Engineering Mesorhizobium loti pyridoxamine-pyruvate aminotransferase for production of pyridoxamine with l-glutamate as an amino donor

Article abstract of DOI:10.1016/j.molcatb.2010.07.013

Pyridoxamine-pyruvate aminotransferase (PPAT), a novel pyridoxal 5′-phosphate-independent aminotransferase, reversibly catalyzes the transfer of an amino group between pyridoxamine and pyruvate to generate pyridoxal and l-alanine. The enzyme can be used for synthesis of pyridoxamine, a promising candidate for prophylaxis and treatment of diabetic complications. A disadvantage of PPAT for industrial application to the synthesis is that it requires an expensive amino acid l-alanine as an amino donor. Here, mutated PPATs with a high activity toward 2-oxoglutarate (and hence toward l-glutamate) were prepared by a rational design plus random mutagenesis of the wild-type PPAT because l-glutamate is readily and economically available. The PPAT(Y35H/V70R/F247C) showed 9.1-fold lower K_m and 4.3-fold higher k_cat values than those of the wild-type PPAT. The model of the complex of mutated PPAT and pyridoxyl-l-glutamate showed that γ-carboxyl group of l-glutamate was hydrogen-bound with an imidazole group of His35. The production of pyridoxamine from pyridoxal with transformed Escherichia coli cells expressing the mutated PPAT did not correlate with the k_cat value or catalytic efficiency of the mutated PPAT but with K_m value at a low level. E. coli cells expressing the PPAT(M2T/Y35H/V70K/E212G) could be used for in vitro conversion of pyridoxal into pyridoxamine at 30 °C with l-glutamate as an amino donor.

Full text of DOI:10.1016/j.molcatb.2010.07.013

Journal of Molecular Catalysis B: Enzymatic 67 (2010) 104–110
Contents lists available at ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Engineering Mesorhizobium loti pyridoxamine–pyruvate aminotransferase for
production of pyridoxamine with l-glutamate as an amino donor
Yu Yoshikanea, Asuka Tamuraa, Nana Yokochia, Khalil Ellouzea, Eitora Yamamurab, Hanae Mizunagab,
Noboru Fujimotob, Keiji Sakamotob, Yoshihiro Sawac, Toshiharu Yagia,∗
a Faculty of Agriculture and Agricultural Science Program, Graduate School of Integrated Arts and Sciences, Kochi University, Monobe-Otsu 200, Nankoku, Kochi 783-8502, Japan
b Technical Department, Daiichi Fine Chemical Co., Ltd., 530 Chokeiji, Takaoka, Toyama 933-8511, Japan
c Department of Life Science and Biotechnology, Shimane University, Matsue, Shimane 690-8504, Japan
a r t i c l e i n f o
a b s t r a c t
Pyridoxamine–pyruvate aminotransferase (PPAT), a novel pyridoxal 5ꢀ-phosphate-independent amino-
transferase, reversibly catalyzes the transfer of an amino group between pyridoxamine and pyruvate to
generate pyridoxal and l-alanine. The enzyme can be used for synthesis of pyridoxamine, a promising
candidate for prophylaxis and treatment of diabetic complications. A disadvantage of PPAT for industrial
application to the synthesis is that it requires an expensive amino acid l-alanine as an amino donor. Here,
mutated PPATs with a high activity toward 2-oxoglutarate (and hence toward l-glutamate) were pre-
pared by a rational design plus random mutagenesis of the wild-type PPAT because l-glutamate is readily
and economically available. The PPAT(Y35H/V70R/F247C) showed 9.1-fold lower Km and 4.3-fold higher
kcat values than those of the wild-type PPAT. The model of the complex of mutated PPAT and pyridoxyl-l-
glutamate showed that ␥-carboxyl group of l-glutamate was hydrogen-bound with an imidazole group
of His35. The production of pyridoxamine from pyridoxal with transformed Escherichia coli cells express-
ing the mutated PPAT did not correlate with the kcat value or catalytic efficiency of the mutated PPAT but
with Km value at a low level. E. coli cells expressing the PPAT(M2T/Y35H/V70K/E212G) could be used for
in vitro conversion of pyridoxal into pyridoxamine at 30 ◦C with l-glutamate as an amino donor.
© 2010 Elsevier B.V. All rights reserved.
Article history:
Received 31 May 2010
Received in revised form 24 July 2010
Accepted 26 July 2010
Available online 1 August 2010
Keywords:
Pyridoxamine
Pyridoxamine–pyruvate aminotransferase
Substrate specificity
Rational design
Random mutagenesis
1. Introduction
mal model, pyridoxamine indeed inhibits early renal disease and
dyslipidemia in the streptozotocin-diabetic rats [9]. Phase 2 studies
Vitamin B6 consists of 6 natural forms, pyridoxamine, pyridoxal,
pyridoxine, pyridoxamine 5ꢀ-phosphate, pyridoxal 5ꢀ-phosphate
(PLP), and pyridoxine 5ꢀ-phosphate. PLP serves as a cofactor of many
reactions involved in amino acid metabolism, such as transamina-
tion, racemization, decarboxylation, and deamination [1]. Recently,
addition to this cofactor action, vitamin B6 and its derivatives
have been shown to be effective antioxidants and quenchers of
singlet oxygen [2–6]. Furthermore, pyridoxamine interferes with
the formation of advanced glycation end-products (AGEs) and
advanced lipoxidation end-products (ALEs), and thus inhibits the
non-enzymatic modification of proteins. AGEs and ALEs are the
major pathogenic factors in diabetic complications [7,8]. In an ani-
of patients with type 1 and type 2 diabetes and overt nephropathy
showed that pyridoxamine provided for a foundation for further
evaluation of AGE inhibitor in diabetic nephropathy [10]. Thus,
pyridoxamine is a promising candidate for a prophylactic and/or
a remedy for diabetic complications. The amount of consumption
of pyridoxamine should become massive because it has been esti-
mated that there are about 180 million diabetics in the world.
Pyridoxamine is currently synthesized by chemical methods.
Two routes (oxidative or non-oxidative method) for the chemical
synthesis are known, where the starting material is the readily and
economically available pyridoxine. The oxidative method, which
has been used for industrial production, changes pyridoxine to pyri-
doxal, and then pyridoxal to pyridoxamine through the pyridoxal
oxime, which is reduced with Pd/C catalyst to pyridoxamine [11].
The non-oxidative one changes pyridoxine to pyridoxamine by the
Gabriel synthesis [12]. Although the latter method does not use
Pd/C catalyst, it takes several steps including production of a pri-
mary alkyl halide derivative of pyridoxine. Enzymatic or microbial
synthetic method is generally preferable in the environmental and
energetic aspects, if they are available. As far as we know, no enzy-
matic method for production of pyridoxamine has been reported.
Abbreviations:
PPAT, pyridoxamine–pyruvate aminotransferase; PLP,
pyridoxal 5ꢀ-phosphate; IPTG, isopropyl-␤-d-thiogalactopyranoside; X-gal, 5-
bromo-4-chloro-3-indolyl-␤-d-galactopyranoside; AGEs, advanced glycation
end-products; ALEs, advanced lipoxidation end-products; CHES, N-cyclohexyl-2-
aminoethanesulfonic acid.
∗
Corresponding author. Tel.: +81 88 864 5191; fax: +81 88 864 5191.
E-mail address: yagito@kochi-u.ac.jp (T. Yagi).
1381-1177/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcatb.2010.07.013

Y. Yoshikane et al. / Journal of Molecular Catalysis B: Enzymatic 67 (2010) 104–110
105
Fig. 1. (A) Reaction catalyzed by pyridoxamine–pyruvate aminotransferase. (B) Superimposition of active site residues found in PPAT–pyridoxyl-l-alanine complex (C–C
bonds are shown in a cyan color. PDF code, 2Z9X) and Thermus thermophilus aspartate aminotransferase–malate complex (C–C bonds are shown in a green color. 1BKG).
Amino acid residues in the latter are shown in italics. The figures were drawn with WebLab. (For interpretation of the references to color in this figure legend, the reader is
referred to the web version of the article.)
Recently, we have identified the gene encoding pyridoxamine–
pyruvate aminotransferase (PPAT) in a nitrogen-fixing symbi-
otic bacterium, Mesorhizobium loti MAFF303099, characterized the
over-expressed recombinant enzyme [13], and determined its ter-
tiary structure [14]. PPAT shows high catalytic efficiency and
thermal stability, and reversibly catalyzes the transfer of an amino
group between pyridoxamine and pyruvate to produce pyridoxal
and l-alanine (Fig. 1A). Thus, PPAT may be used for an indus-
trial enzymatic production of pyridoxamine if its activity toward
l-glutamate, which is much cheaper than l-alanine, could be
increased to use l-glutamate as an amino donor.
PPAT is classified into class V aminotransferases of fold type I of
the PLP-dependent enzymes even though it is not PLP-dependent
[13]. A representative enzyme in the fold type I is aspartate amino-
transferase which shows activity towards l-glutamate. Thus, a
comparison of the active site architecture of PPAT to that of aspar-
tate aminotransferase will give us a rational design of mutated
PPAT with a high activity towards l-glutamate. Fig. 1B shows
superimposed active site residues of PPAT binding pyridoxyl-l-
alanine (PDB code, 2Z9X) and Thermus thermophilus aspartate
aminotransferase binding maleate (1BKG). In aspartate amino-
transferase, one carboxylate in l-glutamate analog, maleate, is
hydrogen-bonded to Lys101 and the other forms charged hydro-
gen bonds with Arg361. Arg361 of aspartate aminotransferase
corresponds to Arg345 of PPAT and binds to ␣-carboxyl group
of substrates. Lys101 in aspartate aminotransferase binds to
␥-carboxyl group of substrate l-glutamate. In PPAT, the corre-
sponding amino acid residue is Val70. Thus, if Val70 is changed
to basic amino acid residue, PPAT may show reactivity towards
l-glutamate.
2. Materials and methods
2.1. Bacterial strains, cultivation, plasmids, and reagents
E. coli strains, JM109 and BL21(DE3), were purchased from
TaKaRa Bio (Tokyo, Japan) and Novagen-Merck Japan (Tokyo,
Japan), respectively. E. coli cells were cultured in a modified LB
medium (1% polypeptone, 0.5% yeast extract, and 1% NaCl) con-
taining 50 ␮g/ml ampicillin at 37 ◦C. Plasmids, pET6806 carrying
wild-type PPAT gene [13], pET21a (Novagen-Merck Japan), pUC119
(TaKaRa Bio), and pQE-30 (QIAGEN Japan, Tokyo, Japan), were
used. M. loti MAFF303099, which has the PPAT gene, was obtained
from MAFF (Ministry of Agriculture, Forestry and Fisheries) DNA
bank (Tsukuba, Japan). Plasmids carrying various mutated PPATs
were prepared in this study and are available on request. Pyri-
doxal and pyridoxamine were purchased from Wako (Tokyo, Japan)
and Tokyo Chemical Industry (Tokyo, Japan), respectively. Sodium
pyruvate, L-alanine, and l-glutamate were purchased from Wako.
Potassium 2-oxoglutarate and lysozyme from chicken egg white
were purchased from Sigma Japan (Tokyo, Japan).
2.2. Site-directed mutagenesis and expression of mutated PPAT
Genes encoding PPAT mutants were prepared with the plasmid
pET6806 as a template, as described previously [13]. The primers
used were: for PPAT(V70K), 5ꢀ-GCGAGCCGAAGCTCGGCC-3ꢀ
and
5ꢀ-GGCCGAGCTTCGGCTCGC-3ꢀ;
for
PPAT(V70R),
5ꢀ-
GCGAGCCGAGGCTCGGCC-3ꢀ and 5ꢀ-GGCCGAGCCTCGGCTCGC-3ꢀ.
The underlined letters indicate mismatched sites. Presence of the
desired mutation and the absence of other mutation in the mutated
enzyme genes were confirmed by the DNA sequencing with an ABI
PRISM 3100-Avant Genetic Analyzer (Applied Biosystems Japan,
Tokyo, Japan). The constructed plasmids containing the mutated
PPAT genes were introduced into BL21(DE3) cells, and then the
transformed cells were aerobically grown in 5 ml of LB medium
containing ampicillin at 37 ◦C for 16 h. The cells were harvested by
centrifugation at 8400 × g for 10 min at 4 ◦C, and then washed with
0.9% NaCl, and stored at −20 ◦C until use.
Here, PPAT was engineered to increase reactivity toward l-
glutamate or 2-oxoglutarate by the rational design plus random
mutagenesis; because PPAT catalyzes reversible reaction as shown
in Fig. 1A, the enzyme with higher activity toward l-glutamate
should show high reactivity toward the corresponding amino
receptor 2-oxoglutarate. The transformed Escherichia coli cells
expressing PPAT(M2T/Y35H/V70K/E212G) could convert pyridoxal
into pyridoxamine in a yield of almost 100%.

106
Y. Yoshikane et al. / Journal of Molecular Catalysis B: Enzymatic 67 (2010) 104–110
2.3. First error-prone PCR for preparation of PPAT(Y35H/V70K)
and PPAT(V70R/F247C)
essentially the same method as that described above. The DNA
fragments were digested with BamHI and PstI, and then ligated
into the BamHI/PstI sites of pQE-30. The plasmids were introduced
into JM109 cells, and then the cells were grown on LB agar plates
containing ampicillin. The colonies on the plate were picked up,
and each was inoculated into LB medium containing ampicillin
and 1 mM IPTG, then the production of pyridoxamine with l-
glutamate as an amino donor was measured as described later.
Two clones, which showed a higher production of pyridoxamine
than PPAT(V70K)-expressing cells, were selected, and then the
plasmids in the cells were sequenced. The plasmids expressed
PPAT(V70K/E212G) and PPAT(M2T/V70K).
Random mutagenesis of an mlr6806 (PPAT) gene was per-
formed by the error-prone PCR. The BamHI–EcoRI fragment
of pET6806(V70K) carrying the V70K-mutated PPAT gene was
inserted into the BamHI/EcoRI sites of pUC119 to construct
pUC6806(V70K). The pUC6806(V70K) was used as a template
for error-prone PCR. Error-prone PCR was performed in a reac-
tion mixture (50 ␮l) consisting of GC buffer (TaKaRa Bio), 2 ng
pUC6806(V70K), 0.2 mM dNTPs, 4 mM MgCl2, 20 pmol each primer,
and 2.5 units BIOTAQTM DNA polymerase (BIOLINE, London, Eng-
land). Primers, 5ꢀ-CCCGGATCCGAGCTGATGTACTCGCACGACAT-3ꢀ
and 5ꢀ-CCCGAATTCTCAGGCGTCGGCGTCGATTACG-3ꢀ, with a BamHI
site (underlined in the former) and an EcoRI site (underlined in the
latter), respectively, were used. The PCR conditions were: heating
at 94 ◦C for 2 min; 50 cycles at 94 ◦C for 15 s, 57.5 ◦C for 30 s, and
72 ◦C for 180 s.
2.6. Construction of PPATs mutated at three or four sites
For preparation of PPAT(Y35H/V70R/F247C), site-specific
mutagenesis of Tyr35 to His was performed with a KOD-Plus-
Mutagenesis kit (Toyobo, Tokyo, Japan) according to the instruction
manual.
3ꢀ (the underlining indicates mismatched sites) and
5ꢀ-AAGCACCGTACGGCCGAGGCC-3ꢀ,
and template,
Primers,
5ꢀ-CATGATTACGATCCTGCATTCCAGCTCC-
2.4. Screening of clones expressing mutated PPAT with a high
activity toward 2-oxoglutarate
a
pUC6806(V70R/F247C), were used. PPAT(Y35H/V70R/F247C)
was expressed by pUC6806(Y35H/V70R/F247C) in JM109
The randomly mutated PCR fragments were subjected to 1%
agarose gel electrophoresis, and purified with an E. Z. N. A.
Gel Extraction Kit (Omega Bio-Tek, Norcross, GA, USA). Iso-
lated DNA fragments were digested with BamHI and EcoRI, and
then ligated into the BamHI/EcoRI sites of pUC119 treated with
alkaline phosphatase. The plasmids harboring the isolated DNA
fragments were introduced into JM109 competent cells, and
then the cells were grown on LB agar plates containing ampi-
cillin, 0.1 mM isopropyl-␤-d-thiogalactopyranoside (IPTG), and
40 ␮g/ml 5-bromo-4-chloro-3-indolyl-␤-d-galactopyranoside (X-
gal). White colonies on the plate were picked up, and each was
inoculated into a separate well containing 0.9 ml of LB-ampicillin
in a deep 96-well plate (BIO–BIK, Tokyo, Japan). The plate was
incubated at 37 ◦C for 16 h under shaking, and then centrifuged at
1200 × g for 10 min at 4 ◦C. The precipitated cells in a deep 96-well
plate were suspended in 55 ␮l of 10 mM Tris–HCl (pH 7.0) contain-
ing 0.1 mg/ml lysozyme. The cell suspensions were incubated for
1 h at 37 ◦C to lyse the cells. Supernatants (50 ␮l each) obtained by
centrifugation at 1200 × g for 10 min at 4 ◦C were separately trans-
ferred to a well of another 96-well plate. The enzyme reaction in
each well was done for 2 h at 30 ◦C by the addition of 150 ␮l of
reaction mixture consisting of 50 mM Tris–HCl, pH 9.0, 1 mM pyri-
doxamine, and 50 mM 2-oxoglutarate (or 5 mM sodium pyruvate).
Then, absorbance at 405 nm of a Schiff base between the produced
pyridoxal and Tris base of the reaction mixture in each well was
measured with a Bio-Rad Model 550 microplate reader (Tokyo,
Japan). Two clones expressing a mutated PPAT with a higher activity
toward 2-oxoglutarate than that of PPAT(V70K) were selected. Plas-
mids were prepared from the selected clones, and then the mutated
base was determined by DNA sequencing. PPAT(Y35H/V70K) and
PPAT(V70R/F247C) were consequently obtained.
cells.
pQE6806(M2T/V70K/E212G)
which
expressed
PPAT(M2T/V70K/E212G) in JM109 were prepared by a PCR in
which pQE6806(V70K/E212G), 5ꢀ-TCAGGCGTCGGCGTCGATTAC-
3ꢀ and 5ꢀ-ATGACGCGCTATCCCGAACATGCCG-3ꢀ were used
as
a
template and primers, respectively. The amplified
DNA fragment was then inserted into SmaI site of pQE-
30. Similarly, PCR was done with pQE6806(Y35H/V70K) as
a
template. The amplified DNA fragment was inserted into
SmaI site of pQE-30 to prepare pQE6806(M2T/Y35H/V70K),
which
To
expressed
prepare pQE6806(M2T/Y35H/V70K/E212G),
PPAT(M2T/Y35H/V70K/E212G) in
PPAT(M2T/Y35H/V70K)
in
JM109.
which
JM109,
expressed
pQE6806(M2T/Y35H/V70K) and pQE(V70K/E212G) were sep-
arately digested with XhoI, and applied to an agarose gel
electrophoresis. Then, the DNA fragment with 0.5 kbp from
the former and that with 4.1 kbp from the latter were ligated. The
ligated DNA fragment was inserted into pQE-30. The presence of
mutation at the desired site and absence of other mutations was
confirmed by DNA sequencing with an ABI PRISM 3100-Avant
Genetic Analyzer.
