Ni2 +-activated glyoxalase i from Escherichia coli: Substrate specificity, kinetic isotope effects and evolution within the βαβββ superfamily

Article abstract of DOI:10.1016/j.jinorgbio.2011.11.008

The Escherichia coli glyoxalase system consists of the metalloenzymes glyoxalase I and glyoxalase II. Little is known regarding Ni ^{2 +}-activated E. coli glyoxalase I substrate specificity, its thiol cofactor preference, the presence or absence of any substrate kinetic isotope effects on the enzyme mechanism, or whether glyoxalase I might catalyze additional reactions similar to those exhibited by related βαβ ββ structural superfamily members. The current investigation has shown that this two-enzyme system is capable of utilizing the thiol cofactors glutathionylspermidine and trypanothione, in addition to the known tripeptide glutathione, to convert substrate methylglyoxal to non-toxic d-lactate in the presence of Ni^{2 +} ion. E. coli glyoxalase I, reconstituted with either Ni^{2 +} or Cd^{2 +}, was observed to efficiently process deuterated and non-deuterated phenylglyoxal utilizing glutathione as cofactor. Interestingly, a substrate kinetic isotope effect for the Ni ^{2 +}-substituted enzyme was not detected; however, the proton transfer step was observed to be partially rate limiting for the Cd ^{2 +}-substituted enzyme. This is the first non-Zn ^{2 +}-activated GlxI where a metal ion-dependent kinetic isotope effect using deuterium-labelled substrate has been observed. Attempts to detect a glutathione conjugation reaction with the antibiotic fosfomycin, similar to the reaction catalyzed by the related superfamily member FosA, were unsuccessful when utilizing the E. coli glyoxalase I E56A mutein.

Full text of DOI:10.1016/j.jinorgbio.2011.11.008

Journal of Inorganic Biochemistry 108 (2012) 133–140
Contents lists available at SciVerse ScienceDirect
Journal of Inorganic Biochemistry
journal homepage: www.elsevier.com/locate/jinorgbio
Ni2+-activated glyoxalase I from Escherichia coli: Substrate specificity, kinetic isotope
effects and evolution within the βαβββ superfamily
1
2
3
4
Kadia Y. Mullings , Nicole Sukdeo , Uthaiwan Suttisansanee , Yanhong Ran , John F. Honek ⁎
Department of Chemistry, University of Waterloo, 200 University Avenue West, Waterloo, Ontario, Canada N2L 3G1
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 21 July 2011
Received in revised form 11 October 2011
Accepted 11 November 2011
Available online 28 November 2011
The Escherichia coli glyoxalase system consists of the metalloenzymes glyoxalase I and glyoxalase II. Little is
known regarding Ni -activated E. coli glyoxalase I substrate specificity, its thiol cofactor preference, the
2+
presence or absence of any substrate kinetic isotope effects on the enzyme mechanism, or whether glyoxa-
lase I might catalyze additional reactions similar to those exhibited by related βαβββ structural superfamily
members. The current investigation has shown that this two-enzyme system is capable of utilizing the thiol
cofactors glutathionylspermidine and trypanothione, in addition to the known tripeptide glutathione, to con-
Keywords:
Glyoxalase
Escherichia coli
Thiol cofactor
2
+
vert substrate methylglyoxal to non-toxic D-lactate in the presence of Ni
ion. E. coli glyoxalase I, reconsti-
tuted with either Ni2+ or Cd , was observed to efficiently process deuterated and non-deuterated
phenylglyoxal utilizing glutathione as cofactor. Interestingly, a substrate kinetic isotope effect for the Ni
2+
2
+
-
Substrate specificity
Kinetic isotope effect
Fosfomycin resistance protein
substituted enzyme was not detected; however, the proton transfer step was observed to be partially rate
limiting for the Cd2+-substituted enzyme. This is the first non-Zn -activated GlxI where a metal
ion-dependent kinetic isotope effect using deuterium-labelled substrate has been observed. Attempts to
detect a glutathione conjugation reaction with the antibiotic fosfomycin, similar to the reaction catalyzed
by the related superfamily member FosA, were unsuccessful when utilizing the E. coli glyoxalase I E56A
mutein.
2+
©
2011 Elsevier Inc. All rights reserved.
1
. Introduction
This thioester is then hydrolyzed by the second enzyme, GlxII, to
produce a non-toxic 2-hydroxycarboxylic acid in the form of D-lactate
and the regenerated thiol cofactor. There is evidence that the inter-
mediate, S-D-lactoylglutathione, can control intracellular pH in
bacteria through interaction with the KefB potassium efflux system
[8].
There are several different thiol cofactors found in various organ-
isms. GSH is the most prominent intracellular thiol in eukaryotes and
some prokaryotes [9–12]. This thiol is involved in many biological
functions including aromatic metabolism, maintenance of cellular
redox potentials and several detoxification mechanisms [11,13]. In
E. coli, a novel GSH and spermidine (spd) conjugate was discovered
and designated as glutathionylspermidine (GspdSH, Fig. 1). This
complex glutathione and its disulfide form (glutathionylspermidine
The intracellularly generated metabolite methylglyoxal (MG,
-oxopropanal) acts as a potent electrophile causing irreparable cel-
2
lular damage if allowed to build to cytotoxic concentrations [1–6].
The glyoxalase (Glx) system is an enzyme couple critical for the de-
toxification of MG. The system consists of two enzymes, glyoxalase I
(
(
GlxI, S-D-lactoylglutathione lyase; EC 4.4.1.5) and glyoxalase II
GlxII, S-2-hydroxyacylglutathione hydrolase; EC 3.1.2.6) (Fig. 1) [7].
The first enzyme, GlxI, converts the hemithioacetal, a product formed
by the non-enzymatic reaction of MG and a thiol cofactor/cosubstrate
such as glutathione (GSH; γ-Glu-Cys-Gly), to S-D-lactolyglutathione.
⁎
Corresponding author. Tel.: +1 519 8884567x35817; fax: +1 519 7460435.
E-mail addresses: kymullings@gmail.com (K.Y. Mullings), sukdeo_n@gmx.com
2
disulfide, (GspdS) ), under anaerobic conditions or stationary phase,
compose approximately 80% of all produced GSH in E. coli [14,15]. An
elevated level of GspdSH production is found to be in proportion to bac-
2
terial cell density [14,15], while the amount of (GspdS) is decreased
(
(
N. Sukdeo), nuuthaiwan@mahidol.ac.th (U. Suttisansanee), tranyh@jnu.edu.cn
Y. Ran), jhonek@uwaterloo.ca (J.F. Honek).
1
Present address: GreenLight Biosciences, 196 Boston Avenue, Suite 2400, Medford,
when cellular stress conditions prevail. GspdSH is also recognized to
act more efficiently toward protecting cells against radical or oxidant-
induced damage than its disulfide form [14,16]. Analysis of these exper-
imental results suggests that GspdSH could be a reasonable alternate
cofactor for the Glx system under these conditions. Interestingly, the
glyoxalase enzymes are also conserved in organisms that produce
MA 02155, United States.
2
Present Address: Department of Microbiology and Immunology, University of
British Columbia, 2350 Health Sciences Mall, Vancouver, BC, Canada V6T 1Z3.
3
Permanent Address: Institute of Nutrition, Mahidol University, 25/25 Phutthamonthon
4
Rd., Salaya, Phutthamonthon, Nakhon Pathom 73170, Thailand.
Permanent Address: College of Life Science and Technology, Jinan University.
4
Guangzhou, China.
0162-0134/$ – see front matter © 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.jinorgbio.2011.11.008

134
K.Y. Mullings et al. / Journal of Inorganic Biochemistry 108 (2012) 133–140
Fig. 1. The two-enzyme glyoxalase system, which consists of glyoxalase I (GlxI) and glyoxalase II (GlxII), using methylglyoxal (MG) and different thiols including glutathione (GSH),
glutathionylspermidine (GspdSH) and trypanothione (T(SH) ).
