Altering the substrate specificity of methyl parathion hydrolase with directed evolution

Article abstract of DOI:10.1016/j.abb.2015.03.012

Abstract Many organophosphates (OPs) are used as pesticides in agriculture. They pose a severe health hazard due to their inhibitory effect on acetylcholinesterase. Therefore, detoxification of water and soil contaminated by OPs is important. Metalloenzymes such as methyl parathion hydrolase (MPH) from Pseudomonas sp. WBC-3 hold great promise as bioremediators as they are able to hydrolyze a wide range of OPs. MPH is highly efficient towards methyl parathion (1 × 10⁶ s^-1 M^-1), but its activity towards other OPs is more modest. Thus, site saturation mutagenesis (SSM) and DNA shuffling were performed to find mutants with improved activities on ethyl paraxon (6.1 × 10³ s^-1 M^-1). SSM was performed on nine residues lining the active site. Several mutants with modest activity enhancement towards ethyl paraoxon were isolated and used as templates for DNA shuffling. Ultimately, 14 multiple-site mutants with enhanced activity were isolated. One mutant, R2F3, exhibited a nearly 100-fold increase in the k_cat/K_m value for ethyl paraoxon (5.9 × 10⁵ s^-1 M^-1). These studies highlight the 'plasticity' of the MPH active site that facilitates the fine-tuning of its active site towards specific substrates with only minor changes required. MPH is thus an ideal candidate for the development of an enzyme-based bioremediation system.

Full text of DOI:10.1016/j.abb.2015.03.012

Archives of Biochemistry and Biophysics 573 (2015) 59–68
Contents lists available at ScienceDirect
Archives of Biochemistry and Biophysics
journal homepage: www.elsevier.com/locate/yabbi
Altering the substrate specificity of methyl parathion hydrolase with
directed evolution
a
b
b
a,
⇑
Tee-Kheang Ng , Lawrence R. Gahan , Gerhard Schenk , David L. Ollis
a
Research School of Chemistry, Australian National University, Canberra, ACT 0200, Australia
School of Chemistry and Molecular Biosciences, The University of Queensland, St Lucia, Brisbane, QLD 4072, Australia
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Many organophosphates (OPs) are used as pesticides in agriculture. They pose a severe health hazard due
to their inhibitory effect on acetylcholinesterase. Therefore, detoxification of water and soil contaminated
by OPs is important. Metalloenzymes such as methyl parathion hydrolase (MPH) from Pseudomonas sp.
WBC-3 hold great promise as bioremediators as they are able to hydrolyze a wide range of OPs. MPH
Received 2 December 2014
and in revised form 10 March 2015
Available online 19 March 2015
6
ꢁ1
ꢁ1
is highly efficient towards methyl parathion (1 ꢀ 10 s
M ), but its activity towards other OPs is more
modest. Thus, site saturation mutagenesis (SSM) and DNA shuffling were performed to find mutants with
Keywords:
Organophosphate pesticides
Bioremediation
Methyl parathion hydrolase
Substrate specificity
Directed evolution
3
ꢁ1
ꢁ1
improved activities on ethyl paraxon (6.1 ꢀ 10 s
M ). SSM was performed on nine residues lining the
active site. Several mutants with modest activity enhancement towards ethyl paraoxon were isolated and
used as templates for DNA shuffling. Ultimately, 14 multiple-site mutants with enhanced activity were
isolated. One mutant, R2F3, exhibited a nearly 100-fold increase in the kcat/K
m
value for ethyl paraoxon
5
ꢁ1
ꢁ1
(5.9 ꢀ 10 s
M ). These studies highlight the ‘plasticity’ of the MPH active site that facilitates the fine-
tuning of its active site towards specific substrates with only minor changes required. MPH is thus an
ideal candidate for the development of an enzyme-based bioremediation system.
Ó 2015 Elsevier Inc. All rights reserved.
Introduction
oxime reactivation appears to depend on the identity of the OP
compounds [7]. Excess OPs in the bloodstream will also re-inhibit
Organophosphates (OPs)1 constitute the most commonly used
pesticides in agriculture and account for approximately 40% of total
pesticide usage [1]. OP are generally toxic as they inhibit acetyl-
cholinesterase, a serine hydrolase that catalyzes the breakdown of
acetylcholine at cholinergic synapses, leading to hyperstimulation
of nerve cells and death [2]. It is estimated that OPs are responsible
for up to 3,000,000 poisonings and 220,000 deaths annually [3].
Oximes, a group of strong nucleophiles, were shown to be capable
to reactivate inhibited acetylcholinesterase [4,5], and current
treatments for OP poisoning rely on chemical antidotes such as
cholinolytics, oximes, atropine and anticonvulsants to minimize
toxic manifestations [6]. However, these antidotes have to be
administered soon after exposure, as they are impotent against aged
inhibited acetylcholinesterase. Furthermore, the effectiveness of
acetylcholinesterase if they are not removed. Therefore, there is an
urgent need to develop (i) methods to remove OPs from contami-
nated soils, as well as (ii) therapies to treat OP poisoning and (iii)
biosensors for OP detection.
The use of OP-degrading enzymes is an effective approach for
OP decontamination. Methyl parathion hydrolase (MPH; E.C.
3.1.8.1),
a
metal ion-dependent enzyme isolated from
Pseudomonas sp. WBC-3, has attracted considerable attention due
to its ability to hydrolyze a broad range of OPs [8,9]. The crystal
structure of MPH reveals an
abba fold, characteristic for
metallohydrolases [10]. Interestingly, MPH is structurally similar
to metallo-b-lactamases (MBLs) from the B3 subgroup, including
the enzyme L1 from Stenotrophomonas maltophilia and AIM-1 from
2
+
Pseudomonas aeruginosa. MBLs are Zn -dependent enzymes that
have emerged as major threat to global health since they are cap-
able of inactivating most of the commonly used antibiotics [11,12].
MPH is able to hydrolyze methyl parathion (MPS) at a rapid
⇑
Corresponding author.
E-mail address: ollis@rsc.anu.edu.au (D.L. Ollis).
Abbreviations used: OP, organophosphate; MPH, methyl parathion hydrolase; SSM,
6
ꢁ1 ꢁ1
rate, with a kcat/K
m
ratio of >10 M
s
[10]. Limited information
1
about the substrate preference and reaction mechanism of MPH is
currently available. Residues Leu65, Leu67, Phe119, Trp179,
site saturation mutagenesis; MBls, metallo-b-lactamases; EPO, ethyl paraoxon; LB,
Luria–Bertani; StEP, staggered extension process; MCO, methyl chlorpyrifos oxon;
ECO, ethyl chlorpyrifos oxon; TCPy, 3,5,6-trichloro-2-pyridinol.
http://dx.doi.org/10.1016/j.abb.2015.03.012
0
003-9861/Ó 2015 Elsevier Inc. All rights reserved.

6
0
T.-K. Ng et al. / Archives of Biochemistry and Biophysics 573 (2015) 59–68
Phe196, Leu258 and Leu273 are part of the substrate binding
pocket [10]; while the substitutions of Phe196 and Leu273 by ala-
nines enhanced enzymatic activity towards the substrate p-nitro-
phenyl diphenylphosphate [13], replacing Phe119, Trp179 and
Phe196 by alanines is detrimental for the catalytic activity towards
MPS [10].
An ideal bioremediator exhibits significant levels of activity for
a broad range of OPs. Considering the close structural similarity
between MPH and enzymes from the MBL superfamily, we
hypothesized that the substrate specificity of MPH can be altered
easily. Here, site saturation mutagenesis (SSM) was performed to
generate mutant libraries that were screened for improved activity
towards ethyl paraoxon (EPO), a substrate that is approximately
and Sanchis et al. [24]. The PCR reactions were performed using
30 ng of template DNA (WTMPH-pJWL1030), 1ꢀ HF buffer,
2
1.5 mM MgCl , 0.1 lM of each mutagenic and flanking primer,
1 U of Phusion polymerase (Thermo Scientific, Waltham, MA,
USA) and 0.2 mM of each dNTP. The amplification program for
mutagenesis was as follows: initial denaturation for 3 min at
98 °C, followed by 18 cycles of 1 min at 98 °C, annealing for
1 min at 55 °C and extension at 72 °C. The second stage consisted
of 25 cycles of 1 min at 98 °C and extension for 5 min at 68 °C
and a final extension for 10 min at 68 °C. Up to four reactions were
pooled and treated with 20 U of DpnI twice at 37 °C for 1 h each.
The PCR product was subsequently purified and 2
lL of the product
was transformed into E. coli DH5 using electroporation. Primers
a
1
00-fold less efficient than MPS in the wild-type enzyme.
P2 and P3 were used to generate the L67P68R72-NDT library; P2
and P4 to generate the R72-NNK library; P2 and P5 to generate
the F119-NDT library; P2 and P6 to generate the P150-NDT library;
P2 and P7 to generate the W179-NNK library; P1 and P8 to gener-
ate the F196-NDT library; P1 and P9 to generate the L258-NDT
library and P1 and P10 to generate the L273-NDT library. The
sequences of the primers are shown in Table S1. NDT (N = A, C, G
or T; D = A, G or T) randomisation was used in all libraries except
R72-NNK and W179-NNK.
Mutants isolated from each library were used as parents for DNA
shuffling to obtain further enhancement towards EPO. Several
mutants with significant improvement also towards ‘‘poor’’ sub-
strates (e.g. chlorpyrifos) were thus found. The results highlight
the active site ‘plasticity’ of MPH and establish this enzyme as a
promising agent to be employed in bioremediation applications.
Materials and methods
Shuffling of mutations isolated from the SSM library was per-
formed using the staggered extension process (StEP) [25]. The
Materials, bacterial strains, growth conditions and plasmids
StEP reaction (25
30 ng of template DNA, 1
l
L) was performed in a solution containing
M of the P1 and P2 primers, 0.2 mM
Pesticides were purchased from either Chem Service (West
Chester, PA, USA) or Sigma Aldrich (St. Louis, MO, USA).
Molecular biology reagents and enzymes were obtained from
New England Biolabs (Ipswich, MA, USA), Thermo Scientific
l
of each dNTP, 1ꢀ HF buffer and 1 U of Phusion polymerase. The
amplification program for StEP was as follows: initial denaturation
for 5 min at 98 °C, followed by 120 cycles for 10 s at 98 °C, anneal-
ing for 10 s at 55 °C, extension for 2 s at 72 °C, final annealing for
2 min at 55 °C and final extension for 5 min at 72 °C. Up to four
reactions were pooled, purified and digested with AseI and PstI.
The digested PCR product was cloned into pJWL1030 as described
for the construction of MPH-pJWL1030.
(
Waltham, MA, USA) or Sigma–Aldrich (St. Louis, MO, USA). DNA
primers (Table S1 in Supplemental material) were synthesized by
Geneworks (Hindmarsh, SA, Australia) or Integrated DNA
Technologies (Coralville, IA, USA). Plasmid extraction and purifica-
tion kits were obtained from Qiagen (Limburg, Netherlands) or
Promega (Madison, WI, USA). Protein purification columns were
purchased from GE Healthcare (Buckinghamshire, UK).
Library screening
Escherichia coli strain DH5a was used for all aspects of the work
described. Cells harboring pJWL1030 plasmids were grown at
The library screening procedures consisted of a primary and
secondary screen. The primary screen master plate was prepared
either by picking single colonies that were then inoculated in
96-well plates or by culture dilution methods developed by
Stevenson et al. [26]. The former was used for most of the gener-
ated SSM and StEP shuffling libraries, while the latter was used
for the L67P68R72-NDT library due to its much larger size.
3
5
7 °C on Luria–Bertani (LB) broth or agar plates supplemented with
g/mL kanamycin. Construction of pJWL1030 plasmid has been
0
l
described previously [14]. mpd, the gene coding for MPH, was syn-
thesized by DNA 2.0 (Menlo Park, CA, USA). The recombinant plas-
mid containing the mpd gene (MPH-pJWL1030) was constructed
by amplification from the MPH-pET47b plasmid using primers P1
and P2. The nucleotides that encode the first 35 amino acid resi-
dues that constitute the signal peptide were excluded in this
amplification. The PCR product was digested with AseI and PstI,
and ligated into the NdeI and PstI sites, and transformed into
Mutants were subsequently transformed into E. coli DH5
colonies were picked and inoculated in 96-well round-bottomed
plates (Sarstedt, Nümbrecht, Germany) containing 100 L LB broth
supplemented with 0.1 mM ZnCl2 and 50 g/mL kanamycin. For
the culture dilution method, the transformants were diluted with
LB-kanamycin supplemented with 0.1 mM ZnCl and dispensed
into 96-well round-bottom plates at as density of 3–4 cells per
00 L. The plates were incubated for 24 h at 37 °C. Prior to activity
assays, the cultures were resuspended by gentle pipetting and
10 L of the 100 L culture from each well was aliquoted into
96-well flat-bottom plates (Sarstedt, Nümbrecht, Germany) and
lyzed with 10
of 0.5ꢀ BugBuster (Novagen, Darmstadt,
Germany) for 15 min. The lysate was assayed with 119 M ethyl
paraoxon (EPO) in 50 mM HEPES, 100 mM NaCl, 0.1 mM ZnCl2,
pH 7.6, in a final volume of 200 L. The activity was monitored
a; single
l
l
E. coli DH5a using electroporation.
2
Structural superimposition studies
1
l
The superimposition of MPH with MBL superfamily enzymes
was done in two stages in PyMOL. PyMOL’s Cealign command
was initially used for the superimpositions. Minor adjustments
were then added using PyMOL’s pair fitting function, by using
the two active site metals, the bridging water and metal coordina-
tion residues of MPH and a target protein as points of alignment.
The PDB IDs used in the superimposition study were 1P9E [10],
l
l
lL
l
l
4
AWY [15], 1BVT [16], 1X8G [17], 1SML [18], 1QH5 [19], 2QED
by the release of p-nitrophenolate at 405 nm using a Spectramax
M2e Microplate reader (Molecular devices, Sunnyvale, CA, USA)
at 30 °C. The assay procedure was adapted and modified from
reference [27]. Mutants that showed improved EPO activity were
streaked on fresh LB-kanamycin plates and single colonies were
picked for a secondary screen. The secondary screen assay was per-
formed exactly as described for the primary screen. For the culture
[
20], 2BR6 [21] and 2CBN [22].
Library generation
The SSM libraries were constructed using a megaprimed, ligase-
free site specific mutagenesis method reported by Tseng et al. [23]

T.-K. Ng et al. / Archives of Biochemistry and Biophysics 573 (2015) 59–68
61
dilution primary screen method, up to five single colonies were
Homology modelling and substrate docking
picked from the streaking of each chosen well for the secondary
screen. Improved mutants were streaked on fresh LB-kanamycin
plates and single colonies were picked and grown for plasmid iso-
lation and DNA sequencing.
SWISS-MODEL [28–30] was used to generate models of the
structures of MPH mutants isolated in the evolution experiments.
Subunit B of wild-type MPH (PDB ID: 1P9E_b) was used as the tem-
plate for the modelling. AutoDock Vina [31] and Swiss-Dock
[
32,33] were used to perform molecular docking of the substrate
Expression and purification
MPS into the active site of subunit B of MPH. The structure of
MPS was obtained from the ZINC database (http://zinc.docking.
org/; ZINC ID: 02040890).
All MPH variants were expressed constitutively in the
pJWL1030 vector. A starter culture consisting of 300
lL LB-kana-
mycin broth supplemented with 0.1 mM ZnCl was inoculated
2
with single colonies of cells harboring the MPH-pJWL1030 plasmid
and grown at 37 °C for 16 h. The starter culture was then inocu-
lated onto 30 mL LB-kanamycin broth, supplemented with
Results and discussion
Assessing the substrate profile of wild-type MPH
0
2
.1 mM ZnCl and was grown at 37 °C for 24 h. The cells were har-
vested by centrifugation and were kept at ꢁ80 °C until use.
In order to assess the substrate preference of wild-type MPH,
specific activities towards various substrates were measured
(Fig. 1). The structures of the substrates used are illustrated in
Fig. 2. Of the various OPs investigated, MPH shows a distinct pre-
ference towards dimethyl-substituted substrates such as methyl
parathion (MPS), methyl paraoxon (MPO) and methyl chlorpyrifos
oxon (MCO), when compared to diethyl-substituted analogues
such as ethyl parathion (EPS), ethyl paraoxon (EPO) or ethyl
chlorpyrifos oxon (ECO). The enzyme also prefers substrates
containing the 3,5,6-trichloro-2-pyridinol (TCPy) rather than the
p-nitrophenol (pNP) leaving group. MPH has no detectable activity
towards the monoester p-nitrophenyl phosphate and diester bis-
(p-nitrophenyl) phosphate, and only very low activity towards
the carboxylester p-nitrophenyl acetate. EPO, which is turned over
ꢂ100-fold slower than MPS, was chosen as the substrate for library
screening. EPO possesses two features that are common among
poor substrates of MPH – it has a diethyl substituent and a
phosphoryl group (P@O). The reason for the preference of MPH
for thionates (thiophosphoryl containing OPs) is unknown but is
unique among phosphotriesterases [34].
All purification steps were carried out at 4 °C. The harvested cells
were resuspended in 10 mL, 20 mM HEPES, pH 7.0, 0.1 mM ZnCl
2
,
1
mM phenylmethylsulfonyl fluoride, and were disrupted with a
French Press (12,000 psi cell pressure, two repeats). The cell debris
was removed by centrifugation and the soluble fraction was loaded
onto a 5 mL SP Sepharose Fast Flow column that has been equili-
2
brated with 20 mM HEPES, pH 7.0, 0.1 mM ZnCl . The bound pro-
tein was eluted with a linear 0–1 M NaCl gradient in the same
buffer. SDS–PAGE analysis of the eluted MPH showed >90% purity.
For storage MPH was dialyzed against 50 mM HEPES, pH 7.6,
1
00 mM NaCl, 0.1 mM ZnCl
oprotectant for storage at ꢁ20 °C. The concentration of MPH was
determined by measuring A280. The extinction coefficients (
were estimated using the ProtParam tool (http://web.expasy.org/
2
. 50% (v/v) glycerol was included as cry-
e
280
)
ꢁ1
ꢁ1
protparam/); for most variants
e280 is 20,400 M cm , except for
the F119Y, F196Y, R1A6, R2D2, R2A2, R2F3 and R2D3 mutants,
ꢁ1
ꢁ1
where e280 is estimated to be 21,800 M cm .
Kinetic studies
Site saturation mutagenesis and DNA shuffling
The kinetic constants of wild-type MPH and its mutants were
obtained by varying the concentrations of six substrates in 50 mM
MPH is part of the MBL superfamily that covers a diverse range
of catalytic functions. Seven structures from that group were
obtained from the Protein Data Bank (http://www.rcsb.org/pdb/
home/home.do) and superimposed onto subunit B of MPH to select
the residues to mutate. The metal ions and metal ion binding
ligands in the binuclear active sites superimpose generally with
good agreement. It is also noted that the MPH active site is the
most crowded of all the structures compared – all the other active
sites are more open than that of MPH. This difference is illustrated
in Fig. 3, in which MPH is displayed in cartoon mode with side
HEPES, pH 7.6, 100 mM NaCl, 0.1 mM ZnCl
performed in either a 96-well flat-bottomed plate (200
volume), or a in micro-cuvette (1 mL). The 96-well plate assays
were monitored with a Spectramax M2e Microplate reader and a
Cary 1E UV–Vis spectrophotometer was used for the micro-cuvette
assays. The assayed substrates (and concentration ranges used)
2
. The assays were
L reaction
l
were methyl parathion (MPS; 7.6–189.9
MPO; 15.2–379.8 M), ethyl parathion (EPS; 15.2–60.8
ethyl paraoxon (EPO; 15.2–379.8 M), methyl chlorpyrifos
oxon (MCO; 30.4–379.8 M) and ethyl chlorpyrifos oxon (ECO;
0.4–379.8 M). The hydrolyses of MPS, MPO, EPS, EPO were moni-
tored via the release of the p-nitrophenolate product at 405 nm
lM), methyl paraoxon
(
l
l
M),
l
l
3
l
ꢁ
1
ꢁ1
(e
405 = 14,696.56 M cm
at pH 7.6), while the hydrolyses of
MCO and ECO were monitored via the release of the 3,5,6-
ꢁ
1
ꢁ1
trichloro-2-pyridinol product at 310 nm (e310 = 5800.67 M cm
at pH 7.6) using a Cary 1E UV–Vis spectrophotometer (Varian,
Palo Alto, CA, USA). All kinetic measurements were done in a final
volume of 1 mL at 30 °C. The kinetics constants were determined
by fitting the data to the Michaelis–Menten equation (Eq. (1)) using
KaleidaGraph 4.1 software (Synergy, Reading, PA, USA):
v
o
¼ VmaxS=ðK
m
þ SÞ
ð1Þ
where is the initial velocity, Vmax (=kcat[E]) is the maximum
v
o
velocity, S is the substrate concentration and K is the Michaelis
m
constant. The error represents the standard deviation from three
independent experiments.
Fig. 1. Catalytic activity of wild-type MPH towards various substrates.

6
2
T.-K. Ng et al. / Archives of Biochemistry and Biophysics 573 (2015) 59–68
Fig. 2. Structures of organophosphate substrates used in this work.
overlaps with that of the representative MBL member, suggesting
that the relevant amino acid residues fulfil a structural role and
are not important for discriminating between substrates. Thus,
the residues in the MPH active site that do not overlap with resi-
dues of GloB may be important for determining the substrate
specificity of MPH. These residues are L67, P68, R72, F119, P150,
F196, L258 and L273. W179 was also selected since it was pre-
viously suggested as a residue involved in substrate binding [10].
The location of these residues in the structure is shown in Fig. 4.
Eight SSM libraries were generated and screened for mutants
that showed improved activity towards EPO (Table 1). NDT
(
N = A, C, G or T; D = A, G or T) randomization was used in most
libraries to produce small and efficient libraries, an approach that
was particularly useful for the L67P68/R72-NDT library, where
three residues were simultaneously mutated [35]. NDT produces
a balanced mix of polar and non-polar, aliphatic and aromatic, as
well as negatively and positively charged representatives, while
excluding most of the structurally similar amino acids through
1
2 codon combinations that encode 12 amino acids without
Fig. 3. (a) Structural superimposition of the metal ion ligands of MPH and GloB. The
amino acid numbering scheme from MPH was used in labelling the metal ion-
binding ligands. The MPH ligands are colored yellow while GloB ligands are colored
in white. (b) GloB is shown in surface representation mode and MPH in cyan. MPH
residues that line the active site and are different from those in GloB are colored in
maroon. (For interpretation of the references to color in this figure legend, the
reader is referred to the web version of this article.)
chains displayed in lines, while Salmonella typhimurium GloB (PDB
ID: 2QED), a glyoxalase II, is displayed in surface representation
mode. It is evident that the major portion of the MPH active site
Fig. 4. Location of the MPH residues selected for site saturation mutagenesis.

T.-K. Ng et al. / Archives of Biochemistry and Biophysics 573 (2015) 59–68
63
Table 1
Summary of site saturation libraries generated, randomization used and the number of mutants screened. The number of possible variants and oversampling required for 95%
coverage was calculated using CASTer 2.0.
SSM
library
Codon
randomisation
Number of possible
variants
Library size to achieve 95%
coverage
Number of mutants
screened
Beneficial
mutations
Detrimental
mutations
L67P68R72 NDT–NDT–NDT
1278
32
12
12
32
12
12
12
5175
94
34
34
94
34
34
34
5800
80
40
40
80
40
40
40
None
None
Y
None
None
I, L, Y
I, S, N, H
I, V
Did not sample
A, T, G
G, N
Y, H, G
S, H, Y, M
R, D
R72
NNK
NDT
NDT
NNK
NDT
NDT
NDT
F119
P150
W179
F196
L258
L273
R
C, D
redundancy [35]. A small number of mutants with diminished
activity was also sampled randomly from each library to examine
the mutations that caused the decrease in activity. Overall, only
mutations affecting F119, F196, L258 and L273 resulted in
enhanced activity.
was used in library screening) to probe possible additional
alterations to the substrate preference of MPH. For multiple site
variants the crude lysate was used in initial screens to pre-select
the mutants of interest. R1A6, R2B2, R2C2, R2D2, R2A2, R2F3 and
R2D4 were thus selected for purification and detailed kinetic
studies. Mutants that were not selected either have only moderate
catalytic improvement, or are represented by other mutants that
exhibit similar catalytic properties. The kinetics parameters for
the various MPH variants are summarized in Table 3.
In general, each of the single site mutants selected from the EPO
m
screen displays modest enhancement in kcat/K towards that sub-
strate, with the best single site mutant, L258H, having a sevenfold
improvement. Furthermore, apart from L258I, they have decreased
A quick relative activity validation test was carried out on
mutants that showed improvement towards EPO hydrolysis during
screening in order to establish a shortlist of promising mutants for
more detailed kinetic studies (data not shown). Ten mutants,
namely F119Y, F196L, F196I, F196Y, L258S, L258I, L258H, L258N,
L273I and L273V, were thus picked for protein expression, purifica-
tion and kinetic studies. Further activity enhancement (via shuf-
fling) was deemed possible as each of the single site mutants has
enhanced activity against EPO. Together with wild-type MPH these
mutants were thus employed as parents in further shuffling
experiments using the staggered extension process (StEP)
approach [36]. The resulting mutants were subsequently tested
for further enhancement in EPO activity. Two rounds of StEP shuf-
fling were carried out, with multiple site variants isolated from the
first round being used as parents in the second; the results or these
screens are summarized in Table 2. After two rounds of StEP, only
double- and triple-site mutants were isolated, possibly due to the
distance between some of the mutations. For example, L258 and
L273 are separated by only 45 nucleotides. The recombination effi-
ciency decreases as the distance between mutations decreases
kcat/K
m
ratios towards MPS, the preferred substrate for wild-type
MPH. Each of the multiple site variants displays further improve-
ments towards EPO hydrolysis in terms of both kcat and kcat/K
The most efficient mutant, R2F3, has a nearly 100-fold improve-
ment in the kcat/K ratio towards EPO. Each of the multiple site
m
.
m
variants also recorded improved efficiency towards ECO (in con-
trast to the single site variants, where only a subset displayed an
increase). Most of the multiple site mutants have, however, a lower
m
kcat and kcat/K towards MPS and EPS, indicating a shift in substrate
specificity away from substrates with the P@S group. The catalytic
parameters of L258S, L258H, R2B2, R2C2, R2F3 and R2D4 for the
EPS hydrolysis were not determined due to their low activity.
[
36].
F119 and F196 – aromatic residues are required for high activity
Kinetic analysis of MPH variants
F119 appears to be more important than F196 for MPH activity.
The activity distribution profiles of F119 and F196 libraries show
different mutational tolerances (Fig. S1 in the Supplemental mate-
rial); over 70% of the F119 library but less than 50% of the F196
library display an activity towards EPO that is at least half as high
as that of wild-type MPH. F119Y is the only mutant with improved
reactivity towards EPO from the F119 library, while F196L, F196I
and F196Y from the F196 library all have improved reactivity, sug-
gesting that an aromatic ring in position 119, but not in position
196 is necessary for efficient catalysis. This hypothesis is further
supported by a comparison of the active sites of MPH and OPHC2
from Pseudomonas pseudoalcaligenes. OPCH2 is a promiscuous
hydrolase that shares near identical protein structure with MPH
despite only 48% amino acid sequence similarity (Fig. S2 in the
Supplemental material) [37]. In OPHC2 (PDB ID: 4LE6), the residue
equivalent to F196 is replaced by a methionine (M188), while F119
is conserved (F111). Thus, F119 may play an important role in
Each of the single site variants with improved hydrolytic activ-
ity towards EPO was expressed, purified and its catalytic properties
were investigated with six OP substrates (other than EPO, which
Table 2
Summary of StEP libraries generated, number of mutants screened and improved
mutants selected from the screen.
StEP
Number of mutants
screened
Selected
mutants
Residue
F119 F196 L258 L273
Round
1
400
R1B2
R1D4
R1A6
R1B6
R2B2
R2C2
R2D2
R2F2
Y
Y
N
I
V
I
S
N
H
I
I
I
Y
H
H
substrate binding, possibly via the formation of
interactions with the aromatic ring of OP substrates.
p–stacking
Y
Round
2
400
R2G4
R2A2
R2E3
R2F3
R2C4
R2D4
L
I
I
I
Y
I
H
Y
Y
Y
Y
Y
V
I
N
H
H
L258 – flexible position that is key to altering MPH specificity
L258 is of particular interest as four different mutations were
found to lead to significantly improved activity towards EPO.

6
4
T.-K. Ng et al. / Archives of Biochemistry and Biophysics 573 (2015) 59–68
Table 3
Catalytic parameters for (A) single-site MPH mutants and (B) multiple-site MPH mutants. The errors for kcat, K
experiments. ND: Not determined.
m
m
and kcat/K represent the standard deviation of three independent

T.-K. Ng et al. / Archives of Biochemistry and Biophysics 573 (2015) 59–68
65
Only 35% of the library has an activity that is less than 50% of that
of the wild-type enzyme (Fig. S1 in the Supplemental material).
L258 forms part of the base of the active site pocket in MPH. The
beneficial mutations are largely hydrophilic (L258H, L258N and
L258S; the L258I leads to some modest improvements in catalytic
efficiency). The effect of the L258S and L258N mutations is similar:
both variants display an increase in kcat for oxon OPs but a decrease
for thiolphosphate OPs (Table 3A). Both variants also show a large
increase in kcat for chlorpyrifos-based OPs (i.e. MCO and ECO);
L258S is 11-fold and 32-fold more reactive towards MCO and
ECO, respectively. The corresponding increase in activity for the
L258N mutant is 19-fold and 91-fold. The increase in kcat is also

6
6
T.-K. Ng et al. / Archives of Biochemistry and Biophysics 573 (2015) 59–68
Fig. 5. Active site conformations of the two MPH monomeric subunits. The metal ion center superimposition of the two MPH subunits is shown in (a). The superimposition of
active site-lining residues is shown in (b). The active site-lining residues for subunit A are shown in (c) and those for subunit B in (d). The surface represents the solvent-
excluded surface of the active site.
reflected in an enhancement of the catalytic efficiency (kcat/K
m
) of
L67, P68, R72, P150 and W179 are sites intolerant to mutations
L258S and L258N towards these substrates to an extent that they
are more efficient than wild-type MPH towards MPS, its optimal
The roles of L67, P68 and R72 are unclear at this point. However,
activity distributions of L67P68R72 NDT and R72 NNK libraries
suggest that these sites are intolerant towards mutations. For both
libraries, more than 80% of mutants screened have shown less than
50% of the EPO activity measured for wild-type MPH and no vari-
ants with improved catalytic efficiency were found (Fig. 3A). For
R72, mutations R72A, R72T and R72G resulted in relative activities
of 36%, 14% and 11% respectively, suggesting an important role in
either catalysis or substrate binding or both. The crystal structure
of wild-type MPH (PDB ID: 1P9E) does not indicate relevant inter-
actions R72 may be involved in. However, a closer comparison of
the active sites of the two subunits in the homodimer reveals con-
formational differences in some of the residues, particularly R72,
F119 and L273 (Fig. 5). Judged from the B-factors extracted from
the crystal structure these residues are located in a flexible region.
Furthermore, docking studies with the two subunits of MPH and
the substrate MPS (using SwissDock [32,33] and AutoDock Vina
[31]) indicate that the conformation observed in subunit B is the
form adopted by the enzyme during catalysis as MPS can only be
successfully docked into this subunit (Fig. S3 in the Supplemental
material). This also suggests that the conformation observed in
subunit A restricts substrate access and thus impedes catalysis.
In the P150 NDT library, the P150Y and P150H mutations were
found to result in a large drop in activity, with the mutants having
only 27% and 8%, respectively, of the activity of wild-type MPH.
This is likely to be due to the steric obstruction introduced by
the relatively large side chains of tyrosine and histidine. P150 is
surrounded by L118, F119 and F196. Mutating P150 into residues
with larger side chain will thus introduce clashes with aforemen-
tioned residues, especially F119 and F196, causing a loss in activity.
In the W179 NNK library, W179S, W179H, W179Y and W179M
were the deleterious mutations sampled. The deterioration in
activity may be due to the disruption of hydrogen bonding
between W179 and E175. E175 appears to act as a bridging residue
substrate (see Introduction). The kcat/K
m
values towards MCO and
6
ꢁ1
ꢁ1
6
ꢁ1
ꢁ1
ECO are 2.42 ꢀ 10 s
M
and 0.96 ꢀ 10 s
M
, and 3.73 ꢀ
6
ꢁ1
ꢁ1
6
ꢁ1
ꢁ1
1
0 s
M
and 1.13 ꢀ 10 s
M
for L258S and L258N MPH,
respectively. Although the reasons for these considerable improve-
ments are currently unknown the increased hydrophilicity in the
active sites of the mutants may provide the basis for a more effi-
cient mode of substrate binding. While the L258H mutation also
leads to some (minor) improvement in catalytic efficiency towards
MCO and ECO (Table 3A) its main effect is a sixfold increase in the
catalytic efficiency towards EPO. In contrast, the L258R mutation is
detrimental for the catalytic efficiency of MPH, likely to be due to
the large size of this side chain.
L273 mutations – conservative but beneficial
L273I and L273V are mutations that lead to an improvement in
the catalytic efficiency of MPH towards EPO (Table 3A). L273I and
L273V mutations are conservative mutations and probably repre-
sent the rearrangement of side chains to allow better access to sub-
strates. This interpretation is in agreement with an earlier study on
MPH where the L273A mutation was shown to have a similar cat-
alytic effect [13]. For both L273I and L273V, an increase in kcat/K
m
for all oxon substrates is observed, while the catalytic efficiency
towards MPS and EPS is reduced by these mutations. Since L258
and L273 are located near each other and mutations to these sites
improve the catalytic efficiencies towards oxon substrates, it is
likely that this part of the substrate-binding pocket determines
the specificity in favor of oxonate rather than thionate substrates.
It is worth noting, however, that although L273 mutations lead to
an increase in efficiency towards oxonate substrates and a signifi-
cant decrease towards thionates, the measured catalytic efficien-
cies for thionates are still higher than those for their oxonate
analogues.

