Angewandte
Chemie
One of the problems we faced was a presence of the
stereogenic nitrogen atom in 2, thus increasing the number
of possible stereoisomeric products. A critical conceptual
breakthrough came with realization that the application of
chiral C2-symmetric amines might significantly simplify the
overall stereochemical background. In contrast, the ligands of
this type are novel and their stereochemical performance is
unknown. After extensive experimentation using various C2-
symmetric secondary amines we identified the ligand 3,
derived from bis(naphthyl) amine,[14] which can be prepared
in both S- and R-enantiomeric forms on a large scale (see the
Supporting Information). Optimization of the reaction con-
ditions for 3 with racemic a-AAs has been systematically
studied, including various bases, solvents, reaction temper-
ature, and time as well as stoichiometry of all starting
materials involved. The optimized reaction conditions and
representative examples are summarized in Table 1.
As an illustrative example to discuss the reaction progress
and the stereochemical outcome, we selected one of the
polyfunctional a-AAs, tryptophan (4a; Table 1, entry 1).
Using HPLC analysis we monitored the reaction as a function
of time. As one can see from Figure S1 in the Supporting
Information, the chiral ligand (S)-3 reacted notably faster
with the R enantiomer of 4a, thus rendering the diastereomer
(Sa,Rc)-5a[15] as the kinetically favored product. Over the time
the relative amounts of the diastereomer (Sa,Rc)-5a steadily
increased while that of (Sa,Sc)-5a was proportionally
decreased, thus suggesting that (Sa,Rc)-5a is also thermody-
namically preferred. Final thermodynamic control was ach-
ieved after nearly 4 hours, thus affording (Sa,Rc)-5a as
virtually a single reaction product which was isolated in
95% yield (entry 1). The (Sa,Rc)-5a/(Sa,Sc)-5a ratio was
99.445:0.380 (see the Supporting Information) and did not
change over time (up to 24 h). We can reasonably assume that
the transformation of (Sa,Sc)-5a into the more thermodynami-
cally stable (Sa,Rc)-5a is an epimerization process occurring
by base-catalyzed formation of the corresponding intermedi-
ate enolate.[16]
Other aromatic (rac)-AAs such as Phe (4b; Table 1,
entry 2) and Tyr (4c; entry 3) also easily reacted with (S)-3,
thus giving rise to the thermodynamically controlled (Sa,Rc)-
configured products 5b and 5c, respectively, in excellent yield
and diastereomeric purity. Next, we demonstrated the use of
(S)-3 for the DKR of a series of racemic w-functionalized
amino acids such as methionine (4d; entry 4), glutamine (4e;
entry 5), glutamic acid (4 f; entry 6), and lysine (4g; entry 7).
Interestingly, in these cases the thermodynamic control was
noticeably faster, possibly suggesting some substituent effect
of the side-chain w-functional groups on the intermediate
enolate formation rate. Importantly, the stereochemical out-
come in this series was consistently excellent, thus allowing
isolation of the resultant products (Sa,Rc)-5d–g in both
chemical yields and diastereomeric excesses well above
90%. To complete this part of the study, three racemic
aliphatic AAs, leucine (4h; entry 8), alanine (4i; entry 9) and
valine (4j; entry 10) were tested. Similar to the aromatic
series, the thermodynamic control was achieved in about
24 hours and the expected diastereomers (Sa,Rc)-5h–j were
obtained virtually as the sole reaction products in excellent
chemical yields.
Table 1: Reaction of ligands (S)- and (R)-3 with racemic a-AAs.
As one may assume, use of the (R)-3 might lead to the
formation of the corresponding thermodynamically con-
trolled diastereomeric products 5 having the Ra,Sc absolute
configuration. To confirm this assumption experimentally, we
performed the reactions of (R)-3 with three (rac)-AAs (4c,d,i)
representing different structural types used in the study with
(S)-3. While quite expected, we were pleased to see a flawless
reproducibility of the data previously obtained for prepara-
tion of (Sa,Rc)-configured products (Table 1, entry 11 versus
3; entry 12 versus 4; entry 13 versus 9).
To conclude this part of the study, we would like to
emphasize that very efficient DKR of racemic amino acids
was achieved using the ligands (S)- or (R)-3, the (rac)-AAs
4a–j, and Ni(OAc)2 in nearly stoichiometric ratios. The choice
to use excess (1.1 equiv) (rac)-AA 4 relative to the chiral
ligand 3 is based on the fact that in most cases the former is
much less expensive than the latter. However, in some
exceptional cases when a rare (rac)-AA is difficult to obtain, 3
can be used in 10 mol% excess, thus affording the optimal
stereochemical outcome. This option was demonstrated using
(S)-3 and (rac)-4b (Table 1, entry 14). Furthermore, the
process is conducted under operationally convenient condi-
tions and can be easily scaled up. To demonstrate this
advantageous feature, we performed the DKR of (rac)-4b on
a 10 gram scale using (S)-3. As shown in entry 15 in Table 1,
full thermodynamic control was accomplished under 22 hours
Entry
3
rac-AA 4
Yield [%]
(S,R)/(S,S)[a]
t [h]
1
2
3
4
5
6
7
8
(S)-3
(S)-3
(S)-3
(S)-3
(S)-3
(S)-3
(S)-3
(S)-3
(S)-3
(S)-3
(R)-3
(R)-3
(R)-3
(S)-3
(S)-3
Trp (4a)
Phe (4b)
Tyr (4c)
Met (4d)
Gln (4e)
Glu (4 f)
Lys (4g)
Leu (4h)
Ala (4i)
95
94
99
97
92
92
91
94
97
93
98
98
96
95
94
>97:3
>97:3
>97:3
>97:3
>96:4
>96:4[b]
>96:4[c]
>96:4
>97:3
>97:3
>97:3[d]
>97:3[d]
>97:3[d]
>97:3
>97:3
24
24
24
3
3
10
4
24
24
24
24
5
24
24[e]
22[f]
9
10
11
12
13
14
15
Val (4j)
Tyr (4c)
Met (4d)
Ala (4i)
Phe (4b)
Phe (4b)
[a] Ratio was based on reverse-phase HPLC analysis. [b] Ratio was
1
determined on the basis of H NMR analysis. [c] Ratio was determined
on that of the free AA after disassembly. [d] (R,S)/(R,R) ratio. [e] The
ligand (S)-3 was used in 10 mol% excess. [f] Large-scale (10 g) reaction.
Angew. Chem. Int. Ed. 2014, 53, 12214 –12217
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim