Increasing the synthesis/hydrolysis ratio of aminoacylase 1 by site-directed mutagenesis

Article abstract of DOI:10.1016/j.biochi.2009.09.017

Aminoacylase-1 from pig kidney (pAcy1) catalyzes the highly stereoselective acylation of amino acids, a useful conversion for the preparation of optically pure N-acyl-l-amino acids. The kinetic of this thermodynamically controlled conversion is determined by maximal velocities for synthesis (V_mS) and hydrolysis (V_mH) of the N-acyl-l-amino acid. To investigate which parameter affects maximal velocities, we focused on?the proton acceptor potential of the catalytic base, E146, and studied the influence of the active site architecture on its contribution to the pKa of residue E146. The modeled structure of pAcy1 identified residue D346 as having the strongest impact on the electrostatic features of the catalytic base. Substitutions of D346 generally decreased enzymatic activities but also altered both the pH-dependency of hydrolytic activity and the V_mS/V_mH ratio of pAcy1. A reduced theoretical pKa value and a lowered experimental pH optimum of hydrolytic rates for the D346A mutant were associated with a 9-fold increase in V_mS/V_mH. This?supports the importance of electrostatic contributions of D346 to the acid-base properties of E146 and demonstrates for the first time the possibility of engineering the V_mS/V_mH ratio of pAcy1.

Full text of DOI:10.1016/j.biochi.2009.09.017

Biochimie 92 (2010) 102–109
Contents lists available at ScienceDirect
Biochimie
journal homepage: www.elsevier.com/locate/biochi
Research paper
Increasing the synthesis/hydrolysis ratio of aminoacylase 1
by site-directed mutagenesis
Rainer Wardenga , Holger A. Lindner , Frank Hollmann , Oliver Thum , Uwe Bornscheuer a
a
b
c
,
1
c
,
*
a
Institute of Biochemistry, Dept. of Biotechnology & Enzyme Catalysis, Greifswald University, Felix-Hausdorff-Str. 4, D-17487 Greifswald, Germany
Biotechnology Research Institute, National Research Council of Canada, 6100 Royalmount Avenue, Montr e´ al, Qu e´ bec H4P 2R2, Canada
Evonik Goldschmidt GmbH, Goldschmidtstrasse 100, 45127 Essen, Germany
b
c
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 10 July 2009
Accepted 30 September 2009
Available online 9 October 2009
Aminoacylase-1 from pig kidney (pAcy1) catalyzes the highly stereoselective acylation of amino acids,
a useful conversion for the preparation of optically pure N-acyl- -amino acids. The kinetic of this thermo-
dynamically controlled conversion is determined by maximal velocities for synthesis (VmS) and hydrolysis
mH) of the N-acyl- -amino acid. To investigate which parameter affects maximal velocities, we focused
L
(V
L
on the proton acceptor potential of the catalytic base, E146, and studied the influence of the active site
architecture on its contribution to the pKa of residue E146. The modeled structure of pAcy1 identified
residue D346 as having the strongest impact on the electrostatic features of the catalytic base. Substitutions
of D346 generally decreased enzymatic activities but also altered both the pH-dependency of hydrolytic
activity and the VmS/VmH ratio of pAcy1. A reduced theoretical pKa value and a lowered experimental pH
Keywords:
Aminoacylase
Pig kidney
Computer modeling
N-acyl-L-amino acids
optimum of hydrolytic rates for the D346A mutant were associated with a 9-fold increase in VmS/VmH
.
This supports the importance of electrostatic contributions of D346 to the acid-base properties of E146 and
demonstrates for the first time the possibility of engineering the VmS/VmH ratio of pAcy1.
Ó 2009 Elsevier Masson SAS. All rights reserved.
1. Introduction
while maintaining 1.6% of its synthetic activity (ATP-synthesis),
presumably, through alterations in the electrostatic environment of
The use of hydrolytic enzymes as catalysts in organic synthesis
the active site [5]. This demonstrates that engineering of the ratio
of maximal velocities for synthesis and hydrolysis in the absence of
a covalent reaction intermediate is also attainable.
is widespread. Applications of these enzymes in reverse hydrolysis
reactions often relies on their stability in non-aqueous solvents
that favor product formation [1,2]. Random approaches to optimize
hydrolases for such applications, e.g. error-prone PCR, necessitate the
construction of large libraries and thus require efficient selection or
screening strategies to identify clones with desired properties. In the
absence of such strategies, available structure-function information
may guide the improvement of library qualityand hence decrease the
library size. For processes in aqueous systems, optimization or even
reversal of enzymatic synthesis over hydrolysis ratios has been ach-
ieved for reactions involving an acyl-enzyme intermediate using e.g.,
penicillin acylases or glycosyl enzymes by semi-random approaches
N-acyl-L-amino acid amidohydrolases (Aminoacylases, EC 3.5.1.14)
from mammalian tissues [6], plants [7] and microorganisms [8,9]
allow a broad spectrum of applications in biocatalysis. Due to their
exquisite stereoselectivity they are frequently used as catalysts for the
kinetic resolution of racemic N-acyl-D,L-amino acids [10] but also in
synthetic applications by reversal of the hydrolytic reaction [11–13].
The advantage of an aminoacylase-catalyzed process consists of the
possibility to obtain the desired N-acyl-L-amino acid directly in
a single step from the corresponding free -amino acid and acyl donor,
L
respectively. Strategies to facilitate this reaction for certain amino
acid and acyl donor combinations include the use of water miscible
solvents in combination with an excess of one substrate and contin-
uous precipitation of the product [12–14]. Aminoacylase-1 from pig
kidney (pAcy1) displays a marked preference for short-chain acyl
[
3,4]. Yet, optimization of hydrolytic enzymes not forming a covalent
intermediate has remained a considerable challenge. As an example,
mutation of a non-essential histidine to aspartate in the active site of
chicken creatinine kinase abolished the enzyme’s hydrolytic activity
moieties and non-branched aliphatic L-amino acids [15,16] and has
gained considerable attention as chiral catalyst. A number of studies
have reported on the use of pAcy1 in reverse hydrolysis, e.g., in
the synthesis of short-chain and surface-active amphipathic N-acyl-
*
Corresponding author. Tel.: þ 49 3834864367; fax: þ49 3834 86 80066.
