Stereoselective hydrolysis of organophosphate nerve agents by the bacterial phosphotriesterase

Article abstract of DOI:10.1021/bi101056m

Organophosphorus compounds include many synthetic, neurotoxic substances that are commonly used as insecticides. The toxicity of these compounds is due to their ability to inhibit the enzyme acetylcholine esterase. Some of the most toxic organophosphates

Full text of DOI:10.1021/bi101056m

7978 Biochemistry 2010, 49, 7978–7987
DOI: 10.1021/bi101056m
Stereoselective Hydrolysis of Organophosphate Nerve Agents by the Bacterial
Phosphotriesterase†
Ping-Chuan Tsai,‡ Andrew Bigley,‡ Yingchun Li,‡ Eman Ghanem,‡ C. Linn Cadieux,§ Shane A. Kasten,§ Tony E. Reeves,§
Douglas M. Cerasoli,§ and Frank M. Raushel*,‡
§
‡Department of Chemistry, P.O. Box 30012, Texas A&M University, College Station, Texas 77842, and U.S. Army Medical
Research Institute of Chemical Defense, 3100 Ricketts Point Road, Aberdeen Proving Ground, Maryland 21010-5400
Received July 1, 2010; Revised Manuscript Received August 6, 2010
ABSTRACT: Organophosphorus compounds include many synthetic, neurotoxic substances that are commonly
used as insecticides. The toxicity of these compounds is due to their ability to inhibit the enzyme acetylcholine
esterase. Some of the most toxic organophosphates have been adapted for use as chemical warfare agents; the
most well-known are GA, GB, GD, GF, VX, and VR. All of these compounds contain a chiral phosphorus
center, with the SP enantiomers being significantly more toxic than the RP enantiomers. Phosphotriesterase
(PTE) is an enzyme capable of detoxifying these agents, but the stereochemical preference of the wild-type
enzyme is for the RP enantiomers. A series of enantiomerically pure chiral nerve agent analogues containing
the relevant phosphoryl centers found in GB, GD, GF, VX, and VR has been developed. Wild-type and
mutant forms of PTE have been tested for their ability to hydrolyze this series of compounds. Mutant forms
of PTE with significantly enhanced, as well as relaxed or reversed, stereoselectivity have been identified. A
number of variants exhibited dramatically improved kinetic constants for the catalytic hydrolysis of the more
toxic SP enantiomers. Improvements of up to 3 orders of magnitude relative to the value of the wild-type
enzyme were observed. Some of these mutants were tested against racemic mixtures of GB and GD. The
kinetic constants obtained with the chiral nerve agent analogues accurately predict the improved activity and
stereoselectivity against the authentic nerve agents used in this study.
Organophosphorus compounds have been utilized for more
than 50 years as insecticides for the protection of agricultural
crops (1), and similar compounds have been developed as
chemical warfare agents (2). The structures of these latter com-
pounds are presented in Scheme 1 and include tabun (GA), sarin
(GB), soman (GD), cyclosarin (GF), VX, and VR. GA has a
cyanide leaving group; the three remaining G agents (GB, GD,
and GF) have a fluoride leaving group, and the two versions of
VX have a thiolate leaving group. The toxicity of these organo-
phosphonates is due to the inactivation of acetylcholinesterase
(AChE),1 an enzyme that catalyzes the hydrolysis of acetylcho-
line at neural synapses, through the phosphonylation of an active
site serine residue (3). GA, GB, GF, VX, and VR contain a chiral
phosphorus center, and thus, each of these nerve agents has two
stereoisomers; soman has four stereoisomers because of an
additional chiral center within the pinacolyl substituent. The
enantiomers are differentially toxic; the SP stereoisomer of sarin
reacts with AChE approximately ∼104 times faster than the
RP stereoisomer, and the two SP stereoisomers of soman react
∼105 times faster than the two RP isomers. Similarly, the SP
stereoisomer of VX is ∼100-fold more toxic than the RP
stereoisomer (2).
The detoxification of organophosphorus nerve agents can be
catalyzed by organophosphate degrading enzymes such as
human paraoxonase 1 (PON1), squid DFPase, organopho-
sphorus acid anhydrolase (OPAA), and phosphotriesterase
(PTE). Human paraoxonase is capable of hydrolyzing GB and
GD, but the overall catalytic activity is relatively low (4). The
DFPase from Loligo vulgaris is able to hydrolyze GA, GB, GD,
GF, and DFP (diisopropyl fluorophosphate). The value of kcat
/
Km for the hydrolysis of DFP is ∼1.3 ꢀ 106 M-1 s-1, but the
catalytic activities for the hydrolysis of GB and GD are sig-
nificantly lower (5, 6). Organophosphorus acid anhydrolase from
Alteromonas sp. JD6.5 is capable of hydrolyzing a wide variety of
organophosphorus compounds, including GB, GD, and GF but
not VX (7). Phosphotriesterase was first isolated from soil
microbes (8). The best substrate identified to date for this enzyme
is the agricultural pesticide paraoxon, and the value of kcat/Km
approaches the diffusion-controlled limit of ∼108 M-1 s-1 (9).
The enzymatic reaction for the hydrolysis of paraoxon to
p-nitrophenol and diethyl phosphate is shown in Scheme 2.
The substrate specificity of PTE is quite broad, and this enzyme
is capable of hydrolyzing GA, GB, GD, GF, VR, and VX (10).
PTE is a homodimeric protein that contains a binuclear metal
center embedded within a (β/R)8-barrel structure and is a member
of the amidohydrolase superfamily (11). The native enzyme
contains two Zn2þ ions, and these metal ions can be substituted
with Cd2þ, Co2þ, Ni2þ, or Mn2þ without a loss of catalytic
activity (9). Previous investigations have identified three subsites
within the active site of PTE that help to define the substrate
specificity for this enzyme (12). The small pocket is defined by the
†This work was supported by National Institutes of Health Grant
GM 68550.
*To whom correspondence should be addressed. Telephone: (979)
845-3373. Fax: (979) 845-9452. E-mail: raushel@tamu.edu.
