Thermally driven asymmetric domino reaction catalyzed by a thermostable esterase and its variants

Article abstract of DOI:10.1016/j.tetlet.2013.01.080

We have developed a thermally driven domino reaction for the synthesis of (S)-a-arylpropionates (profens) using a thermostable esterase from Sulfolobus tokodaii strain 7. Stereoselectivity was improved considerably by engineering of the active site. Stereoselective decarboxylation at the active site of an esterase is a new reaction for the synthesis of optically active carboxylic acids. Crown Copyright.

Full text of DOI:10.1016/j.tetlet.2013.01.080

Tetrahedron Letters 54 (2013) 1921–1923
Contents lists available at SciVerse ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Thermally driven asymmetric domino reaction catalyzed by a thermostable
esterase and its variants
Reina Wada a, Takashi Kumon a, Robert Kourist b, Hiromichi Ohta a, Daisuke Uemura a, Shosuke Yoshida a,
Kenji Miyamoto a,
⇑
a Department of Biosciences and Informatics, Keio University, 3-14-1 Hiyoshi, Kohoku-ku, Yokohama, Kanagawa 223-8522, Japan
b Junior Professorship for Microbial Biotechnology, Ruhr Universität Bochum, Universitätstr. 150, 44780 Bochum, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
We have developed a thermally driven domino reaction for the synthesis of (S)-a-arylpropionates
Received 14 December 2012
Revised 15 January 2013
Accepted 18 January 2013
Available online 29 January 2013
(profens) using a thermostable esterase from Sulfolobus tokodaii strain 7. Stereoselectivity was improved
considerably by engineering of the active site. Stereoselective decarboxylation at the active site of an
esterase is a new reaction for the synthesis of optically active carboxylic acids.
Crown Copyright Ó 2013 Published by Elsevier Ltd. All rights reserved.
Keywords:
Domino reaction
Thermostable esterase
a-Arylpropionates
Thermal spontaneous decarboxylation
Active site
Climate change and the future limitation of energy and re-
sources create a high demand for efficient, sustainable, synthetic
strategies under the green chemistry philosophy.1 One approach
to this is the minimization of resources dedicated to the work-up
and purification of compounds. Domino reactions, where several
bond-forming and bond-breaking steps take place in one pot under
the same reaction conditions, provide a highly efficient strategy.
These reactions improve the synthetic efficiency by decreasing
the number of operations and thus the amount of chemicals and
solvents required. The application of enzymes in multistep reac-
tions permits the use of excellent stereoselectivity and mild reac-
tion conditions usually associated with enzymes.2 Akai et al.
reported the first example of lipase-catalyzed enantioselective
esterification followed by an intramolecular Diels–Alder reaction,
a sequence which gave products with up to 99% enantiomeric ex-
cess (ee).3
enantioselectivity (not published). Moreover, it is known that sev-
eral of these esterases have excellent enantioselectivity toward
bulky malonate diesters (1).6
The hydrolysis reaction can be used to synthesize malonic acid
monoesters (2) in asymmetrization with 100% theoretical yield. In
contrast to mesophilic enzymes, Est0071 can catalyze this reaction
at temperatures above 70 °C. Under these reaction conditions, the
monoester is spontaneously cleaved to the a-arylpropionic acid es-
ter (4). If decarboxylation occurs at the active site it may be stere-
oselective. This prompted us to attempt an enantioselective
domino reaction to produce the pharmacologically important com-
pound naproxen (5b). If the chiral environment of the enzyme
causes the product to be optically pure, the stereospecific outcome
should depend to some degree on the choice of amino acids at the
active site, thus creating the opportunity to optimize enantioselec-
tivity by protein engineering (see Table 2).
We have recently identified a thermostable esterase gene
(ST0071) from the thermophilic archeon Sulfolobus tokodaii by an
in silico screening of the total genome.4 Esterase (Est0071) ex-
pressed from the putative gene is highly similar to esterases with
a so-called GGG(A)X motif in their active sites. This motif has been
associated with the ability of GGG(A)X hydrolases to convert steri-
cally demanding tertiary alcohols.5 Indeed, Est0071 shows activity
toward several arylaliphatic tertiary alcohols, albeit with low
Initially, we examined whether this esterase could catalyze the
hydrolysis of prochiral malonate diesters (1a) at room tempera-
ture. As anticipated, cell-free extract of Est0071 catalyzed the
hydrolysis of prochiral diesters to the corresponding monoesters
((S)-2a) with excellent stereoselectivity (99% ee) and 96% yield.