T.-K. Ng et al. / Archives of Biochemistry and Biophysics 573 (2015) 59–68
67
Table 4
The R2F3 mutant (F119Y, F196I, L258H) is the most efficient
mutant towards EPO (Table 3B). The mutant has an over 73-fold
Comparison of the catalytic efficiencies of wild-type and selected MPH mutants
towards a range of substrates. ND: Not determined.
increased kcat and nearly 100-fold improvement of kcat/K
EPO when compared to wild-type MPH. This mutant also has
, respec-
m
for
ꢁ
1
ꢁ1
Substrates
k
cat/K
m
(s
M )
WT
R1A6
R2F3
R2D2
ꢂ121-fold and 77-fold improvements in kcat and kcat/K
m
6
5
3
4
4
7
6
4
5
6
5
4
5
6
5
tively, for the reaction with ECO, thus making it a very efficient
catalyst for the hydrolysis of diethyl oxon substrates.
MPS
MPO
EPS
EPO
MCO
ECO
1.0 ꢀ 10
6.2 ꢀ 10
1.1 ꢀ 10
6.2 ꢀ 10
3.2 ꢀ 10
2.0 ꢀ 10
3.1 ꢀ 10
6.2 ꢀ 10
4.6 ꢀ 10
9.5 ꢀ 10
1.5 ꢀ 10
5.1 ꢀ 10
2.9 ꢀ 10
1.3 ꢀ 10
ND
6.0 ꢀ 10
4.1 ꢀ 10
1.6 ꢀ 10
1.3 ꢀ 10
5.6 ꢀ 10
8.9 ꢀ 10
3.1 ꢀ 10
1.8 ꢀ 10
7.7 ꢀ 10
4
5
3
5
4
The R2D2 mutant (F119Y, L258H) is a ‘generalist’, with kcat
improvements for all substrates except MPS (Table 3B). While
5
5
6
the kcat for MPS has decreased, the kcat/K
wild-type MPH (due to a decrease in K ). Of the six substrates
examined, five have kcat/K
values of at least 1 ꢀ 10 s
two of these five are at least 1 ꢀ 10 s
m
is similar to that of
m
5
ꢁ1
ꢁ1
m
M ;
6
ꢁ1
ꢁ1
between the side chain of W179 and the backbone of M148, sug-
gesting a structural role for W179. However, Dong et al. found that
the mutation W179F retains catalytic parameters (kcat and K )
m
similar to those of wild-type MPH, suggesting that the interaction
between W179 and E175 is not important and any aromatic side
chain may replace W179 for effective substrate binding [10].
M , representing a signifi-
cant improvement compared to wild-type MPH (Table 4); this
variant is thus a good candidate for application in bioremediation.
A comparison of the catalytic efficiencies of R1A6, R2F3 and R2D2
with that of wild type MPH is summarized in Table 4.
Concluding remarks
Multiple site mutants obtained from shuffling show interesting profiles
The work presented here describes a site saturation mutagene-
sis study of residues lining the active site of MPH and the directed
evolution of this enzyme for improved EPO hydrolysis. To the best
of our knowledge, this is the most extensive substrate profile
characterization of MPH and related OP-degrading enzymes to date
and provides essential clues about the development of specific
mutants useful for bioremedial applications. The substrate profiles
of single site mutants were characterized with various OP sub-
strates in an attempt to understand the roles of these residues in
substrate interaction(s). All the selected single site mutants have
modest improvements for EPO hydrolysis and, interestingly, unex-
pected but impressive improvements were observed for L258
mutants for the hydrolysis of non-selected substrates (MCO and
ECO). Subsequently, two rounds of directed evolution were con-
ducted with the aim to further enhance the EPO-hydrolyzing
The mutant R1A6 (F119Y, L258S) displays improvements for all
substrates examined except MPS, the preferred substrate for wild-
type MPH (Table 3B). The improvements in kcat and kcat/K
the hydrolysis of EPO, the OP used for screening, are approximately
5-fold. However, the largest improvement observed for this
mutant was towards MCO and ECO; for MCO kcat and kcat/K
m
towards
1
m
improved 36- and 46-fold, respectively, and for ECO the
corresponding increases were 304- and 252-fold. The catalytic
efficiencies of the mutant for MCO and ECO, respectively, are thus
7
ꢁ1
ꢁ1
6
ꢁ1
ꢁ1
ꢁ1
ꢂ15-fold (1.47 ꢀ 10 s
M
) and ꢂfivefold (5.13 ꢀ 10 s
M
M
)
6
ꢁ1
larger than that of wild-type MPH for MPS (1.01 ꢀ 10 s
).
The kinetics data obtained for R1A6 are also consistent with the
observation that the L258S mutation introduces a preference for
chlorpyrifos-type substrates (vide infra).
Fig. 6. Comparison of the active site pockets of selected MPH variants and the wild-type enzyme. (a) wild-type MPH, (b) R1A6 mutant, (c) R2F3 mutant, and (d) R2D2 mutant.

6
8
T.-K. Ng et al. / Archives of Biochemistry and Biophysics 573 (2015) 59–68
efficiency of MPH; numerous multiple site mutants with improved
activity were obtained. These results highlight the robustness and
[10] Y.-J. Dong, M. Bartlam, L. Sun, Y.-F. Zhou, Z.-P. Zhang, C.-G. Zhang, Z. Rao, X.-E.
Zhang, J. Mol. Biol. 353 (2005) 655–663.
[
11] A.F. Neuwald, J.S. Liu, D.J. Lipman, C.E. Lawrence, Nucleic Acids Res. 25 (1997)
665–1677.
[12] H. Daiyasu, K. Osaka, Y. Ishino, H. Toh, FEBS Lett. 503 (2001) 1–6.
‘
plasticity’ of the MPH active site to accommodate and turn over
1
different OP substrates – a few amino acid changes in the active
site pocket result in drastic changes in OP specificity. Most of the
beneficial mutations observed can be attributed to changes in the
shape of active site pocket, as shown by MPH mutants modelled
with SWISS-MODEL (Fig. 6). Thus, in summary, the ease of altering
[
13] Y.-S. Jeong, S.-L. Choi, H.-H. Kyeong, J.-H. Kim, E.-J. Kim, J.-G. Pan, E. Rha, J.J.
Song, S.-G. Lee, H.-S. Kim, Des. Sel. 25 (2012) 725–731.