1Abbreviations: PTE, phosphotriesterase; AChE, acetylcholine ester-
ase; ITC, isothermal titration calorimetry; PON1, human paraoxonase 1;
OPAA, organophosphorus acid anhydrolase; DFP, diisopropyl fluoro-
phosphate.
r
pubs.acs.org/Biochemistry
Published on Web 08/11/2010
2010 American Chemical Society

Article
Biochemistry, Vol. 49, No. 37, 2010 7979
Scheme 1: Structures of Chemical Warfare Agents
Scheme 2: Hydrolysis of Paraoxon by Phosphotriesterase
MATERIALS AND METHODS
Materials. Escherichia coli BL21(DE3) and XL1-blue cells
were purchased from Stratagene. Expression plasmid pET20b(þ)
was obtained from Invitrogen. The bacterial phosphotriesterase
and the site-directed mutants (G60A, I106G, F132G, S308G,
I106G/H257Y, I106G/F132G/H257Y, I106A/F132A/H257Y,
I106A/H257Y/S308A, and H254G/H257W/L303T) were puri-
fied to homogeneity as previously described (13, 14, 18). The
H254Q/H257F mutant was isolated from a H254X/H257X li-
brary (18, 19), and the H257Y/L303T mutant was identified
within a H254X/H257X/L303T library (17). In each case, H254
and H257 were completely randomized to generate 400 possible
permutations with 20 different amino acids at each of the two
targeted sites. Nearly 1000 colonies from each library were isolated
and screened for catalytic activity. The H254Q/H257F mutant was
initially identified as being beneficial by screening with Demeton-
S, an analogue of VX. The H257Y/L303T mutant was isolated by
screening with the SP enantiomers of compounds 2 and 4.
side chains of Gly-60, Ile-106, Leu-303, and Ser-308. The large
pocket is formed by the interactions of His-254, His-257, Leu-271,
and Met-317. The leaving group pocket is surrounded by four
aromatic residues: Trp-131, Phe-132, Phe-306, and Tyr-309. The
substrate binding site of PTE is graphically presented in Figure 1.
Wild-type PTE is stereoselective for the hydrolysis of chiral
organophosphates (13), and the degree of stereoselectivity de-
pends on the substituents attached to the central phosphorus
core (14, 15). For example, wild-type PTE hydrolyzes the RP
enantiomer of phenyl 4-acetylphenyl methylphosphonate appro-
ximately 2 orders of magnitude faster than the SP enantiomer
(15). The stereoselectivity of PTE can be manipulated through
the mutation of residues that contact the substrate prior to
catalytic turnover. Thus, the G60A mutant hydrolyzes the
RP enantiomer of phenyl 4-acetylphenyl methylphosphonate
∼5 orders of magnitude faster than the SP enantiomer (15).
Unfortunately, the inherent stereoselectivity of wild-type PTE
does not match that of AChE because PTE preferentially
hydrolyzes the least toxic enantiomer of chiral organopho-
sphates (16). However, the stereoselectivity of PTE can be
inverted through multiple mutations to the small and large
pockets in the active site (15). For example, the I106G/F132G/
H257Y mutant catalyzes the hydrolysis of the SP enantiomer of
phenyl 4-acetylphenyl methylphosphonate ∼3 orders of magni-
tude faster than the RP enantiomer (15). This represents a change
in stereoselectivity of nearly 8 orders of magnitude relative to that
of the G60A mutant with only four amino acid changes.
Synthesis of Racemic Substrates. The corresponding alco-
hol (1 equiv) was dissolved in ethyl ether (∼100 mL) and cooled
in a dry ice/acetone bath. Butyllithium (1 equiv) was added to the
solution while the mixture was stirred. A solution of methylpho-
sphonic dichloride (1 equiv) in ethyl ether (100 mL) was cooled in
a dry ice/acetone bath for 10 min and added to the alcohol
solution, and then the mixture was stirred at room temperature
for an additional 4 h. One equivalent of 4-acetylphenol and
triethylamine (1.2 equiv) were added to the reaction mixture, and
the mixture was stirred for 8 h. The solution was washed with
water (2 ꢀ 50 mL and 1 ꢀ 10 mL) and subsequently dried with
anhydrous magnesium sulfate. After removal of the solvent, the
crude product was purified by silica gel chromatography to yield
the 4-acetylphenyl methylphosphonates (compounds 1-5) in
yields of 38-87%. The structures of compounds 1-5 were veri-
fied by mass spectrometry and 1H NMR and 31P NMR spectros-
In this paper, we describe the synthesis of chiral analogues of
GB, GD, GF, VX, and VR with a 4-acetylphenol leaving group
(Scheme 3). We have utilized these compounds to test a series of
mutant forms of PTE that have been previously shown to have
stereochemical preferences differing from those of the wild-type
enzyme (14, 15, 17). This analysis has allowed the identification of
variants enhanced in their ability to hydrolyze the more toxic SP
enantiomers, and selected mutants were tested with GB and GD
to assess the utility of the analogues in predicting the properties
against authentic nerve agents.
1
copy. The H and 31P NMR spectra were recorded on a Varian
Unity INOVA 300 spectrophotometer with residual CDCl3 as the
internal reference for 1H NMR and H3PO4 (85% in water) as the ex-
ternal reference for 31P NMR. Exact mass measurements were taken
on a PE Sciex APJ Qstar Pulsar mass spectrometer by the Labo-
ratory for Biological Mass Spectrometry at Texas A&M University.
Isolation of SP Enantiomers. The SP enantiomers of
compounds 1-5 were obtained by kinetic resolution of the

7980 Biochemistry, Vol. 49, No. 37, 2010
Tsai et al.
corresponding racemic compounds using the G60A mutant of
PTE. The reactions were conducted at room temperature in a
methanol/water solution (20%) with 50 mM CHES buffer (pH
9.0). The concentration of the racemic phosphonate was approxi-
mately 1.0 g/L. The reactions were monitored by measuring the
formation of 4-acetylphenol at 294 nm and were stopped when no
further increase in the absorbance was observed. The SP en-
antiomers were extracted from the reaction mixture with chloro-
form. The chloroform solutions were dried with sodium sulfate
and condensed to dryness to yield the final product. The SP
enantiomer of 1 was obtained in a yield of 83%, and the SP
enantiomers of 2-5 were obtained in yields greater than 91%.
The enantiomeric excess (ee) was determined using the relative
intensities of the 31P NMR signals in the presence of Fmoc-
Trp(Boc)-OH (19). The ee value for the isolation of the SP
enantiomer of 1 was 95%, and for compounds 2-5, the ee values
were 99% (20).
were conducted at room temperature in a methanol/water
solution (20%) containing 50 mM CHES buffer (pH 9.0). The
concentrations of the racemic phosphonates were ∼1.0 g/L. The
RP enantiomer of compound 1 was obtained using 4.2 mg of
the PTE mutant H257Y/L303T, and the isolated yield was 52%
with an ee value of 98%. The RP enantiomer of compound 2 was
obtained using 4.2 mg of the PTE mutant H257Y/L303T. The
yield was 74%, and the ee value was 98%. The RP enantiomer of
compound 3 was obtained using 7.2 mg of the mutant I106A/
H257Y/S308A. The yield was 48%, and the ee value was 99%.
The RPRC and RPSC enantiomers of compound 4 were obtained
using 6.3 mg of the mutant H257Y/L303T. The yields were 98
and 99%, respectively. The ee values were 99% for both the RPRC
and RPSC enantiomers of compound 4. The RP enantiomer of
compound 5 was obtained using 1.8 mg of the mutant I106G/
F132G/H257Y. The yield was 48%, and the ee value was 97%.
The ee values were determined on the basis of the relative
intensities of the 31P NMR signals in the presence of Fmoc-
Trp(Boc)-OH (20).
Isolation of RP Enantiomers. The RP enantiomers of
compounds 1-5 were obtained by kinetic resolution of the
racemic compounds using various PTE mutants. The reactions
Expression, Growth, and Preparation of Wild-Type and
Mutant Enzymes. The gene for the expression of PTE was
cloned between the NdeI and EcoRI sites of a pET20b(þ)
plasmid. The plasmids for the expression of the wild-type and
mutant enzymes were transformed into BL-21(DE3) cells (9). The
cells were inoculated in Luria-Bertani (LB) broth and grown
overnight at 37 °C. The overnight cultures were incubated in
Terrific Broth (TB) containing 100 μg/mL ampicillin and 1.0 mM
CoCl2 at 30 °C. The expression of PTE was induced by the
addition of 1.0 mM IPTG when the OD600 reached 0.4. The cells
were harvested at 4 °C by centrifugation after the culture was
allowed to reach the stationary phase after 36-42 h at 30 °C. All
subsequent steps were conducted at 4 °C. The isolated cells were
suspended with 10 mM HEPES (pH 8.5) and lysed with a 5 s
pulsed sonication for 24 min at 0 °C at the medium power setting
using a Heat System-Ultrasonics, Inc. (Farmington, NY), model
W830 ultrasonic processor. The lysed cell suspensions were
combined and centrifuged at 13000g for 15 min. The supernatant
fluid was decanted, and then a solution of protamine sulfate (2%,
w/v) was added dropwise over a 20 min period while the mixture
F
IGURE 1: Graphic representation of the binding pockets within
the active site of PTE. The small pocket consists of Gly-60, Ile-106,
Leu-303, and Ser-308. The large pocket consists of His-254, His-257,
Leu-271, and Met-317. The leaving group pocket is surrounded by
Trp-131, Phe-132, Phe-306, and-Tyr 309.
Scheme 3: Structures of Organophosphate Substrates

Article
Biochemistry, Vol. 49, No. 37, 2010 7981
was stirred until the protamine sulfate concentration reached
0.4%. The supernatant solution was subjected to ammonium
sulfate fractionation by the addition of solid ammonium sulfate
to 60% saturation (371 mg/mL) while the mixture was stirred for
30 min and then kept stirring for an additional 45 min. The pellet
was collected by centrifugation at 13000g for 30 min, and the
protein was recovered by dissolving the precipitate in 3 mL of
buffer. The protein was loaded onto a 5.0 cm ꢀ 150 cm gel
filtration column containing Ultrogel AcA 54 (IBF, Columbia,
MD) and eluted at a flow rate of 1.0 mL/min. The fractions were
pooled on the basis of the enzymatic activity and absorbance at
280 nm and then applied to a 5 cm ꢀ 25 cm column, containing
DEAE-Sephadex A-25 previously equilibrated with 10 mM
HEPES (pH 8.5). SDS-polyacrylamide gel electrophoresis de-
monstrated that all of the mutant enzymes were the same size as
the wild-type PTE and that the purity of all proteins was greater
than 95%.
Spectrophotometric Enzyme Assays. The kinetic constants
for each substrate were determined by monitoring the formation
of p-acetylphenol at 294 nm (ε = 7710 M-1 cm-1) at 30 °C using
a SpectraMax plate reader (Molecular Devices Inc., Sunnyvale,
CA). The assay mixtures contained 50 mM CHES (pH 9.0),
100 μM CoCl2, and various concentrations of substrates. The
reactions were initiated by the addition of enzyme. Because of the
limited solubility of compounds 4 and 5, 10% methanol (v/v) was
added to the assay mixtures.
Estimates of kcat/Km for racemic GB and GD were obtained by
transformation of the rate of heat applied for each ITC experi-
ment. The ITC instrument records the instantaneous rate of heat
applied to the reference cell to maintain thermal equilibrium at 5 s
intervals. The rate of heat released during the enzymatic hydro-
lysis of these substrates was obtained by subtracting the experi-
mental values from the baseline value at each time point. The
total heat given off as a function of time was obtained by manual
integration using a geometric approximation method according
to eq 2
t ¼i
X
Hi ¼
dUi ꢀ 5s
ð2Þ
t ¼o
where Hi is the total heat observed at time i and dUi is the
instantaneous rate of heat produced by the reaction at each time
period. The heat of complete hydrolysis of GB (-730 μcal) or
GD (-930 μcal) was determined according to eq 2 as the value
of Hi as the reaction catalyzed by variants of PTE went to
completion. The fractional hydrolysis catalyzed by individual
variants as a function of time was determined by dividing the
total heat observed at each time period by the total expected heat.