2
major intracellular thiols other than GSH. Recent studies on Leishmania
and Trypanosoma have indicated that their corresponding Glx systems
have evolved to use the bis-glutathionyl derivative of spermidine,
trypanothione (T(SH) , N ,N -bisglutathionylspermidine) (Fig. 1), the
2
predominant thiol of these organisms, as a cofactor [17–21].
The Glx system has been investigated with respect to mechanistic,
structural and metabolic aspects. GlxI enzymes are metalloenzymes
that can be divided into two classes according to their metal activation
with different biological functionality within this superfamily
[29,30,32]. Among these structural members, FosA is mechanistically
close to GlxI, as both are metalloenzymes that utilize GSH as a
cofactor/cosubstrate, and both enzymes form an active site octahedral
geometry around the metal center (Fig. 2) [24–26, 33, 34]. However,
FosA employs three water molecules and three metal binding ligands
that coordinate to the metal center via vicinal chelation [33,35], while
GlxI utilizes two water molecules and four metal binding residues
around the active metal [26]. FosA employs two metal binding subsites,
1
8
2
+
2+
2+
profile, Zn -activation (i.e., H. sapiens GlxI [22]) or Ni /Co -activa-
2
+
2+
tion (non-Zn -activated i.e., E. coli GlxI [23]). The active metallated
forms of these enzymes require a metal environment that is octahedral
in nature, with four protein residues serving as metal ligands, along
with two water molecules completing the geometry [24–26]. This oc-
one for the catalytically critical metal Mn and another for a monova-
+
lent K (Fig. 3) [36–38]. GlxI, on the other hand, utilizes only one metal
binding site (Fig. 3) [24–26].
Fosfomycin, the substrate of FosA, is an antibiotic useful in the
treatment of lower urinary tract and gastrointestinal infections due
to its stability and broad spectrum of activity [39–42]. This antibiotic
affects both Gram-positive and Gram-negative bacteria by interfering
with MurA, the first enzyme in the bacterial cell wall biosynthetic
pathway [42], a target that is lacking in humans. FosA likely evolved
from “housing keeping” gene(s) within bacterial cells to resist the cyto-
toxicity of the natural product fosfomycin. FosA activates the epoxide
moiety in fosfomycin for nucleophilic attack by glutathione, resulting
in the formation of an inactive GS-fosfomycin adduct (Fig. 2). We
hypothesize that FosA activity might be engineered from some proteins
in the same superfamily, more specifically GlxI, due to similarities be-
tween these enzymes as stated above.
2
+
curs for the Zn -bound active human GlxI as well as the catalytically
active Ni -bound E. coli GlxI. However, the inactive Zn2+-bound form
of the E. coli GlxI approximates a trigonal bipyramidal geometry around
the Zn2 ion with only one water molecule present [26]. Mechanisti-
cally it is suggested, based on analysis of the X-ray structures of
inhibitors bound to the H. sapiens GlxI, that the arrival of the hemi-
thioacetal substrate at the catalytic site results in the replacement of
the two active site water ligands by the two oxygens from the
hemithioacetal substrate. This is likely followed by a series of deproto-
nation, reprotonation steps (Fig. 1), resulting in thioester product for-
mation [24,25]. This mechanism is supported by reported solvent
isotope incorporation studies as well as observations of a kinetic
isotope effect using deuterated α-ketoaldehydes with yeast GlxI
2+
+
2
+
(
Zn -activated enzyme) [27,28]. To date, no studies have been
reported that investigate the Ni -activation class of GlxI in this
manner.
2
+
GlxI is a member of the βαβββ superfamily of proteins, consisting
of fosfomycin resistance protein (FosA), methylmalonyl-CoA epimer-
ase (MMCE), extradiol dioxygenase (DIOX), mitomycin C resistance
protein (MRP) and bleomycin resistance protein (BRP) [29–31]. These
proteins are believed to have evolved from a common one-module
ancestor, which, upon gene duplication and fusion as well as through
accumulated mutagenesis, resulted in high structural similarity but
Fig. 2. The reaction catalyzed by the metalloenzyme fosfomycin (FM) resistance pro-
tein (FosA), in the presence of both divalent and monovalent metals. The enzyme cat-
alyzes the formation of the product glutathione-fosfomycin adduct (GS-FM).

K.Y. Mullings et al. / Journal of Inorganic Biochemistry 108 (2012) 133–140
135
Fig. 3. A three-dimensional figure of the superimposed structures of P. aeruginosa FosA (cyan, PDB ID: 1LQP) and E. coli GlxI (magenta, PDB ID: 1F9Z) showing the active site metals
including K+ (green) and Mn
2+
(orange) in FosA and Ni
2+
(blue) in GlxI (wall-eyed view).
Herein, we report an investigation of the cofactor specificity of the
E. coli Glx system using a set of thiol cofactors (GSH, GspdSH and
2.2. DNA cloning and manipulation
T(SH)
2
) to probe the extent of cofactor promiscuity by these E. coli
In order to mimic the active site of FosA, the multiple sequence
alignment of GlxI from various organisms was performed in compar-
ison with FosA from Serratia marcescens (Supplementary Information
Fig. S1). There are four conserved metal binding residues in the active
site of GlxI; however, only three exist in FosA. The active site residue,
enzymes and to obtain additional insight into the molecular interac-
tions that occur between substrate and metalloprotein. As well, we
report our studies on evaluating the deuterium kinetic isotope effect
(
KIE) on E. coli GlxI catalysis by utilizing α-deuteriophenylglyoxal (α-
56
48
deuterioPG) or non-deuterated phenylglyoxal (PG) as substrates and
GSH as the cofactor. To determine if the extent of any KIE could be
Glu , in E. coli GlxI corresponds to a smaller residue, Cys , in the S.
marcescens FosA, which has only three protein residues ligating the
metal center based on X-ray analysis. The lack of one active site
metal binding residue in FosA compared to GlxI supports the two-
substrate reaction of FosA, which requires more catalytic space around
the metal ion than the one-substrate reaction of GlxI. Thus, Glu56 in E.
coli GlxI was mutated to Ala to produce a somewhat similar ligand ar-
rangement around the metal center as found in FosA, as well as to
avoid side reactions that might occur with an oxidizable thiol group if
a Cys56 mutation was used instead. The site-directed mutagenesis of
E. coli GlxI-E65A was generated utilizing a QuikChange® polymerase
chain reaction (PCR) protocol with an expression vector containing
the E. coli glxI gene (pET22b-glxI) as previously described [23] and
the following primers; (+) 5′-ACCGAAGAAGCGGTGATTGCCCT-
GACCTACACCTGGGGC-3′ and (−) 5′-GCCCCAGTTGTAGGTCAGGGCAA
TCACCGCTTCTTCGGT-3′. The plasmid was transformed into heat
shocked competent E. coli DH5α cells and its sequence was verified
(Molecular Biology Core Facility, University of Waterloo, Waterloo,
ON) before transforming the plasmid into heat shocked E. coli BL21
(DE3) cells for protein expression purposes.
metal ion dependent, both Ni2 - as well as Cd -activated GlxI were
utilized in these studies. Lastly, intrigued by the mechanistic related-
ness between the enzymes GlxI and FosA, we prepared the E56A
mutant of E. coli GlxI in order to produce a GlxI mimicking the metal
coordination geometry of FosA. Fosfomycin conjugation competency
was evaluated for this GlxI mutein in the presence of various M2+
metal ions.
+
2+
2
. Experimental
2
.1. Materials
Chemical reagents and enzymes were acquired from Sigma Chemi-
cal Company (St. Louis, MO) as follows: L- and D-lactate dehydroge-
nases, fosfomycin (disodium salt), bovine serum albumin (BSA),
methylglyoxal (MG), phenylglyoxal (PG) and reduced glutathione
(
2
GSH). The thiols, T(SH) and GspdSH, were purchased in their
disulfide forms from Bachem Bioscience Inc. (King of Prussia, PA).
