reacting enantiomer)9 ) 1.8] (entry 3). The use of 1.2 equiv of
pivaloyl chloride increased the reactivity without loss of en-
antioselectivity (entry 4). Interestingly, the addition of pivalic
acid (20 mol %) improved the reactivity and enantioselec-
tivity (S ) 3.1, entry 5). Pivalic acid would probably serve
as a Brønsted acid (ammonium proton) to activate 6a or a
nucleophilic base10 or Brønsted base (pivalate anion) to assist
the equilibrium between 6a and its acylammonium salt 7a
in the presence of collidine.11 However, the enantioselectivity
was still moderate. Surprisingly, the use of tert-butyl alcohol
gave high enantioselectivity (S ) 31, entry 8). Pivalic acid
promoted the esterification with tert-butyl alcohol as well
as that with isopropyl alcohol (entry 4 vs entry 5, entry 7 vs
entry 8). The dramatic improvement in enantioselectivity
upon switching from isopropyl alcohol to tert-butyl alcohol
suggests that equilibrium between 4a and 7a may be
important for attaining a high level of kinetic resolution.12
These experimental results suggest that the first kinetic
resolutionat the generation step of 7a would occur at a low
level. However, when the esterification step was much slower
than the generation step of 7a, such as in entry 5, higher
enantioselectivity was observed with the second kinetic
control.In fact, higher asymmetric induction was observed
with the dropwise addition of isopropyl alcohol (entry 5 vs
entry 6). Catalyst 1b was inferior to 1a with regard to
enantioselectivity (entry 9).
absence of pivalic acid and dried MS 4Å, the enantioselec-
tivity and reactivity were further increased (entry 3). Thus,
the esterification proceeded even at -40 °C with the use of
DCC without the addition of pivalic acid to give (+)-5a in
38% yield with 94% ee (S ) 56, entry 3).13
To explore the generality and scope of the above 1-induced
kinetic resolution, the esterification of several structurally
diverse carboxylic acids was examined according to method
A (conditions in entry 8 in Table 1) or method B (conditions
in entry 2, Table 2) which were optimized for (()-4a (Table
3). The esterification of not only 4a but also other O-
Table 3. Generality and Scope of the 1-Induced Kinetic
Resolution of Racemic Carboxylic Acids (Method A or B)a
Next, other condensing agents were examined for the
above reaction. As shown in Table 2, the esterification of
Table 2. Kinetic Resolution of (()-4a Induced by 1aa
a Unless otherwise noted, (()-carboxylic acids (0.25 mmol) were in
14
CCl4 (1.5 mL) according to method A or B (see text). b For step 2.
c Isolated yield of esters. The conversion, which was calculated by using
the ee’s of esters and acids, is shown in brackets. d HPLC analysis. e The
S was calculated by using the yield and ee of esters. The S, which was
calculated by using the ee’s of esters and acids, is shown in brackets. See
ref 9. f Toluene was used as solvent. g i-PrOH (0.6 equiv) was used at 0
°C. h The condensation of 8 (0.25 mmol) with 2-oxazolidinone (0.6 equiv)
in CCl4 (1 mL) was carried out in the presence of N,N-diisopropylethylamine
(2 equiv) and t-BuCOCl (1.2 equiv, method A′) or DCC (1.2 equiv, method
B′) at room temperature for 2 h.
temp (°C), yield (%)c
ee (%)d of
entry additive
t (h)b
of (+)-5a (+)-5a, (-)-4a
Se
1
t-BuCO2H -20, 17
41
89, -
92, 49
94, 56
28
2
-
-
-20, 3
[35]
[37]
57 [56]
3f
-40, 38
38 [37]
a Unless otherwise noted, (()-4a (0.50 mmol) was used in CCl4 (1.5
mL). b For step 2. c Isolated yield. The conversion, which was calculated
by using the ee’s of 5a and 4a, is shown in brackets. d HPLC analysis.
e The S was calculated by using the yield and ee of (+)-5a. The S, which
was calculated by using the ee’s of (+)-5a and (-)-4a, is shown in brackets.
See ref 9. f Toluene (1.0 mL) was used instead of CCl4.
protected R-hydroxycarboxylic acids 4b-d gave high S
values (entries 1-6). (()-N-Boc phenylalanine (8),15 (()-
N-Cbz leucine (9), and (()-syn-6-(pyrrolidine-1-carbonyl)
cyclohex-3-enecarboxylic acid (4e) were also suitable sub-
strates (entries 7-11). Although the reaction conditions were
not optimized for each substrate, methods A and B were both
effective for racemic carboxylic acids bearing a Brønsted
base site. On the other hand, the present protocol was not
effective for simple racemic carboxylic acids such as
2-phenylpropanoic acid (10) (entry 12). Nevertheless, the
kinetic resolution of 10 was observed in the condensation
(()-4a with tert-butyl alcohol proceeded more smoothly with
the use of N,N′-dicyclohexylcarbodiimide (DCC) instead of
pivaloyl chloride at -20 °C under the same conditions as
for entry 5 in Table 1 (entry 1). When DCC was used in the
(9) Kagan, H. B.; Fiaud, J. C. Top. Stereochem. 1988, 18, 249.
(10) Pivalate anion might promote the conversion of 7a to 6a.
(11) The results of entries 3 and 4 indicated that unreacted pivaloyl
chloride did not inactivate 1a.
(12) (a) Spivey, A. C.; Arseniyadis, S. Angew Chem. Int. Ed. 2004, 43,
5436. (b) Xu, S.; Held, I.; Kempf, B.; Mayr, H.; Steglish, W.; Zipse, H.
Chem.-Eur. J. 2005, 11, 4751.
(13) A mixed anhydride of 4a and t-BuCO2H was formed in situ (entry
1). However, anhydride of 4a was not formed as a major species under
conditions of entries 2 and 3.
Org. Lett., Vol. 10, No. 15, 2008
3193