2
D. Gérard et al. / Tetrahedron: Asymmetry xxx (2017) xxx–xxx
Several enzymatic routes have been explored for obtaining
(RS)-2-bromophenyl acetic acid ester racemate.52,54 Enzyme vari-
ants with increased or totally inverted enantioselectivity, concomi-
tant with a remarkable increase in velocity, were also obtained.55
The performances of the Y. lipolytica wild-type lipase is compared
to wild-type lipases (CRL 1 and 4) from C. rugosa, which has been
described in the literature as being one of the most efficient
enzymes for the resolution of the considered racemates.16,19,34,44
In view of the results, the library of Lip2p variants from Y. lipolytica,
previously built for the resolution of the (RS)-2-bromophenyl
acetic acid ester racemate, was tested on the three molecules and
molecular modelling was used to identify new targets for site-
directed mutagenesis and to understand the role of amino acid
changes on the selective recognition of the (RS)-enantiomers of
each racemate.
enantiopure profens and profenols,6 such as use of alcohol dehy-
drogenases,7–10 arylmalonate decarboxylases, ene reductases11
and nitrilases, either based on the kinetic resolution of racemates
or asymmetrization of prochiral precursors.
In the case of 2-arylpropionic acids, lipases have been shown to
be good candidates to catalyze the kinetic resolution of (RS)-race-
mates, either by hydrolysis of an ester or by esterification of the
acid form. Several lipases from various origins (plant and micro-
bial) have already been reported for the resolution of ibuprofen,
naproxen and ketoprofen esters or the corresponding acids.
Lipases from Rhizomucor miehei,12 Carica papaya13 or evolved
Candida antarctica lipase,14,15 were previously tested for the resolu-
tion of (RS)-ibuprofen racemate by either hydrolysis or esterifica-
tion reactions. However, all three of them led to low
enantioselectivity (E < 200). The lipase from C. rugosa (formerly C.
cylindracea) was shown to be the best enzyme to discriminate
(RS) ibuprofen racemate.16–28 The best result, considering the
enantioselective hydrolysis of ibuprofen racemate using free
enzyme, presents an enantioselectivity of 247.16 Nevertheless, this
high E value was obtained by the addition of N,N-dimethylfor-
mamide, which is highly toxic.
C. rugosa, C. papaya and R. miehei free lipases were also studied
for the hydrolysis of naproxen ester racemate.13,17,29–35 Among
them, lipases from C. rugosa were found to be the most enantiose-
lective enzymes for this reaction (E = 397).32 Lysophospholipase
and carboxylesterase were also studied for the resolution of the
naproxen ester:36,37 the carboxylesterase NP produced by Bacillus
subtilis gave a high enantioselectivity (E = 500), nevertheless, this
high enantioselectivity was obtained by the addition of formalde-
hyde, which is highly toxic.37
2. Results and discussion
2.1. Resolution of (RS)-ibuprofen naproxen and ketoprofen
ethyl ester racemates using wild-type lipases
The reaction scheme of the enantioselective lipase-catalyzed
hydrolysis of the racemic mixture of three different esters is pre-
sented in Figure 1. Wild-type Lip2p from Y. lipolytica was compared
to C. rugosa lipases (CRL 1 and 4), the most efficient enzymes
known to date for the resolution of ibuprofen, naproxen and keto-
profen racemates. CRL lipases were produced in recombinant form
by a strain of Y. lipolytica as pure isoform of CRL1 and CRL4.51
All three lipases were produced using the Y. lipolytica strain
JMY1212,56 in which the lipase encoding gene is introduced in
the genome at the zeta docking platform, leading to good repro-
ducibility of the enzyme expression and enabling a comparison
of enzyme variant activities directly from the supernatant.55 In
addition, this strain is deleted for the main extracellular protease
(XPR2) and the main extracellular lipases (Lip2p, 7 and 8), which
in turn enables for the supernatant to be obtained with high pro-
tein purity.57–59 It was checked that no activity with this strain
was obtained whatever the used racemate (data not shown).
The protein contents of Y. lipolytica supernatant containing
Lip2p, CRL1 or CRL4 were compared by SDS page (data not shown).
Enzyme concentration was estimated as 5 times lower for CRL1
and CRL4 expressed in Y. lipolytica compared with Lip2p. The two
enzymes of C. rugosa were then 10 times concentrated for compar-
ison. Enzymes were then tested during the hydrolysis of ibuprofen,
naproxen and ketoprofen ethyl esters (Table 1).
Our results confirmed that C. rugosa lipase CRL1 was an efficient
enzyme with regards to the enantioselectivity to discriminate
between the enantiomers of ibuprofen, naproxen and ketoprofen
ethyl esters, with a total preference for the (S)-enantiomer. Never-
theless, whatever the racemate and even with a 10 times concen-
tration of the supernatant, the rate of reaction was low. Despite
sharing 81 % of identity with CRL1, lipase CRL4 was only active
on ketoprofen ethyl ester with a total preference for the (R)-enan-
tiomer. This reverse enantioselectivity of CRL1 and CRL4 was
already observed during the resolution of bromophenyl acetic acid
ester racemates.51
Lipases from Thermogota maritima,38 Aspergillus niger, Aspergil-
lus terreus, Fusarium oxysporum,39 Mucor javanicus,40,41 R.
miehei,40,41 Trichosporon laibacchii,42 Pseudomonas cepacia,4 C.
papaya,13 B. subtilis43 and from C. rugosa19 were previously tested
for the resolution of (RS)-ketoprofen racemate. However, all of
them led to poor enantioselectivity (E < 200). In order to improve
upon the enantioselectivity of the lipase from C. rugosa (E = 27),19
various strategies were employed, such as enzyme immobiliza-
tion44–46 or two step acetone treatment18 but only an enantioselec-
tivity of 153 was reached.44
The enantioselectivity of Serratia marcescens lipase was
improved (from 63 to 1084) by the addition of a surfactant Brij
92V.47 Similarly, the enantioselectivity of Pseudomonas sp.
KCTC10122BP lipase48 and Acinetobacter lipase49 were found to
be superior to 200 (absolute and 752, respectively) but with use
of triton X-100. However, addition of a surfactant generally leads
to a complexification of the purification process, and to high pro-
duction costs.
Variants of C. Antarctica lipase CalB,14 and of the recently meta-
genome-isolated esterase Est2550 enabled an enantioselectivity
higher than 200, but very low concentrations of the substrate were
used. It appears that the lipases from C. rugosa are good candidates
for the resolution of the three substrates considered. Despite their
significant enantioselectivity, the catalytic efficiencies of the
lipases from C. rugosa are relatively low, almost one order of mag-
nitude lower than that of Lip2p lipase from the oleaginous yeast Y.
lipolytica51,52 during the resolution of 2-bromophenyl acetic acid
ester.
Wild-type Lip2p lipase from Y. lipolytica showed a clear prefer-
ence for the (S)-enantiomer in the hydrolysis of ibuprofen ethyl
ester racemate (E = 52, Table 1). This result is in agreement with
the E value of 56 obtained in previous studies53 of esterification
of ibuprofen using immobilized Lip2p lipase. The positive influence
of the para substitution of the phenyl group on the E value of Lip2p
lipase had already been observed for the transesterification of 2-
bromo-phenyl acetic ethyl and 2-bromo-p-tolylacetic ethyl ester.60
The presence of a methyl group at the para-position of the phenyl
group led to an improvement in the selectivity of Lip2p lipase
Herein we tried to identify new enzymes able to highly discrim-
inate between the (R)- and (S)-enantiomers of ibuprofen, naproxen
and ketoprofen ethyl esters, and with high catalytic efficiency. The
Y. lipolytica Lip2p lipase was reported to catalyze the resolution of
ibuprofen racemate with
a
low enantioselectivity (E = 56).53
Moreover, variants of this lipase have been shown to have a high
catalytic efficiency and enantioselectivity for the resolution of