Chemical Mechanism of the Phosphotriesterase from Sphingobium sp. Strain TCM1, an Enzyme Capable of Hydrolyzing Organophosphate Flame Retardants

Article abstract of DOI:10.1021/jacs.5b12739

The mechanism of action of the manganese-dependent phosphotriesterase from Sphingobium sp. strain TCM1 that is capable of hydrolyzing organophosphate flame retardants was determined. The enzyme was shown to hydrolyze the R_P-enantiomer of O-meth

Full text of DOI:10.1021/jacs.5b12739

Communication
pubs.acs.org/JACS
Chemical Mechanism of the Phosphotriesterase from Sphingobium
sp. Strain TCM1, an Enzyme Capable of Hydrolyzing
Organophosphate Flame Retardants
Andrew N. Bigley,† Dao Feng Xiang,† Zhongjie Ren,‡ Haoran Xue,†,§ Kenneth G. Hull,†,§ Daniel Romo,†,§
and Frank M. Raushel*,†,‡
†Department of Chemistry, Texas A&M University, College Station, Texas 77843, United States
‡Department of Biochemistry & Biophysics, Texas A&M University, College Station, Texas 77843, United States
S
* Supporting Information
leaving group.10 Sb-PTE has been shown to bind manganese in
ABSTRACT: The mechanism of action of the manga-
nese-dependent phosphotriesterase from Sphingobium sp.
strain TCM1 that is capable of hydrolyzing organo-
phosphate flame retardants was determined. The enzyme
was shown to hydrolyze the RP-enantiomer of O-methyl O-
cyclohexyl p-nitrophenyl thiophosphate with net inversion
of configuration and without the formation of a covalent
reaction intermediate. These results demonstrate that the
enzyme catalyzes the hydrolysis of substrates by activation
of a nucleophilic water molecule for direct attack at the
phosphorus center.
the active site and the lack of a strong dependence on the pKa
of the leaving group suggests that the mechanism for the
delivery and removal of protons in the active sites of these two
enzymes is different.2
While relatively little is currently known about the protein
structure of Sb-PTE, homology modeling suggests that the
protein folds as a 7-bladed β-propeller.2 Two other known
phosphotrieserases, human paraoxonase 1 (PON1) and squid
diisopropyl fluorophosphatase (DFPase), are also known to
have a β-propeller fold, and structural alignments among Sb-
PTE, PON1 and DFPase suggest that the active sites of these
enzymes are similar.11,12 DFPase has been shown to utilize
covalent catalysis, with an active site aspartate serving as the
initial nucleophile resulting in a covalently bound intermediate,
which is subsequently hydrolyzed by an attack of water on the
carboxylate carbon of the aspartate side chain rather than the
phosphorus center.12 However, the overall stereochemical
outcome at phosphorus for the hydrolysis of phosphate esters
by either PON1 or DFPase has not apparently been
experimentally determined. Since Sb-PTE can hydrolyze a
broad range of toxic compounds that are not accessible as
substrates by Pd-PTE, PON1, or DFPase, it is important to
more clearly understand the molecular differences and
similarities in the chemical mechanisms for substrate hydrolysis
employed by these enzymes. Here we probe the mechanism of
hydrolysis of phosphate esters by Sb-PTE by determination of
the stereochemical course at phosphorus using a chiral
thiophosphate substrate and by measurement of 18O-incorpo-
ration into the product from solvent under single and multiple
turnover conditions.
he recently discovered phosphotriesterase (PTE) from
TSphingobium sp. strain TCM1 (Sb-PTE) is catalytically
active against a broad range of substrates including organo-
phosphate plasticizers and flame retardants.1,2 The more
commonly known neurotoxic organophosphate insecticides
and nerve agents utilize a labile bond to facilitate their
neurotoxic effects.3 The flame retardants and plasticizers, such
as triethylchlorophosphate (TECP) and tricresylphosphate
(TCP), are extremely stable compounds that persist for long
periods of time.4,5 The use of these compounds as additives in
plastics, foams and lubricants has led to significant concerns
about toxicity and widespread dispersal in the environment.
