2
(25) (a) Yamamoto, T.; Yamamoto, A.; Ikeda, S. J. Am. Chem. Soc. 1971, 93,
3350-3359. (b) Yamamoto, T.; Abla, M.; Murakami, Y. Bull. Chem. Soc.
Jpn. 2002, 75, 1997-2009.
(29) Reactions that are sluggish often result in the formation of ester byproducts
presumably arising from the slow oxidation of the alkyl zinc reagent by
adventitious oxygen; for a similar example, see: Katritzky, A. R.; Luo, Z.
Heterocycles 2001, 55, 1467-1474.
(30) On the basis of our previous report, trifluorotoluene seems to be an effective
promoter if used in higher concentration (30 mol % trifluorotoluene and 1
mol % catalyst).
(31) Schueler, P. E.; Rhodes, Y. E. J. Org. Chem. 1974, 39, 2063-2069.
(32) Qualitative observations suggest bpy is the generally preferred ligand for
succinic anhydrides and pyphos is the preferred ligand for glutaric
anhydrides.
(26) Giovannini, R.; Studemann, T.; Devasagayaraj, A.; Dussin, G.; Knochel,
P. J. Org. Chem. 1999, 64, 3544-3553.
(27) (a) Giovannini, R.; Knochel, P. J. Am. Chem. Soc. 1998, 120, 11186-
11187. (b) Piber, M.; Jensen, A. E.; Rottlander, M.; Knochel, P. Org. Lett.
1999, 1, 1323-1326. (c) Jensen, A. E.; Knochel, P J. Org. Chem. 2002,
67, 79-85.
(28) Reaction times were determined by the disappearance of starting material
as observed by TLC. Each reaction was assayed every 5 min for the first
30 min, then every 30 min for an additional 3 h. After 3 h each reaction
was assayed every 6 h to an ultimate time of 42 h.
9
J. AM. CHEM. SOC. VOL. 127, NO. 1, 2005 249

A R T I C L E S
Bercot and Rovis
Table 3. Succinic Anhydride Substrate Scope
a Reactions conducted in the presence of Ni(COD)2 (5 mol %), bpy (6 mol %), 4-F-sty (10 mol %), and 1.2 equiv of Et2Zn at 0 °C in THF unless
otherwise stated. b Reaction conducted using Ni(COD)2 (10 mol %), bpy (12 mol %), and 4-F-sty (20 mol %). c 4-CF3-sty used as promoter. d Isolated
as the corresponding methyl ester.
respectively as single diastereomers (entries 7 and 8). Alkylation
of polyoxygenated bicyclic anhydride 58 proceeds uneventfully
to provide the pentasubstitued cyclopentane 59, containing
several labile elements, in 88% yield (entry 9).
Nucleophile Scope. The reaction is not limited to the use of
diethyl zinc as the nucleophilic coupling partner (Table 5).
Subjection of anhydride 1 to the reaction conditions using Me2-
Zn and Ph2Zn provides the corresponding methyl and phenyl
ketones in excellent yields (entries 1 and 2). The secondary
diorganozinc reagent, diisopropyl zinc, smoothly provides the
isopropyl ketone with no signs of iso-propyl to n-propyl
isomerization (entry 3). Alkylzinc halides are also competent
coupling partners. Although the bpy derived catalyst system is
less efficient, providing the cross-coupled product in 66% yield
using 20 mol % catalyst loading, we found that changing the
ligand to the bidentate phosphine DPPE resulted in a more
effective catalyst for alkylzinc halide coupling (entries 4 and
5). Functionalized zinc reagents were also found to be compat-
ible with this reaction manifold, allowing for incorporation of
reactive functionality in both coupling partners (entry 6). Diaryl
and dialkyl zinc reagents33 formed in situ from the corresponding
(33) Currently, alkenyl and alkynyl zinc reagents are beyond the scope of this
transformation.
9
250 J. AM. CHEM. SOC. VOL. 127, NO. 1, 2005

Ni-Catalyzed Cross-Coupling of Anhydrides
A R T I C L E S
Table 4. Glutaric Anhydride Substrate Scope
Scheme 1. Use of an Air Stable Precatalyst
(Ni(acac)2) was used as a catalyst precursor affording an active
nickel (0) catalyst upon in situ reduction by diethyl zinc. Taking
advantage of the most active reductive elimination promoter,
4-trifluoromethyl styrene, catalyst loading was reduced to 1 mol
% in the case of cyclohexylsuccinic anhydrides 1 and 8
providing the desired keto acid products 2 and 9 in excellent
yield with a longer reaction time (Scheme 1, eq 7). Each reaction
was executed on multigram scale (15 mmol) and upon comple-
tion, simple acid-base workup provided analytically pure
material circumventing the need for chromatographic separation.
