Regiospecific decarboxylative allylation of nitriles

Article abstract of DOI:10.1021/ol902065p

[Chemical Equation Presented] Palladium-catalyzed decarboxylative α-allylation of nitriles readily occurs with use of Pd₂(dba) ₃ and rac-BINAP. This catalyst mixture also allows the highly regiospecific α-allylation of nitriles in the presence of much more acidic α-protons. Thus, the reported method provides access to compounds that are not readily available via base-mediated allylation chemistries. Lastly, mechanistic investigations indicate that there is a competition between C- and N-allylation of an intermediate nitrile-stabilized anion and that N-allylation is followed by a rapid [3,3]-sigmatropic rearrangement.

Full text of DOI:10.1021/ol902065p

ORGANIC
LETTERS
2009
Vol. 11, No. 24
5630-5633
Regiospecific Decarboxylative Allylation
of Nitriles
Antonio Recio III and Jon A. Tunge*
Department of Chemistry, The UniVersity of Kansas, Lawrence, Kansas 66045
tunge@ku.edu
Received September 5, 2009
ABSTRACT
Palladium-catalyzed decarboxylative r-allylation of nitriles readily occurs with use of Pd2(dba)3 and rac-BINAP. This catalyst mixture also
allows the highly regiospecific r-allylation of nitriles in the presence of much more acidic r-protons. Thus, the reported method provides
access to compounds that are not readily available via base-mediated allylation chemistries. Lastly, mechanistic investigations indicate that
there is a competition between C- and N-allylation of an intermediate nitrile-stabilized anion and that N-allylation is followed by a rapid [3,3]-
sigmatropic rearrangement.
Decarboxylative allylation reactions are a powerful method
for the allylation of a wide variety of nucleophiles under
neutral conditions.1-3 While the decarboxylative allylation
of enolates has received the most attention, relatively little
attention has been paid to nitrogen-containing carbon nu-
cleophiles.2,3 Given the prevalence of nitrogen in biologically
active molecules, we are interested in extending decarboxy-
lative couplings to allow facile incorporation of nitrogen. In
Tsuji’s pioneering work on decarboxylative allylation, he
showed that R-allylation of a nitrile could occur at 100 °C
in dioxane, albeit with substantial amounts of undesirable
decarboxylative protonation.3a That said, Tsuji provided the
proof-of-principle that we have chosen to build upon to
develop milder, regiospecific allylation of nitriles that we
report herein.
The work of Tsuji,3a Saegusa,3b Darensbourg,4 and
Shibasaki3d has shown that the decarboxylation of R-cyano
acetates can provide access to metalated nitriles5 without
the need for strongly basic reagents. We posited that the
absence of basic proton shuttles would allow us to generate
nitrile-stabilized carbanions regiospecifically in the pres-
ence of more acidic functional groups.6 Before we could
approach that problem it was critical that we optimize the
reaction conditions to promote allylation and prevent
unwanted protonation. Toward this end, a variety of
catalyst/ligand combinations were evaluated for their
(1) (a) Shimizu, I.; Yamada, T.; Tsuji, J. Tetrahedron Lett. 1980, 3199.
(b) Rayabarapu, D. K.; Tunge, J. A. J. Am. Chem. Soc. 2005, 127, 13510.
(c) Waetzig, S. R.; Tunge, J. A. J. Am. Chem. Soc. 2007, 129, 4138. (d)
Weaver, J. D.; Tunge, J. A. Org. Lett. 2008, 10, 4657. (e) Mohr, J. T.;
Behenna, D. C.; Harned, A. W.; Stoltz, B. M. Angew. Chem., Int. Ed. 2005,
44, 6924. (f) Trost, B. M.; Bream, R. N.; Xu, J. Angew. Chem., Int. Ed.
2006, 45, 3109. (g) Singh, O. V.; Han, H. J. Am. Chem. Soc. 2007, 129,
(4) (a) Darensbourg, D. J.; Longridge, E. M.; Holtcamp, M. W.;
Klausmeyer, K. K.; Reibenspies, J. H. J. Am. Chem. Soc. 1993, 115, 8839.
(b) Darensbourg, D. J.; Longridge, E. M.; Holtcamp, M. W.; Khandelwal,
B.; Klausmeyer, K. K.; Reibenspies, J. H. J. Am. Chem. Soc. 1995, 117,
318. (c) Pletnev, A.; Larock, R. C. J. Org. Chem. 2002, 67, 9438.
(5) (a) Fleming, F. F.; Zhang, Z.; Liu, W.; Knochel, P. J. Org. Chem.
2005, 70, 2200. (b) Okugawa, S.; Masu, H.; Yamaguchi, K.; Takeda, K. J.
Org. Chem. 2005, 70, 10515. (c) Fleming, F. F.; Liu, W.; Ghosh, S.;
Steward, O. W. Angew. Chem., Int. Ed. 2007, 46, 7098. (d) Yasui, H.;
Yorimtsu, H.; Oshima, K. Chem. Lett. 2007, 36, 32. (e) Gaudin, J.; Millet,
P. Chem. Commun. 2008, 588. (f) Fleming, F. F.; Liu, W.; Ghosh, S.;
Steward, O. W. J. Org. Chem. 2008, 73, 2803.
774. (h) Wang, C.; Tunge, J. A. J. Am. Chem. Soc. 2008, 130, 8118
.
(2) (a) Burger, E. C.; Tunge, J. A. J. Am. Chem. Soc. 2006, 128, 10002.
(b) Yeagley, A. A.; Chruma, J. J. Org. Lett. 2007, 9, 2879. (c) Bi, H.;
Chen, W.; Liang, Y.; Li, C. Org. Lett. 2009, 11, 3246
.
(3) For decarboxylative couplings of nitrile-stabilized anions see: (a)
Tsuji, J.; Yamada, T.; Minami, I.; Yuhara, M.; Nisar, M.; Shimizu, I. J.
Org. Chem. 1987, 52, 2988. (b) Tsuda, T.; Chujo, Y.; Nishi, S.-i.; Tawara,
K.; Saegusa, T. J. Am. Chem. Soc. 1980, 102, 6381. (c) Waetzig, S. R.;
Rayabarapu, D. K.; Weaver, J. D.; Tunge, J. A. Angew. Chem., Int. Ed.
2006, 45, 4977. (d) Yin, L.; Kanai, M.; Shibasaki, M. J. Am. Chem. Soc.
(6) (a) Darvesh, S.; Grant, A. S.; MaGee, D. I.; Valenta, Z. Can. J. Chem.
1991, 69, 712.
2009, 131, 9610
.
10.1021/ol902065p  2009 American Chemical Society
Published on Web 11/18/2009

ability to promote decarboxylative allylation of 1a at the
1
expense of protonation. As can be seen from Table 1, H
Table 2. Decarboxylative Couplings of Nitriles
Table 1. Results of Ligand Screen
a Contains 5% impurity. b Reaction performed at 25 °C. c Reaction
1
performed at 90 °C. However, H NMR analysis reveals smooth reaction
at 25 °C. d 5 mol % of Pd(PPh3)4 at 25 °C.
Scheme 1. Regiospecific Nucleophile Generation
NMR spectroscopic analysis of the crude reaction mixtures
shows that the monodentate ligand, triphenylphosphine,
produces only protonation product 3a even in thoroughly
dried toluene.7 The palladium dppe complex employed
by Tsuji also generates substantial protonation product.3a
In contrast, the rac-BINAP ligated palladium catalyst
produced allylation product 2a exclusively. This observa-
tion is in line with our previous observation that rac-
BINAP minimized protonation products in the decarbox-
ylative allylation of R-sulfonyl anions.1d
Having identified rac-BINAP as the optimal ligand for
decarboxylative allylation of nitriles, we proceeded to
investigate the scope of substrates that were compatible
reaction partners. Simple R,R-dialkyl nitriles undergo ally-
lation in good yield at 100 °C in toluene (2a-c, Table 2).
The R-aryl nitriles 2d-f undergo decarboxylative coupling
at room temperature since they are more activated toward
decarboxylation. Thus, the qualitative rates of the reaction
correlate with the pKa values of putative nitrile-stabilized
carbanions; such a correlation typically suggests that decar-
boxylation is rate limiting.1d
To test this idea, a variety of substrates containing
moderately acidic C-H bonds were synthesized and sub-
jected to decarboxylative allylation. In short, the anion
generation and subsequent allylation reactions are highly
regiospecific when pendant aromatic ketones (2g-i, Table
3), aliphatic ketones (2k, 2l), and esters (2j, 2m) were
present; no regioisomeric products were ever observed by
1H NMR spectroscopic analysis of the crude reaction
mixtures. Thus, nitrile-stabilized anions (pKa ≈ 32 in DMSO)
can be generated and selectively allylated in the presence of
the more acidic R-hydrogens of ketones (pKa ≈ 26 in DMSO)
and esters (pKa ≈ 30 in DMSO).8 This suggests that if a
nitrile-stabilized carbanion is formed, it is trapped by
allylation more rapidly than it undergoes proton transfer to
generate the more stable anion.
Next, we turned our attention to the investigation of
the regiospecificity of the allylation when it is conducted
in the presence of adjacent acidic R-hydrogens. Since
decarboxylation allows the kinetic, site-specific generation
of carbanion equivalents, we were curious whether we
could generate and allylate nitrile-stabilized carbanions
in the presence of more acidic functional groups. To do
so would require that we circumvent the thermodynami-
cally favored proton shift to form the more stable
carbanion (Scheme 1).
To probe the limits of the regiospecificity of the anion
generation and allylation further, a substrate (1n) containing
a diethylmalonate fragment was synthesized (Scheme 2).
Subjecting 1n to the standard reaction conditions produced
the R-allylated nitrile 2n without any regioisomerization of
(8) (a) Bordwell, F. G.; Van Der Puy, P.; Vanier, N. R. J. Org. Chem.
1976, 41, 1885. (b) Matthews, W. S.; Bares, J. E.; Bartmess, J. E.; Bordwell,
F. G.; Conforth, F. J.; Drucker, G. E.; Margolin, Z.; McCallum, R. J.;
McCollum, G. J.; Vanier, N. R. J. Am. Chem. Soc. 1975, 97, 7006.
(9) Scheme 3 shows potential mechanistic “cartoons” that ignore the
ligation state of palladium. We presume that the bidentate BINAP ligand
remains bound throughout the catalysis.
(7) Similar protonation products have been observed in the decarboxy-
lative allylations of enolates: Mohr, J. T.; Nishimata, T.; Behenna, D. C.;
Stoltz, B. M. J. Am. Chem. Soc. 2006, 128, 11348.
(10) Tsuda, T.; Okada, M.; Nishi, S.-i.; Saegusa, T. J. Org. Chem. 1986,
51, 421.
Org. Lett., Vol. 11, No. 24, 2009
5631

nisms can be formulated to fit the observed reaction, there
are two mechanisms that allow us to think about several
critical issues.9 First, the allylation reaction could occur
via palladium-catalyzed decarboxylation followed by
allylation (Scheme 3, upper path) or by allylation of the
nitrile followed by decarboxylation and [3,3]-rearrange-
ment (Scheme 3, lower path). While the latter mechanism
can explain the regiospecificity of the allylation, it does
not readily explain the competing protonation reaction that
is observed with many ligands (Table 1). However, the
observation of protonation products is more readily
explained through protonation of the more basic interme-
diates in the top pathway. Thus, the feasibility of the top
pathway was investigated by determining whether pal-
ladium was a competent catalyst for the decarboxylation
of R-cyanoacetates; palladium(II) is known to catalyze
decarboxylation of ꢀ-ketoacids3a,10 and propiolates,1b but
does not influence the rates of decarboxylation of R-sulfonyl
acetic esters.1d
Table 3. Regiospecific Decarboxylative Couplings
Toward this end, the acid 4 was heated with and without
palladium catalyst (Table 4). The results clearly show that Pd(II)
the anion to form the malonate anion. This is remarkable
given that the malonate is ca. 1015 times more acidic than a
typical aliphatic nitrile. Generation of product 2n via more
standard base mediated allylation would be difficult, if not
impossible.
Table 4. Palladium-Catalyzed Decarboxylation
Next, the mechanism of the decarboxylative allylation
was briefly probed. While a number of different mecha-
catalyst
base
temp (°C)
time (h)
conversion (%)
Scheme 2
none
Et3N
Et3N
none
none
115
80
80
15
5
5
<5
>95
55
Pd(OAc)2
Pd(OAc)2
Pd(OAc)2
80
18
>95
is an effective catalyst for the decarboxylation of R-cyano acids.
This is not surprising given the elegant mechanistic studies of
metal-catalyzed decarboxylations of R-cyanoacetates that were
performed by Darensbourg.4a Moreover, the palladium-
catalyzed decarboxylation is kinetically competent to be taking
part in our decarboxylative allylation reaction. Thus, the upper
Scheme 3. Potential Mechanisms for Decarboxylative Allylation
5632
Org. Lett., Vol. 11, No. 24, 2009