Metals were purchased from J.T. Baker Chemical Co. (Phillipsburg, NJ)
including zinc chloride (assay, 99.3%), cobalt chloride hexahydrate
2.3. Protein purification
(
assay, 100.4%), manganese chloride tetrahydrate (assay, 98.8%) and
cadmium chloride (assay, 99.4%). Magnesium chloride hexahydrate
assay, 99.2%) and calcium chloride dihydrate (assay 75.5%) were
Protein purifications, apo-enzyme preparations and metal analyses
including the 4-(2-pyridylazo)resorcinol (PAR) assays and Inductively
Coupled Plasma Mass Spectrometry (ICP-MS) of protein samples
were performed according to those previously described [23,43,44].
The concentrations of samples of purified enzymes were calculated
using the Bradford Assay with bovine serum albumin (BSA) as standard
[45]. Protein molecular masses were determined by analyses of SDS-
PAGE, gel permeation chromatography (Superdex75 HR 10/30 column
with flow rate of 0.5 mL/min) results, and analysis of positive ion
mode-electrospray ionization mass spectrometry (ESI-MS) results
(Micromass Q-TOF Global Ultima mass spectrometer at the Mass
Spectrometry Facility, University of Waterloo, Waterloo, ON).
(
purchased from BDH Chemicals Ltd. (Poole, England). Nickel chloride
hexahydrate (99.9999% pure) and cupric chloride hydrate (>99.0%)
were supplied from Sigma Aldrich (St. Louis, MO). Water used in all
experiments was Milli-Q water (18 MΩ-cm conductivity, Waters Asso-
ciates, Milford, MA).
Protein visualization on sodium dodecyl sulfate polyacrylamide
gel electrophoresis (SDS-PAGE) using 8–15% acrylamide gradient gel
was performed on a PhastSystem (Pharmacia, Etobicoke, ON). The
enzymatic assay of GlxI was performed on a SpectraMax 190
9
6-well UV–visible spectrophotometer (Molecular Devices, Sunnyvale,
CA) using a Soft Max Pro software (Molecular Devices). The activity as-
2.4. GlxI activity assays
says for GlxII were performed on a Cary 3 UV–vis spectrophotometer
(
Varian, Mississauga, ON) using the Kinetics module of CaryWinUV
Reduction of the cofactor disulfides to their required thiol forms,
version 3.0 (Varian, Mississauga, ON).
2
T(SH) and GspdSH, were carried out using a tris[2-carboxyethyl]

136
K.Y. Mullings et al. / Journal of Inorganic Biochemistry 108 (2012) 133–140
phosphine hydrochloride (TCEP) immobilized reducing gel (Thermo
Fisher Scientific Inc., Rockford, IL) [17] under argon with metal-free
HEPES buffer (pH 7.4) [23]. The thiol solution was extracted in degassed
and argon sparged water and titrated to pH 7.0 using 0.1 mM HCl, also
under inert atmosphere. The free thiol was quantitated by Ellman's
reagent (5′,5′-dithiobis-(2-nitrobenzoic acid) or DTNB) using the detec-
tion at 412 nm and the ε412 of 14,150 M cm [46]. This free thiol so-
lution was immediately frozen under argon and stored at −20 °C.
The enzyme assays were performed as previously reported
26/40 column using gravity flow and fractions were monitored at
240 nm. The purified GS-fosfomycin adduct was identified by
ESI-MS. This reference compound was chemically synthesized using
a reported method [38] with and without microwave assistance in a
Biotage Initiator system (Biotage, LLC; Charlotte, NC, USA).
−
1
−1
2.7. Substrate kinetic isotope effect (KIE) experiments
The synthesis of α-deuteriophenylglyoxal was performed as pre-
viously described in the literature [28]. The chemical structure of
the purified product was confirmed using ESI-MS and proton nuclear
[
17,23,43,44] using the hemithioacetal formation between MG and
thiols (GSH, GspdSH and T(SH) ) in KBP (pH 7.0 or stated otherwise)
with dissociation constant of 3.1 mM [47–52] on a SpectraMax 190
6-well UV–visible spectrophotometer and Soft Max Pro software
Molecular Devices, Sunnyvale, CA). The initial rate data were fitted
2
1
magnetic resonance ( H NMR) on a Bruker 300 MHz spectrometer.
9
(
This deuterated substrate was used in the GlxI assay in comparison
with its non-deuterated counterpart to observe the proton transfer
in the enzyme reaction as was previously undertaken for studies on
to a Michaelis–Menten equation with least squares fit parameters
using GraphPad Prism software version 5.00 for Windows (GraphPad
Software, Inc., San Diego, CA, www.graphpad.com) or GraFit software
version 3.01 (Erithacus Software).
2
+
Zn -activated yeast GlxI [28]. The hemithioacetal was permitted to
form (K of 0.6 mM [28]) between these substrates (PG and
d
α-deuterioPG) and GSH as cofactor in 50 mM KPB (pH 7.0) at 25 °C
for 10 min prior to performing the assay (1 mL). E. coli GlxI
(16.8 nM), which had been reconstituted with either 2.5 molar
2
.5. GlxII activity assays
2 2
equivalents of NiCl or CdCl was added into the reaction mixture,
Assays to measure GlxII activities were performed as previously
described [53,54]. The substrate for GlxII, S-D-lactoylthioester, was
prepared by incubating E. coli GlxI with the hemithioacetal formed
and the activity was monitored as the disappearance of the hemi-
thioacetal at 263 nm, the isobestic wavelength for the free PG and
the hemithioacetal [28]. The kinetic parameters were determined
−
1
−1
between MG and thiol cofactors (GSH, T(SH)
2
and GspdSH) for 1 h
using the extinction coefficients for PG (ε240 =5690 M
cm
)
−
1
−1
at room temperature. The enzyme was then removed by
centrifugation using a Nanosep device (Millipore, Billerica, MA)
with 3 kDa molecular weight cut-off. The thioester in the flow
through fraction was quantitated by detection at 240 nm with ε240
and α-deuterioPG (ε240 =4740 M
cm ) [28,57].
2.8. Molecular modeling
−
1
−1
of 2860 M
cm
[23].
Molecular graphics images and superimpositions of protein struc-
tures (using default settings) were produced using the UCSF Chimera
package (version 1.5.3) from the Resource for Biocomputing, Visualiza-
tion, and Informatics at the University of California, San Francisco (sup-
ported by NIH P41 RR001081).
L- and D-lactate dehydrogenase (LDH) assays were used to determine
whether the D-stereo isomer of lactic acid was produced by the E. coli Glx
system using S-lactoylGSH, S-lactoylTSH and S-lactoylGspdSH as sub-
strates. The GlxII assays using these substrates were performed as stated
above, and the completion of the reactions was monitored at 240 nm.