The nickel (II) pre-catalyst can also be applied to glutaric
anhydrides. Methylglutaric anhydride 42 undergoes smooth
alkylation in the presence of 5 mol % pyphos derived catalyst
providing the desired δ-keto acid 43 in excellent yield (Scheme
1, eq 8).
Mechanistic Considerations. Two potential mechanistic
pathways for this reaction are depicted in Scheme 2. Pathway
A involves oxidative addition of the low-valent nickel complex
to the electron-deficient C-O bond of the cyclic anhydride to
yield the cyclic acyl-nickel carboxylate I. Transmetalation of I
provides acyl-alkyl nickel species III which could then undergo
reductive elimination to provide the desired product along with
regeneration of the active catalyst. Another potential route,
depicted as pathway B, involves the initial formation of an alkyl
nickel intermediate II from direct alkyl group transfer from the
zinc reagent to the catalyst.36 Intermediate II can then react with
the starting anhydride through addition of the alkyl group to
the anhydride directly providing tetrahedral intermediate IV
which upon collapse would provide product. In our previous
work on the decarbonylative cross-coupling of cyclic anhy-
drides,37 we observed the corresponding direct addition product
in minor amounts in some reactions, suggesting that the
alkylation product in the present reaction manifold may arise
from a similar mechanistic pathway, namely a discrete oxidative
addition of the low-valent nickel catalyst (Pathway A). Fur-
thermore, a wealth of precedent exists for the stoichiometric
interaction of electron-rich nickel complexes and cyclic anhy-
a Reactions conducted in the presence of Ni(COD)2 (10 mol %), pyphos
(12 mol %), and Et2Zn (1.2 eq.) at 0 °C in THF unless otherwise stated.
b Reaction conducted using Ni(COD)2 (5 mol %), pyphos (6 mol %), and
4-F-sty (10 mol %).
lithium or Grignard reagents also participate in the reaction
demonstrating that byproduct lithium and magnesium salts are
benign (entries 7-12). It is noteworthy that the aryllithium
reagents (entries 10-12) were generated from the aryl bromide
by treatment with n-butyllithium, producing butyl bromide as
a byproduct which does not seem to affect the outcome of the
reaction.34
Application of an Air-Stable Precatalyst. The nickel (0)
pre-catalyst, Ni(COD)2, although commercially available, is
extremely air-sensitive requiring the use of an inert atmosphere
glovebox. In the interest of making the reaction protocol more
amenable in a practical sense, we sought the development of
an air-stable catalyst precursor thereby obviating the need for
glovebox manipulation.35 Anhydrous nickel (II) acetylacetonate
(36) Alkyl-transition metal intermediates arising from alkyl or aryl group transfer
from a main-group organometallic have been invoked in related reaction
manifolds; see: (a) Bogdanovic, B.; Schwickardi, M. Angew. Chem., Int.
Ed. 2000, 39, 4610-4612. (b) Fu¨rstner, A.; Leitner, A. Angew. Chem.,
Int. Ed. 2002, 41, 609-612. (c) Feringa, B. L. Acc. Chem. Res. 2000, 33,
346-353. (d) Hayashi, T.; Yamasaki, K. Chem. ReV. 2003, 103, 2829-
2844.
(34) It should be noted that generation of aryllithium reagents from aryl iodides
by treatment with n-butyllithium leads to no observed product.
(35) Although all of our observations suggest that Ni(acac)2 and Ni(COD)2 are
completely interchangeable in this reaction, Ni(COD)2 was used for
consistency in our studies regarding the substrate scope.
(37) O’Brien, E. M.; Bercot, E. A.; Rovis, T. J. Am. Chem. Soc. 2003, 125,
10498-10499.
9
J. AM. CHEM. SOC. VOL. 127, NO. 1, 2005 251

A R T I C L E S
Bercot and Rovis
Table 5. Nucleophile Scope
a All reactions conducted in the presence of Ni(COD)2 (10 mol %), indicated ligand (12 mol %), and RZnX (1.2 eq.) at 0 °C in THF unless otherwise
stated. b Reaction conducted using Ni(COD)2 (5 mol %), bpy (6 mol %), and 4-F-sty (10 mol %). c Reaction conducted using Ni(COD)2 (20 mol %), bpy
(22 mol %), and 4-F-sty (40 mol %). d sCF3-sty used as promoter. e Isolated as the corresponding methyl ester.
been characterized by crystallography.39 Nevertheless, the slight
Scheme 2. Proposed Mechanistic Pathways
difference in our optimized reaction conditions compounded
with the caveat that stoichiometrically generated species are not
always intermediates in catalytic cycles spurred us to investigate
the mechanism of this reaction in some detail.