Plots of the fraction GB or GD hydrolyzed over time appear as
exponential time courses. Depending on the agent or variant
used, one or two phases could be observed and fit to eq 3 or 4,
respectively
F ¼ að1 - e- k t
Þ
ð3Þ
Enzymatic Assays for GB and GD. Wild-type PTE and the
variants G60A, S308G, H257Y/L303T, and H254G/H257W/
L303T were purified to homogeneity and stored at -80 °C prior
to use. The time courses for the enzymatic hydrolysis of GB and
GD were obtained by following the heat of reaction using a
MicroCal (Northampton, MA) iTC200 isothermal titration cal-
orimeter (ITC), with a 200 μL reaction cell. All reactions were
conducted by injection of the phosphonofluoridate into the
enzyme solution at 25 °C with a reference offset of 5 μcal/s, a
syringe stirring speed of 1000 rpm, a preinjection delay of 180 s,
and a 5 s recording interval. The enzymes were diluted into
10 mM potassium phosphate (pH 7.4) for GB or 30 mM potas-
sium phosphate (pH 7.4) for GD. The final concentration of GB
for all experiments was 250 μM. The final concentration of GD
for all experiments was 300 μM. The total heat of hydrolysis for
GB (50 nmol) was determined by following the reaction catalyzed
by 20 nM G60A and 80 nM H254G/H257W/L303T to comple-
tion. The total heat of hydrolysis of GD (60 nmol) was deter-
mined by following the reaction catalyzed by 200 nM wild-type
PTE, 200 nM G60A, and 600 nM H254G/H257W/L303T to
completion.
To isolate the deferential hydrolysis of individual enantiomers
with PTE variants, the hydrolysis reactions were conducted with
high and low concentrations of the enzyme. The hydrolysis
of racemic GD with each variant was also analyzed by gas
chromatography with a chiral separation column as previously
described (21). Reaction solutions were removed from the ITC
cell and mixed with an equal volume of dry ethyl acetate to
extract the remaining agent. Aliquots (1 μL) of the organic phase
were injected into the gas chromatograph for analysis.
1
F ¼ að1 - e- k tÞ þ bð1 - e- k t
Þ
ð4Þ
1
2
where F is the fraction hydrolyzed, a and b are the magnitudes of
the exponential phases, t is time, and k1 and k2 are the exponential
components for each phase. The magnitude of the exponential
components was allowed to float. Assuming the concentration of
substrate is less than Km, the value of kcat/Km was estimated from
the exponentials according to eq 5
kx=Et ¼ kcat=Km
ð5Þ
where kx is the exponential component and Et is the enzyme
concentration.
RESULTS
Selection of Mutants. Specific residues in the substrate bind-
ing pocket of PTE were targeted for mutagenesis. The smallest
residue in the small subsite of PTE, Gly-60, was mutated to
alanine to reduce the size of the small pocket. The small pocket
and adjacent areas were also systematically enlarged by mutation
of Ile-106, Ser-308, and Phe-132 to glycine or alanine. Other
mutations combined alterations of the small and large subsites
simultaneously. The catalytic activities of the wild-type and
mutant enzymes with the compounds presented in Scheme 3
were determined, and the kinetic parameters are listed in Tables 1
and 2. The stereoselectivities of these enzymes using the isolated
chiral enantiomers are listed in Table 3.
Catalytic Activity and Stereoselectivity of Wild-Type
PTE. The values of kcat and kcat/Km for wild-type PTE with
compounds 2-5 demonstrate that this enzyme preferentially
hydrolyzes the RP enantiomers of these organophosphonates.
There is, however, a slight preference for the SP enantiomer of
compound 1. As the size of the substituent attached to the
phosphorus center becomes larger, the preference for the RP
Data Analysis. The kinetic constants (kcat and kcat/Km) were
obtained from a fit of the data to eq 1
v=Et ¼ kcat½Aꢁ=ðKm þ ½AꢁÞ
ð1Þ
where v is the initial velocity, kcat is the turnover number, [A] is
the substrate concentration, and Km is the Michaelis constant.

7982 Biochemistry, Vol. 49, No. 37, 2010
Tsai et al.

Article
Biochemistry, Vol. 49, No. 37, 2010 7983
Table 3: Ratios of kcat/Km for the Hydrolysis of Chiral Substrates by the Wild-Type and Mutant Forms of PTE
RP:SP for 1
RP:SP for 2
RP:SP for 3
RPRC:RPSC:SPRC:SPSC for 4
RP:SP for 5
wild-type
1:2
22:1
84:1
18:1
11:1
34:1
1:2
25:1
7600:1
1:1
400:63:35:1
3600:700:20:1
38:5:1.5:1
64:8:16:1
340:73:17:1
3:1:29:1
760:1
23000:1
6:1
G60A
4:1
I106G
1:1
F132G
1:9
3:1
35:1
280:1
16:1
1:3
S308G
1:2
19:1
1:5
H254Q/H257F
H257Y/L303T
I106G/H257Y
I106G/F132G/H257Y
I106A/F132A/H257Y
I106A/H257Y/S308A
H254G/H257W/L303T
1:5
1:64
1:12
1:18
1:51
1:12
1:140
1:44
1:3
1:6
4:1:120:16
8:1:3:1
1:64
1:200
1:270
1:41
1:120
1:7
1:5
7:1:11:4
1:89
1:14
2:1
1:22
1:4
19:1:120:19
13:1:54:7
2:1:65:14
1:33
1:110
enantiomer becomes stronger. For example, with compound 2,
the ratio for R/S(kcat/Km) is 22, whereas for compound 5, the
enantiomeric preference increases to 760.
Enhancement of Stereoselectivity. Gly-60 was originally
mutated to an alanine residue in an attempt to decrease the size of
the cavity for the binding of substrates in the active site of PTE.
The values of kcat and kcat/Km for the slow SP enantiomers with
G60A are decreased up to 340-fold relative to those of the wild-
type enzyme. In addition, the kinetic constants for the RP enan-
tiomers of 4 and 5 are increased relative to those of the wild-type
enzyme. The highest values of kcat/Km were obtained for the two
RP enantiomers of 4. The G60A mutant has an enhanced stereo-
selectivity relative to that of the wild-type enzyme. The most
extreme case is for compound 5, where the RP enantiomer is
preferred by a factor of 2 ꢀ 104.
