Kinetic and mechanistic characterization of the glyceraldehyde 3-phosphate dehydrogenase from Mycobacterium tuberculosis

Article abstract of DOI:10.1016/j.abb.2013.10.007

Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) is a glycolytic protein responsible for the conversion of glyceraldehyde 3-phosphate (G3P), inorganic phosphate and nicotinamide adenine dinucleotide (NAD⁺) to 1,3-bisphosphoglycerate (1,3-BPG) and the reduced form of nicotinamide adenine dinucleotide (NADH). Here we report the characterization of GAPDH from Mycobacterium tuberculosis (Mtb). This enzyme exhibits a kinetic mechanism in which first NAD⁺, then G3P bind to the active site resulting in the formation of a covalently bound thiohemiacetal intermediate. After oxidation of the thiohemiacetal and subsequent nucleotide exchange (NADH off, NAD⁺ on), the binding of inorganic phosphate and phosphorolysis yields the product 1,3-BPG. Mutagenesis and iodoacetamide (IAM) inactivation studies reveal the conserved C158 to be responsible for nucleophilic catalysis and that the conserved H185 to act as a catalytic base. Primary, solvent and multiple kinetic isotope effects revealed that the first half-reaction is rate limiting and utilizes a step-wise mechanism for thiohemiacetal oxidation via a transient alkoxide to promote hydride transfer and thioester formation.

Full text of DOI:10.1016/j.abb.2013.10.007

Archives of Biochemistry and Biophysics 540 (2013) 53–61
Contents lists available at ScienceDirect
Archives of Biochemistry and Biophysics
journal homepage: www.elsevier.com/locate/yabbi
Kinetic and mechanistic characterization of the glyceraldehyde
3-phosphate dehydrogenase from Mycobacterium tuberculosis
⇑
Brett Wolfson-Stofko, Timin Hadi, John S. Blanchard
Department of Biochemistry, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) is a glycolytic protein responsible for the conver-
sion of glyceraldehyde 3-phosphate (G3P), inorganic phosphate and nicotinamide adenine dinucleotide
(NAD+) to 1,3-bisphosphoglycerate (1,3-BPG) and the reduced form of nicotinamide adenine dinucleotide
(NADH). Here we report the characterization of GAPDH from Mycobacterium tuberculosis (Mtb). This
enzyme exhibits a kinetic mechanism in which first NAD+, then G3P bind to the active site resulting in
the formation of a covalently bound thiohemiacetal intermediate. After oxidation of the thiohemiacetal
and subsequent nucleotide exchange (NADH off, NAD+ on), the binding of inorganic phosphate and phos-
phorolysis yields the product 1,3-BPG. Mutagenesis and iodoacetamide (IAM) inactivation studies reveal
the conserved C158 to be responsible for nucleophilic catalysis and that the conserved H185 to act as a
catalytic base. Primary, solvent and multiple kinetic isotope effects revealed that the first half-reaction is
rate limiting and utilizes a step-wise mechanism for thiohemiacetal oxidation via a transient alkoxide to
promote hydride transfer and thioester formation.
Received 11 September 2013
and in revised form 8 October 2013
Available online 23 October 2013
Keywords:
Glyceraldehyde 3-phosphate
dehydrogenase
Glycolysis
Enzyme kinetics
Kinetic isotope effects
Tuberculosis
Ó 2013 Elsevier Inc. All rights reserved.
Introduction
dehydrogenase is also unusual in that it utilizes a covalent thiohem-
iacetal intermediate to promote hydride transfer and catalysis [3].
The etiological agent of tuberculosis, Mycobacterium tuberculosis
(Mtb), has infected nearly one-third of the human population [1].
Approximately 10% of TB-infections lead to an active, symptomatic
infection that resulted in nearly 1.4 million deaths in 2011 [1]. In
addition, multi-drug resistant strains have been reported in every
country surveyed by the World Health Organization [1]. Yet some
of the most basic metabolic enzymes of this bacterium have yet to
be characterized.
Glyceraldehyde 3-phosphate dehydrogenase is a highly con-
served enzyme that is utilized in central carbon metabolism by
some of the most ancient forms of life [2]. GAPDH is best known
for its role in glycolysis, catalyzing the reversible conversion of
glyceraldehyde 3-phosphate (G3P), inorganic phosphate and
NAD+ to 1,3-bisphosphoglycerate (1,3-BPG)1 and NADH [3]. This
The reaction of GAPDH is essential for the regeneration of the two
molecules of ATP used to phosphorylate the hexose carbon source,
glucose. The cleavage of fructose-1,6-bisphosphate yields the two
triose phosphates that are interconverted into G3P. The oxidation
of the aldehyde and substrate-level phosphorylation catalyzed by
GAPDH generate NADH and the high energy carboxy-phosphoric
anhydride containing 1,3-bisphosphoglycerate (1,3-BPG) that is used
in the subsequent reaction catalyzed by 3-phosphoglycerate kinase
to regenerate the two molecules of ATP used earlier in the glycolytic
sequence. The very reactive nature of the product of GAPDH, 1,3-
BPG, has recently been shown to be capable of non-enzymatic mod-
ification of proteins, including GAPDH [4].
Recent studies have also found GAPDH to be involved in a vari-
ety of cellular processes in addition to its major role in glycolysis.
GAPDH has been shown to play a role in transcription, assisting in
the formation of both DNA and RNA binding complexes as well as
acting as a transcription factor co-activator [5–7]. Additionally,
GAPDH has been identified as a microtubule-binding protein, a lac-
toferrin receptor, and as an apoptosis-inducer [8–11]. More infor-
mation on the extra-glycolytic roles of GAPDH can be found in
the review by Nichollis et al. [12].
⇑
Corresponding author. Fax: +1 718 430 8565.
E-mail addresses: blanchar@aecom.yu.edu, john.blanchard@einstein.yu.edu (J.S.
Blanchard).
1
Abbreviations used: 1,3-BPG, 1,3-bisphosphoglycerate; DTT, dithiolthreitol; D-G3P,
D-glyceraldehyde 3-phosphate; 1-[2H]D-G3P, 1-[2H]D-glyceraldehyde 3-phosphate;
G3P, Glyceraldehyde 3-phosphate; GAPDH, Glyceraldehyde 3-phosphate dehydroge-
nase; HEPES, N-(2-hydroxyethyl)piperazine-N0-(2-ethanesulfonic acid); IAM, iodoa-
cetamide; LB, Luria–Bertani; MES, 2-(N-morpholino)ethanesulfonic acid; MKIE,
multiple kinetic isotope effect; NAD+, nicotinamide adenine dinucleotide; NADH,
reduced nicotinamide adenine dinucleotide; Ni-NTA, nickel nitriloacetic acid; PCR,
polymerase chain reaction; PKIE, primary isotope effect; SDS–PAGE, sodium dodecyl
sulfate–polyacrylamide gel electrophoresis; SKIE, solvent kinetic isotope effect; TAPS,
N-[tris(hydroxymethyl)methyl]-3-aminopropanesulfonic acid.
Despite decades of work on GAPDH’s from prokaryotic and
eukaryotic sources, no work has been conducted on the GAPDH
from M. tuberculosis. It was discovered early on in our work that
this enzyme had significant solubility issues. This obstacle was
overcome by co-transforming the Mtb-GAPDH plasmid along with
0003-9861/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.abb.2013.10.007

54
B. Wolfson-Stofko et al. / Archives of Biochemistry and Biophysics 540 (2013) 53–61
O
O
H
H
DIBAL-H
CH2Cl2
-78 οC
H2, Pd/C
MeOH
O
P
OBn
O
P
OH
O
OBn
O
OH
1)
N
P(OBn)2
BnO
HO
O
O
O
DOWEX 50WX4
anh. MeOH
OMe
OBn
(4)
(6)
OMe
OBn
OMe
triazole, CH2Cl2
2) H2O2, THF
O
P
OBn
O
OTr
OH
BnO
OBn
(2)
(1)
(3)
O
O
D
D
H2, Pd/C
MeOH
O
P
O
P
DIBAL-D
CH2Cl2
-78 οC
OH
OBn
O
OH
O
OBn
HO
BnO
(5)
(7)
Scheme 1. Synthesis of
D
-glyceraldehyde 3-phosphate and [1-2H]
D-glyceraldehyde 3-phosphate.
a plasmid expressing the chaperones GroEL/GroES [13]. These
chaperones are believed to create an environment suitable for
proper folding yielding soluble and active Mtb-GAPDH [13]. In this
study, we report the first successful purification and mechanistic
evaluation of Mtb-GAPDH using steady-state kinetics, pH-rate pro-
files, isotope effects and mutagenesis to elucidate both the kinetic
and chemical mechanism of Mtb-GAPDH.
Construction and expression of Mtb-GAPDH mutants C158A, C162A
and H185A
Mtb-GAPDH/pET28a was used as a template to generate C158A,
C162A and H185A mutants. The mutants were constructed by
overlap mutagenesis [14]. The mutation has been underlined in
the following sequences. The forward primer for C158A was 50-
CTCCAATGCGTCGGCCACCACGAACTGCC-30 and the reverse primer
was 50-GGCAGTTCGTGGTGGCCGACGCATTGGAG-30. The forward
primer for C162A was 50-GTGCACCACGAACGCCCTTGCGCCGCTGG-
30 and the reverse primer was 50-CCAGCGGCGCAAGGGCGTTCGTGG
TGCAC-30. The forward primer for H185A was 50-GATGACCAC-
CATCGCCGCCTACACTCAGG-30 and the reverse primer was 50-CCTG
AGTGTAGGCGGCGATGGTGGTCATC-30. DNA sequencing was used
to confirm the described mutations. Each mutant was expressed
and purified in the same manner as wild-type Mtb-GAPDH.
Materials and methods
Materials
All chemicals were purchased from Sigma–Aldrich unless other-
wise noted. The Mtb-GAPDH gene was cloned into the Novagen
pET-28a(+) vector. The GroEL/GroES plasmid was a gift from the
Shrader Lab [13]. Primers were purchased from Invitrogen. All
cloning enzymes and T7 competent Escherichia coli were purchased
from New England Biolabs. Complete EDTA-free protease inhibitor
cocktail and DNase were purchased from Roche. 99.9% deuterated
water was purchased from Cambridge Isotope Laboratories.
Protein concentration
Protein concentration was determined using Bio-Rad Bradford
protein assay using bovine serum albumin as a standard.
Measurement of enzymatic activity
Cloning, expression and purification of Mtb-GAPDH
Enzymatic activity was measured spectrophotometrically by
monitoring the conversion of NAD+ to NADH at 340 nm. Reactions
were conducted in 50 mM HEPES, pH 8 including substrates NAD+,
The M. tuberculosis gap gene (Rv1436) was PCR amplified from
the Erdman strain with a forward primer 50-GGAATTCCATATGGTG
ACGGTCCGAGTAGGC-30 and a reverse primer 50-GTCGGCAAGTCGC
TCTAGAAGCTTGGG-30. NdeI and HindIII restriction sites were used
for forward and reverse primers, respectively. The PCR fragment
was ligated into the pET28a(+) vector encoding for a N-terminal
His6-tag. The plasmid was then sequenced and confirmed. The
Mtb-GAPDH-containing plasmid was co-transformed along with
the GroEL/GroES plasmid into T7 Express E. coli competent cells.
Na2AsO4, and DL-G3P in a total volume of 500 lL. Reactions were
initiated by the addition of Mtb-GAPDH. Initial velocities were
measured before 10% conversion of substrate to product had oc-
curred and initial rates were calculated using the molar extinction
coefficient of NADH (
concentration.
e
340 = 6220 Mꢁ1 cmꢁ1) and the total enzyme
The measurements of kinetic isotope effects were conducted
in the same manner with a few differences. We synthesized and
Kanamycin (35
selection. Cultures were grown in LB broth at 30 °C and induced
with 500
M IPTG at an A600 of ꢀ0.6–0.8 and then grown overnight
lg/mL) and tetracyclin (6 lg/mL) were used for
utilized the individual stereoisomers, [1-1H]
D
-glyceraldehyde 3-
phosphate (1-[1H]D-G3P) and [1-2H]
D-glyceraldehyde 3-phosphate
l
(1-[2H]D-G3P) for these studies. Concentrations of 1-[1H]D-G3P and
1-[2H]D-G3P were determined enzymatically using excess NAD+
and arsenate.
at 18 °C. Cells were harvested by centrifugation and stored at
ꢁ20 °C. The pellets were resuspended in 25 mM HEPES (pH 7.5)
containing 300 mM NaCl, 10 mM imidazole, and 1 mM NAD+. Cells
were lysed using an EmulsiFlex-C3 and centrifuged to remove cel-
lular debris. The clear supernatant was then added to a Ni2+-NTA
agarose column and eluted with a linear imidazole gradient
(10 mM-250 mM). Fractions containing Mtb-GAPDH were pooled
and dialyzed into 25 mM HEPES (pH 7.5) containing 300 mM NaCl,
1 mM NAD+ and 5% glycerol then concentrated and stored at
ꢁ20 °C in 12.5% glycerol.
Synthesis of
3-phosphate
D D-glyceraldehyde
-glyceraldehyde 3-phosphate and [1-2H]
Briefly, the known compound, methyl (R)-2-benzyloxy-3-trityl-
oxypropanoate (1), was synthesized from commercially available
methyl 2,3-O-isopropylidene-D-glycerate according to previously

B. Wolfson-Stofko et al. / Archives of Biochemistry and Biophysics 540 (2013) 53–61
55
estimated by saturating two substrates while varying the concen-
tration of the third. These rates were fit to Eq. (2)
m
¼ ðVSÞ=ðkm þ SÞ
ð2Þ
where V is the maximal velocity and S is the concentration of the var-
ied substrate. Initial velocity studies were conducted by saturating
Fig. 1. Michaelis–Menten experiments with Mtb-GAPDH utilizing phosphate (A) or
arsenate (B) as the varying substrates with 50 mM HEPES, pH 8, and saturating
concentration of G3P and NAD+. Data were fit to Eq. (2). Substrate inhibition profiles
utilizing phosphate (inset A) or arsenate (inset B) with 50 mM HEPES, pH 8 and
saturating concentration of G3P and NAD+. Data were fit to Eq. (1).
published procedures [15]. The 3-O-trityl ether was removed to
generate compound 2 bearing a free hydroxyl group at the C3 po-
sition (Scheme 1). This hydroxyl group was phosphorylated using a
phosphoramidite coupling reagent to generate the common inter-
mediate compound 3 that was used for the synthesis of both
D
-glyceraldehyde 3-phosphate and [1-2H]D-glyceraldehyde 3-phos-
phate. To generate the aldehyde at the C1 position, compound 3
was treated with DIBAL-H or DIBAL-D to generate compounds 4
and 5, respectively.
D-Glyceraldehyde 3-phosphate (6) and
[1-2H]
D
-glyceraldehyde 3-phosphate (7) were then generated
through hydrogenolysis (1 atm H2, Pd/C catalyst) of the benzyl-
protected compounds. Further experimental detail and NMR spec-
tra can be found in the Supplementary Information.
Initial velocity studies
Data were fit using SigmaPlot 11.0. All points are the mean of
experimental duplicates and the error bars are plus/minus one half
of the difference between the experimental duplicates. Substrate
inhibition curves were determined and fitted to Eq. (1) for linear
substrate inhibition
Fig. 2. Initial velocity studies. (A) Initial rate data were obtained by varying
concentrations of G3P and fixed, variable concentrations of NAD+ at 24 (d), 30 (N),
40 (j), 60 (ꢀ) and 120 lM (.) and high non-inhibitory concentrations of Na2AsO4.
The data were fit to Eq. (4). (B) Initial rate data were obtained by varying
concentrations of Na2AsO4 and fixed, variable concentrations of NAD+ at 20 (d), 30
m
¼ VS=ðKm þ S þ S2=KiÞ
ð1Þ
(N), 40 (j), 60 (ꢀ) and 100 lM (.) at saturating concentrations of G3P. The data
were fit to Eq. (3). (C) Initial rate data were obtained by varying concentrations of
G3P and fixed, variable concentrations of Na2AsO4 at 2 (d), 3 (N), 4 (j), and 8 mM
(ꢀ) at saturating concentrations of NAD+. The data were fit to Eq. (4).
where
v
is the velocity, S is the concentration of substrate and Ki is
the substrate inhibition constant. Initial kinetic parameters were

56
B. Wolfson-Stofko et al. / Archives of Biochemistry and Biophysics 540 (2013) 53–61
where V is the maximal velocity, A and B are substrate concentra-
tions, Ka and Kb are the respective Michaelis constants for each of
the substrates, and Kia is the inhibition constant for substrate A.