The well-known PTE from Pseudomonas diminuta (Pd-PTE)
lacks measurable catalytic activity against the organophospho-
rus flame retardants and plasticizers.2
Sb-PTE is able to efficiently hydrolyze organophosphate
triesters that Pd-PTE cannot hydrolyze.1,2 The relative rates of
hydrolysis of organophosphorus substrates by Pd-PTE exhibit a
strong dependence on the pKa of the leaving group, but the
substrate profile for Sb-PTE does not.6 For example, with Pd-
PTE the kcat/Km for the hydrolysis of substrates with a
phenolate leaving group is ∼6-orders of magnitude lower than
for a substrate with a p-nitrophenolate leaving group. The
difference in reaction rates with Sb-PTE is less than 2-fold, and
the best substrate currently known for Sb-PTE is triphenyl-
phosphate.2
There are a variety of reaction mechanisms that are possible
for the enzyme-catalyzed hydrolysis of organophosphate
triesters. The most straightforward mechanism utilizes the
direct nucleophilic attack of an activated water/hydroxide on
the phosphorus center (Scheme 1a). For substrates with highly
activated leaving groups, such as p-nitrophenol, it is also
possible for the initial nucleophilic attack to occur on the
aromatic ring of the leaving group with subsequent C−O bond
cleavage through a nucleophilic aromatic substitution mecha-
nism (not shown in detail), rather than breakage of the P−O
Pd-PTE utilizes a binuclear metal center to activate a bridging
hydroxide for direct nucleophilic attack on the phosphorus
center of the substrate.7−9 However, the active site of Pd-PTE
does not appear to provide a mechanism for protonation of the
Received: December 6, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/jacs.5b12739
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
Scheme 1. Potential Chemical Mechanisms for Sb-PTE
bond (Scheme 1b). Alternatively, the reaction may be initiated
by an enzyme-based nucleophile that results in the formation of
a covalent intermediate that is subsequently hydrolyzed by an
activated water/hydroxide (Scheme 1c). If the enzyme-based
nucleophile is the side chain carboxylate from either glutamate
or aspartate, then hydrolysis of the covalent intermediate can
occur either by nucleophilic attack at phosphorus or attack on
the carboxylate carbon proceeding through a typical tetrahedral
intermediate (Scheme 1d). These mechanisms can be
distinguished from one another by determination of the
stereochemical outcome at phosphorus and measurement of
the isotopic composition of the products when the reaction is
conducted in 18O-labeled water.
To determine whether the initial nucleophilic attack is
directed at the leaving group with C−O bond cleavage
(Scheme 1b), Sb-PTE was used to catalyze the hydrolysis of
paraoxon (diethyl p-nitrophenyl phosphate) in 18O-labeled
water under conditions of multiple turnovers. If C1 of the p-
nitrophenyl group is attacked by water/hydroxide then the
isolated p-nitrophenolate will have a nominal molecular mass of
140, rather than 138. Alternatively, if nucleophilic attack occurs
at the phosphorus center, then the diethyl phosphate product
will have a nominal mass of 155, rather than 153. As expected,
when the hydrolysis reaction was conducted in unlabeled water
the p-nitrophenolate was isolated with an m/z of 138 and
diethyl phosphate at an m/z of 153 (Figure 1a). However,
when the hydrolysis reaction was conducted under multiple
turnover conditions in ∼60% 18O-labeled water, the p-
nitrophenolate was produced with an m/z of 138, whereas
the diethyl phosphate was isolated as a mixture of species with
m/z values of 153 and 155 (Figure 1b). These results eliminate
the mechanism depicted in Scheme 1b. Therefore, the reaction
catalyzed by Sb-PTE occurs via the direct attack of a
nucleophile at the phosphorus center of the substrate.
Figure 1. Mass spectra of p-nitrophenol and diethlyphosphate from
the Sb-PTE catalyzed hydrolysis of paraoxon. (a) Control reaction
conducted in unlabeled water. (b) Multiple turnover conditions where
500 μM paraoxon was hydrolyzed with 50 μM Sb-PTE in ∼60% 18O-
labeled water. (c) Single turnover conditions in ∼60% 18O-labeled
water where 50 μM paraoxon was hydrolyzed with 150 μM Sb-PTE.
The peaks identified are unlabeled p-nitrophenolate (m/z = 138.02),
unlabeled diethylphosphate (m/z = 153.03), and 18O-labeled
diethylphosphate (m/z = 155.03).
outcome is predicted to be retention of configuration (Scheme
1c). However, if the initial nucleophilic attack is via the
carboxylate side chain of aspartate or glutamate, the net
stereochemical outcome is either retention or inversion,
depending on whether the mixed anhydride intermediate is
ultimately attacked at phosphorus (Scheme 1c) or carbon
(Scheme 1d).
The mechanism depicted in Scheme 1d can be interrogated
in two ways. If the reaction is conducted in 18O-labeled water
under single turnover conditions, then the recovered diethyl
phosphate will not contain an 18O-label. When we utilized a 3-
fold molar excess of Sb-PTE to hydrolyze paraoxon in ∼60%
18O-labeled water, the isotopic content of the recovered diethyl
phosphate was essentially the same as the solvent (Figure 1c).
The utilization of a side chain nucleophile from either
glutamate or aspartate was also probed by measurement of
18O-incorporation into the protein when hydrolysis of paraoxon
was conducted in >95% 18O-labeled water. After the substrate
had been consumed, the enzyme was proteolyzed with trypsin
and the resulting peptides analyzed by mass spectrometry. The
net peptide coverage was 87% and no peptide was identified
that had an M+2 or M+4 peak that differed significantly from
that of the unlabeled control experiment (Table S1). The only
peptides containing either glutamate or aspartate that were not
The attack at the phosphorus center can occur with an
enzyme-activated water molecule or an enzyme-based nucleo-
phile. If there is a direct attack by water, then the net
stereochemical outcome at phosphorus is inversion of
configuration (Scheme 1a). If the initial attack at phosphorus
is via a nucleophile such as serine then the stereochemical
identified by mass spectrometry were 181DPR183
,
373SQPKMRLVGEAK384, and 499IVDLR503. These results are
inconsistent with the mechanism depicted in Scheme 1d.
B
DOI: 10.1021/jacs.5b12739
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
The differentiation between the mechanisms depicted in
parts a and c of Scheme 1 was accomplished by determination
of the net stereochemical outcome at phosphorus using O-
methyl O-cyclohexyl O-(p-nitrophenyl) thiophosphate (1) as
the chiral substrate. Previously, we have shown that Pd-PTE
catalyzes the hydrolysis of chiral substrates with inversion of
configuration at phosphorus.7 We have also shown that for the
hydrolysis of chiral substrates such as compound 1 that wild-
type Pd-PTE will preferentially hydrolyze the RP-enantiomer
relative to the SP-enantiomer and that the stereochemical
preference for mutants such as G60A will be significantly
greater than 100:1 for the RP-enantiomer, relative to the SP-
enantiomer.13 Therefore, the hydrolysis of a racemic mixture of
compound 1 by the G60A mutant of Pd-PTE will result in the
formation of the SP-enantiomer of O-methyl O-cyclohexyl
thiophosphate (2) as depicted in Scheme 2.
racemic compound 1, only ∼50% of the substrate is hydrolyzed
after a period of ∼1000 s (blue trace). Addition of the Pd-PTE
mutant VRN-VQFL-I106G/L308S, a variant known to
preferentially hydrolyze the opposite enantiomer than G60A,
hydrolyzes the remaining SP-enantiomer of compound 1.14
When hydrolysis of racemic 1 was initiated by the addition of
34 nM Sb-PTE, the reaction stopped after ∼50% of the
substrate was consumed (Figure 2, red trace). Addition of the
G60A variant of Pd-PTE resulted in no further hydrolysis of the
substrate. This result demonstrates that Sb-PTE and the G60A
variant of Pd-PTE preferentially hydrolyzed the same
enantiomer of compound 1. When the VRN-VQFL-I106G/
L308S variant of Pd-PTE was added to the reaction mixture,
the remaining portion of the substrate was hydrolyzed, thus
confirming that Sb-PTE and G60A Pd-PTE have strong
stereochemical preferences for hydrolysis of the same
enantiomer of compound 1.
The stereochemistry of the product produced by the
hydrolysis of the RP-enantiomer of compound 1 by Sb-PTE
was determined by 31P NMR spectroscopy using the chiral shift
reagent (S)-α-methylbenzylamine.15 When racemic compound
1 was completely hydrolyzed by a combination of the G60A
and L7ep3aG variants of Pd-PTE, the racemic product, O-
methyl O-cyclohexyl thiophosphate (2) gave separate 31P NMR
resonances for each of the two enantiomers because of the
differential interactions with the chiral shift reagent. Resonances
appeared at 56.54 and 56.49 ppm (Figure 3a). When hydrolysis
of racemic 1 was initiated by the addition of the G60A variant
of Pd-PTE, only a single resonance was observed at 56.54 ppm,
due to the selective formation of the (SP)-O-methyl O-
cyclohexyl thiophosphate product (Figure 3b). The unreacted
substrate from the hydrolysis of racemic 1 by the G60A mutant
of Pd-PTE was isolated after ∼45% of the original reaction
mixture had been hydrolyzed. The substrate was subsequently
hydrolyzed by the L7ep3aG mutant of Pd-PTE. The 31P NMR
spectrum of the reaction products shows a major resonance at
56.49 ppm, which is attributed to the formation of (RP)-O-
methyl O-cyclohexyl thiolphosphate (2) and a minor peak at
56.54 ppm for the (SP)-enantiomer of compound 2 (Figure 3c).
When Sb-PTE was used to hydrolyze a racemic mixture of O-
methyl O-cyclohexyl O-(p-nitrophenyl) thiophosphate (1),
only a single resonance was observed at ∼56.54 ppm, which
corresponds to the formation of (SP)-O-methyl O-cyclohexyl
thiophosphate (Figure 3d). When this sample was mixed in a
1:1 ratio with the reaction product produced by the hydrolysis
of racemic 1 catalyzed by the G60A variant of Pd-PTE, only a
single resonance appeared, confirming that the product
catalyzed by Sb-PTE and the G60A variant of Pd-PTE are
the same (Figure 3e). The identity of the single reaction
product was further confirmed by spiking the reaction mixture
from the Sb-PTE- and G60A Pd-PTE-catalyzed reactions with
the product produced by the reaction catalyzed by the L7ep3aG
variant of Pd-PTE (Figure 3f).
Scheme 2. Hydrolysis of (RP)-O-Methyl O-Cyclohexyl O-(p-
Nitrophenyl) Thiophosphate (1) to (SP)-O-Methyl O-
Cyclohexyl Thiophosphate (2) by Pd-PTE
The stereoselective hydrolysis of racemic 1 by the G60A
mutant of Pd-PTE is graphically illustrated in Figure 2. After
the addition of 12 nM of this mutant to a 25 μM mixture of
Figure 2. Stereoselective hydrolysis of 25 μM O-methyl, O-cyclohexyl
p-nitrophenylthiophosphate (1) by Pd-PTE and Sm-PTE. The blue
trace illustrates the time course for the hydrolysis of 1 by the G60A
variant of Pd-PTE (12 nM), which selectively hydrolyzes the RP-
enantiomer, while the second phase is due to the addition of the VRN-
GVQF mutant of Pd-PTE (770 nM), which hydrolyzes the SP-
enantiomer. The red trace shows the time course for the hydrolysis of
the RP-enantiomer by Sb-PTE (34 μM). At time point A (time
=2600−2900 s) 12 nM of G60A Pd-PTE was added to the reaction
mixture and there was no further hydrolysis. At time point B (time
=2900 s) the VRN-GVQF mutant of Pd-PTE (770 nM) was added
resulting in hydrolysis of the SP-enantiomer.
Sb-PTE selectively hydrolyzes the RP-enantiomer of com-
pound 1 and catalyzes the formation of the SP-enantiomer of
the product 2. This result demonstrates that the enzyme-
catalyzed reaction proceeds with inversion of configuration at
phosphorus. The reaction mechanism must therefore involve
the activation of a solvent water molecule for the direct
nucleophilic attack on the phosphorus center and a covalent
reaction intermediate is not involved in substrate turnover.
In this Communication we have demonstrated that the
reaction catalyzed by Sb-PTE proceeds with a direct attack of
C
DOI: 10.1021/jacs.5b12739
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
Figure 3. 31P NMR spectra of the products of hydrolysis of racemic O-methyl O-cyclohexyl p-nitrophenylthiophosphate (1) in the presence of the
chiral shift agent (S)-α-methylbenzylamine in CDCl3. (a) Total hydrolysis of racemic 1 by Pd-PTE. (b) Hydrolysis product of racemic 1 catalyzed by
the G60A variant of Pd-PTE (RP-enantiomer of compound 2). (c) Product from the hydrolysis of compound 1 as catalyzed by the L7ep3aG variant
of Pd-PTE (SP-enantiomer of compound 2). (d) Product from the hydrolysis of racemic 1 catalyzed by Sb-PTE. (e) A mixture of the hydrolysis
products produced by the hydrolysis of racemic 1 by the G60A mutant of Pd-PTE and of Sb-PTE. (f) A mixture of hydrolysis products of the G60A
mutant of Pd-PTE, Sb-PTE, and the L7ep3aG-Pd mutant of PTE.
