Palladium-catalyzed asymmetric allylic alkylations of polynitrogen- containing aromatic heterocycles

Article abstract of DOI:10.1021/ja205523e

We report the palladium-catalyzed asymmetric allylic alkylation (AAA) reaction of a variety of nitrogen-containing aromatic heterocycles, including pyrazine, pyrimidine, pyridazine, quinoxaline, and benzoimidazole derivatives. The mesityl ester, whose steric bulk prevents competitive deacylation of the electrophile from "hard" nucleophiles, is introduced as a new leaving group in allylic alkylation chemistry. In contrast to our previous studies of AAA reactions with pyridine-based substrates, no precomplexation with a Lewis acid is required before deprotonation with LiHMDS, underscoring the relative acidity of these electron-deficient nucleophiles.

Full text of DOI:10.1021/ja205523e

COMMUNICATION
pubs.acs.org/JACS
Palladium-Catalyzed Asymmetric Allylic Alkylations of
Polynitrogen-Containing Aromatic Heterocycles
Barry M. Trost,* David A. Thaisrivongs, and Jan Hartwig
Department of Chemistry, Stanford University, Stanford, California 94305-5080, United States
S Supporting Information
b
Disappointingly, initial attempts to extend our work beyond
pyridine-based nucleophiles were unsuccessful. Under the previously
optimized conditions, no analogous reactivity was observed
for any other BF3-complexed five- or six-membered nitrogen-
containing aromatic heterocycle, including a variety of seemingly
analogous pyridazine, pyrazine, and pyrimidine nucleophiles.
Metalation with many strong bases was never problematic, as
deuterium quenching studies consistently showed. Instead, it was
clear that lithiated BF3-complexed pyridine-based nucleophiles
possess a unique reactivity toward palladium π-allyl electrophiles
that is not shared with other more electron-deficient heterocycles.
The breakthrough came as we were evaluating the compe-
tency of N-alkylated pyrazinium salts to undergo reaction with
palladium π-allyl species. When an excess of 2,5-dimethylpyr-
azine was treated with para-methoxybenzyl chloride, deprotonated
with LiHMDS, and then treated with racemic cyclohex-2-enyl
pivalate in the presence of a palladium(0) catalyst, the desired pro-
duct was observed in 19% conversion (Scheme 2).11 Unexpectedly,
when the pyrazinium salt was prepared in a separate step, no
reaction was observed (Scheme 2). Contrary to our hypothesis
that an alkylated hetereocycle would prove a suitable nucleophile, it
was the unalkylated species that reacted. Heterocycles such as
pyrazine that are either complexed to a Lewis acid or alkylated
become too electron-deficient to undergo nucleophilic attack
with allylic ligands on palladium. BF3 complexes of pyridines,
however, are not similarly deactivated; in fact, such complexation
is necessary for reactivity. When the reaction was performed in
the absence of an activating species, the desired product was
obtained in 37% conversion, along with significant amounts of
cyclohex-2-enol (Scheme 2). This observation suggested that
deacylation of the electrophile was in part responsible for the low
reaction yield. After examining a variety of allylic leaving groups,
it was discovered that this decomposition could be completely
prevented by employing a more sterically hindered mesityl ester,
a leaving group that, to our knowledge, has never before been
used in an allylic alkyation reaction. Remarkably, despite the
strongly basic conditions, no elimination of the allylic ester to
yield cyclohexadienes was observed.