Product inhibition was performed using reduced nicotinamide ade-
nine dinucleotide (NADH) as an inhibitor. Data was obtained at var-
iable concentrations of NAD+, G3P or arsenate at several fixed
concentrations of NADH. The data were fit to competitive, uncom-
petitive, and noncompetitive inhibition using Eqs. (5)–(7),
respectively
m
m
m
¼ ðVSÞ=½Kmð1 þ I=KisÞ þ Sꢂ
ð5Þ
ð6Þ
ð7Þ
¼ ðVSÞ=½Km þ Sð1 þ I=KiiÞꢂ
¼ ðVSÞ=½Kmð1 þ I=KisÞ þ Sð1 þ I=KiiÞꢂ
where S is the varied substrate concentration, I is the inhibitor con-
centration, Kis is the inhibition constant for the slope, Kii is the inhi-
bition constant for the intercept, and Km is the Michaelis constant
for the substrate S.
pH dependence studies
The pH dependence of the kinetic parameters was determined
in 50 mM concentrations of buffers at various pH values and using
the following structurally related buffers at the pH values noted in
parentheses: MES (pH 5–6.5), HEPES (pH 6.5–8), TAPS (pH 8–10).
The pH profile was determined varying either G3P or Na2AsO4 with
other substrates held saturating. The data were fit to Eq. (8)
log k ¼ log½c=ð1 þ Kb=10ꢁpH þ 10ꢁpH=KaÞꢂ
ð8Þ
where c is the pH-independent plateau value and Ka and Kb are dis-
sociation constants of the ionizing groups.
Inactivation studies
Inactivation studies were performed using iodoacetamide (IAM)
as an inactivator. Enzyme activities were measured for native Mtb-
GAPDH every 15-min and every 20-min for Mtb-GAPDH C162A at
varying concentrations of IAM. Kinetic constants were calculated
as described by Kitz and Wilson [16]. Briefly, relative inhibition
was determined by comparing the rates of time-matched samples
exposed to varying concentrations of IAM to control samples lack-
ing IAM. The results were plotted on a logarithmic scale versus
time and the slopes were determined by linear regression. The
values of the slopes represent the inactivation rate constants and
were used to calculate the half-life of the inactivation
(kinactivation = 0.693/t1/2). A Kitz–Wilson plot was produced using
G3P
E-NAD+
13-BPG
NADH
E-NAD+-G3P
E-NAD+-X
Fig. 3. Product inhibition studies. (A) Initial rate data were obtained at varying
concentrations of NAD+ and fixed, variable concentrations of NADH at 0 (ꢀ), 25 (j),
E-Y
NAD+
50 (N), and 100
obtained at varying concentrations of Na2AsO4 and fixed, variable concentrations of
NADH at 0 (ꢀ), 10 (j), 24 (N), and 40 M (d). The data were fit to Eq. (5). (C)
Varying concentrations of G3P and fixed, variable concentrations of NADH at 0 (ꢀ),
20 (j), 40 (N), and 70 M (d). The data were fit to Eq. (7).
lM (d). The data were fit to Eq. (5). (B) Initial rate data were
E-Y-NADH
l
NAD+
E
l
E-Y-NAD+
E-NAD+-13BPG
E-Y-NAD+-Pi
with one substrate, fixing the concentration of another, and varying
the concentration of the third. The results were fitted for intersecting
and parallel patterns using Eqs. (3) and (4), respectively
m
m
¼ ðVABÞ=ðKiaKb þ KaB þ KbA þ ABÞ
¼ ðVABÞ=ðKaB þ KbA þ ABÞ
ð3Þ
ð4Þ
Pi
Scheme 2. Proposed kinetic mechanism of Mtb-GAPDH.

B. Wolfson-Stofko et al. / Archives of Biochemistry and Biophysics 540 (2013) 53–61
57
the calculated half-lives as the y-axis and the reciprocal of
the inhibitor concentration as the x-axis [16]. Values for Ki and
kinactivation were determined from the negative reciprocal of the
x-intercept and reciprocal of the y-intercept, respectively.
Consequently, concentration values reported herein are for the
DL-G3P mixture and presumed to be two times the value for
D-G3P alone. Double-reciprocal plots yielded a series of parallel lines
when varying G3P concentration at several fixed NAD+ concentra-
tions and an optimal, but not saturating concentration of Na2AsO4
(Fig. 2A). Because the concentration of Na2AsO4 was chosen to
avoid substrate inhibition, reported values of V may slightly under-
estimate the actual value. Fitting these data to Eq. (4) yielded
Kinetic isotope effects
Solvent kinetic isotope effects were determined at varying con-
centrations of either D-G3P or Na2AsO4 while the other substrates
were kept at saturating concentrations in H2O or 90% D2O. A vis-
cosity control of 9% glycerol was performed and revealed no effect
on V or V/Km [17]. The data were fit to Eq. (9)
V = 1670 90 min.ꢁ1, KG3P = 280 30
lM and KNAD+ = 40
4 lM.
Varying Na2AsO4 concentrations at several fixed NAD+ concentra-
tions resulted in a series of intersecting lines that when fit to Eq.