Interestingly, the enantiopreference of Est0071 is the opposite to
that of pig liver esterase, which forms the (R)-enantiomer with
96% ee.6 Under high temperatures of 50–70 °C, the monoester
(2a) was cleaved to the racemic ester (4a).4 As the esterase shows
excellent enantioselectivity, the decrease in optical purity was
attributed to racemization during decarboxylation. Variations in
⇑
Corresponding author. Tel.: +81 45 566 1786; fax: +81 45 566 1783.
E-mail address: kmiyamoto@bio.keio.ac.jp (K. Miyamoto).
0040-4039/$ - see front matter Crown Copyright Ó 2013 Published by Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.tetlet.2013.01.080

1922
R. Wada et al. / Tetrahedron Letters 54 (2013) 1921–1923
pH between 6 and 9 and temperature between 50 and 70 °C had no
influence on the ee of the product. Nevertheless, the optical purity
of 5a was increased to 48% ee when His-tag purified enzyme was
employed in the reaction. In addition, no decarboxylation of 2a oc-
curred without Est0071 (data not shown). These observations led
us to assume that cleavage of the carboxyl group might take place
at the active site of Est0071. Accordingly, the chiral environment of
the active site induces stereoselective decarboxylation. As both
steps occur in an enantioselective manner, the reaction is a one-en-
zyme three-step domino reaction (see Fig. 1).
2) Asymmetric
CH3
CO2H
CO2H
1) Hydrolysis
Decarboxylation
Path A
Ph
Ph
6a
CH3
CO2Et
CO2H
2a
CH3
CO2H
Ph
Ph
H
(S)-5a
CH3
CO2Et
H
1) Asymmetric
Decarboxylation
2) Non-selective
Hydrolysis
Path B
To obtain a better understanding of the reaction mechanisms,
several of the compounds considered to be intermediates were
subjected to reaction with Est0071. We considered two possible
reaction courses (Fig. 2). In the first pathway shown as Path A, a
half-ester is first hydrolyzed to prochiral phenylmalonic acid
(6a). Diacid (6a) is then decarboxylated to the corresponding
monoacid (5a). In this pathway, decarboxylation of malonic acid
is the selectivity-determining step. To confirm this assumption,
we incubated phenylmalonate with Est0071 at 70 °C. We found
that decarboxylation proceeded with or without esterase at almost
the same rate, and that both products (5a) were racemic. More-
over, it is known that the esterase-catalyzed hydrolysis of the sec-
ond ester of malonic acid esters is slow. Based on these
observations, a reaction via a prochiral dicarboxylic acid interme-
diate is unlikely. In the second proposed pathway (Path B), decar-
boxylation of the half-ester (2a) is followed by hydrolysis. In this
pathway, a monoester (2a) is decarboxylated to an optically active
ester 4a. The esterase then hydrolyzes 4a to the corresponding acid
(5a). In order to test this possibility, the reaction intermediate was
analyzed. In the beginning of the reaction, the diester disappeared
immediately, and the half ester (2a) generated. After that, the car-
boxylic acid (5a) was formed gradually. Next, both (R)-4a and
(S)-4a were subjected to the reaction. We found that the hydrolysis
rates of both enantiomers were nearly identical. If Path B is correct,
the decarboxylation product must be an optically active ester.
However, the ester (4a) could not be isolated because the ester
hydrolysis rate was faster than that of decarboxylation. Therefore,
the decarboxylation is a rate-determining step. Adding the racemic
monoester (2a) as a substrate to the reaction led to the formation
of optically enriched (S)-5a (45% ee). Selectivity was almost the
same as the reaction using the optically active half-ester (48% ee).
These results indicated that the intermediate of the domino
reaction should be planar (3). After hydrolysis forms the (S)-enan-
tiomer of the monoester (2a), decarboxylation is stereoselective or
at least preserves the absolute configuration to some extent.