E-mail address: uwe.bornscheuer@uni-greifswald.de (U. Bornscheuer).
Present address: Biocatalysis and Organic Chemistry, Delft University of Tech-
1
nology, Department of Biotechnology, Julianalaan 136, 2628 BL Delft, Netherlands.
L-amino acids. The reaction, however, has an unfavorable equilibrium
0300-9084/$ – see front matter Ó 2009 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.biochi.2009.09.017

R. Wardenga et al. / Biochimie 92 (2010) 102–109
103
in water and proceeds at a low catalytic rate [14,17–19]. Synthesis of
N-acetyl- -methionine (A-L-Met) by pAcy1 was shown to be most
efficient at pH 6 while hydrolysis prevails at higher pH [17,20]. Yet,
a decreased concentration of the nucleophilic free amine and thus
reduced overall synthetic rates, as well as low enzyme stability
precluded robust processes at the acidic pH [21].
The equation was implemented in a computable script using
MATLAB 6.5 (MathWorks Inc., Natick MA, USA) and calculations for
equilibrium conversion were performed using published parame-
ters for the synthesis and hydrolysis of N-acetyl-L-alanine [25,26]
with variations in the VmS/VmH ratio, dissociation constants and
L
enzyme concentration.
pAcy1 belongs to the M20 family of the MH clan of co-catalytic
metallopeptidases [22]. The enzyme is
a
homodimer with
2.2. Homology modeling
a molecular mass of 90 kDa [15] and follows a zinc-based, general
base-like catalytic mechanism [23]. As a special feature of the M20
family of homodimeric enzymes, the active site in pAcy1 is situated
The pAcy1 homology model was build with the program
MODELLER 9V4 (http://salilab.org/modeller/) as described previ-
ously by Liu et al. [27] with slight modifications and with additional
analysis of topological errors and energy minimization as described
in Supplementary material.
at the subunit interface [23,24]. N-acyl-L-amino acid binding to the
enzyme places the scissile bond of the substrate in the vicinity of
a zinc-bound water molecule. Nucleophilic attack of this water on
the amide carbonyl group is facilitated through its deprotonation
by the catalytic base E146 leading to formation of a tetrahedral
transition state intermediate. This collapses and releases the free
amino acid and carboxylic acid. The E146-initiated water activation
is thought to be the overall rate-limiting step of pAcy1-catalyzed
hydrolysis [25]. The relation between the enzyme’s backward
and forward maximal velocities, dissociation constants and the
equilibrium constant of the reaction can be described by the Hal-
dane equation [25,26] (Equation (1)). We hypothesized that relative
populations of enzyme substrate complexes are influenced by the
potential of E146 to build the H-bound between the zinc coordi-
nated water hydrogen (expressed by its pKa) and the dissociation
constants of intermediates. Hence, the proton donor/acceptor
potential of the catalytic base may influence the VmS/VmH ratio of
pAcy1 and a decrease in the pKa by mutational modification of its
electrostatic interactions may be a means to increase the VmS/VmH
ratio of pAcy1. In order to achieve this, we constructed a structural
homology model of pAcy1 from which computational pKa predic-
tions identified D346 as a major determinant of E146 basicity and
subjected both residues to site-directed mutagenesis followed by
investigation of the properties of the mutants.
2.3. Web-based computational prediction of pKa values
Estimation of pKa values and significant contributions of
neighboring groups on the protonation state of E146 were extrac-
ted from results of the Hþþ web server [28–30] (http://biophysics.
cs.vt.edu/Hþþ/) using homology models of wild-type and mutant
pAcy1 enzymes. All protonation states were calculated by solving
finite difference Poisson–Boltzmann equation for pH 7.5 using the
following physical conditions: salinity, 0.15; internal dielectric, 6;
external dielectric, 80.
2.4. Recombinant DNA technologies
All routine DNA technologies were performed according to stan-
dard protocols [31]. Restriction enzymes and other DNA-modifying
enzymes were used as specified by the suppliers (New England Biol-
abs, Beverly, MA, USA; Promega, Madison, WI, USA). DNA-sequencing
reactions were carried out at GATC-Biotech (Konstanz, Germany).
Standard protocols were used for the preparation and transformation
of competent E. coli cells [32].
Ac
AH;max$KAc$KL
-
Am
2.5. Site-directed mutagenesis
K
eq
¼
(1)
AS;max$KAc
-L-Am
Site-directed mutagenesis was performed on the expression
vector pET-52(b) containing the pAcy1 gene [33] using the Quik-
ChangeÔ site-directed mutagenesis kit (Stratagene, Heidelberg,
Germany). Forward versions of mutagenic primers are listed in
Table 1. DpnI-treated DNA was transformed into electro-competent
Equilibrium condition of aminoacylase 1 catalyzed reaction (Hal-
dane relation) according to Biselli (1992) [26], see also Supplementary
material, Equation S1. Keq ¼ equilibrium constant, AH,max ¼ Maximum
activity of the hydrolysis of N-acyl-
constant of the enzyme substrate complex with the acyl compound,
L
-amino acid, KAc ¼ Dissociation
E. coli DH5a cells and mutations were confirmed by DNA sequencing.
Ac
K
¼ Dissociation constant of the enzyme (Ac)(L-Am) substrate
L
-
Am
For expression, competent E. coli BL21 (DE3) cells harboring the
chaperone co-expression vector pTf16 (Takara Bio, Otsu, Shiga,
Japan) were transformed with mutated plasmid.
complex with release of amino acid, AS,max ¼ Maximum activity of the
synthesis of N-acyl-
-amino acid, KAc-L-Am ¼ Dissociation constant of
the enzyme substrate complex with N-acyl- -amino acid.
L
L
2
.6. Expression and purification
2
. Material and methods
Unless stated otherwise chemicals were purchased from Fluka
pAcy1 and its variants were expressed in E. coli BL21 (DE3) and
purified as described before [33]. In brief, 50 mL cultures of E. coli
(
(
Buchs, Switzerland), Sigma (Steinheim, Germany) and Merck
Darmstadt, Germany) at the highest purity available. Escherichia coli
BL21 (DE3) (Invitrogen, Carlsbad, USA) [Fꢀ ompT hsdSB(rB ꢀ mB ꢀ)
Table 1
Forward primers.
gal dcm rne131(DE3) was used for protein expression. E. coli
DH5
80lacZ
for cloning.
a
(Clontech, Saint-Germain-en-Laye, France) [supE44
D
lacU169
Mutant
Oligonucleotide sequencea
(F
DM15) hsdR17 recA1 endA1 gyrA96 thi-1relA1] was used
0
0
0
0
0
0
0
0
0
0
E146A
E146D
E146Q
E146N
D346A
D346Q
D346N
D346E
5 -CC TTT GTG CCG GAT GCG GAA GTG GGC GGC-3
5 -CC TTT GTG CCG GAT GAT GAA GTG GGC GGC-3
0
5 -CC TTT GTG CCG GAT CAG GAA GTG GGC GGC CA-3
0
5 -CC TTT GTG CCG GAT AAC GAA GTG GGC GGC-3
2.1. Modeling of pAcy1 catalyzed N-acylation
0
5 -CCG GCG AGC ACC GCG GCG CGT TAT ATT CGT-3
0
5 -CCG GCG AGC ACC CAG GCG CGT TAT ATT CG-3
0
The published kinetic model of aminoacylase 1 [25,26] was used
for theoretical investigation of the influence of rate constants on the
kinetic of the N-acylation (Supplementary material, Equation S1).
5 -CCG GCG AGC ACC AAC GCG CGT TAT ATT CGT-3
5 -CCG GCG AGC ACC GAA GCG CGT TA
a
Altered nucleotides, which introduced the desired mutations, are underlined.

104
R. Wardenga et al. / Biochimie 92 (2010) 102–109
cells were harvested 24 h after induction of pET52(b)þ harboring
a codon-optimized pAcy1 gene. The molecular chaperone trigger
factor was co-expressed from the beginning of the cultivation
from plasmid pTf-16 (Takara Bio, Otsu, Shiga, Japan) by induction
with arabinose (1 mg/ml). Cells were disrupted by sonification, and
the enzyme was purified by StrepTag chromatography (IBA, G o¨ t-
tingen, Germany) according to the manufacturers protocol but
without EDTA containing buffers. Enzyme preparations were stored
microplate reader (Thermo Electron Corporation, USA) and were
evaluated by regression analysis of the linear interval using Origin
Pro 7.5. Due to variations in absolute activities measured for mutants
from individual preparations, pH-dependent activities were normal-
ized to percent of maximum. Resulting standard errors of the mean
were within 3%. Origin Pro 7.5 was used for sigmoidal data fitting.
3. Results
ꢁ
with 50% glycerol at ꢀ20 C. No loss of activity was observed for
activity measurements within 4 weeks.
In order to understand the parameters influencing the VmS/VmH
ratio of pAcy1-catalyzed reactions, we analyzed the contribution
of the active site architecture to the pKa of E146 and applied site-
directed mutagenesis to the most important residues E146 and
D346 as outlined in the next paragraphs.
2.7. Protein assays
Protein concentrations were determined by the BCA-assay
(
Uptima Interchim, France) using BSA as standard. SDS-PAGE was
performed in 12% (w/v) polyacrylamide gel with a stacking gel (4%)
as described by Laemmli to check the purity and the molecular mass
3.1. Modeling of pAcy1 catalyzed N-acylation
[
34]. The proteins in the low-molecular-weight standard mixture
According to the kinetic model of pAcy1 catalysis by Biselli
(1992) [25,26], maximal velocities and the enzymatic VmS/VmH
ratio determines the time needed to reach equilibrium conversion
Ò
Ò
obtained from Roth (Roti -Mark STANDARD) or IBA (Strep-tag
Protein Ladder) were used as reference.
(Supplementary material, Equation S1). Model-based simulations
2
.8. Circular dichroism spectra
suggest that an increase of VmS by a factor of two at constant VmH
can speed up attainment of the reaction equilibrium by approxi-
mately the same time as a doubling of the amount of enzyme
(Supplementary material, Fig. S1).
Buffer exchanges for purified enzymes to 5 mM sodium phos-
phate buffer pH 7.5 were done using NAP-5 columns (GE Healthcare
Bio-Sciences, Freiburg, Germany). The Near-UV CD spectra of pAcy1
variants were recorded at 20 C in a JASCO J-810 circular dichroism
ꢁ
3.2. Homology modeling
spectropolarimeter under nitrogen using 0.06 mg/ml of protein,
a sample cell path length of 2 mm, and response time and band
width of 1 s and 1 nm, respectively. Ten scans between 190 and
Detailed procedures for the building of structural models for
wild-type and mutant pAcy1 are given in Supplementary material.
