V. B. Birman and X. Yang
gy with the majority of KR substrates discussed above, loss
of enantioselectivity was observed with increasing the bulk
of the a-alkyl group (Table 8, entries 8–10). Furthermore,
the yield of the benzoate ester byproduct increased at the
expense of the desired product in these examples, presuma-
bly because these bulkier acyl groups are transferred more
slowly. In fact, the undesirable benzoyl group transfer
became predominant in the case of substrate 46 containing a
secondary alkyl substituent (Table 8, entry 10).
uct/byproduct ratio (3- to 5-fold) was observed also in the
case of methyl-, n-butyl-, and isopropyl-substituted sub-
strates (36, 45, and 46; Table 9, entries 6–8). In the latter
case, the ee value increased to 77%; however, the yield re-
mained rather low (25%). Application of the standard KR
protocol to this substrate (Table 9, entry 9) raised the yield
to 34%, which, however, was still below practically useful
levels.
In an effort to improve the efficacy of our DKR protocol
in the latter 3 cases, we sought to suppress the benzoyl
group transfer in the mixed anhydride. Replacing benzoic
anhydride with PMBA did improve the product/byproduct
ratio, but the reaction was slow and eventually stopped at
46% yield after 3 days (Table 9, entry 1). Similarly slow re-
Conclusion
HBTM-catalyzed enantioselective esterification described
herein provides an effective approach to the kinetic resolu-
tion of a wide variety of chiral carboxylic acids bearing an
aryl or electron-withdrawing group at the a-posi-
tion. This method is fairly robust, amenable to
scale-up,[35] and does not rely on the use of enzymes
Table 9. Optimization of the DKR protocol for bulky substrates.
Entry[a] Substrate Catalyst Activator Byproduct [%][b] Yield [%][b] ee [%]
or stoichiometric chiral reagents. The di(1-naph-
thyl)methyl ester products can be easily deprotect-
ed via hydrogenolysis or treatment with acid or re-
duced to alcohols with little or no racemization.[36]
Use of half an equivalent of DCC as the condensing
agent is particularly attractive due to its low cost
and ease of separation from reaction products. The
modified protocol employing benzoic anhydride,
however, is more practical in the DKR of a-(arylth-
io/alkylthio)alkanoic acids.
1[c]
2[c]
3
4
5
44
44
44
44
44
36
45
46
46
4
4
4
5
47
47
47
47
47
PMBA
Piv2O
MNBA
Bz2O
Bz2O
Bz2O
Bz2O
Bz2O
DCC
2
0
0
19
7
3
8
16
–
46
45
84
70
81 (78)
93 (90)
84 (83)
25
88
86
50
89
88
90
87
77
78
6
7
8
9[c,d]
34
[a] Reaction conditions: same as in Table 7, entry 5, except as noted. [b] Byproduct
(%) and yield (%) were determined by H NMR using an internal standard. In paren-
1
theses are reported the yields of isolated products. [c] 3 days. [d] Standard KR proto-
col was used. MNBA=2-methyl-6-nitrobenzoic anhydride, PMBA=p-methoxybenzo-
ic anhydride.
Our findings can be summarized as follows. The
dependence of asymmetric induction on the nature
of the alcohol employed clearly points to the
second step of the catalytic cycle being enantiose-
lectivity-determining (Figure 3). Structure–selectivi-
action was observed using pivalic anhydride (Table 9,
entry 2). 2-Methyl-6-nitrobenzoic anhydride (MNBA) react-
ed much more rapidly and without any detectable byproduct
formation. The desired product was formed in good yield,
but with only 50% ee (Table 9, entry 3). Stymied in our at-
tempts to find a better alternative to benzoic anhydride, we
decided to examine the utility of substituted HBTM deriva-
tives, which had previously proved to be more active and
enantioselective in the KR of alcohols than the parent cata-
lyst. HBTM-2 5 (Figure 2) developed in our earlier study,[9e]
Figure 3. Proposed catalytic cycle for the KR of symm-anhydrides.
Figure 2. Additional HBTM analogues tested.
showed slightly enhanced ee values, but virtually the same
yield and product/byproduct ratio (Table 9, entry 4). Its iso-
propyl homologue 47 reported by Smith et al.[14f,16] also gave
similar enantioselectivity, but suppressed the byproduct for-
mation considerably, which led to an improved isolated
yield (Table 9, entry 5). Similar relative increase of the prod-
ty trends observed in all cases studied so far (except azlac-
tones) are qualitatively consistent with a Felkin–Anh-like
predictive model (Figure 1) wherein steric and stereoelec-
tonic factors govern the enantioselectivity. As a general rule,
the enantioselectivity decreases with increasing the steric
bulk of the alkyl group on the substrate (with one apparent
11302
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 11296 – 11304