2.7. Preparation of crude extracts containing mutated PPATs and
their purification
The crude extracts were prepared as described previously [13],
andmutatedPPATs were purified essentiallyby thesame procedure
as that for purification of the wild-type PPAT [14]. Briefly, the cells
expressing the mutated PPATs were suspended in 0.5 ml of 20 mM
potassium phosphate, pH 8.0, containing 1 mM EDTA and 1 mM
phenylmethylsulfonyl fluoride. The cell suspension was sonicated
twice for 15 s on ice, and then centrifuged at 8400 × g for 20 min
at 4 ◦C. The supernatants obtained were used as crude extracts. For
the purification of mutated PPATs, crude extracts were sequentially
applied to a butyl-Toyopearl, QA52, and Sephacryl S-200 column
chromatographies.
2.5. Screening of clones expressing mutated PPATs which produce
a high amount of pyridoxamine with l-glutamate as an amino
donor
Plasmid pUC6806, pUC6806(V70K), or pUC6806(Y35H/V70K)
2.8. Enzyme and protein assays
was
primers
used
for
PCR
as
a
template
together
and
with
5ꢀ-ATGATGCGCTATCCCGAACATGCCG-3ꢀ
5ꢀ-
Pyridoxamine–2-oxoglutarate aminotransferase activity was
determined by the phenylhydrazine method as follows. The reac-
tion mixture (0.4 ml) consisted of 0.1 M borate–KOH (pH 9.0),
50 mM 2-oxoglutarate, 3 mM pyridoxamine, and the crude extract
or the enzyme. The reaction was performed at 30 ◦C for 10–30 min
and stopped by the addition of 66 ␮l of 9 M sulfuric acid. The pyri-
TCAGGCGTCGGCGTCGATTAC-3ꢀ. The amplified DNA fragments
were inserted into SmaI site of a plasmid pQE-30 to prepare
pQE6806, pQE6806(V70K), or pQE6806(Y35H/V70K). Second
error-prone PCR with the pQE plasmids as a template, and
pQE-sequencing primer set (QIAGEN) as primers were done by

Y. Yoshikane et al. / Journal of Molecular Catalysis B: Enzymatic 67 (2010) 104–110
107
doxal produced was quantified by the phenylhydrazine method
[13].
The kinetics was done by measuring the initial velocity of the
enzyme reaction based on an initial increase or decrease in A391
or A360 due to pyridoxal produced or consumed, respectively, at
30 ◦C in 1 ml of a reaction mixture consisting of 0.1 M CHES-KOH
(pH 9.0). Kinetic parameters for 2-oxoglutarate and pyruvate with
the purified enzyme were determined by varying the concentra-
tion of 2-oxoglutarate and pyruvate from 0 to 400 mM and from
0 to 20 mM, respectively, at a fixed pyridoxamine concentration
(3 mM). Kinetic parameters for l-glutamate and l-alanine with the
purified enzyme were determined by varying the concentration
of l-glutamate and l-alanine from 0 to 400 mM and from 0 to
250 mM, respectively, at a fixed pyridoxal concentration (0.6 mM).
Kinetic parameters were determined using a curve fitting soft-
ware (Kaleida Graph) to fit the Michaelis–Menten equation with
the Levenberg–Marquardt algorithm.
Fig. 2. (A) SDS-PAGE of some of purified mutated PPATs. Purified proteins (each
3 ␮g) were applied: lane 1, V70K; 2, V70K/E212G; 3, wild-type; 4, M2T/V70K/E212G;
5, Y35H/V70K; 6, M2T/Y35H/V70K; 7, M2T/V70K; 8, M2T/Y35H/V70K/E212G; 9,
V70R/F247C; and 10, Y35H/V70K/F247C. The standard proteins ware applied to lane
Sd. (B) Expression levels of wild-type and mutated PPATs. Crude extracts (10 ␮g)
from E. coli cells expressing the PPATs were applied as shown in (A).
Protein concentration was measured by the dye-binding
method with bovine serum albumin as the standard [15]. The pro-
tein concentration of the purified enzyme was determined by the
molecular absorption coefficient at 280 nm based on the amino acid
composition [16].
2.9. Synthesis of pyridoxamine with the recombinant E. coli cells
expressing the mutated PPATs
chromatography, indicating that all mutated enzymes existed as a
homotetramer. Thus, the mutations did not change the quaternary
structure of PPAT. Table 1 shows kinetic parameters of the mutated
PPATs. The V70K single mutation resulted in 1.7-fold increase in
the catalytic efficiency (kcat/Km) for 2-oxoglutarate. The increase
was caused by 2.2-fold increase in kcat value. In contrast, the Km
value increased 1.3-fold, showing a slight decrease in the affinity
for the substrate. The V70R mutation resulted in an inactivation
of PPAT. Thus, V70K-mutated PPAT was used for an error-prone
PCR to get mutated PPATs with a higher activity towards the
substrates.
In vitro synthesis of pyridoxamine with the E. coli JM109 cells
was done in the reaction mixture (0.1 ml) consisting of 0.1 M
sodium phosphate buffer, pH 7.0, 100 mM pyridoxal, 400 mM l-
glutamate, and the E. coli cells (50 of A660) at 30 ◦C for 48 h. The
concentration of pyridoxal was at a saturated level. The produced
pyridoxamine was determined by an isocratic reversed-phase HPLC
as describedpreviously [13]. Thecontrol reactionwas donewith the
E. coli cells harboring pQE-30 and expressing no PPAT. The amount
of pyridoxamine produced by the mutated PPATs was calculated by
subtracting amount of pyridoxamine in the control reaction.
3.2. Mutated PPATs obtained by the error-prone PCR
2.10. Simulation of binding of pyridoxyl-l-glutamate to
PPAT(Y35H/V70R/F247C)
The first error-prone PCR showed that 1–6 mutations occurred
in the nucleotide sequence, corresponding to an average of 2
amino acid residue substitutions in the newly mutated PPATs.
Out of about 1000 colonies expressing mutated PPATs, about
20 colonies, crude extracts of which showed higher activity of
pyridoxamine–2-oxoglutarate aminotransferase, were selected at
first. From them, two colonies were selected based on the spe-
cific activity and expression level of the mutated PPAT proteins,
which was estimated from the densities on the SDS-PAGE gel. The
genes of the mutated PPAT proteins encoded PPAT(Y35H/V70K)
and PPAT(V70R/F247C). The mutated PPATs were expressed in
JM109, and then purified homogeneously. They showed a sin-
gle band of 41 kDa on SDS-PAGE, and had a molecular weight
of homotetramers under native conditions. PPAT(Y35H/V70K)
showed 10.2-fold higher reaction efficiency for 2-oxoglutarate than
that of PPAT(V70K); Km and kcat values were 8.1-fold lower and
1.3-fold higher than those of PPAT(V70K), respectively (Table 1).
PPAT(V70R/F247C) showed 3.7-fold higher reaction efficiency for
2-oxoglutarate than that of PPAT(V70K) although PPAT(V70R)
showed no activity as described above. The result suggested that
the substitution of Phe247 with Cys caused a proper adaptation of
guanidino group with positive charge of Arg70 to interact with the
␥-carboxyl group of l-glutamate.
The tertiary structure of PPAT(Y35H/V70R/F247C) was deduced
by the MOE Protein Structure Evaluation program (Chemical Com-
puting Group, Montreal, QC, Canada). Docking simulation was
performed for the crystal structure of wild-type PPAT complexed
with pyridoxyl-l-glutamate (PDB code 2Z9X) with the MMFF94x
force field distribution in the MOE 2008.1001 program with the
Monte Carlo docking procedure of ASE-Dock 2005 (Ryoka System,
Tokyo, Japan). In the simulation experiment, 11 solutions were gen-
erated. Solutions were ranked according to total interaction energy,
and the best-matched solution was considered.
2.11. Other methods
The molecular mass of the recombinant enzymes was deter-
mined by
a size-exclusion chromatography with a Cosmosil
5Diol-120-II HPLC column. The standard proteins were used as
described previously [13].
3. Results and discussion
3.1. Mutated PPATs by the rational design
The error-prone PCR with pQE6806(V70K) as the tem-
plate, in which clones were selected based on the production
of pyridoxamine by the intact cells expressing the mutated
enzymes, gave another two mutated PPATs, PPAT(M2T/V70K)
and PPAT(V70K/E212G). Interestingly, they showed only a slight
The purified V70K and V70R enzymes showed a single pro-
tein band on a SDS-PAGE gel, and their molecular weights were
41 kDa (Fig. 2A). The mutated enzymes eluted at the same reten-
tion time of the wild-type PPAT (154 kDa) on the size-exclusion

108
Y. Yoshikane et al. / Journal of Molecular Catalysis B: Enzymatic 67 (2010) 104–110
Fig. 3. Relationship between catalytic turnover numbers for 2-oxoglutarate and
those for pyruvate determined for the wild-type and mutated PPATs. The linear
regression line was drawn by Excel. The R value was −0.96.
increase (1.3- and 1.1-folds, respectively) in the catalytic efficiency
for 2-oxoglutarate (Table 1).
3.3. PPATs mutated at three or four sites
PPAT(M2T/V70K/E212G) showed 1.2-fold lower catalytic
efficiency for 2-oxoglutarate than that of PPAT(M2T/V70K)
(Table 1), indicating that the additive mutation at Glu212 of
PPAT(M2T/V70K) did not synergically increase the catalytic
efficiency. PPAT(M2T/Y35H/V70K) showed 7.3-fold higher and
2.4-fold lower catalytic efficiency for 2-oxoglutarate than that of
the wild-type PPAT and PPAT(Y35H/V70K), respectively. Thus,
the additional mutation of Met2 of PPAT(Y35H/V70K) to Thr
partially retarded the beneficial effect of the double mutation
(Table 1). PPAT(Y35H/V70R/F247C) showed 6.2-fold higher cat-
alytic efficiency for 2-oxoglutarate than that of PPAT(V70R/F247C);
the Tyr35 substitution with His increased the affinity for 2-
oxoglutarate. Thus, PPAT(Y35H/V70R/F247C) showed 39.8-fold
higher catalytic efficiency for 2-oxoglutarate than that of the
wild-type PPAT (Table 1). The catalytic efficiency increased by
synergic effects on the affinity and the catalytic turnover for
2-oxoglutarate.
Relationship between catalytic turnover numbers of 2-
oxoglutarate and pyruvate is shown in Fig. 3. They showed a
negative correlation and a correlation coefficient was −0.96; as the
values of kcat of 2-oxoglutarate increase, those of pyruvate decrease.
The results clearly showed that the in vitro mutation of PPAT suc-
cessfully changed its substrate specificity.
The catalytic efficiency for l-glutamate of mutated PPATs
with the high catalytic efficiency for 2-oxoglutarate was also
determined.
PPAT(Y35H/V70R/F247C),
PPAT(Y35H/V70K),
PPAT(M2T/Y35H/V70K),
PPAT(M2T/Y35H/V70K/E212G),
PPAT(V70R/F247C), and PPAT(V70K) showed 12.7-, 8.3-, 5.8-,
4.8-, 2.8-, and 1.6-fold higher catalytic efficiency for l-glutamate,
respectively. Although the values were not the same as those for
2-oxoglutarate, the mutated PPATs, which had high catalytic effi-
ciencies for 2-oxoglutarate, also showed high catalytic efficiencies
for l-glutamate as well. The results coincided well with the fact
that PPAT reaction is totally reversible [13], and showed that these
mutated PPATs could be used for synthesis of pyridoxamine with
l-glutamate as an amino donor.
3.4. Deduced mechanism of the enhanced catalytic efficiency of
PPAT(Y35H/V70R/F247C)
The PPAT(Y35H/V70R/F247C) showed the highest reaction
efficiency. Docking simulation was performed for pyridoxyl-l-
glutamate to the model of the PPAT(Y35H/V70R/F247C) (Fig. 4).

Y. Yoshikane et al. / Journal of Molecular Catalysis B: Enzymatic 67 (2010) 104–110
109
Table 2
Location of residues mutated and resultant effect.
Original residue (new residue)
Location
Effect of substitution
Met 2 (Thr)
Tyr35(His)
Val 70 (Lys)
Val 70 (Arg)
Glu212 (Gly)
Phe247 (Cys)
Surface
Increase in pyridoxamine production with cells
Increase in reactivity toward 2-oxoglutarate
Increase in reactivity toward 2-oxoglutarate
Elimination of activity
Increase in pyridoxamine production with cells
Reactivation of V70R mutated enzyme
Active site
Active site
Active site
Surface
Hydrophobic core
The ␣-carboxyl group of the glutamate moiety is hydrogen-bonded
to Arg345 and Tyr95. The ␥-carboxyl group of the glutamate moi-
ety is hydrogen-bonded to N␧ of His35 with a distance of 2.9 A
be concluded that almost all pyridoxal was practically converted
into pyridoxamine. PPAT(M2T/Y35H/V70K), PPAT(M2T/V70K),
PPAT(M2T/V70K/E212G), and PPAT(Y35H/V70R/F247C) produced
40, 32, 31, and 27 mM of pyridoxamine, respectively. Interestingly,
the order of the pyridoxamine production from the transformed E.
coli cells did not correlate with the kcat values of the mutated PPATs
expressed in the E. coli cells; the correlation coefficient between the
kcat and the pyridoxamine production was −0.01. The correlation
coefficient between the catalytic efficiency and the pyridoxamine
production was 0.13 showing no correlation between these values.
In contrast, the correlation coefficient between Km and the pyridox-
amine production was −0.49 showing a low correlation between
these values. Because Km values are lower for enzymes with greater
affinity, the mutated PPATs with higher affinity for 2-oxoglutarate
converted pyridoxal into pyridoxamine with the highest efficiency
in the E. coli cells. Thus, the increase in affinity for the substrate used
as a starting material was more effective than the increase in kcat
for the bio-production with the transformed cells. The crowding
environment in the intact bacterial cells could negate the benefi-
cial effect of the increase in the kcat and catalytic efficiency values
[18] by an unknown reason.
˚
rather than to Arg70, which has been predicted to be neces-
sary for binding of the ␥-carboxyl group of the glutamate moiety
(Fig. 1B). The results were consistent with the high affinity of the
PPAT(Y35H/V70R/F247C) for l-glutamate. The model was also in
good agreement with the result that all mutated PPATs with His35
have high affinities towards l-glutamate (Table 1).
The location of residues whose substitution affected the reac-
tivity of PPAT or productivity of pyridoxamine with cells is
summarized in Table 2. Interestingly, the mutation of residues
located on the surface of PPAT increased the productivity of pyri-
doxamine with E. coli cells expressing the mutated PPATs. The
mutation of the residues located in the active site of PPAT affected
the reactivity towards 2-oxoglutarate. The additional mutation of
the residue located in a hydrophobic core reactivated the enzyme
inactivated by the first mutation. The mechanism of the reactiva-
tion needs further studies.
3.5. Synthesis of pyridoxamine with l-glutamate as an amino
donor by the intact E. coli cells expressing the mutated PPATs
4. Conclusion
Our purpose is to develop an industrial synthetic method of
pyridoxamine with the recombinant E. coli cells expressing the
mutated PPATs. Thus, in vitro synthesis of pyridoxamine was exam-
ined. The recombinant cells expressing the mutated PPATs could
produce higher concentrations of pyridoxamine from pyridoxal
and l-glutamate than those expressing the wild-type PPAT as
shown in Table 1; although the expression levels of PPAT proteins
examined by the SDS-PAGE were almost the same as shown in
Fig. 2B. The cells expressing the quadruple mutated PPAT showed
the best production of pyridoxamine, and could change 66 mM
of pyridoxal in the reaction mixture to pyridoxamine in 48 h,
when the production curve reached to a saturation level; because
about 30 mM of pyridoxamine was produced in the control reac-
tion by the well-known non-enzymatic transamination [17], it can
In this study, the rational design and the random mutagene-
sis were used in combination to generate mutated PPATs with the
high activity toward 2-oxoglutarate and l-glutamate. The trans-
formed E. coli cells expressing PPAT(M2T/Y35H/V70K/E212G) could
be used for in vitro conversion of pyridoxal into pyridoxamine with
l-glutamate as an amino donor.
Acknowledgments
This work was supported in part by a Grant-in-Aid for a Research
Fellowship from the Japan Society for the Promotion of Science
for Young Scientists, and by a Kochi University President’s Discre-
tionary Grant.
References
[1] R.A. John, Biochim. Biophys. Acta 1248 (1995) 81–96.
[2] P. Bilski, M.Y. Li, M. Ehrenshaft, M.E. Daub, C.F. Chignell, Photochem. Photobiol.
71 (2000) 129–134.
[3] H. Chen, L. Xiong, Plant J. 44 (2005) 396–408.
[4] S.A. Denslow, A.A. Walls, M.E. Daub, Physiol. Mol. Plant Pathol. 66 (2005)
244–255.
[5] M. Ehrenshaft, P. Bilski, M.Y. Li, C.F. Chignell, M.E. Daub, Proc. Natl. Acad. Sci.
U.S.A. 96 (1999) 9374–9378.
[6] R. Chumnantana, N. Yokochi, T. Yagi, Biochim. Biophys. Acta 1722 (2005)
84–91.
[7] T.O. Metz, N.L. Alderson, S.R. Thorpe, J.W. Baynes, Arch. Biochem. Biophys. 419
(2003) 41–49.
[8] P.A. Voziyan, T.O. Metz, J.W. Baynes, B.G. Hudson, J. Biol. Chem. 277 (2002)
3397–3403.
[9] T.P. Degenhardt, N.L. Alderson, D.D. Arrington, R.J. Beattie, J.M. Basgen, M.W.
Steffes, S.R. Thorpe, J.W. Baynes, Kidney Int. 61 (2002) 939–950.
[10] M.E. Williams, W.K. Bolton, R.G. Khalifah, T.P. Degenhardt, R.J. Schotzinger, J.B.
McGill, Am. J. Nephrol. 27 (2007) 605–614.
[11] M.V. Balyakina, N.L. Yakovleva, V.I. Gunar, Pharm. Chem. J. 14 (1980) 407–
409.
Fig. 4. A model of the active site of a complex of PPAT(Y35H/V70R/F247C) and
pyridoxyl-l-glutamate. The asterisk means a residue of another subunit. The figures
were drawn with WebLab.

110
Y. Yoshikane et al. / Journal of Molecular Catalysis B: Enzymatic 67 (2010) 104–110
[12] R. Khalifah, R. Keilitz, C. Koellner, T.B. Degenhardt, R. Stephen, U.S. Patent
20,070,161,678 (2007).
[13] Y. Yoshikane, N. Yokochi, K. Ohnishi, H. Hayashi, T. Yagi, Biochem. J. 396 (2006)
499–507.
[15] M.M. Bradford, Anal. Biochem. 72 (1976) 248–254.
[16] S.C. Gill, P.H. Hippel, Anal. Biochem. 182 (1989) 319–326.
[17] D.E. Metzler, E.E. Snell, J. Am. Chem. Soc. 74 (1952) 979–983.
[18] J.R. Ellis, Trends Biochem. Sci. 26 (2001) 597–604.
[14] Y. Yoshikane, N. Yokochi, M. Yamasaki, K. Mizutani, K. Ohnishi, B. Mikami, H.
Hayashi, T. Yagi, J. Biol. Chem. 283 (2008) 1120–1127.