Structural flexibility enhances the reactivity of the bioremediator glycerophosphodiesterase by fine-tuning its mechanism of hydrolysis

Article abstract of DOI:10.1021/ja903534f

The glycerophosphodiesterase from Enterobacter aerogenes (GpdQ) belongs to the family of binuclear metallohydrolases and has attracted recent attention due to its potential in bioremediation. Formation of a catalytically competent binuclear center is prom

Full text of DOI:10.1021/ja903534f

Published on Web 08/04/2009
Structural Flexibility Enhances the Reactivity of the
Bioremediator Glycerophosphodiesterase by Fine-Tuning Its
Mechanism of Hydrolysis
Kieran S. Hadler,† Natasˇa Mitic´,† Fernanda Ely,† Graeme R. Hanson,‡
Lawrence R Gahan,† James A. Larrabee,§ David L. Ollis,| and Gerhard Schenk*,†
School of Chemistry and Molecular Biosciences, The UniVersity of Queensland, St. Lucia,
Queensland, 4072, Australia, Centre for Magnetic Resonance, The UniVersity of Queensland, St.
Lucia, Queensland, 4072, Australia, Department of Chemistry and Biochemistry, Middlebury
College, Middlebury, Vermont 05753, and Research School of Chemistry, Australian National
UniVersity, Canberra, ACT, 0200, Australia
Received May 1, 2009; E-mail: schenk@uq.edu.au
Abstract: The glycerophosphodiesterase from Enterobacter aerogenes (GpdQ) belongs to the family of
binuclear metallohydrolases and has attracted recent attention due to its potential in bioremediation.
Formation of a catalytically competent binuclear center is promoted by the substrate (Hadler et al. J. Am.
Chem. Soc. 2008, 130, 14129). Using the paramagnetic properties of Mn(II), we estimated the Kd values
for the metal ions in the R and ꢀ sites to be 29 and 344 µM, respectively, in the absence of a substrate
analogue. In its presence, the affinity of the ꢀ site increases substantially (Kd ) 56 µM), while that of the
R site is not greatly affected (Kd ) 17 µM). Stopped-flow fluorescence measurements identified three distinct
phases in the catalytic turnover, associated with the initial binding of substrate to the active site (kobs1), the
assembly of a catalytically active binuclear center (kobs2), and subsequent slower structural rearrangements
to optimize catalysis (kobs3). These three phases depend on the concentration of substrate ([S]), with kobs1
and kobs2 reaching maximum values at high [S] (354 and 38 s-1, respectively), whereas kobs3 is reduced as
[S] is increased. The kcat for the hydrolysis of the substrate bis(para-nitrophenyl) phosphate (∼1 s-1) gradually
increases from the moment of initiating the reaction, reaching a maximum when the structural change
associated with kobs3 is complete. This structural change is mediated via an extensive hydrogen-bond network
that connects the coordination sphere with the substrate binding pocket, as demonstrated by mutation of
two residues in this network (His81 and His217). The identities of both the substrate and the metal ion also
affect interactions within this H-bond network, thus leading to some mechanistic variations. Overall, the
mechanism employed by GpdQ is a paradigm of a substrate- and metal-ion-induced fit to optimize catalysis.
and demeton).7 For this reason, GpdQ is of great interest for
its potential application as a versatile enzymatic bioremediator.6,8,9
Crystal structures for various derivatives and complexes of GpdQ
have been reported.10-12 Structural comparison to other enzymes
indicates that GpdQ belongs to the group of R/ꢀ sandwich metallo-
phosphoesterases, which also includes purple acid phosphatases
(PAPs), the 3′-5′ cyclic nucleotide diesterase Rv0805 from
Mycobacterium tuberculosis, and Mre11 nuclease.13-23 The active
site (Figure 1) contains two metal-ion binding sites. The metal
1. Introduction
Glycerophosphodiesterase from Enterobacter aerogenes
(GpdQ) is a binuclear metallohydrolase1-5 that displays remark-
able activity toward all classes of phosphate ester substrates.6,7
Notably, GpdQ is capable of degrading EA 2192, a toxic product
formed by the hydrolysis of VX (a powerful nerve agent),6 as
well as a number of organophosphate pesticides (e.g., paraoxon
† School of Chemistry and Molecular Biosciences, The University of
Queensland.
(8) Ghanem, E.; Raushel, F. M. Toxicol. Appl. Pharmacol. 2005, 207,
S459-S470.
‡ Centre for Magnetic Resonance, The University of Queensland.
§ Middlebury College.
(9) Ely, F.; Foo, J. L.; Jackson, C. J.; Gahan, L. R.; Ollis, D.; Schenk, G.
Curr. Top. Biochem. Res. 2007, 9, 63–78.
| Australian National University.
(1) Cleland, W. W.; Hengge, A. C. Chem. ReV. 2006, 106, 3252–3278.
(2) Lowther, W. T.; Matthews, B. W. Biochim. Biophys. Acta 2000, 1477,
157–167.
(10) Hadler, K. S.; Tanifum, E. A.; Yip, S. H.-C.; Mitic´, N.; Guddat, L. W.;
Jackson, C. J.; Gahan, L. R.; Nguyen, K.; Carr, P. D.; Ollis, D. L.;
Hengge, A. C.; Larrabee, J. A.; Schenk, G. J. Am. Chem. Soc. 2008,
130, 14129–14138.
(3) Mitic´, N.; Smith, S. J.; Neves, A.; Guddat, L. W.; Gahan, L. R.;
Schenk, G. Chem. ReV. 2006, 106, 3338–3363.
(4) Wilcox, D. E. Chem. ReV. 1996, 96, 2435–2458.
(5) Crowder, M. W.; Spencer, J.; Vila, A. J. Acc. Chem. Res. 2006, 39,
721–728.
(11) Jackson, C. J.; Carr, P. D.; Liu, J.-W.; Watt, S. J.; Beck, J. L.; Ollis,
D. L. J. Mol. Biol. 2007, 367, 1047–1062.
(12) Jackson, C. J.; Hadler, K. S.; Carr, P. D.; Oakley, A. J.; Yip, S.; Schenk,
G.; Ollis, D. L. Acta Crystallogr. 2008, F62, 681–685.
(13) Guddat, L. W.; McAlpine, A. S.; Hume, D.; Hamilton, S.; de Jersey,
J.; Martin, J. L. Structure 1999, 7, 757–767.
(6) Ghanem, E.; Li, Y.; Xu, C.; Raushel, F. M. Biochemistry 2007, 46,
9032–9040.
(7) McLoughlin, S. Y.; Jackson, C.; Liu, J.-W.; Ollis, D. L. Appl. EnViron.
Microbiol. 2004, 70, 404–412.
(14) Klabunde, T.; Stra¨ter, N.; Fro¨hlich, R.; Witzel, H.; Krebs, B. J. Mol.
Biol. 1996, 259, 737–748.
9
11900 J. AM. CHEM. SOC. 2009, 131, 11900–11908
10.1021/ja903534f CCC: $40.75  2009 American Chemical Society

Reactivity of Bioremediator Glycerophosphodiesterase
A R T I C L E S
activity. Furthermore, a histidine residue (His81 or His217),
located in the second coordination sphere, was associated with
a catalytically relevant pKa value (∼7.0) and was predicted,
based on crystallographic information, to play a role in
orientating the substrate and/or protonating the leaving group
at low pH.10
The conundrum and challenge of mechanistic studies is the
continuous reconciliation between existing catalytic models and
new experimental data, leading to gradual improvements in the
understanding of the molecular details of a chemical reaction.
Numerous examples illustrate this quest, and the enzyme GpdQ
is a recent and versatile paradigm whose mode of action may
involve an intricate regulatory mechanism. Its efficient use in
bioremediation will be greatly assisted by detailed knowledge
of its mechanism of action, i.e., the various steps of a catalytic
turnover.
Toward this aim, we used stopped-flow fluorescence and
absorbance measurements to monitor structural changes that take
place when substrate is added to the enzyme and correlated these
changes to catalytic properties. Furthermore, we used the
paramagnetic properties of Mn(II) to demonstrate the extent to
which a substrate analogue (phosphate) affects metal ions’
binding affinities to the active site of GpdQ and determined
the dissociation constants of the R and ꢀ metal ion in the absence
or presence of phosphate. In addition, the role of a hydrogen-
bond network that connects the metal ions with the substrate
binding pocket was investigated by (i) site-directed mutagenesis,
whereby His81 and His217 (see above) were replaced by
alanines, and (ii) metal-ion substitution studies. The combined
data demonstrate the mechanistic flexibility of this highly
promiscuous enzyme and may provide the basis for using its
catalytic potential for bioremedial applications. Considering
structural similarities observed in the active sites of GpdQ and
other enzymes of the family of bimetallic metallohydrolases
(e.g., methionine aminopeptidase2,26 or various metallo-ꢀ-
lactamases5), it is possible that the mechanistic scheme described
here applies in a more general sense to other enzymes within
this family. Specifically, it may help to resolve some of the
controversies about the metal-ion content and the role(s) of the
metal ions in catalysis.
Figure 1. Schematic representation of the active site of GpdQ reconstituted
with Co(II).10 The metal ion in the R site (left) is six-coordinate, whereas
the one in the ꢀ site (right) is predominantly five-coordinate. However, a
terminal water ligand (as indicated by the dashed line) is visible in two out
of six subunits in the crystal structure.10 This ligand is within hydrogen-
bonding distance of the amide group of Asn80, the metal-ion-bridging
hydroxide, and the water coordinated to the metal ion in the R site.
ion in the R site is coordinated by four amino acid residues, a
terminal water ligand, and a hydr(oxide) molecule that bridges
the two metal ions. In the ꢀ site, the metal ion may also be
coordinated by four amino acid residues but, based on spec-
troscopic and kinetic data, is predicted to be less tightly bound
(Figure 1).10 In addition, a water molecule coordinated to the ꢀ
metal ion is observed in 2 out of 6 subunits in the crystal
structure (Figure 1). The physiologically relevant metal-ion
composition is unclear; however, the recombinantly expressed,
purified enzyme appears to contain at least one Fe(II) metal
ion (in the R site).12 Metal-ion replacement studies indicate that
GpdQ can operate using a range of divalent transition metal
ions, including Zn(II), Cd(II), and Co(II).10-12,24,25
Through a combination of solid state (X-ray crystallography)
and frozen solution (magnetic circular dichroism (MCD))
studies, possible snapshots of the catalytic mechanism were
obtained that enabled us to propose a reaction mechanism for
the hydrolysis of phosphate esters by GpdQ.10,11,25 It has been
postulated that a catalytically competent binuclear metal center
is assembled only in the presence of substrate or substrate
analogues and that a ligand to the metal ion in the ꢀ site, Asn80
(Figure 1), may play an important role in regulating enzymatic
2. Materials and Methods
Materials. All chemicals were purchased from Sigma-Aldrich,
unless stated otherwise. Bis(para-nitrophenyl) phosphate (bpNPP)
was purchased in the sodium salt form because of its greater
solubility in water. Ethyl para-nitrophenyl phosphate (EtpNPP) was
synthesized by known methods.27 Chemical structures of the
substrates used in this study are shown in Figure S1.
Expression, Purification, and Mutagenesis of GpdQ. GpdQ
(wild-type and mutants) were expressed and purified as described
previously.7 Residues His81 and His217 were replaced by alanine
using the Stratagene QuikChange kit.
Preparation of Metal Ion Derivatives of GpdQ. The Co(II)-
and Cd(II)-derivatives of GpdQ were prepared as previously
described.10,25 The Mn(II) form was prepared by adding a 99.9%
pure metal chloride solution to the apoenzyme (prepared as
described elsewhere10).
Characterization of the Metal Binding Affinity of
Mn(II)-GpdQ. The binding of Mn(II) to apo-GpdQ was monitored
by electron paramagnetic resonance (EPR) at room temperature.
(15) Lindqvist, Y.; Johansson, E.; Kaija, H.; Vihko, P.; Schneider, G. J.
Mol. Biol. 1999, 291, 135–147.
(16) Schenk, G.; Gahan, L. R.; Carrington, L. E.; Mitic´, N.; Valizadeh,
M.; Hamilton, S. E.; de Jersey, J.; Guddat, L. W. Proc. Natl. Acad.
Sci. U.S.A. 2005, 102, 273–278.
(17) Stra¨ter, N.; Beate, J.; Scholte, M.; Krebs, B.; Duff, A. P.; Langley,
D. B.; Han, R.; Averill, B. A.; Freeman, H. C.; Guss, J. M. J. Mol.
Biol. 2005, 351, 233–246.
(18) Schenk, G.; Elliott, T. W.; Leung, E.; Carrington, L. E.; Mitic´, N.;
Gahan, L. R.; Guddat, L. W. BMC Struct. Biol. 2008, 8, 6.
(19) Ho¨pfner, K. P.; Karcher, A.; Craig, L.; Woo, T. T.; Carney, J. P.;
Tainer, J. A. Cell 2001, 105, 473–485.
(20) Hadler, K. S.; Huber, T.; Cassady, A. I.; Weber, J.; Robinson, J.;
Burrows, A.; Kelly, G.; Guddat, L. W.; Hume, D. A.; Schenk, G.;
Flanagan, J. U. BMC Res. Not. 2008, 1, 78.
(21) Shenoy, A. R.; Capuder, M.; Draskovic, P.; Lamba, D.; Visweswariah,
S. S.; Podobnik, M. J. Mol. Biol. 2007, 365, 211–225.
(22) Shenoy, A. R.; Sreenath, N.; Podobnik, M.; Kovacevic, M.; Visweswa-
riah, S. S. Biochemistry 2005, 44, 15695–15704.
(23) Keppetipola, N.; Shuman, S. J. Biol. Chem. 2008, 283, 30942–30949.
(24) Gerlt, J. A.; Whitman, G. J. R. J. Biol. Chem. 1975, 250, 5053–5058.
(25) Mirams, R. E.; Smith, S. J.; Hadler, K. S.; Ollis, D. L.; Schenk, G.;
Gahan, L. R. J. Biol. Inorg. Chem. 2008, 13, 1065–1072.
(26) Lowther, W. T.; Matthews, B. W. Chem. ReV. 2002, 102, 4581–4607.
(27) Hendry, P.; Sargeson, A. M. J. Am. Chem. Soc. 1989, 111, 2521–
2527.
9
J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009 11901

A R T I C L E S
Hadler et al.
Steady-State Kinetic Assays. The phosphatase activity of GpdQ
toward the substrates para-nitrophenyl phosphate (pNPP), EtpNPP,
and bpNPP was measured spectrophotometrically by monitoring
the formation of the product p-nitrophenolate at 405 nm. Reactions
were measured to less than 5% of total substrate hydrolysis and
were performed as a continuous assay for pH 5.5 and above and
as a discontinuous assay below pH 5.5. For the discontinuous
assays, the reactions were quenched by adding 5 µL of the reaction
mixture to 3 mL of 0.1 M NaOH (ε400 ) 18 300 M-1 cm-1 for
p-nitrophenolate). For the continuous assays, the extinction coef-
ficient of the product p-nitrophenolate was determined at each
relevant pH by using a standard solution of p-nitrophenol in the
same buffer used for the enzymatic assays. Kinetic data were treated
by the initial rate method and fit to the Michaelis-Menten equation
(eq 2), where Vmax and KM are the maximal rate and Michaelis
constant, respectively.
Vmax[S]
Figure 2. Monitoring of the metal-ion affinity using EPR. Only the data
collected in the absence of the substrate analogue phosphate are shown.
Qualitatively similar data were obtained in the presence of phosphate, and
both data sets were fit to eq 1.
velocity )
(2)
KM + [S]
The pH dependence of kcat and kcat/KM for Co(II)-GpdQ were
reported previously.10 Here, the pH dependence of the catalytic
parameters for the Mn(II)- and Cd(II)-derivatives of GpdQ were
determined. Where applicable, the following equations, derived for
mono- and diprotic systems, respectively,28 were used to fit kcat
(eq 3) and kcat/KM (eq 4).
Scheme 1
Continuous wave (cw) EPR spectra were recorded with a Bruker
Elexsys E580 continuous wave/pulsed EPR spectrometer fitted with
an X-band Super High Q cavity. The magnetic field and microwave
frequency were calibrated with a Bruker ER 036 M Teslameter
and a Bruker microwave frequency counter, respectively. A Bruker
flowthrough AquaX cell was used to perform the Mn(II) binding
titrations. At 20 °C, the aquated Mn(II) ion displays a typical sextet
centered at g ) 2, and the intensity of the resonances reflects the
concentration of Mn(II) ions. Upon binding Mn(II) to GpdQ, EPR
signals were observed only from aquated (free) Mn(II), as EPR
spectra from protein-bound Mn(II) ions are not observable under
these conditions. The highly anisotropic and slow tumbling of the
large Mn-protein complex broadens out the EPR resonances at
room temperature. Standard curves of [Mn(II)] versus signal
intensity were obtained by recording EPR spectra of samples with
varying Mn(II) concentrations from 10 µM to 1 mM in HEPES
buffer (100 mM) at 20 °C (a titration of Mn(II) into buffer without
enzyme indicated that HEPES did not form complexes with Mn(II)).
The plot of peak-to-peak intensities versus [Mn(II)] for any of the
resonances of the sextet in the characteristic Mn(II) EPR spectrum
resulted in a straight line.
To determine the affinity of GpdQ for Mn(II), EPR spectra were
recorded under experimental conditions identical to those used to
obtain the standard curve (peak-to-peak intensity vs [Mn(II)]). The
protein sample was placed in a Bruker super high-Q cavity and
circulated continuously (by peristaltic pump) through an AquaX
flow cell (1.2 mL) equilibrated at 20 °C. A MnCl2 solution (10
µM to 1 mM) was titrated into the protein sample, and EPR spectra
were recorded after 10 min for each titration point. Because Mn(II)
bound to the enzyme does not contribute to the EPR spectrum, at
each Mn(II) concentration the amount of bound metal ions
([Mn(II)]bound) was calculated from the difference in EPR signal
intensity between the standard solution ([Mn(II)]total) and the relevant
protein sample ([Mn(II)]free). The ratio [Mn(II)]bound/[E]total was
plotted against [Mn(II)]total (Figure 2). The data could not be fit
to a model for two independent binding sites. A meaningful fit
was achieved to eq 1, which was derived based on a mechanism
that assumes two sequential binding events, as shown in Scheme
1.
kcat,max
kcat,app
)
+ kcat,low·pH
(3)
[H+]
1 +
(
)
Kes
kcat
(
)
kcat
KM
[H+]
max
)
(4)
(
)
KM
Ke2
[H+]
app
1 +
+
(
)
Ke1
Here, Kesi represents the ith protonation equilibrium associated
with the enzyme-substrate complex, while Kei is associated with
either the free enzyme or the substrate.28
Stopped-Flow Fluorescence and Absorbance Measurements.
Stopped-flow experiments were performed using an Applied
Photophysics SX-18 spectrometer. The path length of the observa-
tion cell was 2.0 mm, and the excitation and emission slits were
both set to 2.0 mm; all measurements were carried out at least in
triplicate. For fluorescence measurements, the excitation wavelength
was set to 260 nm, and the emission at 310 nm was monitored.
The data were recorded at the photomultiplier voltage at which
intrinsic fluorescence of the protein (in the absence of substrate)
was set equal to 8 V. For absorbance measurements, the protein
sample (without substrate) was used as a zero-absorption reference,
and the change in absorbance was monitored at 405 nm. The
reaction was initiated by injecting equal volumes of substrate
(syringe A) and protein (syringe B) to a total volume of ∼65 µL.
The protein solution contained 5 µM GpdQ and 200 µM
CoCl2 ·6H2O. Both the substrate and protein solutions were buffered
with Chelex-treated 50 mM MES, pH 5.5. The extent to which the
rate of change in fluorescence depended on substrate concentration
was measured under pseudo-first-order conditions ([S] . [E]).
Data in which the observed rate (kobs1 and kobs2) exhibited
saturation-type dependence on substrate concentration were fit to
eq 5.
[Mbound
]
K2[M] + 2[M]2
K1K2 + K2[M] + [M]2
)
(1)
(28) Cleland, W. W. AdV. Enzymol. Relat. Areas Mol. Biol. 1977, 45, 273–
[Etotal
]
387.
9
11902 J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009

Reactivity of Bioremediator Glycerophosphodiesterase
A R T I C L E S
kf[S]
kobs
)
+ kr
(5)
K + [S]
Here, kf and kr represent first-order rate constants (s-1) for
substrate binding and release, respectively, while K is the dissocia-
tion constant for the preceding step.29,30
Data in which an increase in [S] led to a decrease in kobs (kobs3
)
were fit as described in the text.
Magnetic Circular Dichroism. MCD spectra were recorded at
a 2 nm bandwidth with a Jasco J-600 spectropolarimeter equipped
with an Oxford SM-4 magnet/cryostat and an Oxford ITC-4
temperature controller. Protein samples were diluted with glycerol
to a ratio of 3 parts glycerol to 2 parts protein solution and stirred
for at least 1 h. The final concentration of protein was in the range
0.68 to 1.09 mM. Protein solutions were buffered at pH 7.0 with
50 mM HEPES. The samples were added into a brass cell with
quartz windows and a 0.62-cm path length. Spectra were collected
in the region between 280 to 800 nm.
Figure 3. Sample progress curves showing the three fluorescence transient
phases of the reaction between GpdQ and the substrate bpNPP (see
Supporting Information for a similar data set with the monoester substrate
pNPP). Main figure: 5 µM GpdQ was reacted with 1.1 mM bpNPP, and
the progress was monitored over 70 s. Inset: 5 µM GpdQ was reacted with
120 µM bpNPP, and fluorescence changes were monitored over 400 ms.
3. Results and Discussion
Substrate Enhances the Binding Affinity of the Metal Ion
in the ꢀ Site. The paramagnetic properties of Mn(II) at room
temperature (i.e., a characteristic isotropic sextet signal if
unbound, EPR-silent if bound) make this metal ion a suitable
probe to quantitate the metal-ion binding affinities of the R and
ꢀ sites in the catalytic center of GpdQ. Figure 2 shows the
fraction of bound Mn(II) as a function of added Mn(II). The
limiting value is 2, which indicates two binding events. These
data could not be fit to a model for two independent binding
sites but instead required an equation derived for a model that
assumes stepwise binding of the two ions (i.e., the binding of
the first metal ion is required to promote binding of the second;
Scheme 1). A metal ion binds to the more strongly coordinating
R site, providing an additional ligand (a terminal water ligand
that becomes the µ-hydr(oxide) molecule in the binuclear
species) for the ꢀ metal ion, thus enabling the second metal
ion to bind. From the fit to the data (Figure 2), the dissociation
constants are KR ) 29 ((12) µM and Kꢀ ) 344 ((27) µM.
When phosphate (10 equiv) was added, the binding constant of
the ꢀ metal ion changed to 56 ((12) µM, while that of the R
site did not change greatly (17 ((6) µM). The affinity of the
metal ions was also assessed by measuring the catalytic activity
as a function of the concentration of added Mn(II). A single
binding event was obtained with a Kd of ∼80 ((20) µM (see
Figure S2 for more information), in good agreement with the
constant for the ꢀ metal ion in the presence of PO4, as obtained
using room temperature EPR. Using a similar approach, it was
shown previously that the Kd for the ꢀ metal ion of the Co(II)
and Cd(II) derivatives of GpdQ is 29 and 16 µM, respectively.10,25
Note that because the Kd for the PO4-Mn(II) complex is ∼15
mM,31 the substrate (or substrate analogue) likely primes the
active site for binding of the second metal ion rather than
importing this metal in the form of a substrate-metal ion
complex.
to investigate the early processes of the catalytic reaction. The
existing mechanistic model implies that the assembly of a
catalytically competent binuclear center requires substrate-
promoted structural changes in the active site of GpdQ.10
Because the intrinsic fluorescent properties of tryptophan and
tyrosine residues are suitable probes to monitor conformational
changes and/or enzyme-ligand interactions, stopped-flow fluo-
rescence measurements were carried out to investigate the effect
of substrate binding (note that similar measurements with the
Mn(II)- and Cd(II)-derivatives were not possible, because these
forms of GpdQ require higher pH for activity, at which protein
precipitation occurs readily under conditions required for the
stopped-flow measurements). The excitation wavelength char-
acteristic for tryptophan residues (∼290 nm) resulted only in
weak fluorescence changes (data not shown; no Trp residues
are in the vicinity of the active site of GpdQ10). However, at
an excitation wavelength of 260 nm, optimal for tyrosine
residues, a strong response was detected (there are four Tyr in
the immediate vicinity of the active site: Tyr19, Tyr229, Tyr230,
and Tyr263). Control experiments demonstrated that the ob-
served fluorescence changes described below are solely due to
the interaction between GpdQ and substrate and not due to
mixing or dilution effects (data not shown). Qualitatively, similar
results were obtained when bpNPP, EtpNPP, or pNPP were
used, and for illustrative purposes only those results acquired
with bpNPP, one of the diester substrates, are shown here (for
data collected with the monoester substrate pNPP, see Support-
ing Information, Figure S3).
Three distinct changes in fluorescence were observed over a
1-min period in the reaction between Co(II)-GpdQ and the
substrate. An initial, rapid decrease in fluorescence intensity is
followed by a slower increase; finally, a second but slower
decrease is then detected (Figure 3). The first two phases are
complete within ∼200 ms after initiating the reaction (Figure
3, inset). The three regions of the observed time course could
each be fit separately to a single exponential function (f(x) )
a exp(-kobs x) + c), resulting in first-order rate constants kobs(i)
(i ) 1, 2, or 3 for the first, second, or third transient,
respectively). Because substrate is essential to form a catalyti-
cally active site, the effect of substrate concentration on the three
transient phases was investigated by measuring kobs at various
Formation of a Fully Active Binuclear Metal Center in
GpdQ Occurs in Three Stages. Because of the wealth of
structural, spectroscopic, and kinetic data available for the
Co(II)-derivative of GpdQ, this form of the enzyme was used
(29) Fersht, A. In Enzyme Structure and Mechanism; W. H. Freeman: New
York; San Francisco, 1977; Chapter 4, pp 103-133.
(30) Spencer, J.; Clarke, A. R.; Walsh, T. R. J. Biol. Chem. 2001, 276,
33638–33644.
(31) White, D. J.; Reiter, N. J.; Sikkink, R. A.; Yu, L.; Rusnak, F.
Biochemistry 2001, 40, 8918–8929.
9
J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009 11903

A R T I C L E S
Hadler et al.
up to the formation of a binuclear center in the active site of
GpdQ are unlikely to contribute to the rate-limiting step.
In contrast to this behavior of kobs1 and kobs2, an increase of
[S] leads to a decrease in kobs3 (Figure 4).32 Although the
observation of a third transient implies that the EB•S complex
is rearranged further, the inhibitory effect of [S] is also
reminiscent of that observed, for instance, in the reaction
between another binuclear metallohydrolase, uteroferrin, and one
of its substrate analogues (phosphate).33 In that study, the
inhibitory effect of phosphate on kobs was interpreted in terms
of the binding of a second molecule of substrate as shown in
Scheme 2:34
An equation for the [S] dependence of kobs is of the type
k3f A[S]
B[S]2 + C[S] + 1
kobs3
)
+ k3r
(6)
where A, B, and C represent fitting parameters.33 The data for
kobs3 cannot be unambiguously fit to this equation. However,
k3r can be estimated, because, for high levels of [S], eq 6
simplifies to eq 7:
Figure 4. Effect of the substrate concentration ([S]) on the observed rates
(kobs) associated with the three fluorescence transients. The data were fit to
equations described in the text, and the resulting parameters are listed in
Table 1.
Table 1. Kinetic Parameters for the Three Transients Observed in
the Reaction between GpdQ and bpNPP by Stopped-Flow
Fluorescence
k3f A
1
kobs3
)
+ k3r
(7)
B [S]
transient
kf
kr
K
The ratio A/B can be shown to correspond to 1/KI,33 where
KI is the dissociation constant for the binding of the second
substrate molecule to GpdQ (Scheme 2), and thus eq 7 can be
rewritten as
1
2
3
354 s-1
∼0
21 µM
206 µM
-
38 s-1
5 s-1
∼0
a
25 s-1 M-1
a The rate constant represents kf/KI (see text for details).
Scheme 2
k3f 1
KI [S]
kobs3
)
+ k3r
(8)
The parameters from a fit to eq 8, listed in Table 1, indicate
that the structural rearrangement is slow and virtually irreversible
(k3r ≈ 0). Although obtaining an accurate estimate for k3f is not
presently possible, the observed rates for the structural change
(kobs3; Figure 4) suggest that as [S] increases, this rearrangement
may become increasingly more important in the rate-limiting
step.
To correlate the observed fluorescence changes to catalytic
steps in the mechanism of phosphorolysis, the reaction between
GpdQ and bpNPP was monitored by stopped-flow absorbance
measurements. Absorbance changes were measured at 405 nm,
the wavelength optimal for detecting the nitrophenolate product
of the hydrolytic reaction (see Figure S5). The absorbance
changes measured over several minutes were used to calculate
the turnover number for the hydrolysis of bpNPP (kcat; s-1), and
the corresponding stopped-flow fluorescence data measured
for the same reaction are also shown in Figure 5. The kcat
gradually increases, reaching a maximum value after ∼100 s
(2 mM bpNPP). The subsequent decrease in kcat corresponds to
[S]. Figure 4 summarizes the data and shows that kobs1 and kobs2
display a nonlinear behavior, whereby an increase in [S] leads
to a gradual increase in kobs, reaching a maximal value at high
concentration (kobs3 behaves differently and is discussed below).
The data for kobs1 and kobs2 were fit to eq 5, which is derived for
a simple bimolecular reaction, and the resulting parameters are
listed in Table 1. A similar behavior (i.e., a rapid loss of
fluorescence followed by a slower regain in intensity) is also
observed for the binuclear L1 metallo-ꢀ-lactamase from Stenotro-
phomonas maltophilia and was interpreted in terms of a minimal
reaction scheme in which an initial binding of substrate was
followed by slower rearrangements.30 For GpdQ, in Scheme 2
(above) is proposed, in which EM and EB correspond to
mononuclear (M) and binuclear (B) enzyme forms, and kf and
kr are the on and off rates, respectively, for the reactions
associated with the individual changes in fluorescence intensity
(i.e., kobs1, kobs2, and kobs3).
(32) For comparison, the dependence of kobs(i) (i ) 1, 2, 3) on the
concentration of the monoester substrate pNPP is shown in Figure
S4. The overall behavior of the three transients are identical to that
observed for bpNPP.
(33) Twitchett, M. B.; Schenk, G.; Aquino, M. A. S.; Yiu, D. T. Y.; Lau,
T. C.; Sykes, A. G. Inorg. Chem. 2002, 41, 5787–5794.
Following rapid formation of an EM•S complex, the substrate
rearranges in the active center to prime the ꢀ site (Figure 1) for
metal-ion binding (EM′•S), a process that is virtually irreversible
(k1r close to 0; Table 1). Upon binding of the metal ion, a
catalytically competent binuclear enzyme-substrate complex
(EB•S) is formed with a rate of ∼40 s-1 (Table 1), a rate that is
approximately 40-fold higher than the turnover number (kcat)
for hydrolysis of bpNPP (∼1 s-1). Thus, the processes that lead
(34) Note that, in principle, the observed inhibition may also arise from
product inhibition. However, several observations favor the proposed
substrate inhibition: (i) qualitatively similar fluorescence data were
measured for both mono-and diester substrates; (ii) the product of the
diester reaction is the monoester substrate (pNPP), which has low
binding affinity and is turned over slowly; (iii) the product of the
monoester reaction is the rather potent inhibitor PO4.
9
11904 J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009

Reactivity of Bioremediator Glycerophosphodiesterase
A R T I C L E S
transients are observed (Figure S7), supporting the above
interpretation of the individual transients observed in the reaction
catalyzed by wild-type GpdQ. kobs1 and kobs2 are associated with
structural changes that occur (i) in the mononuclear enzyme,
preparing it for binding of the second metal ion, and (ii) during
the binding of the second metal with the concomitant formation
of a binuclear center, respectively; kobs3 arises from the flexible
coordination of ligand Asn80. In the Asn80Asp mutant, the
second metal ion remains tightly bound throughout the catalytic
cycle and the Asp80 ligand does not display coordination
flexibility. Thus, the lack of fluorescence transients suggests that
the active site in this mutant lacks structural flexibility, reflecting
its reduced reactivity.
The above assignment of kobs3 to a gradual structural
rearrangement required to attain optimal activity may also
indicate that the ꢀ metal ion remains bound to the active site as
long as substrate is available for the reaction. While the catalytic
turnover number is on the order of 1 s-1, it takes ∼100 s for
the activity to reach its maximum (Figure 5). Thus, once an
activated EB* complex has been formed it remains present until
substrate is depleted, upon which the occupancy of the ꢀ metal-
ion site gradually decreases and the active site returns to its
mononuclear EM state (Scheme 2).
Figure 5. Progress curves for the reaction between GpdQ and bpNPP (2
mM). Upper curVe: the kcat reaches a maximum after ∼100 s. Lower curVe:
the third fluorescence transient (kobs3) is complete after ∼100 s, the time at
which kcat maximizes. After ∼150 s, the kcat reduces, and the fluorescence
starts to rise again as the substrate is being depleted; the enzyme gradually
returns to its inactive mononuclear form.
Role of the Hydrogen-Bond Network in Regulating the
Coordination of the ꢀ Metal Ion. The three stages described in
the previous section demonstrate that structural changes occur
throughout the catalytic cycle of GpdQ, but the question of how
these rearrangements are mediated remains. The likely mech-
anism involves interactions between substrate and the H-bond
network that connects the first and second coordination spheres.
This network includes two histidine residues, His81 and
His217.10 Measuring the pH dependence of the catalytic
efficiency (kcat/KM) of Co(II)-GpdQ demonstrated the presence
of a catalytically relevant pKa at ∼7.010 that was, tentatively,
ascribed to one of those two histidines. Here, we have shown
that replacement of these residues by alanines does not affect
this pKa value substantially (Figure 6). Furthermore, although
the kcat does not display a strong dependence on pH, the
His81Ala and His217Ala mutations lead to a moderate decrease
and increase in the turnover rate, respectively (e.g., at pH 7.0
with the substrate bpNPP, the kcat values for wild-type, His81Ala,
and His217Ala GpdQ are 1.4, 0.6, and 4.9 s-1, respectively).
Thus, the catalytically relevant pKa of ∼7.0 is not likely to
correspond to one of these two histidine residues. Figure 7 shows
the MCD spectra for the two GpdQ mutants in the absence and
presence of the substrate analogue phosphate (using Co(II) for
reconstitution). For comparison, the previously reported10
spectrum for the wild-type enzyme is also included. In the wild-
type spectrum, when 2 equiv of Co(II) were added to GpdQ,
only one metal ion was bound (a six-coordinate species in the
R site).10 When phosphate was added, both metal-ion binding
sites became occupied. In contrast, both histidine-to-alanine
mutants have considerable affinity for the metal ion in the ꢀ
site, even in the absence of phosphate, as demonstrated by the
the gradual depletion of substrate. Interestingly, changes in the
fluorescence behavior follows an identical time course; in the
first ∼100 s of the reaction, the fluorescence quenches (with a
rate characterized by kobs3) and then gradually regains intensity
as substrate is used up and the enzyme returns to its resting
state. Hence, kobs3 (Scheme 2) is ascribed to the rearrangement
of EB•S, a modestly active complex, to a more reactive one
(EB*•S; Scheme 2).
The combined stopped-flow data thus invoke a mechanistic
model whereby rapid binding of substrate to an inactive form
of GpdQ (EM) is followed by rearrangement of the resulting
EM•S complex to a moderately active EB•S complex. This
complex can undergo further conformational changes to become
more reactive, but the binding of a second substrate molecule,
presumably in the vicinity of the active site, may prevent this
rearrangement. These steps, illustrated in Scheme 2, can be
reconciled with previously invoked catalytic steps. After initial
rapid binding of the substrate to the enzyme (forming the EM•S
complex), kobs1 describes the rearrangement of the still mono-
nuclear (EM) active site to prime it for the binding of the second
metal ion (EM′•S). Following the binding of the second metal
ion, a moderately active binuclear EB•S complex is formed (kobs2
;
Scheme 2). Previous studies using MCD spectroscopy indicated
that flexible coordination by the ꢀ metal-ion ligand Asn80
(Figure 1) may regulate enzymatic activity once the binuclear
active center has been assembled.10 The third transient phase
in the reaction between GpdQ and substrate (kobs3) is thus
interpreted as a break in the coordination bond between the ꢀ
metal ion and Asn80 and a concomitant rearrangement of the
active-site geometry and/or substrate orientation (mediated by
the hydrogen-bond network that connects Asn80 with the
substrate binding pocket). The significance of coordination
flexibility in regulating and optimizing the catalytic efficiency
of GpdQ is also demonstrated when the two mutants Asn80Ala
and Asn80Asp are investigated by stopped-flow fluorescence.
Both mutants have been shown to have greatly reduced catalytic
efficiency in comparison to the wild-type enzyme.10 In the
alanine mutant kobs1 is not observed and kobs2 and kobs3 are slower
than in the wild-type enzyme (Figure S6). In the aspartate
mutant, the least reactive form of GpdQ, no fluorescence
(35) Kaden, T. A.; Holmquist, B.; Vallee, B. L. Biochem. Biophys. Res.
Commun. 1972, 46, 1654–1659.
(36) Larrabee, J. A.; Chyun, S. A.; Volwiler, A. S. Inorg. Chem. 2008, 47,
10499–10508.
(37) Larrabee, J. A.; Leung, C. H.; Moore, R. L.; Thamrong-Nawasawat,
T.; Wessler, B. S. H. J. Am. Chem. Soc. 2004, 126, 12316–12324.
(38) Lever, A. B. P. Inorganic electronic spectroscopy, 2nd ed.; Elsevier:
Amsterdam; Oxford, New York, 1984.
(39) Prescott, J. M.; Wagner, F. W.; Holmquist, B.; Vallee, B. L.
Biochemistry 1985, 24, 5350–5356.
9
J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009 11905

A R T I C L E S
Hadler et al.
Figure 6. Effect of pH on the catalytic efficiency (kcat/KM) of wild-type
GpdQ and its His81Ala and His217Ala mutants, using bpNPP (top) or
EtpNPP (bottom) as substrates. The catalytically relevant pKa at ∼7.0 is
apparent in each data set. Below pH 6.0, the two mutants lose activity.
Hence, the data were fit to equations derived for a monoprotic system for
the mutants and for a diprotic one for wild-type GpdQ, as described
elsewhere.10
Figure 7. MCD spectra of His81Ala-GpdQ (top) and His217Ala-GpdQ
(bottom) in the absence and presence of phosphate. Each sample contains
2 equiv of Co(II). For comparison, the spectrum of the wild-type enzyme
(2 equiv Co(II), no phosphate) is also included. Even in the absence of
phosphate, a metal ion can bind to the five-coordinate ꢀ site only in the
two mutant forms of GpdQ.
presence of the band at ∼570 nm, characteristic of a five-
coordinate species10,35-39 (based on a comparison of the signal
intensities of the ∼570 nm bands, the two mutants have ∼2-
(His81Ala) to ∼8-fold (His217Ala) increased metal-ion binding
affinity in comparison to the wild type enzyme; Figure 7).
Adding PO4 enhances this affinity further. Thus, although none
of these histidine residues interact directly with the metal ions,
they are capable of affecting the coordination environment of
the ꢀ metal ion, thus illustrating the importance of the H-bond
network in regulating coordination of the metal ion in the ꢀ
site. Since the KM values (a measure of substrate affinity) are
not greatly affected by these two mutations, none of these
histidine residues are likely to play a major role in substrate
binding. It thus seems likely that the main role for these residues
is to mediate H-bond interactions that facilitate both the binding
of the second metal ion and the optimal positioning of the
substrate, achieved via flexible coordination of Asn80.
bound water molecule (the pKa of [Co(II)(H2O)6] is ∼9.640),
but the water ligand bound to the ꢀ metal ion is within
H-bonding distance (∼3 Å) of the amide group of Asn80, the
metal-ion-bridging hydroxide, and the water molecule coordi-
nated terminally to the R metal ion, which could lower its pKa
(Figure 1).
Metal Ions and Substrates Modulate the Reaction Mecha-
nism. Although the steps leading up to the formation of a
catalytically competent EB•S complex are likely to be largely
independent of the identity of the metal ions or substrates, the
same may not be the case for the subsequent steps in catalysis.
Rearrangement of the EB•S complex to the fully active EB*•S
complex is associated with flexible coordination of Asn80,
whose time course is described by kobs3 (Vide supra). This time
constant (Figure 4) is of a magnitude similar to that of kcat
(Figure 8) and may thus contribute to the overall rate of the
reaction. For the Co(II)-derivative of GpdQ, the kcat in particular
for the reaction with the monoester substrate pNPP was shown
to be pH independent, suggesting that the rate-limiting step is
not associated with bond breakage but with product release and/
or conformational changes.10 Such conformational changes may
of course be linked to the flexible coordination in the ꢀ metal-
ion site. For the reaction with the diester substrates bpNPP and
EtpNPP, these conformational changes may also be relevant,
Because neither of the two histidines can be assigned to the
catalytically relevant protonation equilibrium with a pKa of ∼7.0,
the question about the identity of the corresponding residue
remains. Analysis of the crystal structure of Co(II)-GpdQ reveals
that in two out of six subunits, an additional water ligand is
terminally bound to the metal ion in the ꢀ site (Figure 1).10 We
tentatively assign pKa ∼7.0 to this ligand. Its deprotonated form
at pH > pKa would slow down ligand exchange at the ꢀ site,
which may thus interfere with substrate binding, a likelihood
consistent with the reduced catalytic efficiency at high pH
(Figure 6). Note that a pKa of ∼7 is low for a terminally, Co(II)-
(40) Baes, C. F.; Mesmer, R. E. The hydrolysis of cations; Wiley: New
York, 1976.
9
11906 J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009

Reactivity of Bioremediator Glycerophosphodiesterase
A R T I C L E S
Figure 8. pH profiles for the hydrolysis of pNPP, bpNPP, and EtpNPP for GpdQ reconstituted with Co(II), Mn(II), and Cd(II). Top: pH dependence of kcat
.
Bottom: pH dependence of kcat/KM.
Table 2. Dissociation Constantsa for the Hydrolysis of bpNPP,
are close to 1, indicating that P-O bond fission may be nearly
fully rate-limiting. The likely nucleophile is a nonactivated
hydroxide bound to the R metal ion. While the reaction between
GpdQ and bpNPP is not greatly affected by metal ion substitu-
tions, a replacement of the metal ion does affect the reaction
when EtpNPP is the substrate, indicating that both the metal
ion and the substrate can affect the catalytic mechanism.
Comparing the monoesterase activity of the Co(II)- and
Mn(II)-derivatives of GpdQ (Figure 8) proved interesting.
Although the kcat values for both derivatives are pH independent
(and their rate-limiting steps thus are not associated with bond
breakage), their kcat/KM ratios display a bell-shaped behavior,
indicating the presence of at least two relevant protonation
equilibria associated with the free enzyme, the substrate, or both.
The data were fit to eq 4 (see Table 2 for parameters). For the
Co(II)-derivative, pKe2 ) 5.1 (Table 2) was previously assigned
to the second pKa of the substrate pNPP (pKa2 ≈ 5), which
indicates that this substrate undergoes catalysis preferentially
in its monoanionic form.10 In the Mn(II)-derivative pKe2 is
shifted by two units (∼7.1; Table 2), indicating that this
protonation equilibrium is not likely to correspond to the
substrate, but rather to the free enzyme. A likely residue is the
water molecule terminally bound to the ꢀ metal ion (as described
in the previous section). However, pKe1 (∼4.8) is close to the
second pKa of the substrate, suggesting that, in Mn(II)-GpdQ,
the monoester substrate undergoes catalysis preferentially in the
dianionic form.
EtpNPP, and pNPP by Co(II)-, Mn(II)-, and Cd(II)-Substituted
GpdQ
pNPP
EtpNPP
bpNPP
pKe1 pKe2 pKes1 pKes2 pKes3 pKe1 pKe2 pKes1 pKe1
Co(II)-GpdQ 3.8 5.1 <3.2 6.2
10.2 3.7 6.9 9.9 7.0
Mn(II)-GpdQ 4.8 7.1
Cd(II)-GpdQ
9.7
9.8
-
-
-
-
-
-
-
9.2
9.4
-
-
-
-
-
a The pKes n and pKe n (n ) 1, 2, or 3) values for the Mn(II)- and
Cd(II)-derivatives of GpdQ were obtained from fits of the pH
dependence of the kcat and kcat/KM (Figure 8), respectively. The data for
the Co(II)-derivative were published previously.10
but P-O bond fission is also at least partially rate-limiting.10
Furthermore, although a minimum of one protonation equilib-
rium is relevant to the hydrolysis of bpNPP, at least three are
required for EtpNPP (Figure 8; Table 2).10 Assigning the three
pKa values has enabled us to propose a catalytic scheme for the
hydrolysis of EtpNPP, whereby the metal-ion-bridging hydrox-
ide activates the R metal-ion-bound, nucleophilic hydroxide.
Consequently, GpdQ can effectively hydrolyze this diester over
a large pH range (from pH 3.0 to 11).10 However, the only
apparent protonation equilibrium in the reaction with bpNPP
(pKa ) 9.9; Table 2) suggests that a nonactivated hydroxide
(i.e., a hydroxide terminally bound to the metal ion in the R
site) may act as the nucleophile.10
To probe the effect of the metal-ion composition on the
mechanism of hydrolysis, the catalytic properties of the Mn(II)-
and Cd(II)-derivatives of GpdQ were investigated over a broad
pH range (Figure 8). The Cd(II)-derivative is not an effective
monoesterase, but for the diester substrates bpNPP and EtpNPP
its catalytic properties are similar to those of Mn(II)-GpdQ.
Although the kcat/KM ratios of the two derivatives for the
hydrolysis of the diester substrates are virtually independent of
pH (in marked contrast to the Co(II)-derivative), their turnover
numbers display monoprotic behavior, reaching maximum
activity at pH >9. The data were fitted to eq 3, and the
parameters are listed in Table 2. The slopes of the pH profiles
Conclusion
This study demonstrates how the bioremediator GpdQ, a
binuclear enzyme of the metallohydrolase family, employs an
intricate mechanism for catalysis, whereby the substrate binds
to the active site of the enzyme, thus promoting the binding of
a second metal ion to the ꢀ site and leading to the formation of
a catalytically relevant binuclear enzyme-substrate (EB•S)
complex. Figure 9 depicts a model for the catalytic cycle
employed by this versatile enzyme. The initial steps of catalysis,
9
J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009 11907

A R T I C L E S
Hadler et al.
Figure 9. Mechanism of hydrolysis of phosphate esters by GpdQ, as determined from stopped-flow fluorescence studies.
i.e., the steps leading to the assembly of the EB•S complex, are
rapid in comparison to the turnover number (kcat), indicating
that they do not contribute to the rate-limiting step. The EB•S
complex is partially active, but further arrangements in the active
site lead to further improvements in reactivity. These rearrange-
ments occur within a time scale similar to that of kcat, suggesting
that they may contribute at least in part to rate limitation. The
conformational changes that take place in the active site can be
described as an induced fit triggered by substrate and also, as
the metal-ion replacement studies indicated, by the metal-ion
cofactors. The H-bond network that links the two metal ions to
the substrate binding pocket plays a pivotal role in mediating
these structural changes and thus in fine-tuning the mechanism
of action employed by GpdQ. The flexible coordination of
Asn80 plays an important part in this mechanism. In this respect,
GpdQ resembles another metalloenzyme, the mononuclear
nonheme iron enzyme lipoxygenase from soybean (sLO).41 In
sLO it was shown using site-directed mutants that the extent of
coordination flexibility directly correlates with reactivity;
the greater the flexibility, the greater the catalytic activity of
the enzyme. It is believed that in lipoxygenase the flexible
coordination allows for an optimal positioning of the substrate
in the active site. The same is likely to be the case for GpdQ.
Furthermore, the catalytic mechanism of GpdQ may also be
related to other binuclear metallohydrolases, such as methionine
aminopeptidase (MetAP)2,26,42,43 and metallo-ꢀ-lactamases
(MBLs).5,44,45 However, MetAP is flexible in that it can operate
as either a mononuclear or binuclear hydrolase, with each form
having a distinct reaction mechanism.2,26,42,43 Similarly, it
appears that some MBLs can operate as mononuclear enzymes,
while others require two metal ions.5,44,45 In contrast, the present
study supports a mechanistic model for GpdQ whereby the
mononuclear form is inactive and a substrate- and metal-ion-
induced fit leads to the formation of a catalytically active
binuclear center, providing an intricate regulatory mechanism
for enzyme activity (“on-off switch”).
GpdQ is a very stable enzyme that has a flexible active site
and displays very broad substrate specificity, hydrolyzing a large
number of phosphate esters over a large pH range. Furthermore,
both the reactivity and the reaction mechanism can be modulated
by changing metal ions and specific amino acids. Together with
the ease with which this enzyme can be recombinantly expressed
and purified in large amounts, these properties and our increased
understanding of their tuneability make GpdQ a promising target
for the development of a bioremediator. Efforts to this extent
are currently in progress.
Acknowledgment. This work was funded by the Australian
Research Council (DP0986613). J.A.L. wishes to acknowledge the
National Science Foundation for financial support through an NSF/
RUI grant CHE-0554083 and an instrument grant NSF/MRI CHE-
0820965. The authors express their gratitude toward Prof. Sunhee
Choi for critically assessing the present study.
Supporting Information Available: Chemical structure of
substrates pNPP, EtpNPP, and bpNPP (Figure S1); determina-
tion of metal binding affinity of M(II) by kinetic assay (M )
Co, Cd, Mn) (Figure S2); reaction progress curves (Figure S3)
and the effect of [S] on kobs(i) (Figure S4) for the reaction of
pNPP with GpdQ; example of absorbance change over time
for the reaction of pNPP with GpdQ (Figure S5); reaction
progress curves for N80A-GpdQ (Figure S6) and N80D-GpdQ
(Figure S7). This material is available free of charge via the
Internet at http://pubs.acs.org.
(41) Schenk, G.; Neidig, M. L.; Zhou, J.; Holman, T. R.; Solomon, E. I.
Biochemistry 2003, 42, 7294–7302.
(42) Cosper, N. J.; D’souza, V. M.; Scott, R. A.; Holz, R. C. Biochemistry
2001, 40, 13302–13309.
(43) D’Souza, V. M.; Bennett, B.; Copik, A. J.; Holz, R. C. Biochemistry
2000, 39, 3817–3826.
(44) Cricco, J. A.; Orellano, E. G.; Rasia, R. M.; Ceccarelli, E. A.; Vila,
A. J. Coord. Chem. ReV. 1999, 192, 519–535.
(45) Heinz, U.; Adolph, H. W. Cell. Mol. Life Sci. 2004, 61, 2827–2839.
JA903534F
9
11908 J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009