F
IGURE 2: Time course for the hydrolysis of racemic 5 using two
mutants of PTE. The total concentration of compound 5 was
determined by the complete hydrolysis using 0.1 M KOH. The
hydrolysis of the SP enantiomer of racemic 5 was initiated with
165 nM H254G/H257W/L303T PTE. After 10 min, the G60A
mutant of PTE was added to hydrolyze the RP enantiomer.
Relaxation of Stereoselectivity. The mutation of single
residues in the small or leaving group pockets of PTE to amino
acids with smaller side chains results in an active site that is
somewhat larger than that of the wild-type enzyme. In general,
this perturbation to the active site results in the relaxation of
stereoselectivity through an enhancement of the rate of the
initially slower SP enantiomers. For example, the I106G mutant
diminishes the stereoselectivity for compounds 1 and 3-5 relative
to wild-type PTE. The F132G mutant relaxes the stereoselectivity
for compounds 2-5 compared with wild-type PTE. Thus, single
mutations in the small or leaving group pockets to glycine
residues relax the stereoselectivity of the mutants relative to
wild-type PTE.
Reversal of the Stereoselectivity. The mutants with a
relaxed enantioselectivity contain an expanded small pocket.
Modification of the large and small pockets simultaneously
inverts the inherent stereoselectivity relative to that of wild-type
PTE. The mutants H257Y/L303T, I106G/H257Y, I106G/F132G/
H257Y, I106A/F132A/H257Y, and I106A/H257Y/S308A all
have residues in the small pocket (Ile-106, Phe-132, and Ser-308)
that are replaced with either glycine or alanine. In addition, these
mutants include changes to His-257 in the large pocket. For
example, wild-type PTE prefers the RP enantiomer of compound 3
by a factor of 25, whereas the I106A/F132A/H257Y mutant
inverts the stereoselectivity with compound 3 and prefers the SP
enantiomer by a factor of 260.
enantioselectivity relative to that of the wild-type enzyme, the
H254G/H257W/L303T mutant exhibits the highest overall cat-
alytic activity for the more toxic SP enantiomers of compounds 4
and 5. The values of kcat/Km for the hydrolysis of the SPRC and
SPSC stereoisomers of compound 4 are enhanced 74- and 540-
fold, respectively, relative to that of the wild-type enzyme. The
most notable case is that for the SP enantiomer of compound 5
with a kcat/Km that is more than 3 orders of magnitude higher
than that of the wild-type enzyme. The reversal in stereoselecti-
vity, relative to that of the G60A mutant, for the hydrolysis of
racemic 5 is shown in Figure 2. The addition of H254G/H257W/
L303T hydrolyzes approximately 50% of the total material in
∼5 min, and then the addition of G60A hydrolyzes the remaining
material.
Hydrolysis of Racemic GB and GD. Wild-type PTE is
known to prefer the less toxic RP enantiomers of nerve agents GB
and GD (16). PTE mutants with differing catalytic properties
with the analogues to these compounds (2 and 4, respectively)
were selected for testing with racemic GB and GD. Because of
regulatory considerations on the maximum allowable concentra-
tions of these two nerve agents, the enzymatic reactions were
followed via ITC for the complete hydrolysis at a single fixed
concentration of agent. The PTE mutants were tested under
identical reaction conditions for their ability to hydrolyze 250 μM
GB and 300 μM GD in 200 μL reaction volumes. The values of
kcat/Km for each mutant were determined from the exponential
phases for hydrolysis of the racemic mixtures and are reported in
The H254G/H257W/L303T mutant contains an extended
small pocket and a modified large pocket. This mutant has
a more distinct effect on the reversal of stereoselectivity than
the other mutants for the hydrolysis of compounds 1 and 5. For
these compounds, the SP enantiomers are preferred by factors
of 140 and 110, respectively. In addition to an inversion of

7984 Biochemistry, Vol. 49, No. 37, 2010
Tsai et al.
Table 4: Values of kcat/Km (M-1 s-1) for Wild-Type and Mutant Forms of PTE against Nerve Agents GB and GD
GB
GD
kcat/Km2
compd
wild-type
k
cat/Km1a
kcat/Km2
kcat/Km1
enantiomeric preferenceb
(9 ( 3) ꢀ 104
(2.6 ( 0.2) ꢀ 103
(2.5 ( 0.3) ꢀ 104
(9.1 ( 0.5) ꢀ 103
(2.2 ( 0.2) ꢀ 103
(8.9 ( 0.9) ꢀ 104
RPSC,RPRC,SPRC . SPSC
RPSC,RPRC > SPSC,SPRC
RPRC,RPSC > SPRC . SPSC
SPSC,SPRC . RPSC,RPRC
SPSC,SPRC > RPSC,RPRC
G60A
(1.2 ( 0.1) ꢀ 105
(3.9 ( 0.9) ꢀ 104
(1.5 ( 0.1) ꢀ 104
(3 ( 1) ꢀ 105
(2.66 ( 0.01) ꢀ 104
(2.2 ( 0.3) ꢀ 103
(8 ( 2) ꢀ 102
S308G
H254G/H257W/L303T
H275Y/L303T
(9.7 ( 0.1) ꢀ 102
(3.65 ( 0.01) ꢀ 104
(6.8 ( 0.2) ꢀ 103
akcat/Km1 and kcat/Km2 were determined from the first and second exponential phases, respectively, during the reaction of GB or GD as followed
by ITC. bFirst phase > second phase . not observed.
Table 4. Representative time courses for the hydrolysis of GB are
given in Figure 3.
The fractional hydrolysis of GB catalyzed by wild-type PTE
fits to a single exponential (Figure 3a and Table 4). Similarly, the
ITC traces for the S308G mutant with GB (data not shown) can
be fit to a single exponential (Table 4). The time courses for three
other mutants (G60A, H254G/H257W/L303T, and H257Y/
L303T) could not be fit to a single exponential, indicating that
the two enantiomers were hydrolyzed at different rates. G60A
has a value of kcat/Km for the first phase that is similar to that
observed with the wild-type enzyme and a second phase that is
∼5-fold slower (Figure 3b and Table 4). The value of kcat/Km for
the H254G/H257W/L303T mutant, obtained from the first
phase, is reduced in value relative to that of the wild-type enzyme,
and a second phase that is reduced by ∼2 orders of magnitude
(data not shown). H257Y/L303T has the highest value of kcat/Km
observed for the hydrolysis of GB (3 ꢀ 105 M-1 s-1) and a clear
second phase that was approximately 10-fold slower (Figure 3c).