The reaction mixture was then filtered using a Nanosep device with
3. Results and discussion
10 kDa molecular weight cut-off, and the flow through fraction contain-
ing lactate was used for the LDH assays. The enzymes (2.75 units), L-LDH
3.1. Efficiency of thiol cofactor utilization by the E. coli Glx system
(
EC 1.1.1.27) and D-LDH (EC 1.1.1.28), were incubated with the solution
+
(350 μL) of isolated lactate and NAD (5 mM) in 100 mM Tris buffer (pH
The discovery of a GlxI isolated from E. coli that was selectively ac-
+
2+
2+
8
.5) at 25 °C for 1 h. The assay activity was monitored at 340 nm, and the
tivated by Ni2 /Co , and not by Zn
ion, was the first indication
−
1
−1
NADH formed was quantitated using the ε340 of 6220 M cm
55,56].
that two different classes of GlxI metalloenzymes existed in nature
[23,26]. Previously only Zn2 -activated GlxI enzymes had been
reported, although these enzymes exhibited wide metal ion promis-
cuity, even activation by Ni2 ion was observed. The ability of E. coli
+
[
+
2
.6. Assay for FosA activity
2
+
GlxI to bind, but not be activated by, Zn
ion and to be maximally
+
The assay for FosA activity was performed as previously reported
using DTNB [33]. Fosfomycin (10 mM) was mixed with GSH
15 mM) in buffer containing 50 mM HEPES (pH 8.0) and 100 mM
activated by Ni2 ion was subsequently shown to occur for certain
other GlxI as well [43,44,53]. To date, the hemithioacetal formed
non-enzymatically between GSH and MG has been the only substrate
studied with the E. coli Glx metalloenzymes. However, significant
amounts of GspdSH are biosynthesized by E. coli under increasing
cell turbidity and stress, which normally occur under growth limiting
conditions [14,15]. As well, other parasitic organisms such as Trypa-
nosoma and Leishmania possess a GlxI system that utilizes alternative
(
KCl that had been degassed under argon for 30 min prior to the reac-
tion. The metal-reconstituted GlxI-E56A (4 μM) that was prepared by
incubating the apo-enzyme with metal chloride (50 μM of NiCl
CoCl , CuCl , ZnCl , MnCl , CdCl , MgCl or CaCl ) overnight at 4 °C
was added to the GSH–fosfomycin mixture (final volume of 100 μL)
before incubating at room temperature under argon for 3 h. The reac-
tion was then quenched by an addition of methanol (300 μL). The free
thiol of GSH was quantitated by mixing 5 μL of the quenched reaction,
2
,
2
2
2
2
2
2
2
thiols, such as T(SH)
are able to utilize the precursor of T(SH)
their Glx reactions. The T(SH) -dependent Glx systems favor the
GspdSH cofactor over T(SH) , while their enzymatic activities with
GSH are very low [17,18]. Interestingly, not all protozoans prefer
GspdSH and T(SH) as cofactors for their Glx systems. The malarial
parasite, Plasmodium falciparum, possesses a Glx system that utilizes
GSH as cofactor [58]. The differences in thiol cofactor preference for
various Glx systems depend on the cellular levels of specific thiols
within a particular organism. Trypanosoma and Leishmania predomi-
2
, instead of GSH (Fig. 1) [18]. These protozoans
2
, GspdSH, as a cofactor for
2
1
0 μL of 18 mM DTNB solution and 185 μL HEPES buffer before mon-
2
−
1
−1
itoring at 412 nm with the ε412 of 14,150 M
cm
[46].
An alternative qualitative assay was performed using the analysis
of ESI-MS experiments. The reaction of metal-reconstituted (NiCl
and MnCl GlxI-E56A (14.4 μM) was mixed with fosfomycin
150 mM) and GSH (150 mM) in HEPES buffer (pH 8.0) containing
00 mM KCl to a final volume of 373 μL under argon. The reaction
was incubated at 30 °C for 3 h before quenching by the addition of
00 μL of methanol. The product was purified on a Sephadex G-10
2
2
2
)
(
1
nantly utilize trypanothione and employ a T(SH)
system. Interestingly, the Leishmania major GlxI enzyme utilizes
2
-dependent Glx
5

K.Y. Mullings et al. / Journal of Inorganic Biochemistry 108 (2012) 133–140
137
trypanothione as its cosubstrate and the enzyme is activated selec-
was detected as an increase in wavelength intensity at 340 nm upon
the formation of NADH. No activity was detected using either L-LDH
with the same final product mixture nor using controls lacking GlxI.
2
+
2+
tively by Ni
(inactive with Zn ), similar to E. coli GlxI [53].
Based on these considerations, it was reasonable to hypothesize
that E. coli GlxI might utilize GspdSH as a cofactor when available in
2
+
Analysis of these results indicated that the Ni -activated class of
GlxI has the same stereospecificity as the Zn2 -activated class of
GlxI, a result consistent with the stereospecificity reported for the
aforementioned enzymes from the Leishmania, H. sapiens, and Trypa-
nosoma organisms, among others [14, 60–62]. An exception is the Glx
system from the African trypanosome, T. brucei [63], where GlxI
appeared to be absent and L-lactate was produced as a major product
in this organism.
+
2
the cell. In this current investigation, GspdSH, and also T(SH) , were
indeed found to be reasonable cofactors for Ni2 -activated E. coli
+
GlxI (Table 1). Analyses of the kinetic data indicated that the activities
2
+
(
k
cat) of Ni -activated E. coli GlxI with the hemithioacetal substrates
formed by reaction of MG with each of the thiol cofactors, T(SH) and
GspdSH, decreased drastically to approximately 18% in the case of
GspdSH and 7% in the case of T(SH) , when compared with the natu-
ral GSH cofactor. However, the K for the hemithioacetal substrates
formed from GspdSH and T(SH) were found to be similar. The cata-
lytic efficiencies (kcat/K ) determined for the hemithioacetal sub-
strates formed from GspdSH and T(SH) were only 5% and 2% of
2
2
m
2
3.2. Metal-dependent substrate deuterium isotope effect for E. coli GlxI
m
2
Primary kinetic isotope effect (KIE) experiments were undertaken
in order to determine the effect of deuterium substitution on the rate
of the GlxI reaction mechanism as catalyzed by a Ni2 -activated GlxI
class enzyme. The enzyme from E. coli was chosen as the prototype of
this class as the apo-enzyme can be readily prepared and easily recon-
that of the GSH-hemithioacetal. Thus, a less sterically demanding
GSH appeared to be more acceptable to the enzyme active site than
the larger GspdSH and T(SH) (Supplementary Information and Fig.
2
+
S2). Although trypanothione is not present in E. coli, it is interesting
to note the capability, albeit low, of E. coli GlxI to utilize this thiol.
In the second step of the Glx reaction, the thioester product is hy-
drolyzed by the metalloenzyme GlxII to produce D-lactate and regener-
ate the thiol cofactor. This enzyme in E. coli is a Zn2 metalloenzyme
2
+
2+
2+
2
stituted with activating metals such as Ni , Co , Mn
and Cd ,
with Ni2
+
substitution producing the most active enzyme
−
1
6
−1 −1
(kcat=338 s ; K
Cd
m
=27± 0.4 μM; kcat/K
m
=12.4×10 M
s
) and
+
2+
−1
producing the least active holoenzyme (kcat=21 s ; K =9±
m
6
−1 −1
[
54]. In trypanosomes, thioesters of GspdSH, T(SH)
2
and GSH are able
0.4 μM; kcat/K
m
=2.4×10 M
s
) using MG and GSH [23,64].
2
+
2+
to act as the substrates for GlxII with the trypanothione thioester
being an efficient substrate for the Trypanosoma brucei and Leishmania
donovani GlxII [56,59]. Similar results were observed for E. coli GlxII in
the present study, where the enzyme's activities for the various S-D-
lactoyl thiol conjugates did not significantly differ from S-D-
lactoylglutathione itself, and suggested that GlxII has lower substrate
Metal removal from the Ni /Co -activated enzyme (i.e., E. coli
GlxI) can be readily accomplished using ethylenediaminetetraacetic
acid (EDTA) or isoelectric focusing; the apo-enzyme being catalytically
inactive [23]. On the other hand, the Zn2 -activated class of enzyme
(i.e., H. sapiens GlxI) requires stronger conditions for metal removal
such as use of dipicolinic acid (DPA) to remove bound metals. Metal
reconstitution of the apo-enzyme from Pseudomonas aeruginosa
GloA3 (a Zn2 -activated GlxI enzyme) for example, only recovers ap-
proximately 80% activity in the presence of Zn2 compared to the
pre-DPA treated enzyme [44]. Thus, E. coli GlxI is a more suitable
model for probing potential metal-ion dependent characteristics of
the catalyzed reaction than a Zn2 -activated class enzyme. As well, a
potential metal-dependent effect on the KIE of a GlxI of any class has,
to our knowledge, never been reported. We therefore chose to investi-
+
m
specificity than GlxI (Table 1). However, the K value of S-D-
+
lactoylT(SH) as substrate with E. coli GlxII was increased to 226%,
2
+
while that of S-D-lactoylGspdSH as substrate was reduced to only 24%
compared to that of S-D-lactoylGSH as substrate. Thus, the highest
efficiency of E. coli GlxII was observed using S-D-lactoylGspdSH as
substrate. Analysis of this data suggested that GspdSH present in E.
coli under various conditions could indeed be used as a cofactor to
detoxify cytotoxic MG in combination with the bacterial glyoxalase
+
2
+
system. The E. coli Glx system was also able to utilize T(SH)
2
even
gate substrate-based isotope effects on both the Ni - as well as the
2
+
though this thiol is not produced within the bacterial cells. This
finding is consistent with the close structural relationship held
among the GlxII enzymes from different organisms (Supplementary
Information and Fig. S3).
Cd -activated E. coli GlxI. For various technical reasons and for com-
parison with the literature when available, phenylglyoxal as well
as its α-deuterioPG analog were used as substrates in these KIE
2
+
investigations. Additionally, the X-ray structures of both the Ni
-
Although the Zn2+-activated class of glyoxalase I enzymes are
only known to produce the S-D-lactoylthioester as product, no
and the Cd -substituted have been determined (Supplementary In-
formation Fig. S4) and have reported highly similar metal ligand ar-
rangements around the metal centers; however, several metal
coordination distances differed between the two metallated forms of
the enzyme (Supplementary Information Table S1).
2+
2
+
study has investigated the stereospecificity of the Ni -activated
2
+
class of GlxI. Based on our experiments with Ni -activated E. coli
GlxI used in tandem with the Zn2 -activated E. coli GlxII, and utilizing
L-LDH as well as the D-LDH to detect the stereochemistry of the
resulting lactic acid product produced from the Glx system using
+
In our experiments, the enzyme kinetics of either Ni2+- or Cd2+
-
reconstituted GlxI were recorded for varying concentrations of the
MG and the various thiols (GSH, GspdSH and T(SH)
2
), only D-lactate
hemothioacetal substrates formed from MG, PG or α-deuterioPG
Table 1
Kinetics of E. coli GlxI and GlxII using different thiols (GSH, GspdSH and T(SH)
2
) in the presence of methylglyoxal.
Cofactora
GlxIb
GlxIIc
k
cat
K
m
k
cat/K
m
Relative
cat/K
k
cat
K
m
k
cat/K
m
Relative
(
s−1)
(µM)
(M−1 s−1
)
k
m
(s−1
)
(µM)
(M−1 s−1
)
kcat/K
m
5.6×106
0.3×106
0.1×106
100
5
2
61± 0
58± 1
70± 1
184± 22
45± 0.1
415± 11
0.3×10
1.3×10
0.2×10
6
23
100
15
GSH
GspdSH
T(SH)
344± 8
61± 3
24± 0
61± 4
204± 39
199± 25
6
6
2
a
2
GSH is commercially available as the reduced-form; however, the disulfide forms of GspdSH and T(SH) require reduction by immobilized TCEP prior to performing the kinetic
assays (see text).
b
The GlxI assays (200 μL) were performed using Ni2+-reconstituted GlxI (3 nM, 48 h incubation at 4 °C with 2.5 equivalents of metal) and 0.04–0.7 mM hemithioacetal substrate
in 100 mM KPB (pH 7.0) at room temperature.
c
The GlxII assays (1 mL) were performed using Zn2+-bound GlxII (70 nM) and 0.02–0.1 mM thioester substrate in buffer containing 100 mM KPB (pH 8.0) and 3.5 mM DTT.

138
K.Y. Mullings et al. / Journal of Inorganic Biochemistry 108 (2012) 133–140
and GSH as the cofactor (Table 2). For the Ni2+-activated E. coli GlxI,
the enzyme efficiencies using non-deuterated and deuterated PG-
GSH substrates were determined to be similar, suggesting that there
was no substrate deuterium kinetic isotope effect on the kcat nor on
ion activating the enzyme. However, both metal-reconstituted forms
exhibited a lower K for MG-GSH than for PG-GSH, indicating that
m
the larger PG aldehyde did not interact as well with the active site of
the enzyme compared to the less bulky MG molecule.
The differences in the kinetic results between Ni2 and Cd2+ for
the E. coli GlxI could be due to the difference in the extent of polar-
ization that may occur with the bound substrate for each metal ion,
+
the K
m
for this metallated form of the E. coli enzyme. These data
were in contrast to the effect of the α-hydrogen substitution for the
same substrates with the yeast GlxI enzyme. The yeast GlxI enzyme
was observed to exhibit substrate deuterium isotope effects of 3.3
2
+
Cd
being less capable of lowering the energy barrier to proton ab-
2
+
and 3.2 on K
transfer from substrate is
determining step for this particular enzyme [28]. Assuming the
m
and kcat, respectively, which suggested that proton-
straction compared to Ni , causing a lower catalytic rate to be ob-
2
+
a
major contributor to the rate-
served for the Cd -GlxI with MG and PG but also allowing for a
KIE to be detected for Cd2 -GlxI upon use of deuterated PG as sub-
strate. An additional factor that may contribute in some fashion to
+
2
+
mechanism of Ni -substituted E. coli GlxI is the same as for the
2
+
2+
2+
Zn -activated GlxI class enzymes, the substrate isotope effect
could have been obscured by another step in the catalytic mechanism.
This step would not have been detectable by the experiments typical-
ly used to evaluate the contribution of the proton transfer to substrate
conversion.
the catalytic differences observed between the Ni
and the Cd
forms of GlxI is that of the difference in ligand substitution rates
for each metal. Specifically, the rate of water–ligand exchange from
2
+
4
Cd
centers is faster (~10 ×) than the exchange rates observed for
2
+
water–ligands associated with Ni
centers [65]. The impact of
The activities of the Cd2+-substituted E. coli GlxI using PG-GSH and
α-deuterioPG-GSH substrates contrasted markedly with the results ob-
water–ligand exchange rates on the overall rate of the GlxI-
catalyzed reaction can be considered under the assumption that sub-
strate binding occurs in the active site with bound metal via vicinal
oxygen chelation with concomitant displacement of the resting-
state water ligands. This has been suggested as a likely stage in the
chemical mechanism of GlxI. Hydroxamate analogs, which are be-
lieved to mimic the enediolate (Fig. 1) intermediate(s) have been
2
+
served for the Ni -holoenzyme. A drastic decrease in catalytic activity
2
+
of the Cd -substituted enzyme was observed using the deuterated al-
dehyde. This significant reduction in substrate conversion rate for the
2
+
Cd -substituted enzyme indicated the presence of a kinetic isotope
effect for this form of the enzyme when handling PG-GSH as substrate.
2
+
The K
m
was also affected modestly by isotopic replacement in the alde-
studied with respect to the Zn
center in H. sapiens GlxI as well as
with the Ni2 center in E. coli GlxI [25,66]. In both cases displacement
of the aqua ligands around the metal center were observed, although
it should be stated that hydroxamates may be poor choices for repre-
sentatives of the reaction intermediates of GlxI due to their strong
metal binding properties which might nullify their application as re-
action intermediate analogs for GlxI. If the rate constant for associa-
tion of the substrate hemithioacetal with the active site metal is low
relative to the rate constant for the chemical deprotonation/reproto-
nation step, the KIE should be suppressed. The active site glutamate
residue which serves as one of the metal ligands has been implicated
+
hyde, while analysis of the observed kcat(H)/kcat(D) value of 2.3 indicated
that the α-proton transfer was partially rate-limiting in the enzymatic
reaction. The effect of deuterium substitution on the ratio of kcat/K
m
was observed to be low (a value of 1.7) and did not indicate that the
enzyme–substrate complex off-rate was influencing the value of the in-
trinsic isotope effect. The result was consistent with the interpretation
that the proton abstraction was partially rate determining in the en-
zyme mechanism, and that other steps contributed to product forma-
tion as well.