We reasoned the use of an unsymmetrical anhydride may
provide information regarding the active mechanistic pathway
based on the regioisomer obtained. The regioselective alkyla-
tion40 and reduction41 of unsymmetrical succinic anhydrides is
well-known. Reaction principally takes place at the carbonyl
group adjacent to the more highly substituted carbon atom. This
observation has been attributed to a more favored nucleophilic
attack trajectory42 that proceeds over the least sterically hindered
R-carbon thereby delivering the nucleophile to the carbonyl
neighboring the more substituted R-carbon (Figure 1).43 In the
event, subjection of 2,2-dimethylsuccinic anhydride 60 to the
standard reaction conditions provides a > 9:1 regioisomeric
mixture of acids 61 and 62 favoring addition at the carbonyl
adjacent to the least substituted R-carbon (Scheme 3). This
observation suggests that the anhydride is activated by the
oxidative addition of the low-valent nickel catalyst as depicted
(39) Fischer, R.; Walther, D.; Kempe, R.; Sieler, J.; Scho¨necker, B. J.
Organomet. Chem. 1993, 447, 131-136.
(40) (a) For the addition of Grignard reagents; see: Canonne, P.; Kassou, M.;
Akssira, M. Tetrahedron Lett. 1986, 27, 2001-2004; Berrier, C.; Bonnaud,
B.; Patoiseau, J. F.; Bigg, D. Tetrahedron 1991, 47, 9629-9640. (b) For
drides indicating that oxidative addition clearly occurs when
the two are admixed.38 In one instance, this intermediate has
the addition of enolates; see: Murray, W. V.; Wachter, M. P. J. Org. Chem.
1990, 55, 3424-3426. (c) For Friedel-Crafts acylation; see: Hagishita,
S.; Kuriyama, K. Bull. Chem. Soc. Jpn. 1982, 55, 3216-3224.
(41) (a) Bloomfield, J. J.; Lee, S. L. J. Org. Chem. 1967, 32, 3919-3924. (b)
Cross, B. E.; Stewart, J. C. Tetrahedron Lett. 1968, 3589-3590. (c) Bailey,
D. M.; Johnson, R. E. J. Org. Chem. 1970, 35, 3574-3576.
(38) (a) Trost, B. M.; Chen, F. Tetrahedron Lett. 1971, 12, 2603-2607. (b)
Uhlig, V. E.; Fehske, G.; Nestler, B. Z. Z. Anorg. Allg. Chem. 1980, 465,
141-146. (c) Komiya, S.; Yamamoto, A.; Yamamoto, T. Chem. Lett. 1981,
193-196. (d) Scho¨necker, B.; Walther, D.; Fischer, R.; Nestler, B.;
Bra¨unlich, G.; Eibisch, H.; Droescher, P. Tetrahedron Lett. 1990, 31, 1257-
1260.
(42) (a) Bu¨rgi, H. B.; Lehn, J. M.; Wipff, G. J. Am. Chem. Soc. 1974, 96, 1956-
1957. (b) Bu¨rgi, H. B.; Dunitz, J. D.; Lehn, J. M.; Wipff, G. Tetrahedron
1974, 30, 1563-1572.
(43) Kayser, M. M.; Morand, P. Tetrahedron Lett. 1979, 695-698.
9
252 J. AM. CHEM. SOC. VOL. 127, NO. 1, 2005

Ni-Catalyzed Cross-Coupling of Anhydrides
A R T I C L E S
material or product with reaction intermediates. To establish
whether the proposed interaction was intra- or intermolecular
in nature, cyclohexanedicarboxylic anhydride 1 was subjected
to standard reaction conditions in the presence of 1 equiv of
cyclohexene (Scheme 4, eq 11). After a 15 min reaction time,
less than 5% of the desired alkylation product was observed,
suggesting that the rate enhancement is a consequence of an
intramolecular coordination event.
Figure 1. Nucleophilic trajectory.
Scheme 3. Regioselective Anhydride Alkylation
With evidence of olefinic backbone acceleration arising from
an intramolecular coordination event we still could not rule out
the possibility that the cyclohexenyl anhydride was more
reactive due to geometric rigidity imparted by unsaturation in
the backbone. We reasoned that if the rate difference is a
consequence of geometric constraint, then substitution of the
olefin should not affect the reaction rate; similarly, if the rate
enhancement is a product of an intramolecular coordination
event, then differences in olefin structure should have a sizable
impact on the course of the reaction. To probe this question, a
competition experiment was performed in which a 1:1 mixture
of cyclohexenyl anhydride 8 and dimethylcyclohexenyl anhy-
dride 10 was subjected to standard reaction conditions in the
presence of 1 equiv of nucleophile (Scheme 5). After 1.5 h, the
in pathway A. The regiochemical outcome can be rationalized
by the steric encumbrance imparted by the geminal dimethyl
substituents effectively shielding the adjacent carbonyl thus
favoring precoordination and subsequent oxidative addition of
the catalyst to the more accessible C-O bond.