The racemic GD used in these experiments contained four
diastereomers that can be separated by chiral gas chromatogra-
phy (Figure 4). In all of the enzymatic time courses, only one or
two phases could be clearly distinguished from one another.
Representative plots are given in Figure 5. To determine which
enantiomers were being degraded in each phase, the substrates
remaining after selected ITC experiments were extracted with
ethyl acetate and separated by gas chromatography to identify
the remaining isomers. At high enzyme concentrations, wild-type
PTE exhibited a single-exponential phase with an estimated value
of kcat/Km that is reduced from that obtained with GB (Figure 5a
and Table 4). The gas chromatographic separation indicated that
the two RP enantiomers were almost completely hydrolyzed
along with the SPRC enantiomer, leaving only the SPSC enantio-
mer in the reaction mixture (Figure 5b). The G60A variant
exhibited two phases in the ITC experiment, with the faster phase
having a kcat/Km value nearly 1 order of magnitude higher than
that obtained with wild-type PTE (Figure 5c). Gas chromato-
graphic analysis revealed that the faster phase corresponded
primarily to the degradation of the two RP enantiomers, while the
two SP enantiomers only begin to be degraded in the second
phase (Figure 5d). The SPSC diastereomer is hydrolyzed more
slowly than the SPRC stereoisomer by G60A. The H254G/
H257W/L303T mutant exhibited a single-exponential phase with
a value of kcat/Km similar to that observed with wild-type PTE
(Figure 5e). However, the gas chromatographic separation in-
F
IGURE 3: Hydrolysis of 250 μM racemic GB followed by ITC. All
dicates that the diastereomeric preference is inverted because the
SPSC and SPRC diastereomers are hydrolyzed significantly faster
than the RPRC and RPSC stereoisomers (Figure 5f). The H257Y/
L303T mutant has the highest value of kcat/Km for the hydrolysis
of GD with the hydrolysis of the two SP enantiomers in the first
phase and the two RP enantiomers in the second phase
reactionswere initiated byinjection of5μL ofGBintoavolume of200
μL and followed with a MicroCal iTC200 isothermal calorimeter. (A)
Hydrolysis of racemic GB by 30 nM wild-type PTE as a function of
time. The dataarefit tosingle exponential. (B) Hydrolysis of GBby60
nM G60A as a function of time. The data are fit to the sum of two
exponentials. (C) Hydrolysis of racemic GB by 10 nM H257Y/L303T
as a function of time. The data are fit to the sum of two exponentials.

Article
Biochemistry, Vol. 49, No. 37, 2010 7985
(Figure 5g). The S308G mutant also exhibited two phases in time
course experiments (data not shown), with the first phase having a
enhanced in the catalytic hydrolysis of the most toxic stereo-
isomers.
The inherent stereoselectivity of wild-type PTE for the RP
enantiomers of compounds 1-5 increases with the size of the
large substituent attached directly to the phosphorus center. Site
specific mutations to the active site of wild-type PTE have
demonstrated that the kinetic constants for the hydrolysis of
individual compounds can be significantly enhanced by the direct
manipulation of the relative size of the large and small subpockets
within the active site of this enzyme. For example, substitution of
an alanine for a glycine residue in the small pocket significantly
enhances the stereochemical preference, R/S(kcat/Km), for all of the
compounds tested. With this mutant, the stereochemical selec-
tivity is enhanced while maintaining a relatively high turnover for
the preferred RP enantiomer.
A significant number of the variants tested for this investiga-
tion exhibited an enhanced ability to degrade the SP stereo-
isomers of the nerve agent analogues. The dramatic improve-
ments in the catalytic constant, kcat/Km, for the SP enantiomers of
compounds 2-5 are graphically illustrated in Figure 6. The best
variant of PTE with the SP enantiomers of compounds 2 and 3,
H254Q/H257F, is enhanced by nearly 2 orders of magnitude
relative to the wild-type enzyme. For compounds 4 and 5, the best
mutant tested is H254G/H257W/L303T. With the SPRC enan-
tiomer of analogue 4, this mutant is enhanced by a factor of 74,
and for the SPSC isomer, the enhancement, relative to the wild-
type enzyme, is 540. For compound 5, the enhancement is even
greater; the H254G/H257W/L303T mutant hydrolyzes the SP
enantiomer of analogue 5 more than 3 orders of magnitude faster
than the wild-type enzyme. For compounds 3-5 the second best
mutant identified to date is the H257Y/L303T mutant.
k
cat/Km value enhanced relative to that of the wild-type enzyme
and a second phase reduced by ∼1 order of magnitude.
DISCUSSION
Phosphotriesterase possesses broad substrate specificity and a
tunable stereoselectivity that can be harnessed for the detoxifica-
tion of chemical warfare agents such as GA, GB, GD, GF, VX,
and VR. These nerve agents contain a chiral phosphoryl center,
and the toxicity of the individual stereoisomers differs substan-
tially. Wild-type PTE preferentially hydrolyzes the less toxic RP
stereoisomers of these organophosphorus nerve agents and the
analogues prepared for this investigation (16, 22). Compounds
1-5 were developed as readily prepared analogues of the G and V
agents that can be isolated with chiral purities of 95-99% ee.
These compounds can also be used in high-throughput screening
trials for the rapid identification of enzyme mutants that are
The utilization of nerve agent analogues simplifies the identi-
fication of enzyme variants with the desired ability to hydrolyze
the most toxic enantiomers of GB, GD, and GF. However, it is
FIGURE 4: Separation of stereoisomers of racemic GD by gas chro-
matography.
F
IGURE 5: Hydrolysis of 300 μM GD. All reactions were conducted in a volume of 200 μL and initiated by the injection of 6 μL of GD into the
enzyme solution. Hydrolysis of GD (panels A, C, E, and G) was followed by monitoring the heat liberated as a function of time. Following each
experiment, the unreacted fraction of GD was analyzed by gas chromatography with a chiral separation column (panels B, D, F, and H). The
order of elution from the gas chromatography column was as follows: SPRC (13 min), RPRC (14.1 min), SPSC (14.5 min), RPSC (15 min). (A)
HydrolysisofGDby800 nMwild-type PTEasa functionoftime. The lineisfittoa singleexponential. (B) Analysisofthe GDremainingfollowing
reaction with wild-type PTE. (C) Hydrolysis of GD by 200 nM G60A as a function of time. The line is a fit to the sum of two exponentials. (D)
Analysis of unreacted GD following reaction with G60A. (E) Hydrolysis of GD by 600 nM H254G/H257W/L303T as a function of time. The line
represents a fit of the data to a singleexponential. (F) Analysis of the uneactedGD remaining following reaction with H254G/H257W/L303T. (G)
Hydrolysis of GD by 100 nM H257Y/L303T as a function of time. The line is a fit to the sum of two exponentials. (H) Analysis of the unreacted
GD remaining following reaction with H257Y/L303T.

7986 Biochemistry, Vol. 49, No. 37, 2010
Tsai et al.
GD were determined for a select group of site-specific mutations
to the active site of PTE. The values of kcat/Km reported in Table 4
were obtained from the time courses for the hydrolysis of the
racemic mixtures at a single fixed concentration of GB or GD. It
must be noted, however, that the values reported in some cases
are derived from the composite hydrolysis of multiple enantio-
mers and thus represent a weighted average of the kinetic
constants. It is also assumed in these studies that the initial
substrate concentrations for these assays are less than the Km.
However, all of the mutants were tested under identical condi-
tions, allowing for a direct comparison of the observed kinetic
properties, and thus, this assumption will have relatively little
influence on the overall conclusions.
Under the experimental conditions employed for this investi-
gation, only a single first-order rate constant for the hydrolysis of
racemic GB could be obtained with wild-type PTE. The mutant
G60A, which shows more stereochemical selectivity against the
analogue of GB (compound 2), clearly shows differential rates of
hydrolysis using racemic GB. The value of kcat/Km for the faster
enantiomer (presumably the RP enantiomer) is approximately
4-fold lower than the value of kcat/Km for the RP enantiomer of
compound 2. The mutant S308G did not display a measurable
rate differential for hydrolysis of the two enantiomers of GB. The
single mutation in this variant is located in the leaving group
pocket, and it is likely that the significant difference in the leaving
group between analogue 2 and GB is a contributing cause for the
differing characteristics. The remaining variants (H257Y/L303T
and H254G/H257W/L303T) exhibit greater stereochemical selec-
tivity with compound 2 than the wild-type enzyme. These two
mutants also exhibited differential rates of hydrolysis of the two
stereoisomers of GB that mirror the differences in the values of
kcat/Km with analogue 2. The best enzyme identified for the
hydrolysis of GB is H257Y/L303T. The value of kcat/Km for the
fastest enantiomer of GB (3 ꢀ 105 M-1 s-1) matches quite
well with the value of kcat/Km for the SP enantiomer of analogue 2
(1.1 ꢀ 105 M-1 s-1).
The experiments with GD are complicated by the presence of
four stereoisomers. Under the conditions used, no more than two
distinct phases could be observed in the time courses for the
enzymatic hydrolysis of GD. Gas chromatography was utilized
concurrently with the ITC measurements to determine the
specific stereoisomers that were preferentially hydrolyzed in the
early phases of these reactions. The time course for the hydrolysis
of GD by wild-type PTE could be fit to a rate equation with a
single exponential that appears to be due to the hydrolysis of
three of the four stereoisomers. The stereoselective preferences
measured with analogue 4 show an only 10-fold difference in the
values of kcat/Km among the three fastest stereoisomers. The
slowest stereoisomer of compound 4 (SPSC) is hydrolyzed by
wild-type PTE ∼35-fold slower than the next fastest stereoisomer
(SPRC). The diastereomeric preference determined by gas chro-
matography indicated that the two RP stereoisomers, along with
the SPRC enantiomer of GD, were the preferred substrates.
Differential rates of hydrolysis were detected in the hydrolysis
of racemic GD by mutant G60A. Hydrolysis of the two RP
enantiomers was preferred. The measured values of kcat/Km for
the two RP enantiomers with G60A were significantly improved
over that of the wild-type enzyme as was also observed with
analogue 4. The hydrolysis reaction catalyzed by S308G exhib-
ited two phases in the time courses, with the first phase due to the
two RP stereoisomers and the second phase dominated by the
hydrolysis of the SPRC enantiomer. The kinetic measurements
F
IGURE 6: Bar graph illustrating changes in the value of kcat/Km for
the SP enantiomers of compounds 2-5 catalyzedbythe wild-type and
mutant forms of PTE.
important that the nerve agent stereoselectivity of mutant
enzymes be determined by assessing hydrolysis of authentic nerve
agent enantiomers. The kcat/Km values for nerve agents GB and

Article
Biochemistry, Vol. 49, No. 37, 2010 7987
with the nerve agent analogues indicated that kcat/Km should be
enhanced over that of wild-type PTE, which was observed. The
H254G/H257W/L303T mutant exhibited a single phase during
the hydrolysis of racemic GD. This phase is dominated by the
hydrolysis of the two SP stereoisomers. The observed value of
kcat/Km was in rough agreement with that obtained for the
hydrolysis of the racemic GD by the wild-type enzyme, but this
mutant is significantly enhanced in the hydrolysis of the two most
toxic stereoisomers.
The reaction catalyzed by the H257Y/L303T mutant exhibited
two exponential phases during the time courses for the hydrolysis
of racemic GD, with the faster phase being the hydrolysis of the
two SP diastereomers and the slower phase being due to the
hydrolysis of the two RP diastereomers. The values of kcat/Km for
the hydrolysis of racemic GD were significantly higher than what
was predicted from the investigations with the nerve agent
analogues.
The use of the nerve agent analogues greatly simplifies the
identification of PTE variants with improved ability to hydrolyze
the more toxic stereoisomers of the authentic nerve agents GB
and GD. While the results with analogues and the specific nerve
agents are not identical, these results indicate that the findings
with analogues accurately predict improvements against GB and
GD. Comparison of the results obtained with the nerve agents
GB and GD and their respective analogues indicates that these
compounds may provide a more stringent test of the stereochem-
ical preferences of PTE based on the phosphoryl center. The
changes in kcat/Km obtained with the analogues are largely
reflected in the values of kcat/Km found with the authentic nerve
agents. These strategies have allowed the isolation of variants
that are significantly improved for the hydrolysis of both GB and
GD and with a stereochemical preference for the more toxic SP
stereoisomers.
