amino ester 6 in high yield (Scheme 2), and then conve-
niently protected as the N-Cbz methyl amino ester 7 with
benzyloxycarbonyl chloride (CbzCl) in the presence of (N,
N)-diisopropylethylamine (DIPEA). Before carrying out
the hydrolase-catalyzed hydrolysis reactions, N-Cbz pro-
tected amino acid 8 was prepared by saponification of 7 in
quantitative yield.
Enzymatic hydrolysis reactions of the methyl ester 7 in
either aqueous medium (phosphate buffer) or organic
solvent (THF) using a 20-fold excess of water were un-
successful, recovering the starting material in all cases
when Candida antarctica lipase type A (CAL-A), Candida
antarctica lipase type B (CAL-B), or Pseudomonas cepacia
lipase (also known as Burkholderia cepacia lipase, PSL-C I)
were used at 28 °C and 200 rpm.
a metal or an oxidation agent, which is typical in DKR
processes, and gives an added value to this methodology.15
Coming back to the description of the enzymatic alkox-
ycarbonylation, it is worth noting that the reaction with
3-methoxyphenyl allyl carbonate (9a) released 3-methox-
yphenol to the reaction medium, making monitoring the
enzymatic processes difficult. Therefore, we decided to
attempt the stereoselective alkoxycarbonylation of race-
mic 6 using the commercially available diallyl carbonate
(9b) that produced volatile allylic alcohol as byproduct.
Initial assays using THF as solvent and variable tempera-
tures and amounts of carbonate led us to fix 10 equiv of 9b
and 50 °C as optimal values to start the process optimiza-
tion. In these conditions, a 72% conversion was achieved,
recovering the carbamate (R)-10 with very high selectivity
and substrate with low enantiomeric excess (entry 1, Table 1).
Esterification processes with the racemic amino acid 8
were attempted in a similar manner to that described
by Pietruszka and co-workers in structurally related
compounds;11 however, neither of the lipases tested displayed
activity using 5 equiv of MeOH or BuOH in dry THF, or
alternatively a biphasic toluene/phosphate buffer system.
Once we explored the unsuccessful modification of the
carboxylic group, we decided to study the lipase-catalyzed
alkoxycarbonylation reaction of racemic methyl 1,3-dihy-
dro-2H-isoindole-1-carboxylate (6).12 Initially 3-methox-
yphenyl allyl carbonate (9a) was selected as alkoxy-
carbonylating agent because of the good results shown in
the lipase mediated kinetic resolution of indoline13 or
isoquinoline derivatives.14 Unfortunately CAL-A, CAL-
B, or Candida cycindracea lipase (CCL) did not display any
activity with 2.5 equiv of carbonate (Scheme 4). On the
other hand, PSL-C I allowed the formation of the desired
enantioenriched allyl carbamate 10 in 28% conversion
after 19 h, although surprisingly the starting amino ester
was recovered in racemic form, which means that racemi-
zation is occurring during the process. Therefore we
propose that the racemization equilibrium consists in a
deprotonationꢀprotonation process through an achiral
intermediate, a stabilized enolate. Amino ester racemiza-
tion occurs without the addition of an acidꢀbase catalyst,
Table 1. DKR of Racemic 6 (0.1 M) Using Diallyl Carbonate
and PSL-C I at 50 °C during 48 h at 250 rpm
entry
solvent
carbonate 9b c (%)a eeP (%)b eeS (%)b
1
2
3
4
5
6
THF
10 equiv
72
29
0
96
15
8
1,4-dioxane 10 equiv
>99
MeCN
toluene
TBME
9b
10 equiv
10 equiv
10 equiv
0.1 M
83
87
85
95
94
91
65
78
ndc
a Conversion values determined by 1H NMR of the reaction crude.
b Enantiomeric excesses determined by HPLC. c Not determined.
In order to increase the conversion values, different
solvents were tested, finding a remarkable influence de-
pending on the reaction medium, yielding enantiopure (R)-
10 and the remaining substrate in very low enantiomeric
excess with 1,4-dioxane (entry 2, Table 1), while a more
polar solvent such as acetonitrile provokes the inactivation
of the enzyme (entry 3, Table 1). Solvents with a lower
polaritiy such as tert-butyl methyl ether (TBME) or to-
luene led to higher conversions (83ꢀ87%), yielding (R)-10
with high stereoselectivities (94ꢀ95% ee, entries 4 and 5,
Table 1), while racemizationfailedatlongerreaction times.
The use of diallyl carbonate as both solvent and alkoxycarbo-
nylating agent did not improve the results previously obtained.
Additionally, other PSL preparations were also employed, but
while no conversion was found for PSL-SD (supported on
diatomite mainly active only in aqueous systems), lower
conversions and similar selectivities were found for PSL-IM
(crude preparation stabilized with cyclodextrins).
Scheme 4. Lipase-Catalyzed Alkoxycarbonylation of Racemic 6
Using Allyl Carbonates 9aꢀb in Dynamic Kinetic Resolutions
Biocatalytic reactions were finally examined in terms of
enzyme loading and temperature by adding double the
amount of enzyme in weight or heating at 60 °C (Table 2);
however, in all cases doubling the amount of enzyme led to
similar results as the ones previously obtained (entries 3
(11) Pietruszka, J.; Simon, R. C.; Kruska, F.; Braun, M. Eur. J. Org.
Chem. 2009, 6217.
(12) Amino ester 6 was highly unstable, leading to the corresponding
aromatic heterocycle in short reaction times. Therefore, it was used
immediately after its preparation or stored as a dichloromethane stock
solution previously to be enzymatically reacted with allyl carbonates.
ꢀ
€
(15) (a) Pamies, O.; Backvall, J.-E. Chem. Rev. 2003, 103, 3247. (b)
€
Martın-Matute, B.; Backvall, J.-E. Curr. Opin. Chem. Biol. 2007, 11, 226.
ꢀ
ꢀ
(13) Gotor-Fernandez, V.; Fernandez-Torres, P.; Gotor, V. Tetra-
(c) Ahn, Y.; Ko, S.-B.; Kim, M.-J.; Park, J. Coord. Chem. Rev. 2008, 252,
647. (d) Lee, J. H.; Han, K.; Kim, M.-J.; Park, J. Eur. J. Org. Chem. 2010,
999. (e) Pellisier, H. Tetrahedron 2011, 67, 3769.
hedron: Asymmetry 2006, 17, 2558.
(14) Breen, G. F. Tetrahedron: Asymmetry 2004, 15, 1427–1430.
1698
Org. Lett., Vol. 14, No. 7, 2012