mechanistic pathway involving palladium-induced decarboxy-
lation is feasible (Scheme 3).
Scheme 5
Ultimately, on the basis of Darensbourg’s related
studies,4 our working hypothesis is that the formation of
intermediate A is critical for decarboxylation. After
formation of an intermediate like B, or its ion-paired
isomer C, there is another key mechanistic consideration.
Does allylation proceed by kinetic C-allylation (C f 2)
or does it proceed via an N-allylation/[3,3]-sigmatropic
rearrangement (C f E f 2) mechanism?1c,11 Analogous
[3,3]-sigmatropic rearrangements are known to occur rapidly,
even at room temperature.12 These possibilities are most
straightforwardly addressed by investigating the linear:
branched product selectivity when unsymmetrical allyl
alcohol derivatives are utilized. The general preference for
nucleophilic attack at the least substituted allyl terminus of
palladium-π-allyl complexes suggests that C-allylation
should give rise to the linear allylated product,13 whereas
the N-allylation/sigmatropic rearrangement mechanism should
give rise to the branched allylated product (Scheme 4).1c
Alternatively, it is possible that the linear and branched
products derive from reductive elimination of σ-allyl pal-
ladium complexes (Scheme 6).14 Such an inner-sphere
Scheme 6
Scheme 4. C-Allylation vs. N-Allylation
mechanism is expected to proceed via the favored linear
σ-allyl complex to favor formation of the branched allyl
product.
With this in mind, substrates 1o and 1p were subjected to
our standard conditions for decarboxylative allylation (Scheme
5). In both cases, mixtures of branched and linear regioiso-
mers were obtained, with the linear regioisomers as the major
products. However, the amount of branched isomer is much
larger than is typically observed in palladium-catalyzed ally-
lations, which suggests competing C- and N-allylation. Similar
treatment of R,R-dialkyl nitriles 1q and 1r shows that they
undergo decarboxylative allylation with a higher degree of
regioselectivity favoring the branched isomer. Thus, we suggest
that C-allylation and N-allylation mechanisms are competing
and that N-allylation/[3,3] rearrangement is favored by larger,9
or less stable, nitrile-stabilized anions.
In conclusion, we have developed a practical ligand/catalyst
combination for the decarboxylative allylation of nitriles. In
addition, we have shown that the generation of the nitrile-
stabilized anion equivalent is regiospecific; no isomerization
to more stable carbanions is observed even when much more
acidic C-H bonds are present. Mechanistic investigations
suggest that palladium is directly involved in the decarboxy-
lation to form a metalated nitrile. Lastly, the product allylated
nitriles are formed by competing C-allylation and N-allylation
followed by rapid [3,3]-sigmatropic rearrangement.
Acknowledgment. We the National Institute of General
Medical Sciences (1R01GM079644) for financial support of
this work.
(11) C-allylation vs. N-allylation see: (a) Newman, M. S.; Fukunaga,
T.; Miwa, T. J. Am. Chem. Soc. 1960, 82, 873. (b) Clarke, L. F.; Hegarty,
Supporting Information Available: Experimental pro-
cedures and characterization data for all new compounds.
This material is available free of charge via the Internet at
http://pubs.acs.org.
A. F. J. Org. Chem. 1992, 57, 1940
.
(12) Walters, M. A.; Hoem, A. B.; McDonough, C. S. J. Org. Chem.
1996, 61, 55.
(13) Trost, B. M. Acc. Chem. Res. 1980, 13, 385.
(14) A similar mechanism has been proposed for the decarboxylative
allylation of enolates: (a) Keith, J. A.; Behenna, D. C.; Mohr, J. T.; Ma, S.;
Marinescu, S. C.; Oxgaard, J.; Stoltz, B. M.; Goddard, W. A., III J. Am.
Chem. Soc. 2007, 129, 11876.
OL902065P
Org. Lett., Vol. 11, No. 24, 2009
5633