Isoenzymes of pig-liver esterase reveal striking differences in enantioselectivities

Article abstract of DOI:10.1002/anie.200703256

(Graph Presented) An esterase toolbox: A set of isoenzymes of pig liver esterases (PLE) is identified, cloned, and overexpressed in E. coli. They show striking differences in enantioselectivity and enantiopreference in the kinetic resolution of acetates o

Full text of DOI:10.1002/anie.200703256

Communications
DOI: 10.1002/anie.200703256
Enzyme Catalysis
Isoenzymes of Pig-Liver Esterase Reveal Striking Differences in
Enantioselectivities**
Anke Hummel, Elke Brüsehaber, Dominique Böttcher, Harald Trauthwein, Kai Doderer, and
Uwe T. Bornscheuer*
Esterases and lipases are frequently used biocatalysts because
they accept a broad range of substrates, are usually stable in
organic solvents, and often show high stereoselectivities even
towards non-natural substrates.[1] While a large number of
lipases is commercially available, there are only few well
explored carboxylesterases, among which pig-liver esterase
(PLE) plays the most important role in industrial processes
owing to its high versatility.[2] One major drawback in the
application of PLE is its natural heterogeneity as it consists of
several isoenzymes.[3] These differ in isoelectric point, molec-
ular weight, sensitivity towards inhibitors and—most impor-
tantly—substrate specificity.[3b]
Several years ago, we reported the cloning and recombi-
nant expression of the g-isoenzyme of PLE (g-PLE) in Pichia
pastoris[4] and more recently in E. coli[5] thus overcoming the
undesirable presence of several PLE isoenzymes and of
interfering other hydrolases in the commercial preparations.
Furthermore, we could demonstrate that the recombinant g-
PLE shows considerable differences in enantioselectivity
towards esters of secondary alcohols in comparison with the
naturally occurring mixture of isoenzymes.[6] This encouraged
us to identify the then unknown sequences encoding the other
isoenzymes of PLE. Initially, we used tandem mass spectrom-
etry[7] of PLE samples separated by 2D gel electrophoresis.
Indeed, this led to the discovery of certain amino acid
positions, such as V236P/A237G, which impart enhanced
enantioselectivity. However, the elucidation of the complete
protein sequences appears impossible using this approach.
To access the genes encoding for unknown isoenzymes of
PLE, first, the cDNA of pig-liver RNA was obtained by
reverse transcriptase-polymerase chain reaction (RT-PCR).
The cDNA served as the template for the amplification of
PLE homologous genes using primers derived from the
known g-PLE sequence (GenBank accession code X63323).
To enable functional expression in E. coli, the N-terminal
signal sequence (18 amino acids) and the C-terminal ER-
retention signal (four amino acids, HAEL; ER = endoplasmic
reticulum) were omitted. Amplification by PCR resulted in a
single DNA band of approximately 1.6 kbp (bp = base pairs)
in the agarose gel, matching the size of the g-PLE gene. The
fragments were cloned first into the pET101/D-TOPO vector
and later, for functional expression, into pET15b, and
sequenced. This resulted in the identification of four novel
sequences (named PLE2 to PLE5), bearing 3–21 amino acid
exchanges[8] compared to g-PLE (now renamed PLE1).
Figure 1 schematically shows that the amino acid exchanges
are not randomly distributed along the protein, but can be
found in distinct regions.
Figure 1. Differences between the isoenzymes are not randomly dis-
tributed, but occur in conserved areas. Black: homologous regions,
white: variations in PLE1 (g-PLE), hashed: variations in PLE5, dotted:
variations, which occur neither in PLE1 nor PLE5; AA=amino acid.
After functional expression in E. coli, we observed that
the novel isoenzymes show distinct differences in their
characteristics, amongst others in the specific activity towards
achiral esters: All of them preferentially cleave tributyrin, but
PLE4 and PLE5 also show a high activity for methyl butyrate
and ethyl caprylate.[8] Similarly, the sensitivity of the iso-
enzymes towards certain inhibitors varied considerably:
PLE3–5 are less sensitive than the others towards sodium
fluoride and physostigmin, but are more strongly inhibited by
phenyl methyl sulfonylfluoride.[8] The ratio in the specific
activities against methyl butyrate and tributyrin as well as the
sensitivity against the chosen inhibitors has been reported to
be characteristic for distinguishing between the main iso-
enzyme fractions in the natural PLE mixture, a-PLE, and g-
PLE,[3b] so that it can be suggested that PLE4 or PLE5
represent the so-called a-PLE.
[*] Dipl.-Biochem. A. Hummel, Dipl.-Biochem. E. Brüsehaber,
Dr. D. Böttcher, Prof. U. T. Bornscheuer
Institute of Biochemistry
Dept. of Biotechnology & Enzyme Catalysis
Greifswald University
Felix-Hausdorff-Strasse 4, 17487 Greifswald (Germany)
Fax: (+49)3834-86-80066
E-mail: uwe.bornscheuer@uni-greifswald.de
Homepage: http://www.chemie.uni-greifswald.de/~biotech
Dr. H. Trauthwein, Dr. K. Doderer
Service Center Biocatalysis
Evonik Degussa GmbH (Germany)
[**] We are grateful to the Deutsche Bundesstiftung Umwelt (DBU,
Osnabrück (Germany), Grants AZ13071 and AZ13141) for financial
support.
Most importantly for organic synthesis the enantioselec-
tivity of the PLE isoenzymes differed substantially as
exemplified for the kinetic resolution of esters of secondary
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
8492
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 8492 –8494