(3) yielded KNa2AsO4 = 3.9 1.5 mM, and KNAD+ = 70 20 lM
(Fig. 2B). Varying G3P concentrations at several fixed concentra-
tions of Na2AsO4 resulted in a series of parallel lines that when
fit to Eq. (4) yielded V = 1710 60 min.ꢁ1 and KNa2AsO4 = 5.1 0.3 -
mM (Fig. 2C). To elucidate the binding order of substrates and
product release, product inhibition studies were performed. NADH
was determined to exhibit competitive inhibition versus both
NAD+ and Na2AsO4 (Fig. 3A and B). Fits of these data to Eq. (5)
m
¼ ðVSÞ=½KAð1 þ FiðEV=K ꢁ 1ÞÞ þ Sð1 þ FiðEV ꢁ 1ÞÞꢂ
ð9Þ
where V is the maximal velocity, S is the substrate concentration,
EV/K is the effect on kcat/Km, EV is the effect of kcat and Fi is the frac-
tion of the isotope (Fi = 0 for H2O and Fi = 0.9 for D2O). A proton
inventory was then conducted using saturating conditions of
substrates while varying D2O concentrations from 0% to 90%. The
results were fit to a linear equation.
yielded Ki values of 14
AsO4, respectively. NADH exhibited noncompetitive inhibition ver-
sus G3P and when fit to Eq. (7) yielded Kii = 24 and
Kis = 31 M (Fig. 3C).
1
l
M and 37
1 l
M versus NAD+ and Na2-
Primary and multiple kinetic isotope effects were determined at
5
varying concentrations of either [1-1H] -G3P or [1-2H]D-G3P while
D
other substrates were kept at saturating conditions in either H2O
or 90% D2O. The results were fitted to Eq. (9).
7
l
Results and discussion
Cloning, expression and purification
Mtb-GAPDH was PCR amplified then cloned into the pET28a
expression vector encoding a N-terminal His6-tagged Mtb-GAPDH.
Sequencing confirmed that no mutations were introduced during
the cloning process. Initial expression studies revealed that Mtb-
GAPDH was insoluble under normal conditions. Co-transformation
of the Mtb-GAPDH plasmid along with a GroEL/GroES plasmid
yielded soluble and active protein. After purification and dialysis,
the final protein preparation was >95% pure as determined by
SDS–PAGE. The addition of 1 mM NAD+ to the sonication and puri-
fication buffers was essential to maintain the activity of GAPDH
(data not shown).
Measurement of enzyme activity
The activity of Mtb-GAPDH was determined using a direct spec-
trophotometric assay measuring the conversion of NAD+ into
NADH at 340 nm. NADP+ was screened as an alternate substrate
for NAD+ but no activity could be demonstrated. Na2AsO4 was
substituted for the normal Mtb-GAPDH endogenous substrate, Na2-
PO4, to prevent the reverse reaction. This substitution yields the
product, 1-arseno-3-phosphoglycerate, which is rapidly hydro-
lyzed. Na2AsO4 exhibited kinetic constants (Km = 6.2 0.6 mM,
V = 1590 40 min.ꢁ1
)
similar
to
Na2PO4
(Km = 6 1 mM,
V = 1450 70 min.ꢁ1) (Fig. 1A and B). These studies also revealed
that both Na2PO4 and Na2AsO4 cause linear substrate inhibition
at concentrations higher than 30 mM (Fig. 1A and B, insets), pre-
sumably due to competition with the phosphate binding site for
G3P. Fits of these data to Eq. (1) yielded Ki values of 60 10 and
93 7 mM for Na2PO4 and Na2AsO4, respectively.
Kinetic mechanism
Fig. 4. pH dependence studies for Mtb-GAPDH. (A) Experiments were performed by
varying concentrations of G3P at saturating conditions for all other substrates at pH
values from 5.5 to 8.5. Individual curves at each pH value were fit to Eq. (2), and the
kcat and kcat/KG3P values plotted. (B) Experiments were performed by varying
concentrations of AsO4 at saturating conditions for all other substrates at pH values
from 5.5 to 8.5. Individual curves at each pH value were fit to Eq. (2), and the kcat
and kcat/KG3P values plotted.
Initial velocity studies were conducted to elucidate the kinetic
mechanism of Mtb-GAPDH. Commercially available DL-G3P was
used as
a substrate for these studies due to the laborious
procedures required to synthesize the single D-G3P stereoisiomer.

58
B. Wolfson-Stofko et al. / Archives of Biochemistry and Biophysics 540 (2013) 53–61
Fig. 5. Inactivation of native Mtb-GAPDH with iodoacetamide. (A) Time- and concentration-dependent inactivation at [IAM] = 0 (d), 1 (N), 2 (j), 4 (ꢀ), 8 mM (.) and (B) Kitz–
Wilson re-plot. Inactivation of C162A Mtb-GAPDH with iodoacetamide. (C) Time- and concentration-dependent inactivation at [IAM] = 0 (d), 1 (N), 3 (j), 5 (ꢀ), 8 mM (.) and
(D) Kitz–Wilson re-plot.
Together, these data are consistent with an unusual kinetic
mechanism in which the free enzyme is in reality the E-NAD+ com-
plex. This is supported by both the parallel line initial velocity pat-
tern observed when G3P and NAD+ are varied (requiring a product
to be released between them) and the requirement of NAD+ for sta-
bility. G3P first binds to the E-NAD+ complex, and reacts with the
active site C158 to generate the thiohemiacetal intermediate (E-
NAD+-X in Scheme 2). Hydride transfer to NAD+ generates NADH
and the enzyme thioester intermediate (E-Y-NADH in Scheme 2),
which undergoes a ‘‘nucleotide exchange’’ reaction to generate
the thioester–NAD+ complex (E-Y-NAD+). Binding of phosphate
and phosphorolysis of the thioester generates the product, 1,3-bis-
phosphoglycerate, leaving the enzyme with NAD+ bound and
ready to react with another G3P molecule. A similar ‘‘nucleotide
exchange’’ kinetic scheme was reported for GAPDH from both
rabbit and pig muscle [3,18,19].
logkcat/KG3P. This experiment yielded no detectable change in
either kinetic parameter over this pH range (Fig. 4A), suggesting
that the ionization of the active site cysteine and histidine residues
[20] implicated in the first chemical reaction (thiohemiacetal
formation and hydride transfer to generate the thioester) were
pH dependence studies
To determine which enzyme residues are required for catalytic
function, pH dependence studies were performed. Control experi-
ments were conducted to determine the stability of Mtb-GAPDH.
These studies limited our profile to the pH range of 5.5–8.5. Exper-
iments varying G3P concentrations at saturating concentrations of
NAD+ were used to evaluate the effect of ionizations on logkcat and
Fig. 6. Primary kinetic isotope effects. Primary kinetic isotope effects were
determined using 1-[1H]D-G3P (d) and 1-[1H]D-G3P (N) at saturating concentrations
of other substrates. The data were fitted to Eq. (9).

B. Wolfson-Stofko et al. / Archives of Biochemistry and Biophysics 540 (2013) 53–61
59
outside of the experimental range. Unfortunately, due to enzyme
effects were observed on
V
(DVG3P = 1.2 0.1) and V/K (DV/
instability outside of this pH range we were unable to probe the
enzymatic reaction at pH’s that were closer to the pK1 and pK2 val-
ues previously reported by Polgár for NAD+-bound GAPDH (5.2 and
8.9, respectively) [20]. Similar experiments varying Na2AsO4 to
probe for ionizations implicated in the phosphorolysis reaction
also yielded no evidence for enzyme side chain participation
(Fig. 4B).
KG3P = 1.5 0.2, Fig. 6). The expression for DV/K includes rate con-
stants from the binding of G3P to the E-NAD+ complex to the first
irreversible step, which we suggest is NADH release (Scheme 2).
The value of DV/KG3P is significantly smaller than one would expect
for a hydride transfer reaction, and indicates that G3P is sticky.
Certainly the formation of
contribute to a high commitment factor for this substrate. This
a covalent thiohemiacetal would
Studies with GAPDH homologs have demonstrated the essential
role of an active site cysteine responsible for the formation of the
thioacyl intermediates [3,18,21]. The active site cysteine has been
reported to form a thiolate-imidazolium ion-pair with an active
site histidine that increases the reactivity of the cysteine
[3,20,22–24]. These residues have been identified through
sequence alignment to correspond to C158 and H185 in Mtb-GAPDH.
The lack of any pH dependence of the kinetic parameters in the pH
range tested for Mtb-GAPDH is likely due to the unusual kinetic
mechanism and the requirement of bound NAD+ (or NADH) for sta-
bility and throughout the catalytic cycle. Crystal structures have
shown the binding of NAD+ to induce a conformational change in
the GAPDH of Bacillus stearothermophilus leading to a repositioning
of active site residues favoring the stabilization of a thiolate-
imidazolium pair [25]. Polgár has theorized that the formation of
the thiolate-imidazolium pair in GAPDH would lead to altered pK
values that would appear outside the pH range tested [20]. Given
our inability to assess the required appropriate ionization state of
residues implicated in catalysis by variations in external pH, muta-
genesis studies were conducted to probe the roles of C158, C162
and H185.
Mutagenesis and inactivation studies
In order to verify the essential role of C158 in catalysis, kinetic
inactivation studies were performed. Iodoacetamide (IAM), an irre-
versible, cysteine-specific alkylator was used to probe the role of
the active site cysteines C158 and C162A in catalysis. Fig. 5A illus-
trates the inhibition of Mtb-GAPDH with IAM in both a time- and
concentration-dependent manner. A Kitz–Wilson re-plot gave a
kinactivation and Ki of IAM of 0.16 min.ꢁ1 and 7.8 mM, respectively
(Fig. 5B). This data supports the proposed function of an active site
cysteine in catalysis, as previously reported in other GAPDH homo-
logs [3,19,21].
The C158A, C162A and H185A mutant forms of Mtb-GAPDH
were constructed and purified in order to determine their role in
catalysis. The C158A mutant was inactive, but the C162A mutant
exhibited a comparable V to native Mtb-GAPDH and only a 2-fold
increase in the value of KG3P (data not shown). Time- and concen-
tration-dependent inactivation studies with IAM were conducted
with C162A to verify that acetamidation of C158, and not C162
was responsible for the loss of activity (Fig. 5C). A Kitz–Wilson re-
plot gives a kinactivation and Ki of IAM of 0.12 min.ꢁ1 and 14 mM,
respectively, comparable to the wild-type enzyme (Fig. 5D). The
H185A mutant was also inactive suggesting that C158 is the active
site nucleophile reacting with the aldehyde group of G3P to gener-
ate the thiohemiacetal and that H185 is additionally required to
either stabilize thiolate anion formation or act as a catalytic acid/
base group.
Kinetic isotope effects
Studies of kinetic isotope effects were performed to identify the
rate-limiting chemical steps. In order to measure the primary
kinetic isotope effect on the Mtb-GAPDH catalyzed reaction, the
synthesis of a single stereoisomer of glyceraldehyde 3-phosphate
([1-1H]D-G3P) and [1-2H]D-glyceraldehyde 3-phosphate ([1-2H]D-G3P)
was performed. Using these compounds, only modest isotope
Fig. 7. Solvent kinetic isotope effects. (A) Proton inventory from 0 to 90 atom% D2O
at saturating conditions of all substrates. (B) Solvent kinetic isotope effects
determined using 1-[1H]D-G3P at saturating concentrations of other substrates in
H2O (d) and 90% D2O (N). The data were fit to Eq. (9). (C) Multiple kinetic isotope
effects determined using 1-[2H]D-G3P at saturating concentrations of other
substrates in H2O (d) and 90% D2O (N). The data were fit to Eq. (9).

60
B. Wolfson-Stofko et al. / Archives of Biochemistry and Biophysics 540 (2013) 53–61
His185
His185
His185
Cys158
N
H
Cys158
S
N
H
Cys158
H
N
N
N
N
H
S
H
O
H
H
O
S
H
O
H
H
H
H
OH
OH
OH
N
O
N
O
N
O
O
O
NH2
NH2
O
P O
O
O
O
NH2
O
P
O
O
O
P
O
His185
NAD+
Cys158
S
N
H
N
NADH
O
H
H
OH
O
O
O
N
O
P
O
O
NH2
O
HO
P
O
His185
His185
His185
HO
Cys158
S
Cys158
S
O
Cys158
H
N
N
H
N
H
O
P
N
N
N
O
H
H
P
O
H
S
H
O
O
O
O
H
O
H
OH
OH
N
O
N
O
N
O
HO
NH2
O
O
O
O
O
NH2
NH2
P
O
P
O
Scheme 3. Proposed chemical mechanism of Mtb-GAPDH.
chemistry also results in there being a small (ca. 1.2) inverse equi-
librium effect on the formation of the thiohemiacetal when using
[1-2H]D-G3P versus [1-1H]D-G3P due to the change in hybridization
from sp2 to sp3. Together these data suggest that the hydride trans-
fer is only partially rate-limiting in the chemical oxidation of the
covalently-bound thiohemiacetal. The even lower value of DV
may reflect additional rate limitation by product release. These re-
sults are quantitatively similar to those obtained under similar
experimental conditions with rabbit muscle GAPDH [18].
In order to probe additional chemical steps in the oxidation
reaction, solvent kinetic isotope effects (SKIE) were used to inves-
tigate the potential rate-limiting nature of proton transfer steps. A
proton inventory was conducted under ‘‘pseudo-V’’ conditions (5–
10 times Km values for G3P and NAD+ and maximal, but not sub-
strate inhibitory concentrations of arsenate) in 10% increments of
D2O from 0% to 90%. This experiment yielded a linear, normal
dependence of the rate on solvent isotopic composition, indicative
of a single solvent-derived proton being involved in the oxidative
half-reaction (Fig. 7A). The extrapolated magnitude of ‘‘D2OV’’ was
approximately 1.7.
has been previously suggested [24]. The smaller magnitude of
D2OV suggests that steps after the oxidative half-reaction contrib-
ute to overall rate limitation, as was discussed for the primary KIE.
Oxidation of secondary alcohols can presumably be carried out
in either a concerted (deprotonation-hydride transfer) or stepwise
manner. Theory has been developed that allows the discrimination
of these two mechanisms using multiple kinetic isotope effects
[26]. We elected to perform multiple kinetic isotope effect (MKIE)
experiments examining the effect of substrate deuteration on the
solvent KIE, since the solvent KIE’s are larger and more likely to re-
veal statistically significant differences, if present. Using 1-[2H]D-G3P
in H2O and 90% D2O at pH 8, the solvent kinetic isotope effects
D2O
on
V
was 1.6 0.1 and D2OV/K1-[2H]
was 1.4 0.2
D-
D-
1-[2H] G3P
G3P
(Fig. 7C). D2OV using either 1-[1H]D-G3P or 1-[2H]D-G3P were equiv-
alent (1.6–1.7) and equal to the extrapolated value from the proton
inventory experiment. However, D2OV/K1-[2H]
was significantly
D-
G3P
smaller than when 1-[1H]D-G3P was used as substrate (1.4 versus
2.5). This suggests that the two isotopic perturbants affect different
steps (alkoxide formation or hydride transfer) and that as the hy-
dride transfer transition state is raised by substrate deuteration
that the proton transfer step becomes less rate-limiting and the
magnitude of D2OV/KG3P is decreased. This argues persuasively that
the primary and solvent kinetic isotope effects are reporting on
separate steps, e.g., hydride transfer occurring after alkoxide for-
mation. These data allow us to propose a detailed chemical mech-
anism for the reaction catalyzed by Mtb-GAPDH.
Experiments performed with 1-[1H]D-G3P in H2O and 90% D2O
at pH 8 yielded values of D2OV of 1.7 0.1 and D2OV/KG3P of
2.5 0.4, respectively, (Fig. 7B). As is the case for the primary
KIE, D2OV/KG3P includes rate constants from the binding of G3P to
the release of NADH. In the steps that represent the oxidation of
the thiohemiacetal (Scheme 3), the formation of the thiohemiac-
etal is accompanied by protonation of the carbonyl oxygen. Oxida-
tion and hydride transfer will occur only after deprotonation of the
alcohol, and it is this step that is the most likely source of the sin-
gle-proton solvent KIE. The magnitude of V/KG3P suggests that this
step is slow, and that the relatively rate-limiting formation of the
alkoxide by His185-assisted deprotonation promotes a more rapid
hydride transfer to NAD+ from the thiohemiacetal alkoxide. This
interpretation is further supported by the inactivity of H185A, as
Chemical mechanism
The chemical mechanism for Mtb-GAPDH shown in Scheme 3 is
supported by our determination of the kinetic mechanism, muta-
genesis and isotope effect studies. The ‘‘free enzyme’’ is actually
present with bound NAD+, and the two important catalytic

B. Wolfson-Stofko et al. / Archives of Biochemistry and Biophysics 540 (2013) 53–61
61
residues, Cys158 and His185, are present as a thiolate-imidazolium
References
ion pair. Binding of glyceraldehyde-3-phosphate yields an initial
Michaelis complex, and the Cys158 thiolate nucleophilicially
attacks the aldehyde to generate the neutral thiohemiacetal after
proton donation from His185. Solvent kinetic isotope effects
support a mechanism for oxidation that requires an initial deproto-
nation of the alcohol to generate the alkoxide that is the largely
rate-limiting step. Alkoxide formation promotes hydride transfer
from C1 of the covalently-bound G3P thiohemiacetal to NAD+. This
generates the covalently-bound thioester and NADH. Nucleotide
exchange and inorganic phosphate binding set up the second
half-reaction, the phosphorolysis of the thioester to generate the
high energy mixed carboxy-phosphoric anhydride of the product
1,3-bisphosphoglycerate (1,3-BPG), which is released in the final
step of the catalytic cycle. The thioester and carboxy-phosphoric
anhydride are of roughly equal energy, and permit the substrate
level phosphorylation reaction catalyzed by GAPDH to drive the
subsequent step in glycolysis catalyzed by phosphoglycerate
kinase: the formation of ATP and phosphoglycerate from ADP
and 1,3-bisphosphoglycerate.
[1] Global Tuberculosis Control, World Health Organization. <http://who.int/tb/
publications/global_report/en/index.html>, 2011.
[2] A.H. Romano, T. Conway, Res. Microbiol. 147 (1996) 448–455.
[3] H.L. Segal, P.D. Boyer, J. Biol. Chem. 204 (1953) 265–281.
[4] R.E. Moellering, B.F. Cravatt, Science 341 (2013) 549–553.
[5] M. Perucho, J. Salas, M.L. Salas, Eur. J. Biochem. 81 (1977) 557–562.
[6] Y. Zhou, X. Yi, J.N.B. Stoffer, N. Bonafe, M. Gilmore-Hebert, J. McAlpine, S.K.
Chambers, Mol. Cancer Res. 6 (2008) 1375–1384.
[7] L. Zheng, R.G. Roeder, Y. Luo, Cell 114 (2003) 255–266.
[8] H. Kumagai, H. Sakai, J. Biochem. 93 (1983) 1259–1269.
[9] J. Andrade, S.T. Pearce, H. Zhao, M. Barroso, Biochem. J. 384 (2004) 327–336.
[10] P. Rawat, S. Kumar, N. Sheokand, C.I. Raje, M. Raje, Biochem. Cell Biol. 90
(2012) 329–338.
[11] H. Tajima, K. Tsuchiya, M. Yamada, K. Kondo, N. Katsube, R. Ishitani,
NeuroReport 10 (1999) 2029–2033.
[12] C. Nicholls, H. Li, J.-P. Liu, Clin. Exp. Pharmacol. Physiol. 39 (2012) 674–679.
[13] I.E. Ichetovkin, G. Abramochkin, T.E. Shrader, J. Biol. Chem. 272 (1997) 33009–
33014.
[14] R. Higuchi, B. Krummel, R. Saiki, Nucleic Acids Res. 16 (1988) 7351–7367.
[15] S. Ladame, M. Willson, J. Périé, Eur. J. Org. Chem. 2002 (2002) 2640–2648.
[16] R. Kitz, I.B. Wilson, J. Biol. Chem. 237 (1962) 3245–3249.
[17] W.E. Karsten, C.-J. Lai, P.F. Cook, J. Am. Chem. Soc. 117 (1995) 5914–5918.
[18] L. Liu, W. Huskey, Biochemistry 31 (1992) 6898–6903.
[19] P. Harrigan, D. Trentham, Biochem J. 135 (1973) 695–703.
[20] L. Polgár, Eur. J. Biochem. 51 (1975) 63–71.
[21] S. Moniot, S. Bruno, C. Vonrhein, C. Didierjean, S. Boschi-Muller, M. Vas, G.
Bricogne, G. Branlant, A. Mozzarelli, C. Corbier, J. Biol. Chem. 283 (2008)
21693–21702.
Acknowledgments
We thank Drs. Subray Hegde and Hector Serrano for the assis-
tance in the purification of Mtb-GAPDH. This work was supported
by the National Institutes of Health Grant A1060899 to J.S.B.
[22] D. Trentham, Biochem. J. 122 (1971) 71–77.
[23] J. Jenkins, J. Tanner, Acta Crystallogr. Sect. D 62 (2006) 290–301.
[24] A. Soukri, A. Mougin, C. Corbier, A. Wonacott, C. Branlant, G. Branlant,
Biochemistry 28 (1989) 2586–2592.
_
[25] T. Skarzyn´ ski, A.J. Wonacott, J. Mol. Biol. 203 (1988) 1097–1118.
[26] J.D. Hermes, C.A. Roeske, M.H. O’Leary, W.W. Cleland, Biochemistry 21 (1982)
5106–5114.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.abb.2013.10.007.