The procedure was exhaustively validated and homology models
of wild-type pAcy1 were verified for structural integrity. The most
reliable model underwent an energy minimization procedure and
a molecular dynamic simulation for 500 ps. After this, the PROCHECK
overall average G factor, which checks the stereochemical quality and
overall structure geometry, was 0.10 with a score less than ꢀ0.5 indi-
cating a poor model. ERRAT was used for identifying incorrectly folded
regions and the calculated overall quality factor of 92.9, was well
within the range of a high quality model (factor >50). Thus, the
backbone conformation and non-bonded interactions of the homology
model are all within a reasonable range. Using VERIFY-3D for evalua-
tion of compatibility between the amino acid sequence and the
environment of the amino acid side chains, 84.28% of the residues had
a 3D-1D score of >0.2 indicating the good compatibility of each amino
acid residue in the local 3D structure. Individual examinations of
residues with a distance of 8 Å to the catalytic base indicated no
structural errors in the active site (data not shown). Derived structures
of D346N, D346Q, D346E and D346A variants showed only minor
variations in overall and individual residue quality scores.
270 nm were averaged to ensure a good signal-to-noise ratio.
2.9. Aminoacylase assays
pAcy1 variants at concentrations ranging from 0.5 to 1 mg/ml
ꢁ
were incubated at 25 C in a 50 mM sodium phosphate buffer, pH 7.5,
in the presence of 300 mM
L-methionine and 30 mM acetate for
measurements of A- -Met synthesis [17] and of 20 mM A-
L
L-Met for
hydrolysis studies. For controls, enzyme was omitted from these
mixtures. Per reaction, five samples were withdrawn and quenched
by adding 6 N HCl at defined intervals over reaction times ranging
from 5 min to 24 h. A-L-MET in these aliquots was separated and
quantified by reverse phase high performance liquid chromatog-
raphy (HPLC) (MERCK-HITACHI, Tokyo, Japan) on a Kromasil C8
column (4.0 ꢂ 60 mm, particle size 5
Germany) using isocratic elution with 95% H
and 0.01% (v/v) trifluoroacetic acid. The retention time of A-
m
m, SYNTREX GbR, Greifswald,
O, 5% (v/v) acetonitrile
-MET,
2
L
monitored at 214 nm, was 1.1 min at a flow rate of 4.5 ml/min. Initial
rates were determined from peak areas by linear regression using
Origin Pro 7.5 software (Northampton, MA, USA). VmS/VmH ratios
were determined and averaged from three independent preparations
of each variant. One unit (U) of pAcy1 activity was defined as the
3.3. Local electrostatic environment of E146 and pKa predictions
amount of enzyme that hydrolyzes 1
The pH-dependencies of enzymatic hydrolysis were assayed
spectrophotometrically with 3-(2-furyl)acryloyl- -methionine (FA-
Met) as a substrate at 25 C. Chemical synthesis of FA- -Met and the
assay were performed according to Giardina et al. [35]. A standard
reaction mixture of 100 l contained 5 ng of enzyme and 1.2 mM FA-
Met in a 20 mM MES-NaOH buffer. In the investigated pH range of
.5–7.2, FA- -Met showed no significant change in absorption, and the
addition of NaOH affected the ionic strength negligibly with 15 mM
from pH 5.5 to pH 7.2. Measurements were set up in quadruplicate on
6-well plates with 5 l of enzyme (1 mg/ml) and 95 l of pH-adjusted
FA- -Met stock solution (pH 4.5, 5.0, 5.5, 5.8, 6.0, 6.5, 6.7, 6.9, 7.2) per
well. Changes in absorption over time were recorded in a Varioscan
m
mol of A-
L
-Met per minute.
We used our homology model of pAcy1 to study the importance
of the local electrostatic environment on the acid-base properties of
the enzyme’s putative catalytic base residue E146. Hþþ calcula-
tions resulted in a pKaE146 of 5.8. The Hþþ server identified 17 pH-
independent contributions of non-titratable groups (background)
and 10 pH-dependent direct interactions of titratable groups with
E146, four of which involve zinc ligands (Supplementary material,
Table SI) (Fig. 1). Background contributions ranged between ꢀ1.76
and 0.65 with a sum of 4.2 and a mean of ꢀ0.25 (ꢃ0.07). Direct
contributions ranged between ꢀ0.6 and 2.18 with a sum of þ7.5
and a mean of þ0.86 (ꢃ0.27). Although a background contribution
of 1.76 from T345 was the highest above the mean, the strongest
individual direct influence on pKaE146 was predicted for D346
L
L-
ꢁ
L
m
L-
4
L
D
9
m
m
L

R. Wardenga et al. / Biochimie 92 (2010) 102–109
105
Fig. 1. Significant contributions to the pKa of E146 by interactions with titratable neighboring groups in the pAcy1 homology model. The protein backbone is shown as green ribbon,
residue side chains are shown with lines and a surface mesh for residues E146 and D346, respectively, is shown in grey. Zinc-binding residues are each marked with an asterisk.
The model was drawn with PYMOL V1.1 [42]. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.).
(
þ2.18), which is located 4.5 Å from the catalytic base. In the refined
crystal structure of human Acy1 (hAcy1), the conserved D346
D348 in hAcy1) is located 4.6 Å from the catalytic glutamate, also
a VmS/VmH ratio of 37 ꢂ 10ꢀ3. Catalytic activities of all mutants were
significantly decreased but more for variation of E146 than of D346.
In the case of the E146 variants, residual activities at pH 7.5 did not
exceed 0.39 U/mg for hydrolysis (E146D) and 4.6 mU/mg for
synthesis (E146A) but no significant shift in the VmS/VmH ratio was
observed (data not shown).
(
supporting an electrostatic interaction. Electrostatic calculations
for models of the D346N, D346Q, D346E, and D346A mutants
of pAcy1 (Supplementary material, Fig. S2) suggested, that only
the alanine variant reduced the basicity of E146 from a pKa of 5.8 to
Mutation of D346 also significantly reduced the rates of both
5
.1, whereas the asparagine, glutamine, and glutamate variants
A-L-Met synthesis and hydrolysis at pH 7.5 (Fig. 4A). Compared to
increased this value by 0.3, 1.5, and 3.3 units, respectively (Table 2).
wild-type pAcy1 catalytic rates, the D346E and D346N mutants
showed two, and the D346Q and D346A mutants showed three
orders of magnitude reductions in synthesis. While D346E, D346N,
and D346Q displayed only between 10- and 25-fold reductions for
hydrolysis, this activity was reduced even 4000-fold with the D346A
variant. VmS/VmH ratios were thus decreased about 5-, 8-, and 15-fold
for D346E, D346N, and D346Q, respectively, but in contrast showed
an approximately 9-fold increase for D346A (Fig. 4B, Table 2).
Mutants of D346 showed bell-shaped pH-dependencies of
maximum catalytic activities towards FA-L-Met (Fig. 5), with the
exception of D346E which showed a distinct shift to more basic
values. Mutation of D346 to either asparagine or glutamine had only
minor effects on the pH-dependencies, whereas mutation to alanine
produced a narrower pH-profile and shifted the pH optimum of
hydrolysis towards more acidic values by ꢀ1.02 units (Table 2).
3.4. Expression & characterization of pAcy1 variants
Based on the results of the computational analysis, the following
mutants were constructed and expressed in E. coli: E146A, E146D,
E146Q, E146N, D346E, D346A, D346Q, and D346N. All mutants of
pAcy1 generated in this work showed the same expression and
purification behavior as the wild-type enzyme with only minor
differences in protein yield (Fig. 2). CD spectra of pAcy1 and its
mutants suggested the presence of well-folded proteins in all cases
and only a very slight decrease in
a-helix for E146N (Fig. 3). Hydro-
lytic activity of wild-type pAcy1 against A-L-Met was w77 U/mg
and the synthesis activity was quantified to w2.9 U/mg resulting in
Table 2
Summary of the effects of D346 substitutions in pAcy1.
4. Discussion
pAcy1 variant
D
pKa E146a
V
(
mS/VmH ratiob
fold change)
pH optimumc
Acy1 from porcine kidney has been considered as a biocatalyst for
the aqueous one-step synthesis of short-chain and surface-active
amphipathic N-acyl-L-amino acids [14,17–19]. Our simulation of the
pAcy1-catalyzed amino acid N-acylation reaction based on the
kinetic model by Biselli (1992) (Supplementary material, Equation S1
and Fig. S1) suggests that, with constant reaction time and product
yield, a twofold increase in the VmS/VmH ratio obtained by an increase
of VmS at constant VmH potentially affords a twofold reduction in the
amount of enzyme applied. Therefore and in light of the known
D346d
–
–
7.2
6.8
7.2
>7.2
6.18
D346Q
D346N
D346E
D346A
þ1.5
þ0.3
þ3.3
ꢀ0.7
15ꢂ Y
8ꢂ Y
5ꢂ Y
9ꢂ [
a
Calculated values.
b
V
mS/VmH ratios are given for A-
L-Met.
c
pH optima are given for maximal rates of FA-
The theoretical pKa of E146 in the wild-type enzyme is 5.8, and the VmS/VmH
L
-Met hydrolysis (n ¼ 3).
d
ratio of 0.037.
M
limitations of Kcat/K as a measure of catalytic efficiency for synthesis

106
R. Wardenga et al. / Biochimie 92 (2010) 102–109
Fig. 3. CD spectra for recombinant pAcy1 mutants of (A) E146 and (B) D346 variants.
this may be achieved by altering the local electrostatic environment.
In order to map electrostatic contributions in the active site of pAcy1
that determine the pKa of E146, we constructed a homology model
for the catalytic domain of the enzyme based on a published
procedure [27] with additional validation of structural integrity.
Individual quality scores for residues within 8 Å of the catalytic base
supported the accuracy of our model. Using the Hþþ web server, we
calculated from our model a pKa for E146 of 5.8, very close to the
value of 5.9 estimated by R o¨ hm and Van Etten [18] based on pAcy1-
catalyzed O exchange between acetate and water. The Hþþ server
identified electrostatic contributions to the pKa of E146 from both
ionizable and non-ionizable groups in the enzyme active site
(Supplementary material, Table SI) (Fig. 1). Amongst the 27 identi-
fied groups, the side chain of D346 contributed the strongest to
a raise in the intrinsic pKa of the side chain carboxyl group of the
E146 from around 4.4 [38] to a predicted value of 5.8. pKa predictions
for E146 from our structural models of D346 mutants suggested that
conservative substitutions, i.e., to asparagine, glutamine, and gluta-
mate, all raise the pKa of E146 (Table 2). The comparatively high
3.3 unit raise in the pKa of E146 through the neutral asparagine to
glutamate mutation is in agreement with an alkaline proximity
effect of the negative charge at position 346 approaching the base
residue by 1 Å [39]. Accordingly, elimination of the direct contri-
bution of D346 through replacement of this residue by alanine is
predicted to indeed lower the basicity of E146 by ꢀ0.7 pKa units. It
has to be mentioned that mutations of D346 likely influence the
actual pKa of E146 by mechanisms other than changes in local
electrostatics, such as altered protein dynamics [39].
Fig. 2. SDS-PAGE analysis of fractions obtained by strep-tag purification of recombi-
nant pAcy1 variants. M – Marker, wt – Wild-type pAcy1, (A) E146X – pAcy1 mutants of
E146, (B) D346X – pAcy1 mutants of D346.
[36] we focused on determination and optimization of the enzymatic
VmS/VmH ratio which may promote synthetic applications of pAcy1.
We sought to identify residues in pAcy1 that may function as target
sites for a semi-random approach to increase the VmS/VmH ratio of the
enzyme. The enzyme follows a zinc-based, general base-like catalytic
mechanism [23]. Deprotonation of a zinc-activated water by E146
generates the hydrolytic nucleophile and is thought to be the rate-
limiting step of the overall reaction [25].
18
In line with previous mutational studies on human and rat Acy1
[
23,37] mutation of the putative catalytic base residue E146 in pAcy1
to alanine, aspartate, glutamine, or asparagine reduced the specific
activities for both A- -Met hydrolysis and synthesis at least 200- and
30-fold, respectively, and did not cause clear changes in enzymatic
L
6
VmS/VmH ratios (data not shown). With the possible exception of the
E146N variant, highly similar CD spectra (Fig. 3A) indicated properly
folded enzymes for all variants of E146 and suggest that reduced
activities were not due to global changes in enzyme structures. In
accordance with enzymatic microreversibility principles, E146 thus
appears to act as an acid/base catalyst in both N-acyl-
L-amino acid
synthesis and hydrolysis.
We rationalized that a reduction in the basicity of E146 may yield
a pAcy1 variant with potentially an improved VmS/VmH ratio, and that

R. Wardenga et al. / Biochimie 92 (2010) 102–109
107
Fig. 4. N-acetyl-L-methionine (A) synthesis and hydrolysis and (B) VmS/VmH ratios for pAcy1 variants at pH 7.5. Means with lower and upper values are based on three parallels from
independent expression and purification experiments. Gray and open bars represent hydrolytic and synthetic activities, respectively.
Purified asparagine, glutamine, glutamate and alanine mutants
of D346 all exhibited wild-type like CD spectra (Fig. 3B) supporting
the structural integrity of these pAcy1 mutants. The mutations
reduced enzymatic activities at least 10-fold for hydrolysis and
60-fold for synthesis of A-L-Met (Fig. 4A). These reductions were an
order of magnitude lower than with mutations of E146 and are
consistent with an important local electrostatic interaction of D346
with the catalytic base residue E146 as suggested by our structural
model of wild-type pAcy1 and Hþþ calculations based on structural
models of D346 mutants (Table 2). Comparison of our pAcy1 model
with the crystal structure of hAcy1 in complex with an amino acid
ligand (PDB ID 1Q7L) indicated a distance of at least 6.4 Å between
D346 and the substrate binding pocket. This suggests that D346
does not participate directly in substrate-binding. Although we can
not exclude that mutation of this residue affected substrate binding
indirectly, it is reasonable to assume that with a A-
tration of 20 mM and a corresponding half-saturation concentration
Michaelis constant, K ) of 2.72 mM [40] hydrolysis was generally
L-Met concen-
(
M
measured very close to saturating concentrations. During synthesis,
an extremely high concentration for methionine of 300 mM
(10-fold excess over acetate) approached the experimental limit of
the assay but approximated saturation conditions as far as possible.
Therefore, our measured VmS/VmH ratios depend on experimental
conditions. Noteworthly – compared to the wild-type – VmS/VmH
ratios for conservative mutations (D346N, D346Q, and D346E) were
reduced 5- to 15-fold while pKa values for E146 were predicted to be
increased by 0.3–3.3 units (Table 2). Contrarily, a by 0.7 units
reduced theoretical pKa value of E146 for the D346A variant was
associated with a 9-fold increase in the VmS/VmH ratio. These
observations support our initial hypothesis that a reduction in the
basicity of E146 by altering the local electrostatic environment
could improve the enzymatic VmS/VmH ratio.
In order to measure pH-dependencies of pAcy1 activity, we
resorted to a direct assay based on measuring the change in
absorption at 340 nm, which could be multiplexed but has a lower
M
sensitivity. pAcy1 was reported to exhibit a K value at pH 8.7 of
about 0.1 mM with FA- -Met [41]. K values for pAcy1-mediated A-L-
L
M
Met hydrolysis in a MES-NaOH buffer reported by Henseling and
R o¨ hm [22] showed a very weak pH-dependency (variations within
a range of no more than 1 mM between pH 5.5 and pH 8.5). We thus
assumed, that our pH-profiles of FA-L-Met hydrolysis at 1.2. mM of
substrate were not dominated by pH-effects on substrate affinities
and thus largely reflect maximal velocities (Fig. 5). However, the
reduced assay sensitivity and low residual activities of our mutant
enzymes precluded the determinations of enzyme specificity
Fig. 5. Hydrolytic activities of pAcy1 variants towards FA-L-Met in dependence of pH.
One representative out of three independent experiments with separate enzyme
preparations is shown.
M
constants (kcat/K ) against the pH and thus a reliable derivation of
pKa values. Also, batch-to-batch variations rendered a comparison of

108
R. Wardenga et al. / Biochimie 92 (2010) 102–109
absolute enzyme activities imprecise. Hence, normalized data for
representative experiments are shown in Fig. 5. The clear shift of the
pH optimum for the D346E mutant to higher pH-values compared to
the wild-type follows the strong calculated increase in the pKa for
the catalytic base (Table 2). Mutation of D346 to asparagine and
glutamine, respectively, only slightly shifted the pH-profiles to lower
pH-values. The corresponding stronger reductions in the VmS/VmH
ratios (Fig. 4B) for these two variants compared to D346E, especially
for D346Q, were determined by a more pronounced loss in synthetic
activity at pH 7.5 than with D346E (Fig. 4A). The same loss in
synthetic activity as for D346Q was also seen for D346A. However,
the profile for the D346A variant indicates a strong shift of the
mutant’s pH optimum by minus 1.02 units. This shift and a narrower
pH-profile largely account for the mutant’s three orders of magni-
tude loss in hydrolytic activity (Fig. 4A) mainly responsible for the
[3] E.M. Gabor, D.B. Janssen, Increasing the synthetic performance of penicillin
acylase PAS2 bystructure-inspired semi-random mutagenesis. Protein Eng.
Des. Sel 17 (2004) 571–579.
[4] Y. Honda, M. Kitaoka, The first glycosynthase derived from an inverting
glycoside hydrolase. J. Biol. Chem. 281 (2006) 1426–1431.
[
5] M. Forstner, A. M u¨ ller, M. Stolz, T. Wallimann, The active site histidines of
creatine kinase. A critical role of his 61 situated on a flexible loop. Protein Sci.
6
(1997) 331–339.
[6] M. Mitta, I. Kato, S. Tsunasawa, The nucleotide sequence of human amino-
acylase-1. Biochim. Biophys. Acta 1174 (1993) 201–203.
[7] J.C. Lugay, J.N. Aronson, Palo Verde (Parkinsonia aculeata L.) seed amino-
acylase. Biochim. Biophys. Acta 191 (1969) 397–414.
[8] I. Gentzen, H.G. L o¨ ffler, F. Schneider, Aminoacylase from Aspergillus oryzae
comparison with the pig kidney enzyme. Z. Naturf. C 35 (1980) 544–550.
[
9] V. Sakanyan, L. Desmarez, C. Legrain, D. Charlier, I. Mett, A. Kochikyan,
A. Savchenko, A. Boyen, P. Falmagne, A. Pierard, Gene cloning, sequence anal-
ysis, purification, and characterization of a thermostable aminoacylase from
Bacillus stearothermophilus. Appl. Environ. Microbiol. 59 (1993) 3878–3888.
[
10] T. Sato, T. Tosa, Optical resolution of racemic amino acids by aminoacylase.
BioprocessTechnol 16 (1993) 3–14.
[11] M.I. Youshko, F. van Rantwijk, R.A. Sheldon, Enantioselective acylation of
chiral amines catalysed by aminoacylase I. Tetrah. Asymmetry 12 (2001)
9
-fold increase in its VmS/VmH ratio (Fig. 4B).
Our work was motivated by the rational that lowering
3267–3271.
the basicity of residue E146 of pAcy1 could improve the enzyme’s
mS/VmH ratio. Structural homology modeling and computational
[
[
12] K. Yokoigawa, E. Sato, N. Esaki, K. Soda, Enantioselective synthesis of
N-acetyl-l-methionine with aminoacylase in organic solvent. Appl. Micro-
biol. Biotechnol. 42 (1994) 287–289.
13] M. Koreishi, D. Zhang, H. Imanaka, K. Imamura, S. Adachi, R. Matsuno,
K. Nakanishi, A novel acylase from Streptomyces mobaraensis that efficiently
catalyzes hydrolysis/synthesis of capsaicins as well as N-Acyl-L-amino acids
and N-Acyl-peptides. J. Agric. Food Chem. 54 (2006) 72–78.
14] E. Wada, M. Handa, K. Imamura, T. Sakiyama, S. Adachi, R. Matsuno,
K. Nakanishi, Enzymatic synthesis of N-acyl-l-amino acids in a glycerol-
water system using acylase I from pig kidney. J. Am. Oil Chem. Soc. 79
(2002) 41–46.
15] G.J. Palm, K.H. R o¨ hm, Aminoacylase I from porcine kidney: identification and
characterization of two major protein domains. J. Protein Chem. 14 (1995)
V
pKa predictions provided theoretical evidence for a major contri-
bution of D346 to the E146 basicity. All mutants of D346 generated
exhibited only lower synthetic and hydrolytic activities. Despite
good overall structural integrity, this is likely a result of multiple
roles that D346 plays within the active site. For example, beyond
influencing the pKa of E146, additional interactions by D346 may be
important for the catalytic competency of pAcy1 e.g. significant
electrostatic interactions with D173 were also indentified (data not
shown). However, the fact that replacement of the D346 side chain
with a methyl group (D346A) led to the most pronounced of the
measured reductions in hydrolytic activity and thus improved
the corresponding VmS/VmH ratio (Fig. 4B), is in accordance with
a decrease in pKa E146 as calculated from our structural model of
the mutant (Table 2). Experimental pH-dependencies of wild-type
pAcy1 steady-state kinetic parameters were previously shown to be
very small [21]. Despite no direct role of D346 in substrate binding
can be assumed, the experimental data are not appropriate for
deriving pKa values for our mutants. Nevertheless, we believe that
the clear pH-shift for D346A by 1.02 pH units toward a more acidic
value and corresponding sharpened pH transitions strongly argue
in favor of a reduction in the pKa of a catalytic group involved in the
rate-limiting step, most likely E146. To the best of our knowledge,
this is the first reported alteration of the enzymatic VmS/VmH ratio
for an aminoacylase-catalyzed reaction. Still, hurdles in preserving
the absolute synthetic enzyme activity remain to be overcome for
instance by a semi-rational approach to enzyme engineering with
iterative saturation mutagenesis in the neighborhood of D346.
[
[
233–240.
[16] S.M. Birnbaum, L. Levintow, R.B. Kingsley, J.P. Greenstein, Specificity of amino
acid acylases. J. Biol. Chem. 194 (1952) 455–470.
17] A. Ferjancic-Biagini, T. Giardina, M. Reynier, A. Puigserver, Hog kidney and
[
intestine aminoacylase-catalyzed acylation of
media. Biocat. Biotrans 15 (1997) 313–323.
L-methionine in aqueous
[18] K.H. R o¨ hm, R.L. Etten, Catalysis by hog-kidney aminoacylase does not involve
a covalent intermediate. Eur. J. Biochem. 160 (1986) 327–332.
19] I. Galaev, V.K. Svedas, A kinetic study of hog kidney aminoacylase. Biochim.
[
Biophys. Acta 701 (1982) 389–394.
[20] Y. Kimura, Y. Kobayashi, S. Adachi, R. Matsuno, Aminoacylase-Catalyzed
synthesis of N-acyl amino acid from fatty acid or its ethyl ester and amino
acid. in: S. Furusaki, I. Endo, R. Matsuno (Eds.), Biochemical Engineering
2001. Springer, Tokyo, 2001, pp. 109–111.
[21] J. Henseling, K.H. R o¨ hm, Aminoacylase I from hog kidney: anion effects and
the pH dependence of kinetic parameters. Biochim. Biophys. Acta 959 (1988)
370–377.
[
[
22] N.D. Rawlings, F.R. Morton, C.Y. Kok, J. Kong, A.J. Barrett, MEROPS: the
peptidase database. Nucleic Acids Res. 36 (2008) 320–325.
23] H.A. Lindner, V.V. Lunin, A. Alary, R. Hecker, M. Cygler, R. Menard, Essential
roles of zincligation and enzyme dimerization for catalysis in the amino-
acylase-1/M20 family. J. Biol. Chem. (2003) 44496–44504.
[24] H.A. Lindner, A. Alary, L.I. Boju, T. Sulea, R. Menard, Roles of dimerization
domain residues in binding and catalysis by aminoacylase-1. Biochemistry
4
4 (2005) 15645–15651.
[
[
25] K. Drauz, H. Waldmann, Enzyme Catalysis in Organic Synthesis. Wiley-VCH,
Weinheim, 1995.
26] M.F. Biselli, Enzyme-calalyzed racemate separation of amino acids with
Acknowledgement
integrated racemization using the conversion of DL-alanine to
example. Ber. Forschungszent. J u¨ lich, J u¨ lich (1992) 192.
27] Z. Liu, Z. Zhen, Z. Zuo, Y. Wu, A. Liu, Q. Yi, W. Li, Probing the catalytic center of
porcine aminoacylase 1 by site-directed mutagenesis, homology modeling
and substrate docking. J. Biochem 139 (2006) 421–430.
[28] J.C. Gordon, J.B. Myers, T. Folta, V. Shoja, L.S. Heath, A. Onufriev, Hþþ:
a server for estimating pKas and adding missing hydrogens to macromole-
cules. Nucleic Acids Res. 33 (2005) 368–371.
[29] D. Bashford, M. Karplus, pKa’s of ionizable groups in proteins: atomic detail
from a continuum electrostatic model. Biochemistry 29 (1990) 10219–10225.
L-alanine as an
This work was supported by the ‘‘Fachagentur f u¨ r Nachwach-
sende Rohstoffe’’ (FNR, G u¨ lzow, Germany, Grant No. 22009405) and
Evonik Goldschmidt GmbH. We thank Anita Gollin for chemical
[
syntheses of FA-
L-Met. This is NRCC publication no. 50656.
Appendix. Supplementary material
[
[
30] R. Anandakrishnan, A. Onufriev, Analysis of basic clustering algorithms for
numerical estimation of statistical averages in biomolecules. J. Comput. Biol.
15 (2008) 165–184.
31] J. Sambrook, D.W. Russel, Molecular Cloning: a Laboratory Manual. Cold
Spring HarborLaboratory Press, 2001.
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.biochi.2009.09.017.
References
[32] C.T. Chung, S.L. Niemela, R.H. Miller, One-step preparation of competent
Escherichia coli: transformation and storage of bacterial cells in the same
solution. Proc. Natl. Acad. Sci. U.S.A 86 (1989) 2172–2175.
[33] R. Wardenga, F. Hollmann, O. Thum, U. Bornscheuer, Functional expression of
porcineaminoacylase 1 in E. coli using a codon optimized synthetic gene and
molecular chaperones. Appl. Microbiol. Biotechnol. 81 (2008) 721–729.
[
[
1] A.M. Klibanov, Enzymatic catalysis in anhydrous organic solvents. Trends
Biochem. Sci. 14 (1989) 141–144.
2] V.A.M.P. Koskinen, A.M. Klibanov, Enzymatic Reactions in Organic Media.
Blackie, London, 1996.

R. Wardenga et al. / Biochimie 92 (2010) 102–109
109
[
[
34] U.K. Laemmli, Cleavage of structural proteins during the assembly of the
head of bacteriophage T4. Nature 227 (1970) 680–685.
35] T. Giardina, A. Biagini, F. Dalle Ore, E. Ferre, M. Reynier, A. Puigserver, The hog
intestinalmucosa acylase I: subcellular localization, isolation, kinetic studies
and biological function. Biochimie 79 (1997) 265–273.
[38] M. Tollinger, J.D. Forman-Kay, L.E. Kay, Measurement of side-chain carboxyl
pKa values of glutamate and aspartate residues in an unfolded protein by
multinuclear NMR spectroscopy. J. Am. Chem. Soc. 124 (2002) 5714–5717.
[39] J.E. Nielsen, T.V. Borchert, G. Vriend, The determinants of {alpha}-amylase
pH-activity profiles. Protein Eng. 14 (2001) 505–512.
[
[
36] R.J. Fox, M.D. Clay, Catalytic effectiveness, a measure of enzyme proficiency
for industrial applications. Trends Biotechnol. 27 (2009) 137–140.
37] A. Durand, T. Giardina, C. Villard, A. Roussel, A. Puigserver, J. Perrier, Rat
kidney acylase I: further characterisation and mutation studies on the
involvement of Glu 147 in the catalytic process. Biochimie 85 (2003)
[40] H.A. Lindner, M. T a¨ fler-Naumann, K.-H. R o¨ hm, N-acetylamino acid utilization
by kidney aminoacylase-1. Biochimie 90 (2008) 773–780.
[41] D. Heese, S. Berger, K.H. R o¨ hm, Nuclear magnetic relaxation studies of the
role of the metalion in Mn2(þ)-substituted aminoacylase I. Eur. J. Biochem.
188 (1990) 175–180.
9
53–962.
[42] W.L. Delano, The PyMOL molecular graphics system. DeLano Scientific (2002).