[
14] J. Liu, Y. Boucher, H.W. Stokes, D.L. Ollis, Protein Exp. Purif. 47 (2006) 258–263.
[15] H.-K.S. Leiros, P.S. Borra, B.O. Brandsdal, K.S.W. Edvardsen, J. Spencer, T.R.
Walsh, O. Samuelsen, Antimicrob. Agents Chemother. 56 (2012) 4341–4353.
[
16] A. Carfi, E. Duée, M. Galleni, J.M. Frère, O. Dideberg, Sect. D: Biol. Crystallogr.
D54 (1998) 313–323.
its substrate specificity make MPH
a strong candidate for
bioremediation and pharmaceutical applications.
[17] G. Garau, C. Bebrone, C. Anne, M. Galleni, J.-M. Frère, O. Dideberg, J. Mol. Biol.
45 (2005) 785–795.
3
[
[
[
18] J.H. Ullah, T.R. Walsh, I.A. Taylor, D.C. Emery, C.S. Verma, S.J. Gamblin, J.
Spencer, J. Mol. Biol. 284 (1998) 125–136.
19] A.D. Cameron, M. Ridderström, B. Olin, B. Mannervik, Structure 7 (1999) 1067–
Acknowledgments
1078.
This work was financially supported by the Australian Research
Council, Discovery Projects Scheme (DP120104263). T.-K.N. is
supported by an Australian National University PhD Scholarship
and an Australian National University Research Scholarship. G.S.
also acknowledges the receipt of an ARC Future Fellowship
20] V.A. Campos-Bermudez, N.R. Leite, R. Krog, A.J. Costa-Filho, F.C. Soncini, G.
Oliva, A.J. Vila, Biochemistry 46 (2007) 11069–11079.
[21] M.H. Kim, W.-C. Choi, H.O. Kang, J.S. Lee, B.S. Kang, K.-J. Kim, Z.S. Derewenda,
T.-K. Oh, C.H. Lee, J.-K. Lee, Proc. Natl. Acad. Sci. U.S.A. 102 (2005) 17606–
1
7611.
22] B. Kostelecky, E. Pohl, A. Vogel, O. Schilling, W. Meyer-klaucke, J. Bacteriol. 188
2006) 1607–1614.
23] W.-C. Tseng, J.-W. Lin, T.-Y. Wei, T.-Y. Fang, Anal. Biochem. 375 (2008) 376–
78.
[
[
[
(
FT120100694).
(
3
Appendix A. Supplementary data
24] J. Sanchis, L. Fernández, J.D. Carballeira, J. Drone, Y. Gumulya, H. Höbenreich, D.
Kahakeaw, S. Kille, R. Lohmer, J.J.-P. Peyralans, J. Podtetenieff, S. Prasad, P. Soni,
A. Taglieber, S. Wu, F.E. Zilly, M.T. Reetz, Appl. Microbiol. Biotechnol. 81 (2008)
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.abb.2015.03.012.
3
87–397.
25] H. Zhao, L. Giver, Z. Shao, J.A. Affholter, F.H. Arnold, Nat. Biotechnol. 16 (1998)
58–261.
[26] B.J. Stevenson, J.-W. Liu, D.L. Ollis, Biochemistry 47 (2008) 3013–3025.
[
2
References
[
[
[
27] G. Fu, Z. Cui, T. Huang, S. Li, Protein Exp. Purif. 36 (2004) 170–176.
28] N. Guex, M.C. Peitsch, Electrophoresis 18 (1997) 2714–2723.
29] T. Schwede, J. Kopp, N. Guex, M. Peitsch, Nucleic Acids Res. 31 (2003) 3381–
[
[
1] POST (1998) Organophosphates. Parliamentary Office of Science and
Technology, 122.
2] F. Ely, J.-L. Foo, C.J. Jackson, L.R. Gahan, D.L. Ollis, G. Schenk, Curr. Top. Biochem.
Res. 9 (2007) 63–78.
3385.
[
[
[
[
30] K. Arnold, L. Bordoli, J. Kopp, T. Schwede, Bioinformatics 22 (2006) 195–201.
31] O. Trott, A.J. Olson, J. Comput. Chem. 31 (2010) 455–461.
32] A. Grosdidier, V. Zoete, O. Michielin, J. Comput. Chem. 32 (2011) 2149–2159.
33] A. Grosdidier, V. Zoete, O. Michielin, Nucleic Acids Res. 39 (2011) W270–
W277.
34] H. Yang, P.D. Carr, S.Y. McLoughlin, J.W. Liu, I. Horne, X. Qiu, C.M.J. Jeffries, R.J.
Russell, J.G. Oakeshott, D.L. Ollis, Protein Eng. 16 (2003) 135–145.
35] M.T. Reetz, D. Kahakeaw, R. Lohmer, ChemBioChem 9 (2008) 1797–1804.
36] H. Zhao, W. Zha, Nat. Protoc. 1 (2006) 1865–1871.
[
[
[
[
[
3] J. Jeyaratnam, World Health Stat. Q. 43 (1990) 139–144.
4] I.B. Wilson, Ann. N. Y. Acad. Sci. 135 (1966) 177–183.
5] I.B. Wilson, S. Ginsburg, Biochim. Biophys. Acta 18 (1955) 168–170.
6] A.A. Weinbroum, Br. Med. Bull. 72 (2004) 119–133.
7] F. Worek, H. Thiermann, L. Szinicz, P. Eyer, Biochem. Pharmacol. 68 (2004)
[
2
237–2248.
[
[
[
[
[
8] A. Ekkhunnatham, B. Jongsareejit, W. Yamkunthong, J. Wichitwechkarn, World
J. Microb. Biot. 28 (2012) 1739–1746.
9] R. Zhang, Z. Cui, X. Zhang, J. Jiang, J.-D. Gu, S. Li, Biodegradation 17 (2006) 465–
37] G. Gotthard, J. Hiblot, D. Gonzalez, M. Elias, E. Chabriere, PLoS One 8 (2013)
e77995.
4
72.