Angewandte
Chemie
Table 1: Enantioselectivity of different recombinant PLE isoenzymes and
a commercial PLE preparation in the kinetic resolution of 2.
alcohols 1–4 (Scheme 1) and the desymmetrization of the
meso-diacetate 5 (Scheme 2).
In previous studies,[6] we reported that the recombinant
PLE1 (g-PLE) isoenzyme showed increased enantioselectiv-
ity (E) towards 3 (E > 100) and 4 (E = 17) in contrast to the
PLE[a]
t[h] eeS [%][b] eeP [%][b] Conv. [%] E[c] Preference
PLE1
PLE2
PLE3
PLE4
PLE5
2
2
74
67
77
81
24
94
95
56
49
45
43
42
45
54
17
19
2
66
94
7
R
R
S
R
R
R
1.5 18
3
2
68
79
Fluka-PLE[d,e] 1.5 65
[a] In all reactions 0.5 U of esterase (based on pNPA assay) were used.
[b] eeS =Enantiomeric excess of the non-converted substrate, eeP =en-
antiomeric excess of the product as determined by GC analysis on a
chiral stationary phase. [c] Calculated according to Chen et al.[9] [d] Com-
mercially available PLE preparation from Fluka. [e] Data for Fluka-PLE
taken from literature.[6a]
Scheme 1. Acetates 1–4 of secondary alcohols used in the kinetic
resolution with the PLE isoenzymes.
isoenzymes preferred the (R)-enantiomers (Figure 2,
Supporting Information). Although enantioselectivity is well
pronounced for all the other isoenzymes towards all acetates
studied, PLE3 shows nearly no preference.
Scheme 2. PLE-catalyzed desymmetrization of 5 yielding 6a or 6b.
Analogously, we found a change in enantiopreference in
the desymmetrization of
5
(Figure 3, Supporting
naturally occurring PLE mixture (E < 5). The comparison of
the enantioselectivities and the enantiopreferences of the
novel PLE isoenzymes (PLE2–5) with PLE1 and the
commercial enzyme from Fluka in the kinetic resolutions of
1–4 clearly shows striking differences in their properties
(Figure 2).
Information). The resulting cyclopentene monoesters are
Figure 2. Enantioselectivity and enantiopreference of the recombinant
PLE isoenzymes and commercial PLE isoenzyme mixture in the kinetic
resolution of acetates 1–4.
Figure 3. The enantiopreference of the recombinant PLE isoenzymes
and of a commercial PLE isoenzyme mixture in the desymmetrization
of 5.
PLE1 (g-PLE) and PLE2 differ by only three amino acids
and it is not surprising, that their selectivities are highly
similar. In contrast, drastic changes are clearly seen for
PLE3–5, which differ by 20 or 21 amino acid exchanges[8]
from g-PLE: For 1 and 2 notably higher enantioselectivities
were found using enzymes PLE4 and PLE5 (Table 1 and
Supporting Information) with the E-value towards 2 increas-
ing from 17 (PLE1, g-PLE) to 66 (PLE4) and 94 (PLE5). In
the kinetic resolution of acetate 1, PLE5 shows a more than
tenfold increase in enantioselectivity compared to PLE1.
For the other two acetates (3 and 4), even a switch in
enantiopreference takes place: while PLE1 and PLE2
preferentially converted the (S)-enantiomer, the other three
important chiral building blocks in the synthesis of prosta-
glandins and their derivatives.[10] The commercial PLE
(mixture) shows pro-(R) selectivity yielding 80% ee.[11]
Figure 3 shows, that the same selectivity was found using
Fluka PLE, but only 60% ee was achieved for 6a. Most
importantly, whereas PLE1–3 show the same preference and
gave up to 80% ee, isoenzymes PLE4 and 5 favored the pro-
(S) acetoxy group yielding monoacetate 6b with 42% ee
(PLE4) and 17% ee (PLE5).
PLE4 and PLE5 do not only show altered enantioselec-
tivities, but also exhibit higher kinetic constants in the
hydrolysis of p-nitrophenyl acetate (Table 2). Owing to a
Angew. Chem. Int. Ed. 2007, 46, 8492 –8494
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 8493