Analysis of the kinetic data obtained for the Ni2+- and the Cd2+-ac-
tivated forms of the E. coli GlxI enzyme with the hemithioacetal sub-
strates MG-GSH and PG-GSH indicated that both metal activation
classes were comparable in efficiently processing the bulkier PG alde-
hyde co-substrate. The activities of the Ni2 -reconstituted enzyme
with the hemithioacetals, MG-GSH and PG-GSH, were found to be sim-
ilar; however, the catalytic efficiency was approximately 4.8-fold lower
1
22
as a potential catalytic base in the enzyme reaction (Glu
in E. coli
GlxI) [66,67]. It would be interesting to determine whether such an
2
+
inhibitor has a different rate of association with a Cd
a Ni
center versus
+
2+
center in E. coli GlxI. The most notable observation from these
kinetic experiments is that different metal ions in the active site of
E. coli GlxI exhibit different kinetic isotope effects when an α-
deuterio-substituted substrate is evaluated.
for the reaction using PG-GSH as the substrate. Interestingly, the Cd2+
-
reconstituted enzyme exhibited lower activity with the hemithioacetal
of MG-GSH than with PG-GSH, which suggested that the substrate
specificity of E. coli GlxI may be dependent upon the specific metal
3.3. Structural relatedness between GlxI and FosA
The gene coding for FosA has been detected in the P. aeruginosa
genome and shares 60% sequence identity with the plasmid encoded
protein from transposon Tn2921 [68]. From this information, it is
possible that genomically encoded FosA might have evolved from a
protein having a similar fold that already existed in an ancestral organ-
ism and likely acted as a “housekeeping” enzyme. GlxI with its high
structural similarity to FosA (Fig. 3) could be a possible candidate as a
progenitor to FosA such that a gene duplication event followed by
multiple rounds of mutations, deletions and insertions could have per-
mitted the evolution of a GlxI molecular framework that developed fos-
fomycin inactivating properties. In an attempt to investigate the
mechanistic relatedness between these two proteins, the deletion of
one of the metal binding ligands in GlxI was performed. Based on a
multiple sequence alignment, the E. coli GlxI-E56A was chosen to pro-
duce a protein containing a similar metal ligand environment around
the active site metal ion to what is found in the active sites of several
FosA enzymes (1× Glu, 2× His and 3 aqua ligands surrounding the
metal center) (Fig. 4; Supplemental Information and Fig. S5). The single
point mutation on E. coli glxI was generated (E56A), and the existence
of the purified mutein was confirmed by analysis of its electrospray
mass spectrum, which exhibited the molecular weight of a single
Table 2
Kinetics of metal reconstitutions (Ni2+ and Cd2+) of E. coli GlxI using the different
hemithioacetal substrates, including MG-GSH, PG-GSH and α-deuterioPG-GSH. The as-
says were performed using 0.02–0.5 mM hemithioacetal substrate in KPB (pH 7.0) at
room temperature.
Metal Substrates
k
cat
K
m
k
cat/K
(M−1 s−1
27± 0.4 12×106
m
k
cat(H)
/
K
m(H)
/
(kcat(H)/Km(H))/
(kcat(D)/Km(D)
k
cat(D)
K
m(D)
)
(
s−1) (μM)
)
Ni
MG-GSHa
PG-GSH
338
–
–
–
322 127± 23 2.5×106
0.85
0.76
1.1
α-deuterioPG- 381 168± 25 2.3×106
GSH
MG-GSH
PG-GSH
α-deuterioPG-
GSH
b
9± 0.4 2.4×106
Cd
21
182 100± 25 1.8×106
80
73± 9 1.1×106
–
–
–
2.28
1.37
1.64
V
max (μmol/min/mg): Ni2+-GlxI (MG-GSH (676± 17), PG-GSH (648± 54), α-deuterioPG-
GSH (766± 69)).
Cd2+-GlxI (MG-GSH (43± 5), PG-GSH (366± 59), α-deuterioPG-GSH (161± 5)).
a
Data obtained from Clugston et al. [23].
Data obtained from Clugston et al. [64].
b

K.Y. Mullings et al. / Journal of Inorganic Biochemistry 108 (2012) 133–140
139
subunit at 14861.5 Da (calculated MW is 14861.8 Da, data not shown).
The mutein's molecular structure was shown to remain unchanged by
a variety of biophysical experiments (Supplementary Information
Fig. S6 and S7).
under the studied conditions. Variation of the divalent metals utilized
(NiCl , CoCl , CuCl , ZnCl , MnCl , CdCl , MgCl and CaCl ) in the incu-
2
2
2
2
2
2
2
2
bation did not result in any detectable fosfomycin-conjugating activity.
As high background levels of detectable GSH were produced in this
reported assay, an alternative method to detect the GS-fosfomycin
adduct was also undertaken. This approach utilized a series of ESI-MS
experiments. The products of the various enzyme incubation reactions
were isolated utilizing Sephadex G-10 gel permeation chromatography
to reduce background contamination. Chemically synthesized
GS-fosfomycin adduct was utilized to determine product retention
characteristics on the Sephadex G-10 column. The purified synthetic
product-containing fractions were confirmed by ESI-MS analysis, for
which the presence of a compound of the expected mass of 444 Da in-
dicated the likely presence of the target GS-fosfomycin adduct. Analysis
of the ESI-MS spectra indicated unresolved purification however,
where GSH was found to co-elute with the compound whose mass cor-
responded to the expected GS-fosfomycin adduct. Detection of the ad-
duct in the ESI-MS of the control experimental fractions (without
addition of the E56A mutein) as well as the enzymatic reaction mixture
were found to be similar, suggesting that small amounts of the
GS-fosfomycin adduct could form without addition of the mutein.
Analysis of the results suggested that the mutein had little or no detect-
able fosfomycin conjugating activity. The presence of varying amounts
The enzymatic activity of the mutein was then studied to evaluate
its glyoxalase I as well as its possible FosA-like activities. The Ni2
+
-
reconstituted E. coli GlxI-E56A protein possessed a similar metal activa-
tion profile for its glyoxalase I activity as does wild-type GlxI, where
enzyme activity was detected for the Ni2 - and Co -reconstituted
forms of the enzyme (Supplementary Information Fig. S8A). The
maximum glyoxalase I activity for the E56A mutein was achieved
+
2+
2
+
−1
with Ni -reconstitution (kcat of 11.5 s
m
and K of 0.68± 0.08 mM)
2
+
and approximately 60% activity was observed with Co -reconstitu-
−
1
tion (kcat of 7.3 s
son to wild-type E. coli GlxI with a reported kcat of 338 s
activity of the mutein was reduced to only b4%. The enzyme reconsti-
m
and K of 0.93± 0.17 mM). However, in compari-
−
1
[23], the
tution with other divalent metals such as Mn2 and Cd
+
2+
resulted in
only trace GlxI activities. Analyses of metal titration profiles indicated
that the enzyme required at least 10 equivalents of metal ion to attain
maximum catalytic activity (Supplementary Information Fig. S8B). This
mutein therefore has reduced metal ion affinity as a result of the muta-
tion of one of the four metal binding residues. These results are consis-
tent with studies undertaken on the H. sapiens (E99N) and
Pseudomonas putida (E93D) GlxI wherein these corresponding muta-
tions reduced enzymatic activities to 1.5% and b0.02% of corresponding
wild type activities, respectively [67,69]. Metal analysis of the human
GlxI E99N mutant indicated that the mutein only contained 0.3 mol
zinc per subunit enzyme [67]. Thus, low metal content in the enzyme
may explain the decrease in glyoxalase I activity for the E. coli
GlxI-E56A mutein observed in the current study; however, saturating
amounts of metal ions produced no more than b4% of the wild type
activity.
The fosfomycin conjugating activity of the E. coli Glx-E56A mutein,
however, was undetectable by use of an established DTNB assay,
which previously had been successfully utilized to measure FosA activ-
ity [33]. The GlxI-E56A mutein was incubated in the presence of GSH,
fosfomycin and various concentrations of a variety of metal ions for
various lengths of time. The reaction was subsequently quenched by
the addition of methanol. The unreacted GSH in the solution was
then monitored by DTNB detection, thus the enzyme kinetics was
examined in terms of the decrease in substrate thiol levels. However,
the absorbance of the quenched reaction mixture were similar to
background detection, suggesting that there was either little or no
formation of the target GS-fosfomycin adduct by the GlxI-E56A mutein
of divalent metals (Ni2 and Mn ) did not alter the negative out-
+
2+
come of the experiments.
Even though the active site of E. coli GlxI was mutated in order to
mimic the metal liganding environment present in FosA, other func-
tional parts in the active site structure were different. Both share
overall structural similarity; however, the differences can be clearly
observed in the connecting loops (Fig. 3). As well, limited catalytic
space in GlxI might not support the binding of the two substrates, fos-
fomycin and GSH. Several critical residues present in FosA but absent
in the E. coli GlxI-E56A mutein likely further explain the absence of
detectable fosfomycin-conjugating activity by GlxI-E56A (Supple-
mentary Information Fig. S5).
4. Conclusions
The current investigation has extended our understanding of the
Ni2+-activated GlxI class of metalloenzyme. The E. coli Glx system,
composed of GlxI and GlxII, has been shown to utilize a variety of
2
thiol cofactors. GSH, GspdSH and T(SH) cofactors can be utilized to
stereospecifically convert MG to D-lactate by the E. coli Glx system.
As well, the metal ion-dependent presence of a deuterium isotope
effect in E. coli GlxI is the first non Zn2 -dependent GlxI enzyme
characterized by such experimental approaches. An observable deu-
+
terium kinetic isotope effect has been detected for the Cd2+
-
substituted E. coli GlxI indicating that α-proton transfer is partially
rate-limiting in the enzymatic reaction.
Due to the high structural similarity between GlxI and FosA, it is
possible that GlxI directed evolution experiments could produce
FosA enzymatic activity in a GlxI enzyme. Although the current inves-
tigation was unable to detect FosA activity by mutation of a key metal
liganding residue in the active site of E. coli GlxI, additional studies on
the evolution of antibiotic (FosA) and toxic metabolite (Glx) resis-
tance proteins from ancestral protoresistance proteins is a subject of
exciting current interest [70] and worthy of further investigation.
Abbreviations
ESI-MS electrospray ionization mass spectrometry
Fig. 4. A three-dimensional figure of the detailed active sites for the superimposed
structures of P. aeruginosa FosA (cyan, PDB ID: 1LQP) and E. coli GlxI (magenta, PDB
FosA
Glx
GSH
fosfomycin resistance protein A
glyoxalase
glutathione
ID: 1F9Z) showing the active site metals Mn2+ (orange) in FosA and Ni
2+
(blue) in
GlxI. The oxygen atoms of the two water molecules bound to Ni2+ in the GlxI active
site are magenta spheres. The mutated Glu56 in the GlxI is colored yellow and the cor-
responding residue in FosA, Cys48, is colored green. A fosfomycin molecule (red)
bound to FosA in the X-ray structure is also shown to elucidate the fosfomycin binding
area (wall-eyed view).
GspdSH glutathionylspermidine
(
GspdD)
2
glutathionylspermidine disulfide
4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid
HEPES

140
K.Y. Mullings et al. / Journal of Inorganic Biochemistry 108 (2012) 133–140
[
20] S.L. Oza, M.P. Shaw, S. Wyllie, A.H. Fairlamb, Mol. Biochem. Parasitol. 139 (2005)
07–116.
21] L. Thomson, A. Denicola, R. Radi, Arch. Biochem. Biophys. 412 (2003) 55–64.
ICP-MS inductively coupled plasma mass spectrometry
1
KIE
kinetic isotope effect
[
D-LDH
L-LDH
MG
D-lactate dehydrogenase
L-lactate dehydrogenase
methylglyoxal
[22] S. Sellin, L.E. Eriksson, A.C. Aronsson, B. Mannervik, J. Biol. Chem. 258 (1983)
091–2093.
2
[
23] S.L. Clugston, J.F. Barnard, R. Kinach, D. Miedema, R. Ruman, E. Daub, J.F. Honek,
Biochemistry 37 (1998) 8754–8763.
PAR
PCR
PDB
PG
Spd
TCEP
4-(2-pyridylazo)resorcinol)
polymerase chain reaction
protein data bank
phenylglyoxal
spermidine
[24] A.D. Cameron, B. Olin, M. Ridderstrom, B. Mannervik, T.A. Jones, EMBO J. 16
(1997) 3386–3395.
[
25] A.D. Cameron, M. Ridderstrom, B. Olin, M.J. Kavarana, D.J. Creighton, B. Mannervik,
Biochemistry 38 (1999) 13480–13490.
[26] M.M. He, S.L. Clugston, J.F. Honek, B.W. Matthews, Biochemistry 39 (2000)
8719–8727.
[
[
[
[
27] S.S. Hall, A.M. Doweyko, F. Jordan, J. Am. Chem. Soc. 98 (1976) 7460–7461.
28] D.L. Vander Jagt, L.P. Han, Biochemistry 12 (1973) 5161–5167.
29] R.N. Armstrong, Biochemistry 39 (2000) 13625–13632.
Tris[2-carboxyethyl]phosphine hydrochloride
trypanothione (reduced state)
T(SH)
2
30] M. Bergdoll, L.D. Eltis, A.D. Cameron, P. Dumas, J.T. Bolin, Protein Sci. 7 (1998)
1661–1670.
[
31] I. Feierberg, A.D. Cameron, J. Aqvist, FEBS Lett. 453 (1999) 90–94.
Acknowledgments
[32] A.P. Saint-Jean, K.R. Phillips, D.J. Creighton, M.J. Stone, Biochemistry 37 (1998)
0345–10353.
1
[
[
33] Z. Beharry, T. Palzkill, J. Biol. Chem. 280 (2005) 17786–17791.
34] L.T. Laughlin, B.A. Bernat, R.N. Armstrong, Chem. Biol. Interact. 111–112 (1998)
41–50.
The authors would like to thank Dr. Richard Smith for electrospray
ionization mass spectrometry analysis of the samples and Dr. Zhengding
Su for sharing his knowledge and suggestions. This research was sup-
ported by the Natural Sciences and Engineering Research Council of
Canada (NSERC) (JFH), University of Waterloo (Ontario, Canada), an
NSERC graduate scholarship (to NS), the Government of Thailand (grad-
uate scholarship to US) and the Government of China (visiting scientist
award to YR). The funding sources had no involvement in any of the as-
pects of the experimental research, the interpretation of results or the
preparation of this article.
[35] C.L. Rife, R.E. Pharris, M.E. Newcomer, R.N. Armstrong, J. Am. Chem. Soc. 124
2002) 11001–11003.
[
[
[38] B.A. Bernat, L.T. Laughlin, R.N. Armstrong, Biochemistry 38 (1999) 7462–7469.
[
(
36] B.A. Bernat, R.N. Armstrong, Biochemistry 40 (2001) 12712–12718.
37] B.A. Bernat, L.T. Laughlin, R.N. Armstrong, Biochemistry 36 (1997) 3050–3055.
39] T. Hidaka, M. Goda, T. Kuzuyama, N. Takei, M. Hidaka, H. Seto, Mol. Gen. Genet.
49 (1995) 274–280.
40] L.J. Higgins, F. Yan, P. Liu, H.W. Liu, C.L. Drennan, Nature 437 (2005) 838–844.
2
[
[41] R.D. Woodyer, Z. Shao, P.M. Thomas, N.L. Kelleher, J.A. Blodgett, W.W. Metcalf,
W.A. van der Donk, H. Zhao, Chem. Biol. 13 (2006) 1171–1182.
42] W.W. Metcalf, W.A. van der Donk, Annu. Rev. Biochem. 78 (2009) 65–94.
43] N. Sukdeo, S.L. Clugston, E. Daub, J.F. Honek, Biochem. J. 384 (2004) 111–117.
[
[
Appendix A. Supplementary data
[44] N. Sukdeo, J.F. Honek, Biochim. Biophys. Acta 1774 (2007) 756–763.
[
[
[
45] M.M. Bradford, Anal. Biochem. 72 (1976) 248–254.
46] P.W. Riddles, R.L. Blakeley, B. Zerner, Anal. Biochem. 94 (1979) 75–81.
47] D.L. Vander Jagt, L.P. Han, C.H. Lehman, Biochemistry 11 (1972) 3735–3740.
Supplementary data to this article can be found online at doi:10.
1
016/j.jinorgbio.2011.11.008.
[48] R. Vince, S. Daluge, W.B. Wadd, J. Med. Chem. 14 (1971) 402–404.
[
[
49] E.E. Cliffe, S.G. Waley, Biochem. J. 79 (1961) 475–482.
50] D.J. Creighton, M. Migliorini, T. Pourmotabbed, M.K. Guha, Biochemistry 27
(1988) 7376–7384.
References
[
51] C.E. Griffis, L.H. Ong, L. Buettner, D.J. Creighton, Biochemistry 22 (1983)
2945–2951.
[
1] I.R. Booth, G.P. Ferguson, S. Miller, C. Li, B. Gunasekera, S. Kinghorn, Biochem. Soc.
Trans. 31 (2003) 1406–1408.
2] R.A. Cooper, Annu. Rev. Microbiol. 38 (1984) 49–68.
3] K.M. Desai, T. Chang, H. Wang, A. Banigesh, A. Dhar, J. Liu, A. Untereiner, L. Wu,
Can. J. Physiol. Pharmacol. 88 (2010) 273–284.
4] M.P. Kalapos, Toxicol. Lett. 110 (1999) 145–175.
[52] D.L. Vander Jagt, E. Daub, J.A. Krohn, L.P. Han, Biochemistry 14 (1975) 3669–3675.
[53] T.J. Vickers, N. Greig, A.H. Fairlamb, Proc. Natl. Acad. Sci. USA 101 (2004)
13186–13191.
[54] J. O'Young, N. Sukdeo, J.F. Honek, Arch. Biochem. Biophys. 459 (2007) 20–26.
[55] W.G. Gutheil, Anal. Biochem. 259 (1998) 62–67.
[
[
[
[
[
[
[
5] M.P. Kalapos, Drug Metabol. Drug Interact. 23 (2008) 69–91.
6] P.J. Thornalley, Biochem. Soc. Trans. 31 (2003) 1372–1377.
7] U. Suttisansanee, J.F. Honek, Semin. Cell Dev. Biol. 22 (2011) 285–292.
8] M.J. MacLean, L.S. Ness, G.P. Ferguson, I.R. Booth, Mol. Microbiol. 27 (1998)
[56] T. Irsch, R.L. Krauth-Siegel, J. Biol. Chem. 279 (2004) 22209–22217.
[57] A. Sommer, P. Fischer, K. Krause, K. Boettcher, P.M. Brophy, R.D. Walter, E. Liebau,
Biochem. J. 353 (2001) 445–452.
[58] M. Akoachere, R. Iozef, S. Rahlfs, M. Deponte, B. Mannervik, D.J. Creighton, H.
Schirmer, K. Becker, Biol. Chem. 386 (2005) 41–52.
563–571.
[
9] C.E. Hand, J.F. Honek, J. Nat. Prod. 68 (2005) 293–308.
[59] P.K. Padmanabhan, A. Mukherjee, R. Madhubala, Biochem. J. 393 (2006) 227–234.
[60] Y.C. Chai, S.S. Ashraf, K. Rokutan, R.B. Johnston Jr., J.A. Thomas, Arch. Biochem. Bio-
phys. 310 (1994) 273–281.
[
10] R.C. Fahey, W.C. Brown, W.B. Adams, M.B. Worsham, J. Bacteriol. 133 (1978)
1126–1129.
[
[
11] A. Meister, M.E. Anderson, Annu. Rev. Biochem. 52 (1983) 711–760.
12] G.L. Newton, K. Arnold, M.S. Price, C. Sherrill, S.B. Delcardayre, Y. Aharonowitz, G.
Cohen, J. Davies, R.C. Fahey, C. Davis, J. Bacteriol. 178 (1996) 1990–1995.
[61] Y.C. Chai, S. Hendrich, J.A. Thomas, Arch. Biochem. Biophys. 310 (1994) 264–272.
[62] J. Melchers, N. Dirdjaja, T. Ruppert, R.L. Krauth-Siegel, J. Biol. Chem. 282 (2007)
8678–8694.
[
[
13] A. Pastore, G. Federici, E. Bertini, F. Piemonte, Clin. Chim. Acta 333 (2003) 19–39.
14] K. Smith, A. Borges, M.R. Ariyanayagam, A.H. Fairlamb, Biochem. J. 312 (1995)
[63] N. Greig, S. Wyllie, S. Patterson, A.H. Fairlamb, FEBS J. 276 (2009) 376–386.
[64] S.L. Clugston, R. Yajima, J.F. Honek, Biochem. J. 377 (2004) 309–316.
[65] S.F. Lincoln, Helv. Chim. Acta 88 (2005) 523–545.
465–469.
[
[
15] H. Tabor, C.W. Tabor, J. Biol. Chem. 250 (1975) 2648–2654.
16] S. Awad, G.B. Henderson, A. Cerami, K.D. Held, Int. J. Radiat. Biol. 62 (1992)
[66] G. Davidson, S.L. Clugston, J.F. Honek, M.J. Maroney, Biochemistry 40 (2001)
4569–4582.
4
01–407.
[67] M. Ridderstrom, A.D. Cameron, T.A. Jones, B. Mannervik, J. Biol. Chem. 273 (1998)
21623–21628.
[68] S. Pakhomova, C.L. Rife, R.N. Armstrong, M.E. Newcomer, Protein Sci. 13 (2004)
1260–1265.
[
[
17] N. Greig, S. Wyllie, T.J. Vickers, A.H. Fairlamb, Biochem. J. 400 (2006) 217–223.
18] A. Ariza, T.J. Vickers, N. Greig, K.A. Armour, M.J. Dixon, I.M. Eggleston, A.H.
Fairlamb, C.S. Bond, Mol. Microbiol. 59 (2006) 1239–1248.
[
19] P.R. Romao, J. Tovar, S.G. Fonseca, R.H. Moraes, A.K. Cruz, J.S. Hothersall, A.A.
[69] Y. Lan, T. Lu, P.S. Lovett, D.J. Creighton, J. Biol. Chem. 270 (1995) 12957–12960.
[70] M. Morar, G.D. Wright, Annu. Rev. Genet. 44 (2010) 25–51.
Noronha-Dutra, S.H. Ferreira, F.Q. Cunha, Braz. J. Med. Biol. Res. 39 (2006)
355–363.