Communications
even at a lower temperature, but the enantioselectivity was
aryl and alkyl substitutents at the a position of 1 have less of
an effect on the yields and enantioselectivities, and they were
indeed good substrates to afford 2a–e in high enantioselec-
tivity (91–98% ee; Table 2, entries 1–5). The hydroxylation of
1 f with a sterically demanding iBu group also proceeded
nicely to give 2 f with 88% ee (Table 2, entry 6). We next
investigated the scope, limitation, and ability of the present
method to discriminate between ester moieties of different
steric bulk. As shown in Table 2, the (R,R)-DBFOX/NiII
catalyst efficiently discriminated between the Me/tBu ester
system (Table 2, entries 1–6), the Et/tBu ester system
(Table 2, entries 7–10), as well as the Pr/tBu ester system
(Table 2, entry 11), and gave at least 90% ee. Notably, the
discrimination between both sterically hindered iPr and tBu
esters was possible, and gave 2l with 82% ee (Table 2,
entry 12). We then attempted the discrimination between
Me and Et ester substituents using 1m. However, the desired
product 2m was obtained in 98% yield with only 12% ee
(Table 2, entry 13). These results indicate that the tBu ester
moiety in 1 is responsible for the enantioselective trans-
formation. Discrimination between the Ph and Et esters of 1n
was also possible with the (R,R)-DBFOX/NiII catalyst, and
gave 2n with 66% ee (Table 2, entry 14), however, the
ee value was not satisfactory. When oxazoline 3b was used,
2n was obtained with 60% ee (Table 2, entry 15). To improve
the enantioselectivity, we attempted the same reaction of 1n
using 2 equivalents of 3a or 3b, but the results were slightly
worse (Table 2, entries 16 and 17). This outcome could be
explained in light of the “matched/mismatched” concept
between two chiral molecules, namely, (R,R)-DBFOX-Ph and
the oxaziridine derivatives. To identify the “matched/mis-
matched” effect of 3 more clearly, we investigated the
reaction of 1n with an alternative oxaziridine 3c, which is
available in both enantiomeric forms. Although the reactivity
of 3c was very low, the reaction of 1n with (+)-3c proceeded
with a high enantioselectivity of 81% ee to produce 2n in the
matched case (Table 2, entry 18). The mismatched case using
not excellent (82–84% ee; Table 1, entries 4–5). Optimization
experiments for both the Lewis acid and the solvent were
carried out to improve both the yield and enantioselectivity of
the transformations (Table 1, entries 6–14), and the combina-
tion of Ni(ClO4)2·6H2O in 1,2-dichloroethane at reflux was
very effective for the dynamic kinetic asymmetric trans-
formation in the a-hydroxylation of malonate derivatives
(Table 1, entry 6). The structure of the oxidant slightly
affected the yield and enantioselectivity of 2a. Changing the
oxidant from cyclic oxaziridine 3a to acyclic 3-(4-nitro-
phenyl)-2-(phenylsulfonyl)-1,2-oxaziridine (3b) resulted in
lower yields with lower ee values (Table 1, entries 15–16).
With the reaction conditions now optimized, we next
explored the range of substrates that could be tolerated
during the hydroxylation reaction (Table 2). First we exam-
ined the malonate derivatives and looked at the effect that
substitutents on the stereogenic center at the a position had
on the hydroxylation reaction. The results showed that the
Table 2: Catalytic enantioselective hydroxylation of racemic 1a–p.
Entry
1
R
R1 R2
t [h]
2
Yield [%] ee [%][a]
1
2
3
4
5
6
7
8
1a CH2Ph Me tBu
48
12
16
36
36
48
48
16
14
62
48
62
14
3
2
3
1
16
24
2
2a
2b
2c
2d
2e
2 f
2g
2h
2i
2j
2k
2l
2m
2n
2n
2n
2n
2n
82
83
84
74
80
65
71
74
72
52
63
48
98
80
55
79
59
24
91
93
98
95
94
88
91
90
94
90
90
82
12
66
60
54
47
81
–
1b Ph
1c Me
1d Et
1e Bu
1 f iBu
Me tBu
Me tBu
Me tBu
Me tBu
Me tBu
1g CH2Ph Et tBu
1h Ph
Et tBu
Et tBu
Et tBu
9
1i
1j
Me
Et
10
11
12
13
14
1k CH2Ph Pr tBu
1l CH2Ph iPr tBu
1m CH2Ph Me Et
(À)-3c provided
a trace amount of product (Table 2,
entries 19). We next carried out the reaction with aryl ethyl
malonate derivatives having larger substituents on the
benzene ring to overcome the poor discrimination between
Ph and Et esters. It should be noted that the enantioselectivity
was finally improved to 88% and 90% ee by the use of larger
aryl substituents, o-fluorophenyl ester 1o and 1-naphthyl
ester 1p, respectively (Table 2, entries 20 and 21). The
configuration of the resulting a-hydroxy malonate 2 can be
explained by the preferential approach of the hydroxylating
agent 3 from the less hindered Si face of the complex formed
between the substrates, a NiII ion, and (R,R)-DBFOX-Ph—
based on the mechanism outlined previously for the enantio-
selective fluorination of malonate derivatives by (R,R)-
DBFOX-Ph/ZnII catalysis (Figure 1).[7] The octahedral com-
plex of 1c/(R,R)-DBFOX-Ph/NiII containing a water mole-
cule was optimized by using the PM3 program (Spartan 06)[8b]
in light of the reported X-ray structure of the (R,R)-DBFOX-
Ph/NiII complex[10] (Figure 2).
1n Me
Et Ph
15[b] 1n Me
Et Ph
16[c]
1n Me
Et Ph
17[b,c] 1n Me
18[d] 1n Me
Et Ph
Et Ph
19[e]
20
21
1n Me
1o Me
1p Me
Et Ph
Et o-FPh
Et 1-naphthyl
2n trace
2o
2p
91
93
88
90
2
[a] Determined by HPLC on a chiral stationary phase. The absolute
configuration of 2n was determined to be S by comparison with the
optical rotation of the known (S)-1-ethyl 3-phenyl-2-hydroxy-2-methyl-
malonate,[1a] and the configuration of 2 was tentatively assumed to be the
same by analogy. [b] 3b was used instead of 3a. The reaction was
performed at RT. [c] 2 equivalents of 3a or 3b was used. [d] (+)-3c was
used instead of 3a. [e] (À)-3c was used instead of 3a.
The utility of a-hydroxy malonate 2 was next demon-
strated by the synthesis of chlozolinate (4; Scheme 2), an
important antifungal agent.[1a,11] Previously, (R)-4 was synthe-
804
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 803 –806