Communications
Table 2: Kinetic data of the different isoenzymes towards pNPA.[a]
Keywords: desymmetrization · enantioselectivity ·
enzyme catalysis · hydrolases · pig-liver esterase
.
Isoenzyme[b]
Vmax [UmgÀ1
]
Km [mm]
kcat/Km [mÀ1 sÀ1 [c]
]
PLE1
PLE3
PLE4
PLE5
149
110
133
217
1.57
0.96
0.81
0.76
3.0ꢀ105
3.6ꢀ105
5.2ꢀ105
9.1ꢀ105
[1] a) K. Faber, Biotransformations in Organic Chemistry, 4th ed.,
Springer, Berlin, 2004; b) U. T. Bornscheuer, R. J. Kazlauskas,
Hydrolases in Organic Synthesis—Regio- and Stereoselective
Biotransformations, 2nd ed., Wiley-VCH, Weinheim, 2006.
[2] a) C. Tamm, Pure Appl. Chem. 1992, 64, 1187; b) L.-M. Zhu,
M. C. Tedford, Tetrahedron 1990, 46, 6587.
[3] a) E. Heymann, W. Junge, Eur. J. Biochem. 1979, 95, 509; b) W.
Junge, E. Heymann, Eur. J. Biochem. 1979, 95, 519.
[4] S. Lange, A. Musidlowska, C. Schmidt-Dannert, J. Schmitt, U. T.
Bornscheuer, ChemBioChem 2001, 2, 576.
[5] D. Böttcher, E. Brüsehaber, K. Doderer, U. T. Bornscheuer,
Appl. Microbiol. Biotechnol. 2007, 73, 1282.
[6] a) A. Musidlowska, S. Lange, U. T. Bornscheuer, Angew. Chem.
2001, 113, 2934; Angew. Chem. Int. Ed. 2001, 40, 2851; b) A.
Musidlowska-Persson, U. T. Bornscheuer, J. Mol. Catal. B 2002,
19–20, 129.
[a] pNPA, p-nitrophenyl acetate; activity measured at pH 7.5 and room
temperature. [b] PLE2 was not measured as its properties are very close
to PLE1. [c] To calculate kcat, the PLE trimer was regarded as one
catalytically active unit.
higher vmax and a lower Km value, the catalytic efficiency (kcat
/
Km) of PLE 5 is about threefold higher than that of PLE1.
These results emphasize that the differences in protein
sequences between the naturally occurring isoenzymes have a
strong impact on the enantioselectivity and enantioprefer-
ence of pig-liver esterase. The availability of individual PLE
isoenzymes now provides a versatile source for the applica-
tion of this very important esterase. Thus, well-defined
biocatalysts with distinct properties can be selected for a
given synthetic problem.
[7] E. Brüsehaber, D. Böttcher, A. Musidlowska-Persson, D.
Albrecht, M. Hecker, K. Doderer, U. T. Bornscheuer, Appl.
Microbiol. Biotechnol. 2007, 76, 853.
[8] See the Supporting Information.
[9] C. S. Chen, Y. Fujimoto, G. Girdaukas, C. J. Sih, J. Am. Chem.
Soc. 1982, 104, 7294.
[10] M. Harre, P. Raddatz, R. Walenta, E. Winterfeldt, Angew. Chem.
1982, 94, 496; Angew. Chem. Int. Ed. Engl. 1982, 21, 480.
[11] Y. Wang, C. S. Chen, G. Girdaukas, C. J. Sih, J. Am. Chem. Soc.
1984, 106, 3695.
Received: July 20, 2007
Published online: September 27, 2007
8494 www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 8492 –8494