During the course of our olefinic promoter studies we noted
that subjection of anhydride 8 to standard reaction conditions
using 4-fluorostyrene as the promoter provided the desired
product 9 in less than 5 min. To our surprise, when the same
reaction was carried out in the absence of styrenic additive, the
product keto acid 9 was isolated in excellent yield in 15 min
constituting a striking rate enhancement when compared to the
corresponding saturated anhydride derivative (Scheme 4, eq 10).
Scheme 5. Competition Experiment
Scheme 4. Backbone of Olefin Effects
reaction was quenched and the unpurified reaction mixture was
esterified and analyzed by gas chromatography to obtain ratios
of all four possible reaction components. In the absence of
styrenic promoter, virtually complete consumption of anhydride
8 in preference to anhydride 10 is observed. Even in the presence
of 4-fluorostyrene anhydride 8 is preferentially consumed,
although keto ester 11 is observed in minor amounts compared
to ester 9. These observations suggest that intramolecular
coordination is the contributing factor to the increase in reaction
rate when unsaturation is present in the anhydride backbone.
The efficiency of coordination is influenced by the nature of
the olefin with tetrasubstituted derivatives being far less effective
than their disubstituted counterparts.
In light of our mechanistic investigations and based on
previously described literature precedent, we propose that the
present reaction proceeds through a manifold which involves
discrete oxidative addition of the low-valent nickel catalyst
followed by alkyl-group transfer and catalyst turnover as
illustrated in pathway A (Scheme 2). The intermediacy of an
alkyl-nickel intermediate that directly adds to the anhydride
coupling partner is unlikely based on our observations involving
the cross-coupling of a geminally disubstituted succinic anhy-
dride derivative. Furthermore, the rate acceleration imparted by
unsaturation in the anhydride backbone is easiest to rationalize
We attributed the observed acceleration of reaction rate to the
ability of the olefinic backbone to stabilize the oxidative addition
intermediate by an intramolecular coordination event (A) or
facilitate the subsequent reductive elimination of the proposed
acyl-alkyl nickel intermediate (B).44 However, we could not
rule out the possibility of an intermolecular promotion of the
reaction by interaction of the olefinic backone of the starting
(44) A similar effect has been observed by Knochel in the context of nickel-
catalyzed Csp3-Csp3 cross-coupling reactions, see: Refs 23 and 26.
9
J. AM. CHEM. SOC. VOL. 127, NO. 1, 2005 253

A R T I C L E S
Bercot and Rovis
by invoking an intramolecular coordination event, with the
presence of the olefin stabilizing intermediates and/or transition
states in the catalytic cycle.
alkylation of unsymmetrical anhydrides as well as examination
of the unexpected impact of unsaturation present in the
anhydride backbone.
Conclusion. A nickel-catalyzed alkylation of cyclic anhy-
drides with alkylzinc reagents has been developed. The reaction
proceeds rapidly under mild conditions providing the product
keto acids containing up to six contiguous stereocenters in good
yields. A dramatic rate enhancement has been observed by the
addition of styrenic olefins. The substrate scope has been
extended to succinic and glutaric anhydride derivatives bearing
a range of functionality as well as a variety of zinc nucleophiles
including functionalized zinc reagents. An air-stable nickel
precatalyst has also been developed obviating the need for air-
sensitive nickel (0) and allowing for the reduction of catalyst
loading to 1 mol %. Insight into the possible mechanism has
also been gained by the examination of the regioselective
Acknowledgment. We thank Colorado State University, the
donors of the Petroleum Research Fund (administered by the
American Chemical Society), Merck, GlaxoSmithKline, Amgen,
Johnson and Johnson, and Eli Lilly for financial support. E.A.B
thanks Pharmacia for a graduate fellowship. We would also like
to thank Professor Rick Finke (CSU) and Dr. Steve Swallow
(Roche Palo Alto) for helpful discussions.
Supporting Information Available: Experimental procedures
and spectral data for all compounds (PDF). This material is
available free of charge via the Internet at http://pubs.acs.org.
JA044588B
9
254 J. AM. CHEM. SOC. VOL. 127, NO. 1, 2005