Journal of the American Chemical Society
Article
e− (ionic) reactions (Scheme 1A). This method can be
5
2
Table 1. Optimization of the Reaction Conditions
applied to synthesize various α-mono- and disubstituted α-
amino acids V from a glycinate Schiff base bearing
benzophenone imine I or an amino acid Schiff base bearing
aldimine II, respectively. The applicability of alkylating
reagents, however, has been limited to primary and small
secondary alkyl halides, and incorporation of highly congested
tertiary alkyl groups has not been achieved due to the nature of
the nucleophilic substitution reaction.
We thus focused on the use of an α-monosubstituted amino
acid Schiff base bearing benzophenone imine VI, which is
generally introduced to the glycinate. The Schiff base VI could
generate an azaallyl radical species VII via a single electron
6
transfer process, in which the formed radical species would be
7
further stabilized by a captodative effect (Scheme 1B). We
hypothesized that this radical species VII would be
preferentially generated with the assistance of redox active
II
II
I
Cu(II) catalyst (Cu activation: Cu to Cu ). The formed
radical species VII would couple with tertiary alkyl radical IX
generated from tertiary alkyl bromide VIII and Cu(I) catalyst
I
I
II
in proximity to the VII (Cu activation: Cu to Cu ), affording
extremely congested α-amino acid bearing contiguous
8
,9
tetrasubstituted carbon centers X.
Recently, Watson’s
group reported elegant copper catalyzed alkylation of nitro-
a
Conditions: 1a (0.1 mmol), 2a (0.2 mmol), K CO (1.0 equiv), and
1
0
2
3
alkanes. A wide variety of nitroalkanes and α-bromocarbonyl
compounds could be used under mild conditions. Con-
struction of contiguous tetrasubstituted carbon, however, is
very limited. Here, we report that an α-amino acid Schiff base
is a key platform for the synthesis of highly congested
unnatural α-amino acids.
1
tert-butylbenzene (0.5 mL). Yields were determined by H NMR
analysis. 5 mol % of the (CuOTf) ·benzene complex was used.
b
2
by the internal basis of the Cu(II) acetate complex. Other
Cu(I) catalysts afforded product 3aa in low chemical yield. In
addition, other redox active metals, iron and nickel, did not
catalyze the reaction at all.
RESULTS AND DISCUSSION
We selected alkyl bromide 2a as a cross-coupling partner to
construct contiguous tetrasubstituted carbon centers.
■
Having identified the optimal conditions, we evaluated the
scope of the amino acid synthesis (Table 2). The present
catalysis can be run on a gram scale and product 3aa was
isolated in 1.74 g (79% yield). It is noteworthy that bulky tert-
butyl ester substrate could be applicable (3aa′). Various phenyl
glycine derivatives were applicable, including electron-with-
drawing and -donating substituents (Me, OMe, Cl, Ph) (3ba-
3ea). The use of meta-substituted aryl groups (3fa, 3ga)
produced no detrimental effects. An acceptable yield was
obtained when a highly hindered ortho-substituted phenyl
substrate was used under the optimized reaction conditions
1
1,12
Initially we compared the reactivity of a general amino acid-
derived Schiff base bearing aldimine and one bearing
benzophenone imine using a Cu(OAc) /phenanthroline
2
13
complex (Scheme 2). Although the aldimine substrate did
not provide the desired product, benzophenone imine
substrate 1a provided the product in 66% yield.
Scheme 2. Initial Trial Using Amino Acid Schiff Bases
Bearing Aldimine and Benzophenone Imine
(
(
3ha). A naphthyl group was applicable to the present catalysis
3ia). Product 3ja was isolated in 35% yield using an α-thienyl
substrate. Azlactone was also converted to a α-amino acid
derivative bearing contiguous tetrasubstituted carbon centers,
14
although regioisomers were observed in a 4.2/1 ratio (3ka).
We next performed the reaction using various alkyl bromides
. Cyclic scaffolds, cyclohexane, cyclopentane, and cyclobutane
2
were incorporated into amino acid derivatives (3aa−3ad). The
reaction of benzyl ester proceeded without any detrimental
effects (3ae). The aniline functionality survived under the
optimized conditions and product 3af was isolated in 76%
yield as a diastereomixture. Other functionalities, such as alkyl
chloride, aryl iodide, and alkyl bromide, could be used under
slightly modified reaction conditions (3ag, 3ah, and 3aj). The
longer acyclic alkyl chain at the α-position of alkyl bromide 2
decreased the chemical yields due to steric hindrance (3ah−
3aj). Although the chemical yield was moderate when using a
protecting group-free primary hydroxy group (3ak), a tertiary
hydroxy group provided the product in high yield (3al).
Chemoselective α-alkylation of 1 was achieved in the presence
This initial trial prompted us to further optimize the reaction
condition using benzophenone imine. As shown in Table 1,
various bidentate ligands were evaluated. A survey of
phenanthroline derivatives revealed that bathophenanthroline
(
L5) was the optimal ligand, affording product 3aa in 86%
yield. 2,2′-Bipyridyl ligands (L8 and L9) provided product 3aa
in slightly lower yield than L1.
We then examined a series of copper salts. Both cationic
Cu(II) triflate and Cu(II) halides were ineffective in the
present reaction. In contrast, Cu(II) acetate, which can
function as a Lewis acid/Brønsted base cooperative catalyst,
was optimal, indicating that activation of 1a would be achieved
B
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX