Development of a Novel Biocatalyst for the Resolution of rac-Pantolactone

Article abstract of DOI:10.1002/1615-4169(200212)344:10<1103::AID-ADSC1103>3.0.CO;2-R

A novel L-pantolactone hydrolase, Lph, from Agrobacterium tumefaciens Lu681 was characterized, which stereospecifically hydrolyses L-pantolactone to L-pantoic acid yielding D-pantolactone with > 95% enantiomeric excess. The enzyme was found to be a 30 kDa-Zn²⁺-hydrolase with a K_m for L-pantolactone of 7 mM and a V_max of 30 U/mg. The corresponding lph gene was identified as an 807 bp ORF and cloned into E. coli. It was overexpressed under control of P_tac and P_rha yielding enzyme activities of up to 600 U/g dry weight. Resolution of D,L-pantolactone in repeated batches with isolated Lph and enzyme recovery by membrane filtration gave D-pantolactone with 50% yield and 90-95% ee over 6 days. Covalent immobilization to EupergitC led to a stable biocatalyst easy to handle in a repeated batch production of D-pantolactone. Further improvements in the activity of Lph were achieved by directed evolution of the enzyme. Activities of mutants F62S, K197D and F100L were increased 2.3, 1.7, and 1.5 fold, respectively.

Full text of DOI:10.1002/1615-4169(200212)344:10<1103::AID-ADSC1103>3.0.CO;2-R

FULL PAPERS
Development of a Novel Biocatalyst for the Resolution
of rac-Pantolactone
Maria Kesseler,* Thomas Friedrich, Hans Wolfgang Hˆffken, Bernhard Hauer
BASF AG, Research Fine Chemicals, GVF/EÀA30, 67056 Ludwigshafen, Germany
Fax: ( ‡ 49)-621-60-52808, e-mail: maria.kesseler@basf-ag.de
Received: May 9, 2002; Accepted: August 28, 2002
Abstract: A novel l-pantolactone hydrolase, Lph, Lph and enzyme recovery by membrane filtration
from Agrobacterium tumefaciens Lu681 was charac- gave d-pantolactone with 50% yield and 90 ± 95% ee
terized, which stereospecifically hydrolyses l-panto- over 6 days. Covalent immobilization to EupergitC
lactone to l-pantoic acid yielding d-pantolactone led to a stable biocatalyst easy to handle in a repeated
with > 95% enantiomeric excess. The enzyme was batch production of d-pantolactone. Further im-
2‡
found to be a 30 kDa-Zn -hydrolase with a K for l- provements in the activity of Lph were achieved by
m
pantolactone of 7 mM and a Vmax of 30 U/mg. The directed evolution of the enzyme. Activities of mu-
corresponding lph gene was identified as an 807 bp tants F62S, K197D and F100L were increased 2.3, 1.7,
ORF and cloned into E. coli. It was overexpressed and 1.5 fold, respectively.
under control of Ptac and Prha yielding enzyme
activities of up to 600 U/g dry weight. Resolution of Keywords: enzyme catalysis; kinetic resolution; lacto-
2‡
d,l-pantolactone in repeated batches with isolated nohydrolase; d-pantothenic acid; Zn hydrolase
Introduction
ing spontaneous chemical hydrolysis. This paper de-
scribes the characterization and cloning of l-pantolac-
d-Pantolactone is an important intermediate in the tone hydrolase (Lph) from Agrobacterium tumefaciens
production of d-pantothenic acid, also called vitamin Lu681 and its use for kinetic resolution of d-pantolac-
B5. It is used mainly in feed for chicken and pigs and also tone. To our knowledge this is also the first report on
as vitamin supply in human nutrition. While d-panto- directed evolution of a lactonase and it reveals the
thenic acid has been traditionally produced by chemical characteristics of some Lph mutants with increased l-
processes which involve an efficient but troublesome pantolactone hydrolyzing activity.
diastereomer crystallization step, new biotechnical
preparation methods have shown up in the literature
[
1]
during the last years. Apart from microbial synthesis of Results and Discussion
d-pantothenic acid, especially the kinetic resolution of
pantolactone by enantioselective enzymatic hydrolysis
Screening, Purification and Characterization of a
Novel Enantioselective Pantolactone Hydrolase
is an attractive approach since it not only fits into the old
chemistry of d-pantothenic acid synthesis but also gives
access to a second, related product, d-pantothenol.
Today, the aldonolactonase of the fungus Fusarium After screening a variety of 965 bacterial and fungal
oxysporum AKU3702 which hydrolyses d-pantolactone strains, 17 strains were found to produce ꢀ 1 mM l- or
[
2]
to d-pantoic acid is used by Fuji/Daiichi Chemicals d-pantoic acid (ee > 70%) from d,l-pantolactone
for the production of d-pantothenic acid on a multi- (50 mM) in 3 h, which did not belong to the known d-
[
3]
thousand-ton scale. Recently, new pantolactone hy- pantolactone hydrolyzing fungi. The l-selective panto-
drolyzing enzymes with opposite enantioselectivity lactone hydrolase from Agrobacterium tumefaciens
[
4]
were purified from bacteria (for a review of the Lu681 (formerly identified as Burkholderia caryophyl-
[
5]
diversity of lactonases see ).
li) was selected for further characterization (7.6 mM l-
Considering a kinetic resolution process, use of both a pantoic acid, ee 97%). The enzyme was purified as
d- or an l-specific pantolactone hydrolase is possible in described in the experimental part and two peptides
principle (Figure 1). Hydrolyzing the unwanted l-form, were identified after SDS-PAGE, trypsin digestion and
it may take longer to reach a sufficient enantiomeric Edman degradation, (1) YGIEGLNNLEAL and (2)
excess for the remaining d-pantolactone, however, the AKEDANSTIEAED. Interestingly, the cysteine syn-
process is much more robust, e.g., towards the compet- thetases both from Lu681 and from recombinant E. coli
Adv. Synth. Catal. 2002, 344, No. 10
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 1615-4150/02/34410-1103 ± 1110 $ 17.50+.50/0
1103

FULL PAPERS
Maria Kesseler et al.
Figure 1. Production of d-pantothenic acid via resolution of d,l-pantolactone by either a d- or an l-specific lactonohydrolase.
were copurified by this method. Table 1 shows the reaction increases at high pH and temperature, optical
substrate specificity of Lph. d-Pantolactone was not resolution of d,l-pantolactone was performed at pH
[
4a]
converted. Since there was no ester or lactone revealing 7.0 ± 7.5 and 30 8C. The (apparent) K values were
m
a higher activity and l-pantolactone is not an obvious 10 mM and 7 mM for the purified enzymes from Lu681
substrate in natural metabolism, it is not clear what the and E. coli, respectively, and 33 mM for the whole cells
enzyme×s physiological substrate might be. One idea is of recombinant E. coli. Vmax was about 30 U per mg
that it could be involved in the quorum sensing of the protein.
bacteria which typically proceeds via lactones. Although
the enzyme showed neither inhibition by chelating
agents (and a series of other enzyme inhibitors) nor Cloningand Expression of the l-Pantolactone
significant activation by metal ions (Table 2) it turned Hydrolase from Lu681
2‡
out to be a Zn -containing hydrolase by X-ray
fluorescence spectroscopy of crystals. Renaturation of For cloning of the Lph gene, a chromosomal library of
urea/EDTA (8 M/0.1 M)-denatured Lph in the presence Lu681 was constructed in E. coli XL1/pBluescriptSKII
2‡
2‡
2‡
2‡
of 1 mM Zn , Mn , Co or Ni restored 9.8, 6.0, 4.1 and screened for Lph activity by colony lift assay. One
and 2.1% of the original activity (0.2% without addition isolated positive clone carried a 7.5 kB insert with an
of metal ions). The purified enzyme showed broad ORF of 807 bp/269 aa (sequence in Figure 2) which
temperature and pH optima of 60 ± 70 8C and 6.8 ± 7.8, contained the peptides 1 and 2 previously identified by
respectively. Since the competing chemical hydrolysis Edman degradation. The homologies to public protein
1104
Adv. Synth. Catal. 2002, 344, 1103 ± 1110

Novel Biocatalyst for the Resolution of rac-Pantolactone
FULL PAPERS
Table 1. Substrate specificity of Lph from Agrobacterium
tumefaciens Lu681.
Table 2. Effects of various additives on the activity of Lph
from Agrobacterium tumefaciens Lu681.
Substrate
Concentra-
tion (mM)
Activity[a]
[%]
Substance
Concentra- Relative
tion [mM] Activity [%]
l-Pantolactone
d-Pantolactone
g-Butyrolactone
g-Valerolactone
d-Valerolactone
e-Caprolactone
150
150
150
150
50
150
15
75
150
150
150
150
150
150
0.25
100
>0.1
13
4
No addition
EDTA
Citrate pH 6.4
o-Phenanthroline
HgCl2
p-Chloromercuribiphenyl-
sulfonic acid
±
1
30
1
1
1
100[a]
97
97
95
133
108
2
32
n.d.[b]
11
12
40
10
57
9
(
‡ /À )-d-Decanolactone
d-Nonalactone
DTT
1
1
1
1
1%
1
1
1
1
1
1
1
1
1
99
113
115
117
95
102
99
100
99
101
100
97
104
96
113
102
104
104
93
(
(
‡)-Ethyl d-lactate
À)-Ethyl l-lactate
Phenylmethanesulfonyl fluoride
Diisopropyl fluorophosphate
Pepstatine
d-Galactono-g-lactone
l-Galactono-g-lactone
l-(‡)-Gulono-g-lactone
d-(À)-Gulono-g-lactone
H O
2
2
KCN
KCl
9
1
1
,2-O-Dilauryl-rac-glyceryl-
NH Cl
4
3-glutaryl resorufin ester
MgCl2
CaCl2
MnCl2
CoCl2
FeCl2
NiCl2
ZnCl2
SDS
5-Hydroxy-2-coumaranone
2.5
2.5
10
41
2
7
a-Naphthyl acetate
Isatine
[
[
a]
b]
Activity for l-pantolactone was set to 100%.
Although a decrease of ( ‡ / ±)-d-decanolactone indicated
an activity equivalent to 75%, no ( ‡ / ±)-d-decanoic acid
could be detected.
1
1%
0.1%
0.1%
CHAPS
Tritone
2
-Propanol
10%
Acetonitrile
Methanol
10%
10%
119
100
sequences were low except for hypothetical proteins
from Agrobacterium tumefaciens, Sinorhizobium meli-
loti and Nostoc sp. (see Table 3). However, although
isatine is a bad substrate for Lph, isatine hydrolase is
the fully characterized protein most closely related to
Lph. Peptide sequences of the lactonohydrolase from
Agrobacterium tumefaciens AKU316 identified so far
differ only in 10 from 66 residues characterized by
Edman degradation. Despite the striking similarity, this
[a]
Activity was set to 100%.
[
6]
enzyme was described to have a higher pH optimum of
4b]
8
.0 ± 8.5[ than Lph from Agrobacterium tumefaciens
Lu681. Since the spontaneous hydrolysis rate of
pantolactone increases with higher pH, better ee values
are obtained in bioconversions around pH 7.0.
The lph gene was cloned into pKK223-3- and
pDHE19.2[ -expression vectors for IPTG- and rham-
nose-dependent overexpression in E. coli, respectively
4a,7]
(
resulting plasmids: pKK681 and pDHE681). The
portion of soluble and active Lph was increased to
approximately 10% of total protein by coexpression of
groELS (Figure 3). Up to 600 U/g (cell dry weight) were
produced in a 12 L batch fermentation compared to
1
50 U/g without groELS and 3 U/g in Lu681. For
Figure 2. DNA sequence of the lph gene and deduced
peptide sequence of Lph. Alignment with homologous
sequences (Table 3) reveals a very conserved region (con-
sensus sequence htgTHiDAPxHf, bold letters). Peptides
identified by Edman degradation of purified Lph are shown
in italic letters.
convenient production of the biocatalyst with the
pDHE system, a rhamnose-deficient E. coli strain can
be used.
Adv. Synth. Catal. 2002, 344, 1103 ± 1110
1105

FULL PAPERS
Maria Kesseler et al.
Table 3. Homologies of Lph.
Homologous Sequence
Accession No.
or Reference)
Organism
Homology[8]
(% aa)
Identity[8]
(% aa)
(
AGR_L_227p
AAK88681
NP 386248
NP 487809
WO9119175
NP 519956
B69206
D70817
B72430
G69399
S39964
S75497
T06135
T05418
O68500
O86485
Q54196
Agrobacterium tumefaciens
Sinorhizobium meliloti
Nostoc sp.
Pseudomonas putida
Ralstonia solanacearum
M. thermoautotrophicum
M. tuberculosis
T. maritima
Archaeoglobus fulgidus
S. griseus
Synechocystis sp.
A. thaliana
93.3
59.3
46.6
47.9
43.8
42.6
37.1
45.4
35.5
39.6
36.3
31.3
33.0
42.7
43.2
39.6
90.3
52.2
38.6
37.4
34.3
29.8
28.5
31.9
26.3
29.8
26.8
23.1
23.7
32.6
34.9
29.8
Hypothetical transmembrane protein
Hypothetical protein
Isatine hydrolase
Hypothetical protein
Hypothetical protein
Hypothetical protein
Hypothetical protein
Hypothetical protein
Hypothetical protein
Hypothetical protein
Hypothetical protein
Hypothetical protein
Polyketide synthase
Tetracyclin synthase
ORF
A. thaliana
S. peuceticus
S. argillaceus
S. griseus
Figure 3. Expression of lph under control of PRha in E. coli with (Lu9981) and without (Lu9871) coexpression of groELS.
Coomassie- (left) and activity-staining with a-naphthyl acetate and Fast-Red (right). MW-marker (1), Lu9871 cells (37 8C),
whole fraction (2), Lu9781 supermatant (3) and Lu9871 pellet (4), (5) ± (7) Lu9981 (37 8C) analogues, (8) ± (10) Lu9981 (30 8C)
analogues. Lane 6 indicates the increase of soluble and active Lph with coexpression of groELS compared to lane 3. Growth at
30 8C does not show further advantage.
Resolution of Pantolactone with Lph Extract
homogenate gave d-pantolactone in 50 ± 53% yield with
0 ± 95% ee for 6 batches (6 d; Figure 4). Lph and the
9
For an economic resolution process of rac-pantolactone, product solution were separated by membrane filtration
the repeated use of the corresponding catalyst is for each batch. Taking into account that 8% of solution
required. E. coli cells suspended in a reactor vigorously was lost during each filtration step due to the residual
stirred for quick titration of the synthesized pantoic acid volume of the laboratory equipment, a half-life-time of
and containing high concentrations of pantolactone and 12 (instead of 6) days might be estimated. The reason for
Na or NH pantoate tend to lyse and are hard to be the decrease in reaction velocity for each conversion was
4
recovered from the product solution. Therefore, Lph found to be a competitive inhibition by d-pantolactone
was used as isolated enzyme after heat precipitation and (Eq. 1) and a slight product inhibition of Lph. Under the
membrane filtration of the E. coli extract. Repeated same conditions of pH and temperature, l-pantolactone
conversion of d,l-pantolactone (30% w/v) with this was completely converted to l-pantoic acid.
1106
Adv. Synth. Catal. 2002, 344, 1103 ± 1110

Novel Biocatalyst for the Resolution of rac-Pantolactone
FULL PAPERS
Figure 4. Resolution of pantolactone with Lph extract,
repeated batch assay (7 days). The reaction process was
followed by titration of the produced pantoic acid by 4 M
NaOH. *: conversion according to NaOH consumption.
Conversions given in the inset were determined by HPLC
analysis.
Figure 5. Resolution of pantolactone with EupergitC-Lph,
repeated batch assay (6 days). The reaction process was
followed by titration of the produced pantoic acid by 10 M
NaOH. *: conversion according to NaOH consumption.
Conversions given in the inset were determined by HPLC
analysis.
After filtration of a homogenate resolution mixture of Lph was desirable. A library of 11680 lph mutants was
with > 50% conversion and > 90% ee, d-pantolac- prepared by error prone PCR with a mutation rate of
tone can be extracted with ketones or esters of medium 1.9 bp/lph gene (1.36 aa/Lph-variant) and a transition to
[
9]
polarity with good yields. l-Pantoic acid can be transversion rate of 0.89 (corresponding to 53% active
recycled from the aqueous layer by addition of mineral clones). After screening for increased l-pantolactone
acid and heating as indicated in Figure 1.
hydrolyzing activity, 3 isolated mutants F62S, K197D
and F100L with 2.3, 1.7 and 1.5 fold activity in the
screening assay were verified and characterized so far
[
10]
ꢁ
1†
(
Table 4).
While mutant F62S shows excellent activity in the
standard enzyme assay, its performance under process
conditions (high concentrations of d,l-pantolactone)
was poor. Since close adaptation of the high throughput
screening assay to these conditions is very difficult this
Immobilization of Lph and Resolution of Pantolactone
with Immobilized Lph
From several attempts to efficiently and cheaply example shows the challenge of assay development for
immobilize Lph, covalent attachment to EupergitC improvement of industrially valuable catalysts. It also
(
Rˆhm GmbH, Darmstadt, Germany) worked best. shows that the factors of improvement achievable by
Binding of Lph from homogenate to EupergitC pro- only one round of random mutagenesis and screening
ceeded with 55 ± 70% yield achieving specific activities strongly depend on what the starting point is, because
of 24 ± 81 U/g (wet weight). E. coli cells could easily be obviously, like in this case, the specific activity of an
entrapped into various forms of alginate beads with already good biocatalyst is not easily improved by
good yield of activity, however, the performance of these
immobilizates in the repeated batch reaction was poor in
terms of mechanical stability.
Table 4. Comparison of Lph mutants.
The repeated batch conversions of d,l-pantolactone Mutation
with Eupergit-Lph show kinetic characteristics similar
to those of the reactions with free enzyme (Figure 5).
Number
of Clones
Specific
Activity[
Specific Activity
with 30%
a]
d,l-Pantolactone[
b]
Depending on the solid portion (>8 ± 10%), the stirring
geometry and velocity, Eupergit-Lph immobilizate was
slowly ground to fine particles that tended to clog the F100L
filters during product recovery. Nevertheless, even with
this unoptimized system the half-life-time of Lph was
extended to 13 days.
F62S
K197D
4
2
3
227%
171%
149%
>70%
120%
120%
[a]
Activity according to the standard enzyme assay with
.15 M l-pantolactone compared to the activity of the
0
wild-type clone with groELS coexpression (TG1 pKK681
pAgro4 pHSG575).
[
b]
Resolution of d,l-pantolactone was performed as de-
scribed below. 50 ± 54% conversion and ee > 90% were
reached after addition of only 80% of biomass compared
to the wild-type clone with groELS coexpression (TG1
pKK681 pAgro4 pHSG575) under the same conditions.
Directed Evolution of Lph
For reduction of biocatalyst costs in the production
process of d-pantolactone, an improved specific activity
Adv. Synth. Catal. 2002, 344, 1103 ± 1110
1107

FULL PAPERS
Maria Kesseler et al.
factors in the range of 10 ± 100 like in cases where the monitored at 220 nm. d- and l-pantoic acid were determined
[
11]
by chiral HPLC analysis of the filtrates as described previous-
ly[ with modified flow (0.7 mL/min) and slightly increased
acetonitrile concentration (8%).
starting point was substantially lower. Resolution of
d,l-pantolactone by mutants K197D and F100L gave d-
pantolactone with 90.4% ee at 51.8% conversion and
13]
9
7
0.2% ee at 50.1% conversion, respectively, after
20 min using only 80% of the biomass compared to
the unmutated control (90.5% ee at 52.3% conversion
after 900 min). This corresponds to an increase of the
productivity of the recombinant biocatalyst from 136 to
Purification of l-Pantolactone Hydrolase from Lu681
and Activity Stainingafter SDS-PAGE
1
70 g/g (g d-pantolactone per g biomass) and a slight
1
2
128 g of wet pellet were vigorously resuspended in 1.8 L
0 mM Tris/HCl, pH 7.4, buffer to an end volume of 3 L. This
shortening of the reaction time down to 12 h.
solution was freed of small particles by filtration on a bed of
glass beads (0.1 ± 0.2 mm). Thereafter, the cells were homo-
genized in a microfluidizer (1000 bar). After washing of the
system the total volume was 4 L.
Conclusion
The results presented here are promising for industrial
utilization of recombinant Lph from Agrobacterium
tumefaciens Lu681 for the resolution of d,l-pantolac-
tone in a process for the production of d-pantothenic
acid. Further progress might be obtained by saturation
mutagenesis of the aa positions 62, 100 and 197, by
combining the mutations and introducing these mutants
To 4 L homogenate 200 mL 1 M MnCl were added. The pH
was continuously adjusted to pH 7.0. The precipitate was
removed by centrifugation. To the cleared solution we added
2
2
00 mL of a 0.2 M EDTA solution (pH 7.5) and the pH
dropped to 5.0. The newly formed precipitate was also
removed by centrifugation and the pH was readjusted to 7.0.
Then, more protein was precipitated by ammonium sulfate
50% saturation). The precipitate was removed by centrifuga-
tion and the supernatant (3.7 L) was divided in 1.2 L frac-
(
into further evolutionary cycles or a shuffling proce-
dure.[ Future investigations on the crystal structure of tions.
12]
Lph are expected to reveal molecular details of the
For hydrophobic chromatography one 1.2 L fraction was
enzymatic l-pantolactone hydrolysis. Structural infor- applied to a phenyl-sepharose column (Pharmacia, 5 cm
diameter, 25 cm height, 490 mL volume). The column was
mation about Lph might also be used to design an
enzyme not inhibited by excesses of d-pantolactone.
washed with 1 L buffer A (20 mM sodium phosphate, pH 7.4,
0% ammonium sulfate). Active fractions were eluted in a
4
linear gradient to 100% buffer B (20 mM sodium phosphate,
pH 7.4; 120 min, 10 mL/min) and collected (250 mL). They
were diluted to a conductance of less than 7 ms/cm (3 L) and
applied to a Q-sepharose column (Pharmacia Fast Flow, 5 cm
diameter, 25 cm height, 490 mL volume) for ion exchange
chromatography. The column was washed with 1 L buffer B
Experimental Section
Strains and Chemicals
(
10 mL/min) and eluted with buffer C (buffer B with 1 M
Agrobacterium tumefaciens Lu681 (DSMZ 13050, formerly
identified as Burkholderia caryophylli) was grown overnight in
Luria Bertani medium in flask cultures or a 12 L fermenter at
NaCl) in a linear gradient during 2 h to 100% buffer C. Active
fractions were collected (118 mL) and concentrated by ultra-
filtration (10 kDa, Omega Membrane) and then dialysed
against 10 mM Tris/HCl pH 7.0 (21 mL). 6 mL of this prepa-
ration were applied to a Waters Q HR8. The column was
equilibrated with 20 mM MES, pH 6.0. The enzymatic activity
was eluted with the same buffer including 500 mM of NaCl in a
linear gradient at a velocity of 1% per minute to 100%. Active
fractions were dialysed and centrifuged and then applied to
Mono-P chromatography (Pharmacia, 0.5 cm diameter, 5 mL).
The enzymatic activity was again eluted with NaCl and the
active Mono-P fractions were precipitated with cold acetone
and applied to non-reducing SDS-PAGE without heating the
sample (12% Tris/glycine minigels, 2.5 h, 125 V, 50 mA; 0.4%
SDS in the sample buffer).
After SDS-PAGE the gel was briefly washed in Tris buffered
saline and incubated in 50 mL TBS with a-naphthyl acetate
(Sigma N8508, 20 mg in 5 mL acetone ‡ 45 mL water) for
10 min. After washing the gel was incubated with 50 mL Fast
Red solution (Sigma F-8764, 50 mg in 50 mL water). Esterases
(and Lph) form red bands. Coomassie stained bands in a
second set of gels were cut out and digested in the gel slices with
trypsin. Peptides were isolated by C18 HPLC and sequenced
by Edman degradation.
30 8C. All chemicals were purchased from Sigma-Aldrich
Chemie GmbH, Munich, unless otherwise stated. Sodium
pantoate was synthesized by complete hydrolysis of pantolac-
tone: 20 mL of 12 M NaOH were added to 1 mL of 100 mM
pantolactone and heated for 5 min at 95 8C. The solution was
neutralized by addition of 13 mL of concentrated HCl and
dilution in 50 mM Tris/HCl buffer, pH 7.0.
Enzyme Assay
l-Pantolactone hydrolase activity was routinely assayed for 1 h
at 30 8C in a mixture containing 150 mM Pipes pH 7.0, 150 mM
l-pantolactone and an appropriate amount of enzyme or cells
(
0.1 U/mL). The reaction was stopped on ice and the cells/
enzyme filtered off using a 10 kDa membrane spin filter.
Filtrates were analyzed by HPLC using a 2.1 mm  150 mm
Alltima C18-column (5 mm, Alltech GmbH, Unterhaching,
Germany) and a linear gradient of 5 ± 50% acetonitrile in
potassium phosphate (100 mM, pH 2.0) in 6 minutes at 0.3 mL/
min on an HP1090 (oven temperature 65 8C) with diode-array
detection. The pantoic acid and pantolactone signals were
1108
Adv. Synth. Catal. 2002, 344, 1103 ± 1110

Novel Biocatalyst for the Resolution of rac-Pantolactone
FULL PAPERS
Cloningand Sequencingof the l-Pantolactone
Immobilization of l-Pantolactone Hydrolase
Hydrolase Gene
To 300 mL homogenate 17.5 g NaCl were added and the pH
was adjusted to 6.8. After addition of 15 g dry EupergitC
(Rˆhm AG, Darmstadt) the suspension was incubated for 17 h
at room temperature. The Lph-EupergitC immobilizate (60 g
wet wt) was washed with water and stored in 10 mM phosphate
buffer pH 7.5.
Genomic DNA of Agrobacterium tumefaciens Lu681 was
isolated with QIAGEN Genomic Tips, partially digested with
EcoRI and cloned into pBluescriptIISK (Stratagene). After
transformation into E. coli XL1Blue ultracompetent cells
(
(
Stratagene), colonies were replicated on nitrocellulose filter
HAWP09000, Millipore) and screened for activity: The filters
were sequentially soaked in 1% SDS/10 mM Tris/HCl pH 7.0
(
(
0
5 minutes at room temperature), 10 mM Tris/HCl pH 7.0
5 minutes at room temperature) and 150 mM l-pantolactone/
.1% nitrazine yellow/10 mM Tris/HCl pH 7.0 (from 3 minutes
Resolution of d,l-Pantolactone in a Repeated Batch
Process
up to 15 h at 30 8C). Positive colonies turned yellow and were
further characterized by DNA sequencing, subcloning into
expression vectors pKK223-3 (Pharmacia) and pDHE19.2 (a
derivative of pJOE2702, Stuttgart ) and biochemical char-
acterization.
A 20 ± 1000 mL mixture of d,l-pantolactone (300 g/L ˆ
2.3 M), 10 mM NaHCO (optional) and either Lph homoge-
3
nate (16000 U/L ) or immobilized Lph (16000 U/L or max.
80 g/L) was stirred at 30 8C with automatic control of the pH by
titration with 4 or 10 M NaOH (pH 7.5). After complete
hydrolysis of l-pantolactone (d-pantolactone ee ꢀ 90%, yield
[4a,7]
ꢀ
50%), the enzyme was recovered by membrane filtration
homogenate, see above) or by aspiration through an HPLC-
(
filter (solvent filter 655-1395, Merck), washed twice with 1
volume of water and used again.
Growth of Recombinant E. coli and Preparation
of l-Pantolactone Hydrolase Homogenate
For cultivation in flasks or microplates E. coli strains were
grown in Luria Bertani medium with appropriate antibiotics
and inducing agents at 37 8C (ampicillin 100 mg/L, spectino-
mycin 50 mg/L, chloramphenicol 10 mg/L, IPTG 0.1 ± 1 mM,
Generation of Lph Mutants
Lph mutants were generated by error-prone PCR with Taq-
polymerase (2.5 U, Roche) and pKK681 as template (2 ng/mL
using the primers 5'-CCGGAATTCATGTGCAACAAC-
TGC and 5'-CCCAAGCTTCAGACCAGGGCCAGAA.
Other concentrations were 40 mM ATP and GTP each,
rhamnose ¥ H O 2.5 g/L).
2
2
00 mL of rich tryptone-yeast medium with 100 mg/mL
ampicillin, e.g., 13.3 g/L KH PO , 4.0 (NH ) HPO , 1.2 g/L
2
4
4
2
4
MgSO ¥ 7 H O, 0.2 g/L CaCl , 1.7 g/L citric acid, 15.0 g/L
200 mM CTP and TTP each, 0.5 mM MgCl , 1 M GC rich
2
4
2
2
tryptone (Difco), 66.6 g/L yeast lysate 65% (Fould Springer),
1
resolution solution (Roche, contains DMSO), GC rich reaction
buffer (1x, Roche). The PCR program consisted of 3 min 95 8C,
10 min 85 8C, 4 cycles of 0.5 min at 54 8C, 2 min at 72 8C,
0.5 min at 95 8C and 26 cycles of 0.5 min at 58 8C, 2 min at
72 8C, 0.5 min at 95 8C and finally 10 min at 72 8C and cooling
to 4 8C. PCR products were isolated by E-Gel electrophoresis
(Invitrogen) and elution with Qiaquick or GFX-Kit (Qiagen or
Pharmacia, respectively). The isolated PCR products were
cloned with EcoRI/HindIII into pKK223-3. Ligation mixtures
were transformed into XL1 Blue competent cells containing
pAgro4 and pHSG575 for chaperone coexpression. Colonies
were transferred into 96-well microplates for screening.
Controls were prepared by standard PCR (Pfu polymerase)
and transferred into each microplate (A1-D1).
0 mL/L trace metal solution 100,[14] 0.1 g/L Fe(III) citrate,
pH 7.0 with NaOH, were inoculated with E. coli TG1
pDHE681 or TG1 pDHE681 pAgro4 pHSH575 and grown
for 8 h at 37 8C and 200 rpm. The preculture was added to 10 L
of glycine-tryptone-yeast medium (13.3 g/L KH PO , 4.0
2
4
(
NH ) HPO , 1.2 g/L MgSO ¥ 7 H O, 1.7 g/L citric acid,
4 2 4 4 2
1
5.0 g/L tryptone (Difco), 5 g/L yeast lysate 65% (Fould
Springer), 40 g/L glycerine 99.5% (DAB13, Riedel Ha e» n),
[14]
1
0
3
0 mL/L trace metal solution 100, 0.1 g/L Fe(III) citrate,
.1 g/L Tego KS911 antifoam). Culturing conditions were
7 8C, pH 7.0 (25% NaOH), ꢀ 600 rpm (pO ꢀ 25%). For
2
TG1 pDHE681 pAgro4 pHSH575 cultivation, 0.15 mM IPTG
was added after 1 h for induction of chaperone expression. For
induction of Lph expression, rhamnose (2.5 g/L rhamnose ¥
H O) was added after 1 h 15 min and kept at 1.0 to 2.0 g/L in
2
concentration by automatic HPLC control and addition of a
High Throughput Assay for Screening of Lph (and
Other Acid ProducingEnzymes)
10% solution. For rhamnose-deficient E. coli strains, adding
once 0.5 g/L rhamnose is sufficient. After 8 h cells were
harvested for whole cell assays (typical yields were 600 U/g
dry weight and 8000 ± 10000 U/L) or immediately disrupted in
a microfluidizer (2 Â 1000 bar) at 4 ± 10 8C. After centrifuga-
tion (20 min at 9000 g, 10 8C) Lph was concentrated (factor 10)
by cross-flow filtration (H‰moflow F60, Fresenius, 10 kDa
membrane). Heat precipitation of other proteins (20 min
A preculture of the mutants was grown in microplates
(LBamp spec cm) for 18 h at 37 8C with normal shaking at
220 rpm and used for inoculation of a similar, induced culture
containing 0.15 mM IPTG. After growth (18 h at 37 8C) the
cells were evenly suspended and 20 mL automatically diluted
into 80 mL 10 mM Tris/HCl pH 7.0. Growth was checked by
measurement of optical density at 600 nm. To start the assay
100 mL of 0.6 M rac-pantolactone/0.01% nitrazine yellow
pH 4.3 were added and decrease of the extinction was
measured at 590 nm every 10 min for 1 h. Clones were
evaluated by their decrease coefficient determined by linear
6
0 8C; 20 min 10 8C, centrifugation as above) led to a
homogenate reaching ca. 3000 U/g protein and 90000 U/L
>90% yield of activity and 49% protein content compared to
the starting cell material). Thehomogenate was stable for 2 d at
8C and > 3 months at À 20 8C.
(
4
Adv. Synth. Catal. 2002, 344, 1103 ± 1110
1109

FULL PAPERS
Maria Kesseler et al.
regression (™activity∫). Since the standard deviation of positive
controls was 21.5%, only clones showing > 120% activity
compared to the control level were reassayed. Clones showing
again > 120% activity were precultured overnight and grown
in 600 mL cultures for 6 h at 37 8C. Cells were harvested and
tested by an enzyme standard assay. Under these conditions
the wild-type clone reached 1200 U/L. Different amounts of
biomass were used for the resolution of 30% (w/v) d,l-
pantolactone in comparison with resolution batches (20 mL,
see above) with the wild-type clone. The latter gave 52.3%
conversion and ee 90.5% in 15 h with addition of a cell
suspension equivalent of 10.5 g/L dry biomass (ˆ100%; 13000
[3] a) S. Shimizu, M. Kataoka, Ann. N. Y. Acad. Sci.
(Enzyme Engineering XIII) 1996, 799, 650 ± 658; b) K.
Sakamoto, H. Yamada, S. Shimizu (Fuji Yakuhin Kogyo
Kabushiki Kaisha), EP 436730 B1, 1995; c) K. Sakamoto,
H. Yamada, S. Shimizu (Fuji Yakuhin Kogyo Kabushiki
Kaisha), EP 504421 B1, 1995.
[
4] a) M. Kesseler, B. Hauer, T. Friedrich, R. Mattes (BASF
AG), DE 19952501, 2000; b) M. Kataoka, J.-i. Nomura,
M. Shinohara, K. Nose, K. Sakamoto, S. Shimizu, Biosc.
Biotechnol. Biochem. 2000, 64, 1255 ± 1262.
[
[
[
5] S. Shimizu, M. Kataoka, K. Honda, K. Sakamoto, J.
Biotechnol. 2001, 92, 187 ± 194.
6] W. Weyler, T. C. Dodge, J. J. Lauff, D. J. Wendt
U/L).
(
Genencor International Inc.), WO 97/19175, 1997.
7] pJOE2702-derivative pDHE19.2: a) J. N. Volff, C. Ei-
Acknowledgements
chenseer, P. Viell, W. Piendl, Molecular Microbiol. 1996,
2
1, 1037 ± 1047; b) J. Altenbuchner, P. Viell, I. Pelletier,
The authors thank Prof. Dr. R. Mattes for the expression vector
pDHE19.2 and Prof. Dr. B. Bukau for the groELS coexpression
plasmids pAgro4/pHSG575. Dr. Markus Lohscheidt is thanked
for helping with the kinetic studies. We also thank Birgit Bender,
August Blechschmidt, Elke Eichler, Hans-J¸rgen Lang, Ag-
nieszka Peters and Andreas Sch‰dler for their technical
assistance.
Meth. Enzymol. 1992, 216, 457 ± 460.
8] GAP program: S. B. Needleman, C. D. Wunsch J. Mol.
Biol. 1970, 48, 443 ± 453.
9] M Kataoka, K. Shimizu, K. Sakamoto, H. Yamada, S.
Shimizu Appl. Microbiol. Biotechnol. 1995, 44, 333 ± 338.
10] L. Stryer, Biochemie, Spektrum der Wissenschaft Ver-
lagsgesellschaft mbH, Heidelberg, 1990, p. 203.
[
[
[
[
11] a) H. Uchiyama, T. Inaoka, T. Ohkuma-Soyejima, H.
Togame, Y. Shibanaka, T. Yoshimoto, T. Kokubo J.
Biochem. 2000, 128, 441 ± 447; for reviews see, e. g.,
b) F. H. Arnold, P. L. Wintrode, K. Miyazaki, A. Ger-
shenson, Trends Biochem. Sc. 2001, 26, 100 ± 106; c) M. T.
Reetz, K.-E. Jaeger, in Top. Curr. Chem.: Biocatalysis -
from discovery to application, Vol. 200, (Ed.: W. D.
Fessner), Springer, Berlin, 1999, pp. 31 ± 57.
References and Notes
[
1] a) H. Yamada Chimia 1993, 47, 5 ± 10; fermentation of
Corynebacterium: b) H. Sahm, L. Eggeling, Appl. Envi-
ron. Microbiol. 1999, 65, 1973 ± 1979; c) L. Eggeling, G.
Thierbach, H. Sahm (Degussa-H¸ls AG), EP 1006189,
2
000; fermentation of E. coli: d) T. Moriya, Y. Hikichi, Y.
Moriya, T. Yamaguchi (Takeda Chemical Ind.), WO 97/
0340, 1997; e) Y. Hikichi, T. Moriya, I. Nogami, H. Miki,
[
12] For recent successful applications of the diverse tech-
niques, see e. g. a) K. A. Gray, T. H. Richardson, K.
Kretz, J. M. Short, F. Bartnek, R. Knowles, L. Kan, P. E.
Swanson, D. E. Robertson, Adv. Synth. Catal. 2001, 343,
1
T. Yamaguchi (Takeda Chemical Ind.), EP 590857, 1994,
f) N. Dusch, A. P¸hler, J. Kalinowski, Appl. Environ.
Microbiol. 1999, 65, 1530 ± 1539; g) F. Elischewski, J.
Kalinowski, A. P¸hler, N. Dusch, J. Dohmen, M.
Farwick, G. Thierbach (Degussa-H¸ls AG), EP
6
07 ± 617; b) K. A. Powell, S. W. Ramer, S. B. del
Cardayre, W. P. C. Stemmer, M. B. Tobin, P. F. Long-
champ, G. W. Huisman, Angew. Chem. Int. Ed. 2001, 40,
3
948 ± 3959.
1
001027, 2000; h) N. Dusch, J. Kalinowski, A. P¸hler
[
[
13] M Kataoka, K. Shimizu, K. Sakamoto, H. Yamada, S.
Shimizu Appl. Microbiol. Biotechnol. 1995, 43, 974 ± 977.
14] D. J. Korz, U. Rinas, K. Hellmuth, E. A. Sanders, W. D.
Deckwer, Journal of Biotechnology 1995, 39, 59 ± 65.
(Degussa-H¸ls AG), EP 1006192, 2000; i) M. Rieping, G.
Thierbach, W. Pfefferle, N. Dusch, J. Kalinowski, A.
P¸hler (Degussa-H¸ls AG), EP 1006193, 2000.
2] S. Shimizu, M. Kataoka, K. Shimizu, M. Hirakata, K.
[
Sakamoto, H. Yamada, Eur. J. Biochem. 1992, 209, 383 ±
3
90.
1110
Adv. Synth. Catal. 2002, 344, 1103 ± 1110