water on the phosphorus center that results in an inversion of
stereochemistry at phosphorus (Scheme 1a). This phospho-
triesterase is unusual in its ability to hydrolyze organo-
phosphate triesters with unactivated leaving groups. This
observation suggests that the catalytic mechanism for the
protonation of the leaving group in the transition state must be
significantly different from that observed with other PTEs, such
as the well-characterized one from Pseudomonas diminuta.
Efforts are currently underway to determine the three-
dimensional crystal structure by X-ray diffraction methods.
REFERENCES
■
(1) Abe, K.; Yoshida, S.; Suzuki, Y.; Mori, J.; Doi, Y.; Takahashi, S.;
Kera, Y. Appl. Environ. Microbiol. 2014, 80, 5866.
(2) Xiang, D. F.; Bigley, A. N.; Ren, Z.; Xue, H.; Hull, K. G.; Romo,
D.; Raushel, F. M. Biochemistry 2015, 54, 7539.
(3) Maxwell, D. M.; Brecht, K. M.; Koplovitz, I.; Sweeney, R. E. Arch.
Toxicol. 2006, 80, 756.
(4) Reemtsma, T.; Quintana, J. B.; Rodil, R.; Garcia-Lopez, M.;
Rodriguez, I. TrAC, Trends Anal. Chem. 2008, 27, 727.
(5) van der Veen, I.; de Boer, J. Chemosphere 2012, 88, 1119.
(6) Hong, S. B.; Raushel, F. M. Biochemistry 1996, 35, 10904.
(7) Lewis, V. E.; Donarski, W. J.; Wild, J. R.; Raushel, F. M.
Biochemistry 1988, 27, 1591.
(8) Samples, C. R.; Howard, T.; Raushel, F. M.; DeRose, V. J.
Biochemistry 2005, 44, 11005.
(9) Vanhooke, J. L.; Benning, M. M.; Raushel, F. M.; Holden, H. M.
Biochemistry 1996, 35, 6020.
ASSOCIATED CONTENT
■
S
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/jacs.5b12739.
(10) Aubert, S. D.; Li, Y.; Raushel, F. M. Biochemistry 2004, 43, 5707.
(11) Ben-David, M.; Elias, M.; Filippi, J. J.; Dunach, E.; Silman, I.;
Sussman, J. L.; Tawfik, D. S. J. Mol. Biol. 2012, 418, 181.
(12) Blum, M. M.; Lohr, F.; Richardt, A.; Ruterjans, H.; Chen, J. C. J.
Am. Chem. Soc. 2006, 128, 12750.
(13) Tsai, P. C.; Bigley, A.; Li, Y.; Ghanem, E.; Cadieux, C. L.;
Kasten, S. A.; Reeves, T. E.; Cerasoli, D. M.; Raushel, F. M.
Biochemistry 2010, 49, 7978.
Additional methods, and NMR and mass spectral data
(PDF)
AUTHOR INFORMATION
■
Corresponding Author
*raushel@tamu.edu
(14) Bigley, A. N.; Mabanglo, M. F.; Harvey, S. P.; Raushel, F. M.
Biochemistry 2015, 54, 5502.
(15) Mikolajczyk, M.; Omelanczuk, J.; Leitloff, M.; Drabowicz, J.;
Present Address
§H.X., K.G.H., and D.R.: Department of Chemistry and
Biochemistry, Baylor University, Waco, TX 76798.
Ejchart, A.; Jurczak, J. J. Am. Chem. Soc. 1978, 100, 7003.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported in part by the Defense Threat
Reduction Agency (HDTRA1-14-1-0004) in a grant to F.M.R.
The Natural Products LINCHPIN Laboratory at Texas A&M
University was supported by the College of Science and the
Office of the Vice President for Research.
D
DOI: 10.1021/jacs.5b12739
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX