Structure-based design and application of an engineered glutathione transferase for the development of an optical biosensor for pesticides determination

Article abstract of DOI:10.1016/j.bbagen.2018.12.004

In the present work, a structure-based design approach was used for the generation of a novel variant of synthetic glutathione transferase (PvGmGSTU) with higher sensitivity towards pesticides. Molecular modelling studies revealed Phe117 as a key residue

Full text of DOI:10.1016/j.bbagen.2018.12.004

BBA - General Subjects 1863 (2019) 565–576
Contents lists available at ScienceDirect
BBA - General Subjects
journal homepage: www.elsevier.com/locate/bbagen
Structure-based design and application of an engineered glutathione
transferase for the development of an optical biosensor for pesticides
determination
a
b
c
d
Evangelia G. Chronopoulou , Dimitrios Vlachakis , Anastassios C. Papageorgiou , Farid S. Ataya ,
Nikolaos E. Labrou
a
a,⁎
Laboratory of Enzyme Technology, Department of Agricultural Biotechnology, Agricultural University of Athens, 75 Iera Odos Street, GR-11855 Athens, Greece
Laboratory of Genetics, Department of Biotechnology, School of Food, Biotechnology and Development, Agricultural University of Athens, 75 Iera Odos Street, GR-11855
b
Athens, Greece
c
Turku Centre for Biotechnology, University of Turku, Åbo Akademi University, Turku 20521, Finland
Department of Biochemistry, College of Science, King Saud University, P.O. Box 2455, Riyadh 11451, Saudi Arabia
d
A R T I C L E I N F O
A B S T R A C T
Keywords:
In the present work, a structure-based design approach was used for the generation of a novel variant of syn-
thetic glutathione transferase (PvGmGSTU) with higher sensitivity towards pesticides. Molecular modelling
studies revealed Phe117 as a key residue that contributes to the formation of the hydrophobic binding site (H-
site) and modulates the affinity of the enzyme towards xenobiotic compounds. Site-saturation mutagenesis of
position Phe117 created a library of PvGmGSTU variants with altered kinetic and binding properties. Screening
of the library against twenty-five different pesticides, showed that the mutant enzyme Phe117Ile displays 3-fold
higher catalytic efficiency and exhibits increased affinity towards α-endosulfan, compared to the wild-type
enzyme. Based on these catalytic features the mutant enzyme Phe117Ile was explored for the development of an
optical biosensor for α-endosulfan. The enzyme was entrapped in alkosixylane sol-gel system in the presence of
two pH indicators (bromocresol purple and phenol red). The sensing signal was based on the inhibition of the
Glutathione transferase
Protein engineering
Biosensor
Pesticides determination
α-Endosulfan
+
sol-gel entrapped GST, with subsequent decrease of released [H ] by the catalytic reaction, measured by sol–gel
entrapped indicators. The assay response at 562 nm was linear in the range pH = 4–7. Linear calibration curves
were obtained for α-endosulfan in the range of 0–30 μΜ. The reproducibility of the assay response, expressed by
relative standard deviation, was in the order of 4.1% (N = 28). The method was successfully applied to the
determination of α-endosulfan in real water samples without sample preparation steps.
1. Introduction
adsorption, cross-linking, covalent bond formation, entrapment or af-
finity capture [5,9,10]. In the entrapment method, the enzyme is re-
Over the last three decades there has been an intense research in-
stricted within a three-dimensional structure of a matrix such as silicate
materials using the sol-gel process, electropolymerized films, poly-
saccharide or carbon paste [5,11,12]. This method provides simplicity
and stabilization of enzyme activity. However, leaching of enzyme and
possible diffusion barriers are the main disadvantages [5]. The sol-gel
process involves hydrolysis of alkoxide precursors under acidic condi-
tions followed by condensation of the hydroxylated units, which leads
to the formation of a porous gel [5,11–13]. The alkoxide precursor
molecules are hydrolyzed in the presence of water at low pH, resulting
terest in the field of enzyme biosensors. These analytical tools, except of
being accurate, environmentally benign and safe, they can provide real-
time monitoring with minimum sample preparation in a simple and
efficient manner [1–5]. Enzymatic biosensing methods based on en-
zyme activity or inhibition have been widely used [6]. Some examples
include the enzymes cholinesterases, organophosphate hydrolase, al-
kaline and acid phosphatase, ascorbate oxidase, lipase etc. [7,8].
In enzyme-based biosensors the enzymes are immobilized by
Abbreviations: ANNs, Artificial Neural Networks; G-site, glutathione-binding site; GSH, glutathione; GST, glutathione transferase; H-site, hydrophobic-binding site;
MD, molecular dynamics; PTMOS, phenyltrimethoxysilane; TEOS, tetraethyl orthosilicate; PEG, polyethylene glycol; PvGmGSTUG, Synthetic glutathione transferase
from Phaseolus vulgaris and Glycine max; RDS, relative standard deviation
⁎
Corresponding author.
E-mail address: lambrou@aua.gr (N.E. Labrou).
https://doi.org/10.1016/j.bbagen.2018.12.004
Received 30 July 2018; Received in revised form 8 November 2018; Accepted 4 December 2018
Available online 24 December 2018
0304-4165/ © 2018 Elsevier B.V. All rights reserved.

E.G. Chronopoulou et al.
B
B
A- General Subjects 1863 (2019) 565–576
in the formation of silanol (SieOH) groups. In the second step, the
condensation reaction between silanol moieties at alkaline (or acidic)
pH results in the formation of siloxane (SieOeSi) polymers, generating
a matrix in which an enzyme can be entrapped [14,15].
Glutathione transferases (GSTs, EC 2.5.1.18) comprise a family of
enzymes that are involved in the detoxification mechanism of en-
dogenous and xenobiotic electrophile compounds [16–20]. They cata-
lyze the nucleophilic attack of reduced glutathione (GSH) on the elec-
trophilic centre of xenobiotic compounds [21]. Different GST
isoenzymes have already found successful applications in the develop-
ment of enzyme biosensors for the determination of herbicides and
insecticides [3,9,17,22–26].
Enzyme engineering is the process of improving the catalytic,
functional or structural features of an enzyme by modifying its amino
acid sequence [27]. The goal of this technology is to improve or over-
come the potential disadvantages of native enzymes (e.g. low stability,
low specificity and catalytic activity), aiming at maximizing the bio-
catalytic applications of the enzymes [28]. A number of interesting
engineering studies can be found in the literature [29–32]. Rational
design using computational tools and directed evolution are the two
main approaches in enzyme engineering [33–35]. However, the com-
bination of directed evolution and rational protein design is becoming
increasingly useful and effective [36]. The application of enzyme en-
gineering for the development of tailor-made enzymes with improved
properties for application in biosensor technology is gaining particular
interest over the last years [28]. Enzyme mutants with higher sensi-
tivity can improve biosensor's analytical performance such as durability
and sensitivity [8,24].
The present work, through the combination of directed evolution
and rational protein design approach, aims at the development of an
optimized GST mutant for the creation of an optical GST-based bio-
sensor for α-endosulfan. Endosulfan is a toxic insecticide which is
considered as a major Persistent Organic Pollutant (POP). It has been
detected in a variety of environmental samples across the world [37].
The development, therefore, of a new method for direct determination
of endosulfan in environmental samples (e.g. water) has both scientific
interest and practical importance.
template for site-saturation mutagenesis at amino acid position 117,
using the quick-change method [23]. The mutations were introduced
using a set of degenerate synthetic oligonucleotides, in which the mu-
tation site was diversified using a randomized NNN codon. The pairs of
oligonucleotide primers used in the PCR reactions for the saturation
mutagenesis were as follows:
FPrimer117 5′ AAA GCT ACT NNN TCT ATT GAT 3′
RPrimer117 5′ ATC AAT AGA NNN AGT AGC TTT 3′
The PCR contained: 8 pmol of each primer, 5× Kapa High Fidelity
DNA polymerase's buffer, 100 μΜ each dNTP, 5 ng of plasmid DNA-
shuffling and 1 U Kapa HiFi DNA polymerase. The PCR comprised of
30 cycles of denaturation at 94 °C for 2 min, annealing at 40 °C for 2 min
and polymerization at 72 °C for 2 min. A final extension time at 72 °C
for 10 min was performed, after the 30th cycle. After completion of the
PCR, the reaction product was subjected to DpnI digestion. Following
digestion, the mutated plasmids DNA (pEXP5-CT/TOPO/TA) were used
to transform competent E. coli TOP10 cells. Transformed cells were
selected by LB agar plate containing ampicillin (100 μg/mL). The site-
saturation library was screened by measuring the enzyme activities-
using CDNB/GSH substrate system. Transformants were grown at 37 °C
in LB medium (10 mL) containing ampicillin (100 μg/mL).
3.3. Expression and purification of the wild-type and mutants' enzymes
E. coli cells (BL21(DE3) or BL21(DE3)pLysS), harbouring the re-
combinant plasmid pT7PvGmGSTUG were grown at 37 °C in LB medium
containing ampicillin (100 μg/mL) and chloramphenicol (34 μg/mL)
(for the BL21(DE3)pLysS strain). Protein expression was performed as
described by Chronopoulou et al. [39]. Enzyme purification was carried
out using affinity chromatography as described by Chronopoulou et al.
[39]. Protein purity was judged by SDS PAGE.
3.4. Assay of enzyme activity and kinetic analysis
GST assays were performed by measuring the reaction rate (37 °C)
between CDNB (1 mM) and GSH (2.5 mM) at 340 nm
(ε = 9.6 mM ·cm−1) as previously described [17]. Steady-state ki-
netic analysis were carried out as described by Chronopoulou et al.
[17]. Curve-fits were obtained using the GrafPad (GraphPad Software
Inc., Version 7.00) computer program.
−1
2. Materials and methods
2.1. Materials
3
.5. Inhibition analysis and screening of the wild type and mutants
All enzyme substrates, antibiotics, tetraethyl orthosilicate (TEOS),
PvGmGSTUGs
phenyltrimethoxysilane (PTMOS) and polyethylene glycol (PEG) were
purchased from Sigma-Aldrich, (USA). Ampicillin and chloramphenicol
were purchased from Sigma-Aldrich, (USA). KAPA Taq and KAPA High
fidelity DNA polymerases were purchased from KAPA Biosystems
GST inhibition analysis was performed using the CDNB/GSH
system, as described above, in the presence or absence of 100 μM pes-
ticide diluted in acetone. During the course of the assay (30–60 s) no
measurable pesticide/GSH conjugation was observed. The IC50 values
of α-endosulfan were measured using the CDNB/GSH assay system in
the presence of different concentrations of α-endosulfan (0–100 μΜ),
respectively. The IC50 values were determined by fitting the con-
centration–response data to the following Eq. (2):
(
(
USA). Plasmid purification kit was obtained by Macherey–Nagel,
Germany) and QIAquickTM Gel Extraction kit from Quiage (Germany).
The pesticides: fenvalerate, permethrin, diazinon, malathion, carbaryl,
atrazine, diuron, fluorodifen, alachlor, metolachlor, dichlorvos,
omethoate, λ-cypermethrin, dieldrin, spirodiclofen, α-cyhalothrin,
spinosad, deltamethrin, aldrin, spiromesifen, thiacloprid, pirimicarb,
methomyl, chlorpyriphos, endosulfan, carbofuran and fluazifop-p-butyl
were purchased from Riedel de Haen (Germany).
1
00
%inhibition =
1
+ (IC50/[I])
(2)
where [I] is the pesticide concentration. The IC50 values were de-
termined using the program GraphPad Prism version 7.00.
3. Methods
3.1. Protein determination
3.6. Biocomputing analysis
Protein concentration was determined by the Bradford assay using
BSA (fraction V) as a standard [38].
3.6.1. Sequence database search and phylogenetic tree construction
In order to identify homologous PvGmGSTUG protein sequences, the
non-redundant publicly available databases: UniProtKB [40] and Gen-
Bank [41] were searched with the entire amino acid sequences of
PvGmGSTUG applying reciprocal BLASTp and tBLASTn [42]. The entire
PvGmGSTUG amino acid sequence was searched against PROSITE [43],
3
.2. Site-saturation mutagenesis
The expression construct pT7PvGmGSTUG [39] was used as
566

E.G. Chronopoulou et al.
B
B
A- General Subjects 1863 (2019) 565–576
in order to identify protein domains. The retrieved protein sequences
were aligned using CLUSTALW [44]. The resulting multiple sequence
alignment was trimmed by applying Gblocks [45] with default options,
and subsequently was used to reconstruct a phylogenetic tree by em-
ploying the neighbor-joining method [46]. The number of amino acid
substitutions per site was estimated using the JTT model [47].
and left at 4 °C for 24 h. The same procedure was followed for enzyme
by keeping the sol-gel mixture for 1 h at 4 °C. The enzyme (0.95–2.4 U/
mL) was mixed with polyethylene glycol N-hydroxy-hydro-
xysuccinimide ester (1:1 w/v) and left for 1 h at the same temperature,
before mixing. Then, the mixture (200–500 μL) was placed in a glass
cuvette and 30–60 μL KOH (1 M) was added for xerogel formation.
“
Aging” of xerogel was achieved using TEOS (2 or 5 days). The sol-gel
was stored in potassium phosphate buffer (1 mM, pH = 7).
3.6.2. Molecular docking
To in silico establish the complex structures of the PvGmGSTUG and
α-endosulfan or GSH, the docking suite ZDOCK (version 3.0) was used
3.8. Determination of α-endosulfan in natural water samples
[
48]. Docking experiments were conducted on the models that had
been energetically minimized and conformationally optimized using
molecular dynamics simulations. ZDOCK uses a scoring function that
returns electrostatic, hydrophobic and desolvation energies as well as
performing a fast pairwise shape complementarity evaluation. RDOCK
was utilized to refine and quickly evaluate the results obtained by
ZDOCK [48]. RDOCK performs a fast minimization step to the ZDOCK
molecular complex outputs and re-ranks them according to their re-
calculated binding free energies.
Recovery experiments were carried out using drinking water (col-
lected from Athens water supply network) and mineral water samples
(Korpi, NESTLE Hellas), spiked with known amounts (0.8–16 μM) of α-
endosulfan. Spectroscopic measurements were carried out in the region
700–300 nm in potassium phosphate buffer (1 mM, pH = 6.5). For
standard curve creation, the absorbance at 562 nm was measured
(3 min response time for color development).
3.9. Stability analysis of the immobilized enzyme
3.6.3. Energy minimization and molecular dynamics simulations
Energy minimizations were used to remove any residual geometrical
Stability analysis of the wild-type and its Phe117Ile mutant (free
strain in each molecular system, using the Charmm27 forcefield as it is
implemented into the Gromacs suite, version 4.5.5 [49]. All Gromacs-
related simulations were performed though our previously developed
graphical interface [50]. Molecular systems were then subjected to
unrestrained Molecular Dynamics Simulations (MDS) using the Gro-
macs suite, version 4.5.5 [49]. MDS took place in a SPC water-solvated,
periodic environment. Water molecules were added using the truncated
octahedron box extending 7 Å from each atom. Molecular systems were
neutralized with counter-ions as required. For the purposes of this study
all MDS were performed using the NVT ensemble in a canonical en-
vironment, at 300 K and a step size equal to 2 femtoseconds for a total
and sol-gel entrapped) at 4 °C was monitored by periodically mea-
surement of enzyme activity in 1 mM potassium phosphate buffer
pH 6.5 using the CDNB/GSH system [17].
4. Results and discussion
4.1. The interaction of PvGmGSTUG with pesticides
GSTs that belong to the tau class contain a deep binding cleft, which
has been evolved to accommodate a broad range of hydrophobic sub-
strates and non-substrate ligands [18,20]. Non-substrate ligands usually
bind to the L-site (ligandin-binding site), which is located in a distinct
region or overlaps with the G- and H-sites [51,52]. The binding of these
compounds to the L-site of GSTs affects the binding of the normal
substrates and therefore inhibits the enzyme's catalytic activity [53]. In
the present work, the interaction of a wide range of non-substrate li-
gands (pesticides) with PvGmGSTUG was studied to characterize en-
zyme's binding selectivity towards xenobiotics.
100 ns simulation time. An NVT ensemble requires that the Number of
atoms, Volume and Temperature remain constant throughout the si-
mulation.
3.7. Entrapment of enzyme and pH indicators in sol-gel
Sol–gel formation was carried out as previously described [9]. The
hybrid sol-gel contained TEOS and PTMOS in a 2:1 M ratio. In a glass
vial, TEOS (1.10 mL), PTMOS (0.45 mL), CTAB (82 mM), ethanol
PvGmGSTUG is a synthetic isoenzyme that was recently designed
and created in our lab [39] by DNA shuffling of abiotic stress-inducible
GST genes from Phaseolus vulgaris and Glycine max. The interaction was
investigated by measuring the inhibition of the enzyme's catalytic ac-
tivity. As shown in Fig. 1, PvGmGSTUG displays a wide binding speci-
ficity and interacts with a wide range of pesticides. Among all tested
compounds, the most potent inhibitors were selected for screening with
mutants.
(
0.50 mL), water (0.70 mL) and 0.1 M HCl (0.35 mL), were added. The
vial was capped and placed in an ultrasonic bath for 10 min. After so-
nication, the clear solution was left at room temperature, for approxi-
mately 15 h. One mL of the sol-gel was withdrawn and added to a vial
containing 3.0 mg of bromocresol purple and phenol red (1:1 ratio)
indicator. Then, 20 μL of the mixture were placed on a microscope slide
Fig. 1. Inhibition of the wild-type enzyme and the mutants Phe117Ile, Phe117Gly, Phe117His, Phe117Trp, Phe117Ser by pesticides (100 μΜ). In the absence of
pesticides, enzyme activity was taken as 100%. Experiments were performed in triplicate using GSH-CDNB as substrate system.
567

E.G. Chronopoulou et al.
B
B
A- General Subjects 1863 (2019) 565–576
Table 1
Steady-state kinetic analysis of PvGmGSTUG and its mutants for the CDNB/GSH
substrate system using as a variable substrate GSH. The CDNB was used at
saturated concentration.
Enzyme
K
m
(mM) (GSH)
k
(
cat (min−1)
GSH)
kcat/K
m
−
(mΜ min−1
1
)
(
GSH)
Wild-type
1.17 ± 0.090
194.1 ± 4.85
165.89 ± 16.90
(PvGmGSTUG)
Phe117Ser
Phe117Gly
Phe117His
Phe117Trp
Phe117Ile
0.15 ± 0.013
0.89 ± 0.09
0.38 ± 0.038
0.57 ± 0.06
0.22 ± 0.024
75.65 ± 1.30
60.16 ± 2.49
69.44 ± 1.76
194 ± 6.13
511 ± 53.06
67.60 ± 9.63
182.73 ± 22.90
340.35 ± 46.58
530.0 ± 70.54
116.6 ± 2.80
Fig. 2. Dose-response inhibition effect of α-endosulfan on the wild-type
PvGmGSTUG (▲) (IC50 value: 9.84 ± 0.61 μΜ) and mutant Phe117Ile (●)
(
IC50 value: 5.17 ± 0.47 μΜ) enzymes. GSTs were assayed using the
4
.2. Structural determinants that affect pesticide binding to PvGmGST and
CDNB–GSH assay system and the experiments were performed in triplicate.
design of H-site mutants: the role of Phe117
Enzyme-based biosensors that rely on the inhibition of the enzyme
activity by xenobiotic compounds, allow their measurement based on
the determination of the inhibition of the substrate turnover rates
key residue to the GSH coordination and catalysis (Fig. 4C-D). Any
change in position 117 will trigger a cascade that will affect sig-
nificantly the substrate binding and catalysis via the Asp105 →
Arg21 → Phe18 route.
[
5,21,26]. Although the enzyme-based biosensors are in general ap-
propriate and efficient, their sensitivity and specificity depend on the
enzyme's itself catalytic properties [24]. A suitable enzyme that re-
cognize with high affinity the target compound is a requirement [54].
Protein engineering is, therefore, a powerful strategy for the design of
tailor made enzymes with appropriate improvements for the develop-
ment of such analytical applications [55].
Phe117 was subjected to site-saturation mutagenesis aiming at
creating a new enzyme variant with improved catalytic and binding
properties towards α-endosulfan. Following site-saturation mutagen-
esis, the coding sequences of the mutant enzymes were cloned, ex-
pressed in E. coli BL21(DE3) and the library of the mutated enzymes
were screened using activity assays [56] for selecting mutant enzymes
with high inhibition potency towards α-endosulfan. Five clones dis-
playing the desired characteristics were selected for further study. Se-
quencing analysis verified that the five clones contained a single point
mutation at position 117. The substitutions were amino acids with wide
diversity in their physicochemical properties: Ile, His, Trp, Gly and Ser.
Analysis of the molecular model of PvGmGSTUG revealed that
Phe117 seems to be in a rather remote end of the PvGmGSTUG structure
without too much involvement in the overall fold of the enzyme.
However, careful study and exhaustive molecular dynamics simulation
prove that Phe117 is a key residue and plays a central role in binding
and catalysis. Phe117 is strategically located immediately after helix H4
and just before helix H5. It establishes interactions with a set of posi-
tively charged residues nearby. Namely, hydrogen bonds and π-stacking
interactions are formed between Phe117, the preceding Arg113 and the
following Lys122 and Arg124 residues (Fig. 3) That interaction ar-
rangement is pivotal and vital for the spatial arrangement of the H4
helix. Disruption of these interactions results in a rather significant
change in the relative angle between helices H4 and H6 with a poten-
tially profound effect to the H-site. The evolutionary study revealed that
the closest GST structure bearing GSH is the 2CA8 PDB file (Suppl.
Fig. 1). The two structures were superimposed and the GSH coordinates
were copied to the PvGmGSTUG molecular system. The system was
energetically optimized and subjected to molecular dynamics simula-
tions. The interaction pattern revealed a stabilizing mechanism for the
PvGmGSTUG enzyme that is highly dependent on position 117. In more
detail, as mentioned above, the H4 helix is spatially coordinated by the
loop bearing the 117 position. The H4 helix is stabilized by hydrogen
bonding of Asp105 in the midst of the helix to Arg21 (Fig. 4A-B).
Consequently, Arg21 hydrogen bonds and stabilizes Phe18, which is a
4.3. Kinetic analysis of site-saturation enzyme variants
The wild-type PvGmGSTUG enzyme and site-saturations variants
were purified by affinity chromatography and subjected to steady-state
kinetic analysis and the results are shown in Fig. 5 and the kinetic
parameters in Table 1 and Table 2. Kinetic analysis using GSH as
variable substrate (Table 1) showed that all site-saturation variants
obey Michaelis-Menten kinetics and display improved affinity for GSH,
GSH
compared to the wild-type enzyme. The improvement in K
m
value is
higher for the mutants Phe117Ile and Phe117Ser (~5-fold) which is
GSH
translated to 3-fold improvement in catalytic efficiency (kcat/Κ
m
)
(
Table 1). Kinetic analysis using CDNB as variable substrate (Fig. 5,
Table 2), showed that all but Phe117Ser mutant do not obey Michaelis-
Menten kinetics, but rather a sigmoid dependence on substrate con-
centration with Hill coefficient > 1, suggesting
a positive co-
operativity between the two H-sites.
The role of Phe117 was further investigated by in silico mutagenesis
experiments. All mutants were modelled on computer-based
a
Table 2
Steady-state kinetic analysis of PvGmGSTUG and its mutants for the CDNB/GSH substrate system using as a variable substrate CDNB. The GSH was used at saturated
concentration.
cat (min−1) (CDNB)
n
H
(CDNB)
k
m
cat/K )
(mΜ min−1
−1
Enzyme
S
0.5 (mM) (CDNB)
k
Wild-type (PvGmGSTUG)
Phe117Ser
Phe117Gly
Phe117His
Phe117Trp
0.88 ± 0.05
2.59 ± 0.31a
0.94 ± 0.057
0.92 ± 0.07
0.68 ± 0.05
0.71 ± 0.07
–
1.77 ± 0.14
–
2.30 ± 0.23
1.50 ± 0.11
2.09 ± 0.23
1.33 ± 0.13
–
97.98 ± 7.15
37.83 ± 7.28b
–
–
–
–
Phe117Ile
a
m
The mutant Phe117Ser enzyme obeys Michaelis-Menden kinetics and this value should be considered as K (mM).
The mutant Phe117Ser enzyme obeys Michaelis-Menden kinetics and this value should be considered as kcat/K
b
(mΜ−1 min−1).
m
568

E.G. Chronopoulou et al.
B
B
A- General Subjects 1863 (2019) 565–576
Fig. 3. Study of the structural significance of Phe117 on the fold and function of the PvGmGSTUG enzyme. Left: The 3D structure of PvGmGSTUG with the Phe117
residue highlighted in turquoise color. The key interacting residues are shown in stick representation. Right: The 2D interaction diagram for the residue Phe117.
Fig. 4. The structural mechanism that elaborates how the Phe117 residue can influence the catalytic potential of the PvGmGSTUG enzyme. A: The structure of the
PvGmGSTUG enzyme with key residues highlighted in stick representation. B: Zoom-in of A. C: The 2D interaction diagram for the residue Arg21. D: 2D interaction
diagram for the residue Phe18.
molecular system and were subjected to unrestrained, explicitly sol-
vated molecular dynamics (MDs) simulations. The results of the mole-
cular modelling study coincide with the experimental results (Tables 1
and 2) and solidify the proposed mechanism described herein. The
wild-type enzyme and the Phe117His mutant behave quite similarly as
the side chains of the Phe and His residue are capable of establishing
the same interactions. The imidazole ring of His residue can maintain
the interactions of the original benzene ring of the Phe of the wild-type
569

E.G. Chronopoulou et al.
B
B
A- General Subjects 1863 (2019) 565–576
Fig. 5. Kinetic analysis of the wild-type PvGmSTUG and its mutants (Phe117Ile, Phe117Gly, Phe117His, Phe117Trp and Phe117Ser), using the GSH as a variable
substrate and CDNB at a fixed concentration (left graphs) or using the CDNB as a variable substrate and GSH a fixed concentration (right graphs). A: wild-type
enzyme PvGmSTUG, B: mutant Phe117Ile, C: mutant Phe117Gly, D: mutant Phe117His, E: mutant Phe117Trp and F: mutant Phe117Ser.
enzyme. Consequently, the structure of Phe117His mutant remains
unchanged upon MDs (Fig. 6C). The structure of Phe117Gly mutant is
the only one of the five variants that resulted in a relative shift of the H4
helix towards the catalytic site, thus via the Asp105 to Arg21 interac-
tion pushing Phe18 into the conformational space of the catalytic site
and restraining and limiting the spatial flexibility of the ligand
slightly shifted away from the catalytic site. This structural effect has
consequently dragged Arg21 towards Asp105 and has allowed more
space to Phe18. This translates in an increase of the overall volume of
the catalytic site. The increase in volume is most observed in the
Phe117Ile mutant, with the Phe117Ser to follow and the Phe117Trp
mutant with the least increase in the volume of the catalytic site of the
enzyme (Fig. 6A, D and E).
(
Fig. 6B). On the contrary, the Phe117Ser, Phe117Trp and Phe117Ile
variants produced molecular systems where the H4 helix is tilted and
570

E.G. Chronopoulou et al.
B
B
A- General Subjects 1863 (2019) 565–576
Fig. 5. (continued)
4
.4. Inhibition analysis of site-saturated enzyme variants towards pesticides
(IC50 = 5.17 ± 0.47 μΜ) compared to the wild-type enzyme
(
9.836 ± 0.610 μM) (Fig. 2). The ability of α-endosulfan to bind and
To assess the effect of mutations at position 117 on enzyme's in-
inhibit the catalytic activity of mutant Phe117Ile was explored for the
development of analytical enzyme-based biosensor.
hibition by pesticides, kinetic inhibition analysis was carried out
Fig. 1). The results showed that the mutants Phe117Ile, Phe117Trp,
(
In an effort to investigate in silico the binding potential of the
Phe117Ile mutant, α-endosulfan was docked and the resulting mole-
cular system was subjected to MDs alongside the same system of the
wild-type enzyme. The systems reached equilibrium and the molecular
interactions of the docked α-endosulfan compound were identified
(Fig. 7). The results confirm that the Phe117Ile mutant has a rather
and Phe117Ser display higher sensitivity towards organochloride in-
secticides. Among all mutants, Phe117Ile was more sensitive towards α-
endosulfan, reaching 100% inhibition. Concentration-response curves
(
Fig. 2), for the wild-type enzyme and the mutant Phe117Ile, showed
that the latter enzyme is more sensitive to inhibition by α-endosulfan
571

E.G. Chronopoulou et al.
B
B
A- General Subjects 1863 (2019) 565–576
interact with the -S=O moiety and secure higher affinity of the
Phe117Ile mutant for this compound, when compared to the wild-type
enzyme.
4.5. Enzyme immobilization, stability and kinetic analysis
The mutant enzyme Phe117Ile was entrapped in a sol-gel polymer
(
Fig. 8), which was synthesized by condensation reaction between
TEOS and PTMOS at acidic pH in the presence of CTAB, a single
chained cationic surfactant [9,15]. Leaching of the enzyme through
diffusion from the sol-gel matrix was not detected. The use of CTAB
allows the synthesis of a flexible silica sol-gel matrix able to adopt
suitable shape with desirable pore size network [57] and mechanical
strength [58]. The use of alkoxysilane, PTMOS and TEOS are preferred
for enzyme immobilization due to their ability to provide desirable pore
structure, uniform pore-size distribution, high specific pore volumes,
large internal surface area [59] and surface functionalities [60].
After sol-gel synthesis, “aging” of the matrix was achieved for two or
five days. During the “aging” process, cross-linking of the polymer
network increases and the internal solvent is expelled from the matrix,
causing alterations in internal polarity and viscosity. In addition, the
average pore size decreases [61]. This “aging” step was found to be
crucial, as resulted in improvement in enzyme stability, as illustrated in
Fig. 9. As it can be seen from the figure, the activity of the entrapped
enzyme decreased at a significant lower rate, compared to that of the
free enzyme under the same conditions. The immobilized enzyme re-
tained > 60% of its initial activity after about twenty-five days, while
the free enzyme totally lost its activity after twelve days of storage.
These data clearly show that the entrapment benefited significantly the
long-term stability of the enzyme, which improves considerably the
practical viability of this system.
The concentration–response curve of the entrapped enzyme for α-
endosulfan was assessed and the results are shown in Fig. 10. As can be
seen from the figure, the entrapped enzyme notably retains its sensi-
tivity towards α-endosulfan but exhibits increased affinity with an IC50
of 2.60 ± 0.22 μΜ, suggesting that the sol-gel process provides an
appropriate technology for the development of a GST-based biosensor.
4.6. Biosensor assessment and application
The sensing-signal of the GST-based biosensor exploited the in-
hibition by α-endosulfan of the CDNB/GSH conjugation reaction, cat-
alysed by the sol-gel entrapped mutant Phe117Ile.
The concentration of released H+ ions reflects the progress of the
conjugation reaction and therefore pH, changes can be transduced to an
optically measurable signal by the immobilized indicators. The mixture
of bromocresol purple and phenol red provided adequate sensitivity
and reproducibility. Fig. 11A shows the VIS spectrum of sol–gel en-
trapped mixture of two dyes at pH values ranging from 4.0 to 7.0. As
can be seen, the mixture of the two indicators exhibits two distinct
Fig. 6. In silico mutation study for the PvGmGSTUG mutants A: Phe117Ser, B:
Phe117Gly, C: Phe117His, D: Phe117Trp and E: Phe117Ile.
larger catalytic site (> 25% increase in volume) when compared to the
wild-type enzyme. Phe18 interacts with α-endosulfan and acts as a
barrier that balances α-endosulfan to achieve the two hydrogen bonds
with Lys56 and Val57. However, on the Phe117Ile mutant and due to
the increase in volume of the catalytic site, the Phe18 residue has
moved away, dragged from the movement of the H4 helix, as discussed
above. This allowed α-endosulfan to move deeper towards the enzyme
core and establish two more interactions with Phe43 and Met13. This
shift does not only allow the extra two interactions but more im-
portantly it coordinates α-endosulfan in optimal position for Lys56 to
2
absorption maxima, at 444 and 562 nm. Linear dependence (R = 0.98)
of the absorption at 562 nm versus pH was found between pH 4 to 7
(
Fig. 11B). The optical signal at 562 nm of the sol-gel entrapped
Phe117Ile mutant enzyme was assessed using different concentrations
of α-endosulfan (Fig.11C). As can be seen in Fig. 11D a linear depen-
2
dence (R = 0.98) of the absorption at 562 nm versus different con-
centrations of α-endosulfan (0.625–30 μM) was observed, suggesting
that the system can provide a high sensitive analytical method for α-
572

E.G. Chronopoulou et al.
B
B
A- General Subjects 1863 (2019) 565–576
Fig. 7. Molecular docking of α-endosulphan in the PvGmGSTUG structure. A: The wild type PvGmGSTUG with docked α-endosulphan. B: The Phe117Ile mutant
PvGmGSTUG with docked α-endosulphan.
endosulfan determination.
(e.g. α-endosulfan in the present work) in a sample where only this
particular known analyte is present. Alternatively, such biosensors can
be exploited as detection systems able to detect several analytes si-
multaneously. This can be achieved using data processing of Artificial
Neural Networks (ANNs) coupled with a sensor array. This combinan-
tion can substantially improve the selectivity and allow exact de-
termination of pesticides present in a sample. ANN can combines the
response of different enzymes (e.g. different engineered forms of GST)
to find a pattern that relates inhibitor concentrations with the inhibition
percentages measured. Such biosensors have been developed based on
acetylocholinesterase [63–65] or tyrosinase [66] inhibition and che-
mometric data analysis using ANNs. This area represents a very pro-
mising approach in environmental monitoring and screening.
The biosensor was further evaluated using real water samples (fin-
ished drinking water and bottled mineral water) spiked with known
concentrations of α-endosulfan and the results are illustrated in Fig. 12.
The optical signal at 562 nm obeys different response when used
drinking water and mineral water, compared to that observed using
2
ddH O. This may be due to the differences in salt and organic content of
the two water samples. Linearity was observed within the concentration
range of 0.8–16 μM of α-endosulfan, obeying the following equations:
2
Mineral water: y = 0.0249× + 0.338, (R = 0.99)
2
Drinking water: y = 0.0472 x + 0.303, (R = 0.99)
Recovery experiments using drinking water and bottled mineral
water samples spiked with a known amount of α-endosulfan, were also
achieved and the results are listed in Table 3. Recoveries of α-en-
dosulfan in the mineral water were in the range between 80.6 and
1
03% (mean value 93.42 ± 3.25%, Ν = 6). For drinking water, re-
coveries were ranged between 88 and 104% (mean value
8.6 ± 3.46%, Ν = 4). The reproducibility of the assay response, ex-
pressed by relative standard deviation, was in the order of 8.54% and
5. Conclusion
9
In the present work, a structure-based combinational approach was
used for the design of an optimized mutant of PvGmGSTUG, for the
development of an optical biosensor for α-endosulfan determination.
Site-saturation mutagenesis at residue 117 proved to be a successful
approach for the improvement of enzyme's sensitivity towards α-en-
dosulfan. The proposed biosensor provides a low-cost alternative to
current chromatographic-based methods [62]. Since GSTs exhibit wide
inhibition capabilities, the present approach can be further adapted and
optimized to different pesticides, providing a route for developing
analytical tools for environmental and toxicological applications.
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.bbagen.2018.12.004.
7
.02% for the mineral and drinking water, respectively.
The design of enzyme biosensors is mainly based on the knowledge
about the target analyte, as well as the complexity of the biological
sample as well as the matrix in which the analyte has to be determined.
The application of enzymes with wide binding specificity for devel-
oping biosensors, such as acetylcholinesterases [63–65], tyrosinases
[
66] and GSTs [3,9,17,22–26], suffer from poor specificity since many
inhibitors can interfere and therefore one limitation of such biosensors
is the difficulty in discriminating between different inhibitors. Such
biosensors can only be used for the quantification of a known analyte
573

E.G. Chronopoulou et al.
B
B
A- General Subjects 1863 (2019) 565–576
Fig. 8. Schematic illustration of the GST -based biosensor system for the determination of α-endosulfan.
Fig. 9. Stability analysis of the free and the entrapped mutant enzyme
Phe117Ile. The effect of gel aging (2 and 5 days) on the stability was also in-
vestigated.
Fig. 10. Dose-response inhibition of α-endosulfan on the sol-gel entrapped
Phe117Ile mutant enzyme. All assays were accomplished using the CDNB–GSH
assay system and the experiments were performed in duplicate.
574