7. Hill, C. M., Li, W. S., Cheng, T. C., DeFrank, J. J., and Raushel, F. M.
(2001) Stereochemical specificity of organophosphorus acid anhy-
drolase toward p-nitrophenyl analogs of soman and sarin. Bioorg.
Chem. 29, 27–35.
8. Harper, L. L., Mcdaniel, C. S., Miller, C. E., and Wild, J. R. (1988)
Dissimilar plasmids isolated from Pseudomonas diminuta Mg and a
Flavobacterium Sp (ATCC-27551) contain identical Opd genes. Appl.
Environ. Microbiol. 54, 2586–2589.
9. Omburo, G. A., Kuo, J. M., Mullins, L. S., and Raushel, F. M. (1992)
Characterization of the zinc binding site of bacterial phosphotriester-
ase. J. Biol. Chem. 267, 13278–13283.
10. Kolakowski, J. E., DeFrank, J. J., Harvey, S. P., Szafraniec, L. L.,
Beaudry, W. T., Lai, K. H., and Wild, J. R. (1997) Enzymatic
hydrolysis of the chemical warfare agent VX and its neurotoxic
analogues by organophosphorus hydrolase. Biocatal. Biotransform.
15, 297–312.
11. Benning, M. M., Shim, H., Raushel, F. M., and Holden, H. M. (2001)
High resolution X-ray structures of different metal-substituted forms
of phosphotriesterase from Pseudomonas diminuta. Biochemistry 40,
2712–2722.
12. Vanhooke, J. L., Benning, M. M., Raushel, F. M., and Holden, H. M.
(1996) Three-dimensional structure of the zinc-containing phospho-
triesterase with the bound substrate analog diethyl 4-methylbenzyl-
phosphonate. Biochemistry 35, 6020–6025.
13. Chen-Goodspeed, M., Sogorb, M. A., Wu, F. Y., Hong, S. B., and
Raushel, F. M. (2001) Structural determinants of the substrate and
stereochemical specificity of phosphotriesterase. Biochemistry 40,
1325–1331.
14. Chen-Goodspeed, M., Sogorb, M. A., Wu, F. Y., and Raushel, F. M.
(2001) Enhancement, relaxation, and reversal of the stereoselectivity
for phosphotriesterase by rational evolution of active site residues.
Biochemistry 40, 1332–1339.
15. Nowlan, C., Li, Y. C., Hermann, J. C., Evans, T., Carpenter, J.,
Ghanem, E., Shoichet, B. K., and Raushel, F. M. (2006) Resolution of
chiral phosphate, phosphonate, and phosphinate esters by an enan-
tioselective enzyme library. J. Am. Chem. Soc. 128, 15892–15902.
16. Li, W. S., Lum, K. T., Chen-Goodspeed, M., Sogorb, M. A., and
Raushel, F. M. (2001) Stereoselective detoxification of chiral sarin
and soman analogues by phosphotriesterase. Bioorg. Med. Chem. 9,
2083–2091.
17. Lum, K. T. (2004) Directed evolution of phosphotriesterase: Towards
the efficient detoxification of sarin and soman. Ph.D. dissertation,
Department of Toxicology, Texas A&M University, College Station,
TX.
18. Hill, C. M., Li, W. S., Thoden, J. B., Holden, H. M., and Raushel,
F. M. (2003) Enhanced degradation of chemical warfare agents
through molecular engineering of the phosphotriesterase active site.
J. Am. Chem. Soc. 125, 8990–8991.
19. Reeves, T. E., Wales, M. E., Grimsley, J. K., Li, P., Cerasoli, D. M.,
and Wild, J. R. (2008) Balancing the stability and the catalytic
specificities of OP hydrolases with enhanced V-agent activities.
Protein Eng., Des. Sel. 21, 405–412.
REFERENCES
1. Ecobichon, D. J. (2001) Casarett and Doull’s Toxicology: The basic
science of poisons, 6th ed., pp 763-810, McGraw-Hill, New York.
2. Benschop, H. P., and Dejong, L. P. A. (1988)Nerve agent stereoisomers:
Analysis, isolation, and toxicology. Acc. Chem. Res. 21, 368–374.
3. Worek, F., Szinicz, L., Eyer, P., and Thiermann, H. (2005) Evaluation
of oxime efficacy in nerve agent poisoning: Development of a kinetic-
based dynamic model. Toxicol. Appl. Pharmacol. 209, 193–202.
4. Josse, D., Xie, W., Renault, F., Rochu, F., Schopfer, L. M., Masson,
P., and Lockridge, O. (1999) Identification of Residues Essential for
Human Paraoxonase (PON1) Arylesterase/Organophosphatase
Activities. Biochemistry 38, 2816–2825.
20. Li, Y., and Raushel, F. M. (2007) Differentiation of chiral phosphorus
enantiomers by P-31 and H-1 NMR spectroscopy using amino acid
derivatives as chemical solvating agents. Tetrahedron: Asymmetry 18,
1391–1397.
21. Yeung, D. T., Smith, J. R., Sweeney, R. E., Lenz, D. E., and Cerasoli,
D. M. (2007) Direct detection of stereospecific soman hydrolysis by
wild-type human serum paraoxonase. FEBS J. 274, 1183–1191.
22. Lum, K. T., Huebner, H. J., Li, Y. C., Phillips, T. D., and Raushel, F. M.
(2003) Organophosphate nerve agent toxicity in Hydra attenuata. Chem.
Res. Toxicol. 16, 953–957.
5. Hartleib, J., and Ruterjans, H. (2001) High-yield expression, purifica-
tion, and characterization of the recombinant diisopropylfluoropho-
sphatase from Loligo vulgaris. Protein Expression Purif. 21, 210–219.
6. Wang, F., Xiao, M. Z., and Mu, S. F. (1993) Purification and
properties of
Todarodes-Pacificus-Steenstrup. J. Biochem. Toxicol. 8, 161–166.
a
diisopropyl-fluorophosphatase from squid