ABSTRACT: We report the palladium-catalyzed asym-
metric allylic alkylation (AAA) reaction of a variety of
nitrogen-containing aromatic heterocycles, including pyra-
zine, pyrimidine, pyridazine, quinoxaline, and benzoimida-
zole derivatives. The mesityl ester, whose steric bulk
prevents competitive deacylation of the electrophile from
“hard” nucleophiles, is introduced as a new leaving group in
allylic alkylation chemistry. In contrast to our previous
studies of AAA reactions with pyridine-based substrates,
no precomplexation with a Lewis acid is required before
deprotonation with LiHMDS, underscoring the relative
acidity of these electron-deficient nucleophiles.
olynitrogen-containing aromatic heterocycles are ubiquitous
P
structural elements important to a wide range of fine and bulk
chemical fields, including natural products synthesis, medicinal
chemistry, polymer research, and materials science.1 Among the
many protocols that permit the union of preformed nitrogen-
containing aromatic heterocycles with other organic molecules,
transition metal-catalyzed cross-coupling reactions have emerged as
the principal methods that have greatly increased the facility with
which these building blocks can be installed.2 New transforma-
tions that also stereoselectively generate chiral centers provide
access to families of compounds with uniquely defined two- and
three-dimensional structure, a feature with particular relevance to
the needs of pharmaceutical science as well as for other applications.3
Despite progress in transition metal-catalyzed asymmetric cross-
coupling technology,4 relatively few examples exist of such reactions
in which either the nucleophilic or electrophilic coupling partner
in an enantioselective CÀC bond-forming event is a nitrogen-
containing aromatic heterocycle.5
We felt that our recent reports on the palladium-catalyzed
asymmetric allylic alkylation (AAA) reactions of BF3-complexed
2-substituted pyridines (Scheme 1)6,7 would provide a foundation
upon which we might explore the reactivity of other nitrogen-
containing aromatic heterocycles, with the ultimate objective of
developing a general procedure for their use in AAA reactions.
Our aim was to simply deprotonate substituted heterocycles and
react them stereoselectively with palladium π-allyl electrophiles.8
Such an achievement would be particularly significant in stream-
lining the synthetic incorporation of heterocycles with regards to
both atom9 and step economy10 because the metalation event
would not require substrate prefunctionalization, a characteristic
feature of cross-coupling chemistry.
Subsequent optimization experiments revealed that the alky-
lation could be performed comparably well in a variety of ethereal
solvents (Table 1, entries 1À3). Decreasing the amount of base
from 3 to 2 equiv led to a small decrease in yield (entry 4). Little
to no desired reaction was observed when the lithium counterion
was replaced with either potassium or sodium (entries 5 and 6),
nor when LiHMDS was substituted with other strong bases
Received: June 14, 2011
Published: July 20, 2011
r
2011 American Chemical Society
12439
dx.doi.org/10.1021/ja205523e J. Am. Chem. Soc. 2011, 133, 12439–12441
|

Journal of the American Chemical Society
COMMUNICATION
Scheme 1. Attempted Extension of Previous Work with
Pyridines
Table 1. Selected Optimization Experimentsa
entry
solvent
base (equiv)
Nu:Elb
yieldc
eed
1
dioxane
DME
THF
THF
THF
THF
THF
THF
THF
THF
THF
LiHMDS (3.0)
LiHMDS (3.0)
LiHMDS (3.0)
LiHMDS (2.0)
NaHMDS (2.0)
KHMDS (2.0)
NaH (2.0)
1.0:1.0
1.0:1.0
1.0:1.0
1.0:1.0
1.0:1.0
1.0:1.0
1.0:1.0
1.0:1.0
1.3:1.0
1.5:1.0
2.0:1.0
49%
59%
51%
48%
30%
NR
95%
96%
95%
98%
93%
À
2
3
4
5
6
7
NR
À
8
KOt-Bu (2.0)
LiHMDS (2.0)
LiHMDS (3.0)
LiHMDS (4.0)
NR
À
95%
97%
95%
9
53%
71%
67%
10
11
a Reactions run on a 0.109 mmol scale. b Nu = 2-methylpyrazine, El =
cyclohex-2-en-1-yl 2,4,6-trimethylbenzoate. c Isolated yield. d Deter-
mined by chiral HPLC.
Scheme 2. Alkylation of 2,5-Dimethylpyrazine Prevents the
Desired AAA Reaction
(entries 7 and 8), highlighting the critical importance of the
lithium aggregate structure on the course of the desired reaction.
Using 1.5 equiv of the heterocycle proved optimal; when either
more or less was employed, the isolated yield of the desired pro-
duct decreased, although its enantiomeric excess remained largely
unchanged (entires 9À11).
With optimized conditions in hand, the substrate scope of
reaction was explored (Figure 1). In addition to 2-methylpyr-
azine, which could be alkylated with cyclohex-2-en-1-yl 2,4,6-
trimethylbenzoate to give 1 in 71% yield and 97% ee, 2,5-
dimethylpyrazine could be similarly alkylated to give 2 in 66%
yield and 99% ee. Notably, despite the reasonable potential for
bis-alkylation, no such product was observed. An important exten-
sion of the current work is that particularly electron-deficient
nitrogen-containing aromatic heterocycles do not require com-
plexation with an electron-withdrawing Lewis acid such as BF3
to react with palladium π-allyl complexes. This not only renders
the procedure operationally simpler and circumvents potential side
reactions caused by the addition of a reactive Lewis acid, but
also obviates the requirement that the proximal nitrogen atom
be sterically accessible to form a dative bond. In our previous reports,
a substrate such as 2,6-lutidine proved unreactive for this reason. In
this series, where complexation is not required, a substrate such as
2,6-dimethylpyrazine reacts to give 3 in 83% yield and >99% ee.
Figure 1. AAA reactions of nitrogen-containing aromatic heterocycles.
Reactions with both 2-methylquinoxaline and 2,3-dimethyl-
quinoxaline provide the desired products (4 and 5) in 55% and
92% yield, respectively, and in 93% and 98% ee, respectively.
Instead of a 1,4-configuration of nitrogen atoms, as found in both
pyrazines and quinoxalines, a 1,3-configuration is also possible, as
seen in the example of 4-methylpyrimidine, which gives 6in 79% yield
and 80% ee. The reacting center can also be placed between the two
nitrogen atoms of pyrimidine: the reaction of 2-methylpyrimidine
provides alkylated product 7 in a reduced yield of 44% but in high
12440
dx.doi.org/10.1021/ja205523e |J. Am. Chem. Soc. 2011, 133, 12439–12441

Journal of the American Chemical Society
COMMUNICATION
Scheme 3. Diastereo- and Enantioselective Allylic Alkylation
of 5,6,7,8-Tetrahydroquinoxaline
Wiley-Blackwell: New York, 2010. (c) Quin, L. D., Tyrell, J. Funda-
mentals of Heterocyclic Chemistry: Importance in Nature and in the
Synthesis of Pharmaceuticals; Wiley: New York, 2010.
(2) (a) Topics in Current Chemistry; Miyaura, N., Ed.; Springer-
Verlag: New York, 2002; Vol. 219. (b) Metal-Catalyzed Cross-Coupling
Reactions; de Meijere, A., Diederich, F., Eds.; Wiley-VCH: New York,
2004. (c) Billingsley, K. L.; Anderson, K. W.; Buchwald, S. L. Angew.
Chem., Int. Ed. 2006, 45, 3484–3488 and references therein. (d) Kudo,
N.; Perseghini, M.; Fu, G. C. Angew. Chem., Int. Ed. 2006, 15, 1282–1284
and references therein.(e) Li, J. J.; Gribble, G. W. Palladium in Hetero-
cyclic Chemistry: A Guide for the Synthetic Chemist, 2nd ed.; Elsevier
Science: New York, 2007.
enantiomeric excess (97% ee). Reaction of 3-phenyl-6-methyl-
pyridazine, a heterocycle with a 1,2 configuration of nitrogen
atoms, gives 8 in 75% yield and 75% ee.
(3) Roughley, S. D.; Jordan, A. M. J. Med. Chem. 2011, 54, 3451–
3479.
(4) Hayashi, T. In Handbook of Organopalladium Chemistry for
Organic Synthesis; Negishi, E.-i., Ed.; Wiley Interscience: New York,
2002; Chapter III.2.16.
(5) For an example of an enantioselective Negishi cross-coupling
with an indole nucleophile, see: Smith, S. W.; Fu, G. C. J. Am. Chem. Soc.
2008, 130, 12645–12647. For an example of an enantioselective
Kumada cross-coupling with an indole nucleophile, see:Lou, S.; Fu,
G. C. J. Am. Chem. Soc. 2010, 132, 1264–1266.
(6) Trost, B. M.; Thaisrivongs, D. A. J. Am. Chem. Soc. 2008,
130, 14092–14093.
(7) Trost, B. M.; Thaisrivongs, D. A. J. Am. Chem. Soc. 2009,
131, 12056–12057.
(8) For a recent report of a non-asymmetric decarboxylative cou-
pling of heteroaromatic alkanes, see:Waetzig, S. R.; Tunge, J. A. J. Am.
Chem. Soc. 2007, 129, 4138–4139.
(9) Trost, B. M. Science 1991, 254, 1471–1477.
(10) Wender, P. A.; Verma, V. A.; Paxton, T. J.; Pillow, T. H. Acc.
In addition to 6-membered nitrogen-containing heterocycles,
5-membered ring-derived nucleophiles also participate in the
palladium-catalyzed process. Both 1-benzyl-2-methyl-1H-benzo-
[d]imidazole and 1-(4-methoxybenzyl)-2-methyl-1H-benzo[d]imida-
zole react with cyclohex-2-en-1-yl 2,4,6-trimethylbenzoate to pro-
vide the desired products (9 and 10) in 85% and 93% yield,
respectively, and 97% and 99% ee, respectively. Finally, 5-mem-
bered ring-derived electrophiles are also tolerated, as can be seen
in the example of cyclopent-2-en-1-yl 2,4,6-trimethylbenzoate,
which reacts with 1-benzyl-2-methyl-1H-benzo[d]imidazole to
give 11 in 83% yield and >99% ee.
We next asked whether it would be possible to control the
concomitant formation of a second stereocenter on a more
substituted nitrogen-containing aromatic heterocycle, a process
that would then be both diastereo- and enantioselective. To our
delight, submitting 5,6,7,8-tetrahydroquinoxaline to the opti-
mized reaction conditions delivers desired tricycle 12 in 99%
yield, 4.3:1 dr, and >99% ee (Scheme 3).
Chem. Res. 2008, 41, 40–49.
(11) See Supporting Information for assignment of absolute
stereochemistry.
In summary, we have demonstrated that polynitrogen-con-
taining aromatic heterocycles are competent cross-coupling
partners for enantioselective palladium catalysis. Importantly,
no substrate prefunctionalization or preactivation is required
prior to the AAA event. The mesityl ester, a new leaving group in
allylic alkylation chemistry, was singularly effective for the desired
transformation, an observation that highlights its particular
robustness under strongly basic reaction conditions.
’ ASSOCIATED CONTENT
S
Supporting Information. Experimental procedures and
b
analytical data for all new compounds are available free of charge
via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
bmtrost@stanford.edu
’ ACKNOWLEDGMENT
We thank the National Science Foundation for their support
of our programs and Johnson Matthey for generously providing
us with palladium salts.
’ REFERENCES
(1) (a) Eicher, T., Hauptmann, S. The Chemistry of Heterocycles:
Structure, Reactions, Syntheses, and Applications, 2nd ed.; Wiley-VCH:
New York, 2003. (b) Joule, J. A., Mills, K. Heterocyclic Chemistry, 5th ed.;
12441
dx.doi.org/10.1021/ja205523e |J. Am. Chem. Soc. 2011, 133, 12439–12441