E.G. Chronopoulou et al.
B
B
A- General Subjects 1863 (2019) 565–576
Fig. 11. Dependence of optical signal on pH. A. Visible spectra of sol–gel with entrapped bromocresol purple/phenol red at different pH values B. Calibration curve of
the absorption at 562 nm on different pH. C. Visible spectra of the sol-gel co-entrapped Phe117Ile and bromocresol purple/phenol red in the presence of different
concentration of α-endosulfan (0–30 μM). D: Calibration curve of the absorbance at 562 nm on α-endosulfan concentration (0–30 μM).
Table 3
Pesticide recovery experiments in water samples spiked with known amounts of
α-endosulfan.
Sample
Added (μΜ)
Control
Spiked
Recovery (%)
Mineral water
0.8
2
4
0.34
0.41
0.5
0.55
0.7
0.35
0.38
0.44
0.56
0.58
0.71
102.9
94.2
87.6
102.7
83.9
89.3
8.54
104.5
88.8
102.2
98.9
7.02
8
1
1
0
6
0.8
RSDa (%)
Drinking water
0.8
0.33
0.42
0.50
0.68
0.35
0.37
0.51
0.68
1
4
8
.6
Fig. 12. Dependence of the absorbance at 562 nm on α-endosulfan concentra-
tion (0–30 μM) using drinking (●) and mineral (■) water.
RSDa (%)
a
RSD: relative standard deviation.
575

E.G. Chronopoulou et al.
B
B
A- General Subjects 1863 (2019) 565–576
Conflict of interest
None to declare.
[27] V. Tiwari, Front. Chem. 4 (2016) 39.
[
[
[
28] C. Bernal, K. Rodriguez, R. Martínez, Biotechnol. Adv. 36 (2018) 1470.
29] K. Tars, B. Olin, B. Mannervik, J. Mol. Biol. 397 (2010) 332.
30] Y. Ivarsson, A.J. Mackey, M. Edalat, W.R. Pearson, B. Mannervik, J. Biol. Chem. 278
(2003) 8733.
[
[
31] M.A. Norrgård, B. Mannervik, J. Mol. Biol. 412 (2011) 111.
32] N.E. Labrou, M.M. Muharram, M.S. Abdelkader, Biochim. Biophys. Acta 1864
Acknowledgement
(
2016) 1315.
NEL and FSA acknowledge the International Scientific Partnership
Program at King Saud University, Saudi Arabia, for funding this re-
search work through ISPP# 0071.
[33] J. Damborsky, J. Brezovsky, Curr. Opin. Chem. Biol. 19 (2014) 8.
[34] S. Lutz, Curr. Opin. Biotechnol. 21 (2010) 734.
35] R. Verma, U. Schwaneberg, D. Roccatano, Comput. Struct. Biotechnol. J. 2 (2012)
[
[
[
e201209008.
36] M. Bilal, H.M.N. Iqbal, S. Guo, H. Hu, W. Wang, X. Zhang, Int. J. Biol. Macromol.
108 (2018) 893.
37] J. Weber, C.J. Halsall, D. Muir, C. Teixeira, J. Small, K. Solomon, M. Hermanson,
H. Hung, T. Bidleman, Sci. Total Environ. 408 (2010) 2966.
38] M. Bradford, Anal. Biochem. 72 (1976) 248.
References
[1] A.J. Dowd, A. Steven, E. Morou, J. Hemingway, J. Vontas, M.J.I. Paine, Enzym.
Microb. Technol. 45 (2009) 164.
[
[
39] E.G. Chronopoulou, A.C. Papageorgiou, F. Ataya, I. Nianiou-Obeidat, P. Madesis,
[
2] K. Kiilerich-Hansen, C.R. Poulsen, J. Daprà, O. Nikolaj, C. Rozlosnik, N. Rozlosnik,
Polymer based biosensors for pathogen diagnostics, in: V. Somerset (Ed.),
Environmental Biosensors, IntechOpen, 2011, pp. 193–209.
N.E. Labrou, Front. Plant Sci. 9 (2018) 737, https://doi.org/10.3389/fpls.2018.
0
40] M. Magrane, Uniprot Consortium, Database, 2011 (2011) bar009.
41] D.A. Benson, I. Karsch-Mizrachi, D.J. Lipman, J. Ostell, E.W. Sayers, Nucleic Acids
Res. 39 (2011) D32.
1737.
[
[
[3] E. Morou, H.M. Ismail, A.J. Dowd, J. Hemingway, N. Labrou, M. Paine, J. Vontas,
Anal. Biochem. 378 (2008) 60.
[
4] G. Rocchitta, A. Spanu, S. Babudieri, G. Latte, G. Madeddu, G. Galleri, S. Nuvoli,
P. Bagella, M.I. Demartis, V. Fiore, R. Manetti, P.A. Serra, Sensors 16 (2016) 780.
5] A. Sassolas, L.J. Blum, B.D. Leca-Bouvier, Biotechnol. Adv. 30 (2012) 489.
6] B. Bucur, F.-D. Munteanu, J.-L. Marty, A. Vasilescu, Biosensors (Basel) 8 (2018) 27.
7] A. Amine, F. Arduini, D. Moscone, G. Palleschib, Biosens. Bioelectron. 76 (2016)
[
42] S.F. Altschul, W. Gish, W. Miller, E.W. Myers, D.J. Lipman, J. Mol. Biol. 215 (1990)
4
03.
[
[
[
[
43] C.J. Sigrist, E. de Castro, L. Cerutti, B.A. Cuche, N. Hulo, A. Bridge, L. Bougueleret,
I. Xenarios, Nucleic Acids Res. 41 (2013) D344.
44] J.D. Thompson, D.G. Higgins, T.J. Gibson, Nucleic Acids Res. 22 (1994) 4673.
[
[
[
[
1
80.
45] J. Castresana, Mol. Biol. Evol. 17 (2000) 540.
[8] M. Trojanowicz, Electroanalysis 14 (2002) 1311.
46] D.T. Jones, W.R. Taylor, J.M. Thornton, Comput. Appl. Biosci. 8 (1992) 275.
47] D.H. Huson, D.C. Richter, C. Rausch, T. Dezulian, M. Franz, R. Rupp, BMC
Bioinformatics 8 (2007) 460.
48] R. Chen, L. Li, Z. Weng, Proteins 52 (2003) 80.
49] B. Hess, C. Kutzner, D. van der Spoel, E. Lindahl, J. Chem. Theory Comput. 4 (2008)
[9] V.G. Andreou, Y.D. Clonis, Anal. Chim. Acta 460 (2002) 151.
[
[
[
10] S. Andreescu, A. Avramescu, C. Bala, V. Magearu, J.L. Marty, Anal. Bioanal. Chem.
3
74 (2002) 39.
[
[
11] K. Flavin, J. Mullowney, B. Murphy, E. Owens, P. Kirwan, K. Murphy, H. Hughes,
P. McLoughlin, Analyst 132 (2007) 224.
12] C. Kascholke, S. Hendrikx, T. Flath, D. Kuzmenka, H.M. Dörfler, D. Schumann,
M. Gressenbuch, F.P. Schulze, M. Schulz-Siegmund, M.C. Hacker, Acta Biomater. 63
4
35.
[
[
[
50] D. Sellis, D. Vlachakis, M. Vlassi, Bioinform. Biol. Insights 3 (2009) 99.
51] I. Axarli, P. Dhavala, A.C. Papageorgiou, N.E. Labrou, J. Mol. Biol. 385 (2009) 984.
52] P. Reinemer, L. Prade, P. Hof, T. Neuefeind, R. Huber, R. Zettl, K. Palme, J. Schell,
I. Koelln, H.D. Bartunik, B. Bieseler, J. Mol. Biol. 255 (1996) 289.
53] I.A. Axarli, D.J. Rigden, N.E. Labrou, Biochem. J. 382 (2004) 885.
54] E.M. Gillam, Clin. Exp. Pharmacol. Physiol. 32 (2005) 147.
55] R.M. Ferraz, A. Vera, A. Arís, A. Villaverde, Microb. Cell Factories 5 (2006) 15.
56] I. Axarli, A.W. Muleta, E.G. Chronopoulou, A.C. Papageorgiou, N.E. Labrou,
Biochim. Biophys. Acta, Gen. Subj. 1861 (2017) 3416.
(2017) 336.
[
13] A. Kloskowski, M. Pilarczyk, W. Chrzanowski, J. Namieśnikournal, Crit. Rev. Anal.
Chem. 40 (2010) 172.
[
[
[
[
[14] A.E. Danks, S.R. Hall, Z. Schnepp, Mater. Horiz. 3 (2016) 91.
[15] R. Gupta, N.K. Chaudhury, Biosens. Bioelectron. 22 (2007) 2387.
[16] N. Allocati, L. Federici, M. Masulli, C. Di Ilio, FEBS J. 276 (2009) 58.
[17] E. Chronopoulou, A.C. Papageorgiou, A. Markoglou, N.E. Labrou, J. Mol. Cat. B:
Enzymatic. 81 (2012) 43.
[
57] C. Rottman, G. Grader, Y. De Hazan, S. Melchior, D. Avnir, J. Am. Chem. Soc. 121
[
[
18] B. Mannervik, U.H. Danielson, CRC Crit. Rev. Biochem. 23 (1988) 283.
19] J.D. Hayes, J.U. Flanagan, I.R. Jowsey, Annu. Rev. Pharmacol. Toxicol. 45
(1999) 8533.
[
[
[
[
[
58] B. Wang, B. Li, Q. Deng, S. Dong, Anal. Chem. 70 (1998) 3170.
59] M. Hartmann, X. Kostrov, Chem. Soc. Rev. 42 (2013) 6277.
60] A.C. Pierre, Biocatal. Biotransform. 22 (2004) 145.
61] B. Dunn, J.I. Zink, Chem. Mater. 9 (1997) 2280.
62] T. Siddique, Z.A. Zahir, W.T. Frankenberger Jr., J. Liq. Chromatogr. Relat. Technol.
(2005) 51.
[
[
20] D. Sheehan, G. Meade, V.M. Foley, C.A. Dowd, Biochem. J. 360 (2001) 1.
21] E. Chronopoulou, P. Madesis, B. Asimakopoulou, D. Platis, A. Tsaftaris, N.E. Labrou,
Planta 235 (2012) 1253.
[22] J.W. Choi, Y.K. Kim, S.Y. Song, I.H. Lee, W.H. Lee, Biosens. Bioelectron. 18 (2003)
2
6 (2003) 1069.
1
461.
[
63] M. Cortina-Puig, G. Istamboulie, T. Noguer, J.L. Marty, Analysis of pesticide mix-
tures using intelligent biosensors, in: V.S. Somerset (Ed.), Intelligent and Biosensors,
[
[
23] E.G. Chronopoulou, N.E. Labrou, Recent Pat. Biotechnol. 3 (2009) 211.
24] P. Kapoli, I.A. Axarli, D. Platis, M. Fragoulaki, M. Paine, J. Hemingway, J. Vontas,
N.E. Labrou, Biosens. Bioelectron. 24 (2008) 498.
2
010, p. 205.
[
[
64] G. Istamboulie, M. Cortina-Puig, J.L. Marty, T. Noguer, Talanta 79 (2009) 507.
65] K. Mwila, M.H. Burton, J.S. Van Dyk, B.I. Pletschke, Environ. Monit. Assess. 185
[25] E.M. Materon, P.J. Jimmy Huang, A. Wong, A.A. Pupim Ferreira, P. Sotomayor
Mdel, J. Liu, Biosens. Bioelectron. 58 (2014) 232.
(2013) 2315.
[
26] T.I. Oliveira, M. Oliveira, S. Viswanathan, M. Fátima Barroso, L. Barreiros,
O.C. Nunes, J.A. Rodrigues, P. de Lima-Neto, S.E. Mazzetto, S. Morais, C. Delerue-
Matos, Talanta 106 (2013) 249.
[
66] X. Cetó, F. Céspedes, M. del Valle, Talanta 99 (2012) 544.
576