Accordingly, protonation occurs mainly from one side of the planar
enolate intermediate. This mechanism strongly resembles that of
arylmalonate decarboxylase, which catalyzes the enantioselective
(S)-4a
Figure 2. Two possible reaction courses.
decarboxylation of malonate (6a) and proceeds via a prochiral
intermediate.7–9 In this two-step reaction, the formation of the
enolate is the rate determining step, and the direction of the addi-
tion of the proton decides the absolute configuration of the prod-
uct. It is also important to note that the optically active ester
(4a) is hydrolyzed to the corresponding acid in a non-selective
manner. Interestingly, the esterase-catalyzed reaction forms
(S)-enantiomers, while arylmalonate decarboxylase is (R)-specific.
These findings encouraged us to increase the stereoselectivity of
the domino reaction by protein engineering. Directed evolution
efficiently improves the enantioselectivity of esterases.10 Unfortu-
nately, no high-throughput assay for determining the optical pur-
ity of 5 was available. It is rather difficult to predict the amino
acid substitutions for rational engineering of stereoselectivity be-
cause the exact role of active-site amino acids in the decarboxyl-
ation mechanism remains unknown. Therefore, we chose a semi-
rational approach. A combination of structural considerations for
generating small, high-quality libraries has been shown to be effec-
tive for improving esterases.10 A homology model of Est0071 based
Thermal spontaneous
decarboxylation in the
active site of esterase
Esterase-catalyzed
hydrolysis
CH3
CO2Et
CO2Et
1a, b
CH3
CO2Et
CO2H
(S)-2a, b
Ar
Ar
CH3
Esterase-catalyzed
hydrolysis
CH3
CO2Et
H
CH3
CO2H
H
O
Ar
Ar
(S)-4a, b
Ar
(S)-5a, b
OEt
Planar enolate
intermediate (
3
)
a
; Ar =
b
; Ar =
MeO
Figure 3. (a) Residues of the active site pocket of Est0071 with potential impact on
the conversion of phenylmalonate diethyl ester (1a); (b) catalytic triad (D243, H273
and S150) and oxyanion hole (G78, G78, and G80) of Est0071.
Figure 1. Esterase-catalyzed asymmetric domino reaction.

R. Wada et al. / Tetrahedron Letters 54 (2013) 1921–1923
1923
Table 1
putative prochiral intermediate, based on better delocalization of
the charge on the naphthyl ring system. The effects of the amino
acid substitutions differed remarkably from the findings using 1a
as the substrate. Variant A179 M/L198Y, which produced 5a with
80% ee, catalyzed the formation of 5b with 31% ee. In contrast,
A179 K gave rise to (S)-5b with good optical purity (64% ee).
We developed the first enzymatic domino reaction for the syn-
thesis of optically active (S)-profens. A mechanistic analysis of the
reaction showed that it proceeds via enantioselective hydrolysis of
the malonic acid ester to the monoester. Cleavage of the carboxyl
group at the active site of the enzyme leads to the formation of
the prochiral intermediate, which is then protonated to give the
enantioenriched monoacid. Engineering of the active site gave rise
to a variant with good stereoselectivity. As the esterase-catalyzed
domino reaction does not need external cofactors and allows a the-
oretical yield of 100%, it offers clear advantages over alternative
enzymatic routes.11 The esterase-catalyzed domino reaction is thus
an important step on the road toward the sustainable synthesis of
optically active profens.
Enantioselectivity of Est0071 and its variantsa
Variant
Conversion (%)
eeb (%)
Configuration
A179M/L198Y
A179M/L198E
I203R
A179K
I203E
37
43
5
56
15
35
85
80
77
74
70
67
66
48
S
S
S
S
S
S
S
A179M/L198S
Wild-type
a
Reaction conditions: pH 7, 70 °C, 48 h.
The ee values were determined by chiral-phase HPLC.
b
Table 2
Asymmetric synthesis of naproxen (5b)a
Variant
Conversion (%)
ee (%)
Configuration
A179K
I203R
I203E
A179M/198Y
A179M/198E
Wild-type
>99
>99
>99
>99
>99
>99
64
63
52
31
18
21
S
S
S
S
S
S
References and notes
1. Poliakoff, M.; Licence, P. Nature 2007, 450, 810–812.
a
Reaction conditions: pH 7, 70 °C, 48 h.
2. Burda, E.; Hummel, W.; Gröger, H. Angew. Chem., Int. Ed. 2008, 47, 9551–9554.
3. Akai, S.; Tanimoto, K.; Kita, Y. Angew. Chem., Int. Ed. 2004, 43, 1407–1410.
4. (a) Suzuki, Y.; Miyamoto, K.; Ohta, H. FEMS Microbiol. Lett. 2004, 236, 97–102;
Esterase (Est0071) is commercially available from Enzymicals AG (Germany),
http://www.enzymicals.com/.
5. Kourist, R.; Bornscheuer, U. T. Appl. Microbiol. Biotechnol. 2011, 191, 505–517.
6. Domínguez de María, P.; García-Burgos, C. A.; Bargeman, G.; van Gemert, R. W.
Synthesis 2007, 10, 1439–1452.
7. Miyamoto, K.; Ohta, H. J. Am. Chem. Soc. 1990, 112, 4077–4078.
8. Miyamoto, K.; Tsuchiya, S.; Ohta, H. J. Am. Chem. Soc. 1992, 114, 6256–6257.
9. Ijima, Y.; Matoishi, K.; Terao, Y.; Doi, N.; Yanagawa, H.; Ohta, H. Chem. Commun.
2005, 877–879.
10. Kourist, R.; Jochens, H.; Bartsch, S.; Kuipers, R.; Padhi, S. K.; Gall, M.; Bottcher,
D.; Joosten, H. J.; Bornscheuer, U. T. ChemBioChem 2010, 11, 1635–1643.
11. Kourist, R.; Domínguez de Maria, P.; Miyamoto, K. Green Chem. 2011, 13,
2607–2618.
on the X-ray structure of an esterase from Archaeoglobus fulgidus
(PDB 1JJI) was the basis for the identification of three amino acid
residues around the substrate binding site (Fig. 3).
Residues 179, 198, and 203 are situated in close proximity to
the catalytic triad. These positions were simultaneously altered
by mutagenic polymerase chain reaction with primers bearing
the NNK codon. A library of 200 clones was constructed. The num-
ber of clones was not sufficient for full coverage, but helped to en-
sure the potential impact of the residues on the outcome of
decarboxylation. All clones were expressed in Escherichia coli and
purified by His-tag affinity chromatography. After purification, all
variants were active in the hydrolysis of p-nitrophenyl butyrate
at room temperature. In the domino reaction at 70 °C, 69 variants
12. Typical procedure: Two hundred microliters of 50 mM substrate (1a or 1b)
solution in CH3CN was added to 4 ml of 400 mM HEPES–NaOH buffer (pH 7).
To the resulting solution, 1 ml of Est007 solution (25 lg/ml) was added and the
mixture was stirred for 48 h at 70 °C. The reaction was stopped by addition of
2 ml of 2 M HCl. The products were extracted with 5 ml diisopropyl ether. The
organic layer (4 ml) was concentrated in vacuo, and to this was added 400
showed activity. Using these positive mutants, we next confirmed
the stereoselectivity of the domino reaction by HPLC
12
.
ll
diisopropyl ether, 100 l methanol, and an excess amount of Me3SiMeN2 in
l
All variants showed (S)-selectivity. Interestingly, 6 variants
showed increased stereoselectivity compared to the wild-type en-
zyme (Table 1). The best variants (A179 M/L198Y) showed 80% ee,
at conversions around 37%. The variants with improved selectivity
had hydrophobic residues such as leucine replaced with polar argi-
nine, serine, glutamate, or lysine.
The enantioselective domino reaction was applied to the syn-
thesis of (S)-naproxen (5b). The conversion of 1b was much faster
than that of 1a. This can be explained by the higher stability of the
hexane. The solution was vortexed and allowed to stand at room temperature
for 30 min. Enatiomeric excess of methyl ester of 5a and 5b was determined by
HPLC analysis using CHRALCEL OJ (Daicel Corporation, Japan). To determine
the conversion rate, 500
ll of the reaction mixture was transferred to the tube,
and the reaction was stopped by addition of 200
l
l of CH3CN containing 2%
CF3CO2H. The products were extracted with EtOAc which contains diethyl
phthalate as an internal standard. The conversion rate of the products was
determined by reverse phase HPLC using COSMOSIL 5C18-PAQ (Nakarai
Tesque, Inc., Japan).