Stages of the formation of nonequivalence of active centers of transketolase from baker's yeast

Article abstract of DOI:10.1016/j.mcat.2019.01.008

For baker's yeast transketolase (TK), cooperative binding of thiamine diphosphate (ThDP) and substrates in the transferase reaction is known. We show here that the differences in the properties of the active centers of TK are formed already upon the binding of Ca²⁺ in one of two initially identical subunits. When Ca²⁺ is bound in only one of the two active centers its affinity for the second decreases. The absence of a cation in the second active center decreases the affinity of ThDP to the first active center. Ca²⁺ binding increases the thermal stability of apo- and holoTK, i.e. changes the whole structure of the enzyme. Only in the presence of Ca²⁺, but not Mg²⁺, does the thermal stability of holoTK increase. In the one-substrate reaction in the presence of Ca²⁺, two K_m are measured for the binding of xylulose-5-phosphate and hydroxypyruvate. For both substrates, Vmax of the first active center of holoTK, when it binds the substrate alone, is higher than of semiholoTK. When the substrate begins to bind also in the second active center, Vmax of both active centers decreases, which is explained by the previously shown flip-flop mechanism.

Full text of DOI:10.1016/j.mcat.2019.01.008

MolecularCatalysis466(2019)122–129
Contents lists available at ScienceDirect
Molecular Catalysis
journal homepage: www.elsevier.com/locate/mcat
Stages of the formation of nonequivalence of active centers of transketolase
from baker’s yeast
Olga N. Solovjevaa,⁎, Vitaly A. Selivanovb, Victor N. Orlova, German A. Kochetova
a Belozersky Institute of Physico-Chemical Biology, Lomonosov Moscow State University, 119992 Moscow, Russian Federation
b Departament de Bioquímica i Biologia Molecular, Facultat de Biología, Universitat de Barcelona, Barcelona 08028, Spain
A R T I C L E I N F O
A B S T R A C T
Keywords:
Transketolase
Thiamine diphosphate
One-substrate reaction
Circular dichroism
Differential scanning calorimetry
For baker’s yeast transketolase (TK), cooperative binding of thiamine diphosphate (ThDP) and substrates in the
transferase reaction is known. We show here that the differences in the properties of the active centers of TK are
formed already upon the binding of Ca2+ in one of two initially identical subunits. When Ca2+ is bound in only
one of the two active centers its affinity for the second decreases. The absence of a cation in the second active
center decreases the affinity of ThDP to the first active center. Ca2+ binding increases the thermal stability of
apo- and holoTK, i.e. changes the whole structure of the enzyme. Only in the presence of Ca2+, but not Mg2+
does the thermal stability of holoTK increase.
,
In the one-substrate reaction in the presence of Ca2+, two Km are measured for the binding of xylulose-5-
phosphate and hydroxypyruvate. For both substrates, Vmax of the first active center of holoTK, when it binds the
substrate alone, is higher than of semiholoTK. When the substrate begins to bind also in the second active center,
Vmax of both active centers decreases, which is explained by the previously shown flip-flop mechanism.
1. Introduction
dependent enzymes have been determined [9,10].
TK activation requires ThDP and bivalent cations simultaneously.
All thiamine diphosphate (ThDP) dependent enzymes are dimers (or
dimers of dimers) with identical subunits. The active centers are located
at the interface between its subunits [1]. It is known that ThDP is bound
at the subunits interface, for all structurally characterized ThDP en-
zymes. ThDP interacts with the Pyr domain of one monomer and the PP
domain of other, the last via interaction with bivalent cation [2–4].
Thus, each active center is formed by both subunits. For such enzymes
whose catalytic centers are coupled, ping-pong mechanism of substrate
binding and cleavage is known [4]. Due to the structural homology of
ThDP-dependent enzymes, the results obtained for a single enzyme are
found to be applicable to other ThDP-dependent enzymes.
Transketolase (TK; EC 2.2.1.1) from baker's yeast Saccharomyces
cerevisiae is the simplest representative of ThDP-dependent enzymes
and so model enzyme for studying the structure of the active centers
and the mechanism of action of ThDP-dependent enzymes. It is a
homodimer with molecular weight 148.4 kDa [5]. Its subunits are
identical in amino acid composition and tertiary structure [6,7]. Its
three-dimensional structure was determined first in the series of ThDP-
dependent enzymes [8]. Then crystal structures of a number of ThDP-
The bivalent cation serves as a bridge between apoTK and diphosphate
tail of ThDP [6]. In the native TK calcium was found [11]. The two
active centers are localized in a deep cleft between the subunits and
binding of ThDP to the apoprotein involves amino acid residues of both
subunits [8]. First, the pyrophosphate tail of ThDP binds to the cation of
one subunit, then the aminopyrimidine ring binds to the second subunit
[12]. The thiazole (middle) ring of ThDP is located in the cleft between
the PP domain of the one subunit and the Pyr domain of the other
subunit, interacting with amino acid residues of both [8] (Fig. 1). In the
absence of cation, ThDP binding does not occur [13]. Active centers
have equal catalytic activity, but different affinity for coenzyme [6] and
substrates [14] in the transferase reaction.
A variety of kinetic properties of TK essentially depends on the type
of the bivalent cation (Ca2+ or Mg2+) used as a cofactor in the course of
holoTK reconstitution [15]. One of the reasons for the observed dif-
ferences is, apparently, various protein conformation, caused by the
type of used cation. The native holoTK contains one calcium ion per
active center [11]. With one active center of TK, Ca2+ is bound strong
and is removed under fairly hard conditions, and from the other one it
Abbreviations: TK, transketolase; ThDP, thiamine diphosphate; DSC, differential scanning calorimetry; X5P, xylulose 5-phosphate; HPA, hydroxypyruvate; R5P,
ribose 5-phosphate; TMX, tetramethyl murexide; CD, circular dichroism
⁎ Corresponding author.
E-mail address: soloveva_o@list.ru (O.N. Solovjeva).
https://doi.org/10.1016/j.mcat.2019.01.008
Received 6 August 2018; Received in revised form 29 December 2018; Accepted 5 January 2019
2468-8231/©2019ElsevierB.V.Allrightsreserved.

O.N. Solovjeva et al.
MolecularCatalysis466(2019)122–129
Fig. 1. Cofactor-protein interactions in the thiamine diphosphate binding center of transketolase. Conserved residues are underlined and residues from the second
subunit are marked by* [8].
readily released [16].
Ca2+ [15,19] and Mg2+ [15]. When binding substrates in the absence
of free bivalent cations in the medium, the active centers of holoTK are
equivalent. In the presence of added Ca2+ from the outside, the active
centers become nonequivalent in two-substrate reaction in binding both
the donor substrate and the acceptor substrate. Mg2+ does not exert
such influence [14].
In this paper, we show that the difference in the properties of the
active centers of TK is formed already when the calcium ion is bound in
one of the two initially identical subunits.
In the presence of Ca2+ as well as Mg2+, ThDP binding is char-
acterized by negative cooperativity [15,17]. The values of apparent
constants of ThDP dissociation, measured for the two active centers,
differ by approximately one order of magnitude. In the presence of
Mg2+, the affinity of the active centers of TK for the coenzyme is
considerably lower than in the presence of Ca2+ [15,18,19].
TK catalyzes transfer of a two-carbon fragment, a glycolaldehyde
residue, from a ketose (donor substrate) to an aldose (acceptor sub-
strate). For TK from S.cerevisiae donor substrates are D-xylulose 5-
phosphate (X5P), D-sedoheptulose 7-phosphate, D-fructose 6-phosphate,
D-erythrulose 4-phosphate, hydroxypyruvate (HPA); acceptor substrates
are D-ribose 5-phosphate (R5P), D-glyceraldehyde 3-phosphate, D-ery-
throse 4-phosphate, glycolaldehyde. [20,21]. When HPA is used as a
substrate, TK reaction becomes irreversible. TK catalyses also one-
substrate reaction with only ketose substrate. If it is X5P, the reaction
velocity is 2% relative to velocity of transferase reaction [22,23].
It was shown negative cooperativity in cleavage of X5P and R5P in
transferase reaction but only in the presence of Ca2+ in medium. When
Mg2+ serves as a cofactor, the active centers of TK are equivalent in
binding both the donor substrate (X5P) and the acceptor substrate
(R5P). Likewise, there are no differences between the centers in the
catalytic activity. Replacement of Mg2+ by Ca2+ decreases the values of
Km for both substrates with respect to one of the two active centers
(conditionally the first one) and increases the values of Km with respect
to the other center (conditionally the second one). In parallel, the ac-
tivity of center 1 decreases, and the activity of center 2 increases. The
value of the Hill coefficient, determined from the same experimental
data, equaled 0.67, which indicates that substrate binding by the active
centers is characterized by negative cooperativity [24].
2. Experimental
2.1. Materials
ThDP, NAD+, NADH, R5P, HPA (lithium salt), EGTA, glycer-
aldehyde 3-phosphate dehydrogenase, ribose phosphate isomerase, ri-
bulose 3-phosphate epimerase, dithiothreitol, tetramethyl murexide
(TMX), glycylglycine, eriochrome black T, lactate dehydrogenase,
Dowex 50 × 8 were obtained from Sigma, HEPES - from Serva, Chelex
100 resin from Biorad, EDTA from Reanal. Other reactants were of extra
pure grade.
2.2. Preparation of phosphopentose mixture
The barium salt of the phosphopentose mixture (an equilibrium
mixture of X5P, R5P and ribulose 5-phosphate, the last is no substrate
and no inhibitor of TK) was prepared by incubating R5P with ribose
phosphate isomerase and ribulose phosphate 3-epimerase [25]. The
barium salt was converted to the potassium salt with Dowex 50 × 8
before use. Then the concentration of the phosphopentose mixture was
selected in order to determine the maximum catalytic activity.
It was believed that the nonequivalence of the active centers of TK is
formed upon binding of the first ThDP molecule [17]. Binding of ThDP
to one active center (conditionally - to the first) is approximately an
order of magnitude higher than to the other, both in the presence of
123

O.N. Solovjeva et al.
MolecularCatalysis466(2019)122–129
2.3. Preparation of xylulose 5-phosphate
concentration was determined by eriochrome black T titration in the
presence of Mg2+. Prior to the measurement, TK samples were passed
through Sephadex G-50, equilibrated with 20 mM HEPES, pH 7.6. The
concentration of the enzyme was 10–15 mg/ml. The TK samples were
then dropped into boiling water for 3 min, cooled, and the denatured
protein was removed by centrifugation. 0.2-0.6 ml of the prepared
sample was added to a 2 ml solution containing 0.1 mM TMX in 20 mM
HEPES, pH 7.6, and a similar volume of the similarly prepared buffer
was added to the comparison cuvette.
Spectra were recorded at wavelengths from 480 to 550 nm. Then the
standard solutions of CaCl2 (10–50 μM) were added to the cell with the
test sample and the spectra were also recorded. Calcium concentration
was determined by the difference in absorption at 550 and 507 nm, as
well as 550 and 482 nm.
X5P was obtained using TK reaction with HPA and glyceraldehyde
3- phosphate [26]. × 5P concentration was measured enzymatically in
TK reaction with an excess of R5P [16].
2.4. Transketolase purification
ApoTK was isolated from baker’s yeast Saccharomyces cerevisiae on a
column with immobilized antibodies to TK isolated from the serum of
the immunized rabbit, according to the method described previously
[27], aliquoted and stored frozen in 10 mM potassium phosphate buffer
with 50 mM ammonium sulfate, pH 7.6. The enzyme was homogeneous
by SDS–PAGE and has a specific activity in the transferase reaction of
14 U/mg. TK concentration was determined spectrophotometrically
1%
using the absorption coefficient A
of 14.5 at 280 nm [28]. Before
2.8. Differential scanning calorimetry
1cm
use, the TK solution was passed through a Sephadex G-50 column
equilibrated with 50 mM glycylglycine pH 7.6.
DSC measurements were done on a scanning adiabatic micro-
calorimeter DASM4 (Biopribor, Pushchino, Russia) with 0.47 ml pla-
tinum capillary spiral cells at excessive pressure of 2 atm, in the range
20–100 °C (scanning rate 1°/min). In all experiments, a comparative
analysis of the sample and the buffer was performed. The baseline
(determined after each scanning by repeated heating of the sample in
the same temperature range) was subtracted from each curve.
Calorimetric scanning of TK (0.5 mg/ml) was performed in 50 mM
glycylglycine buffer, pH 7.6. In experiments where the effects of biva-
lent cations on apoTK thermostability were studied, the enzyme solu-
tion was supplemented with 0.1 mM CaCl2 or 2.5 mM MgCl2 prior to
scanning. In experiments with a holoenzyme, the solution of apoTK in
50 mM glycylglycine (pH 7.6) was supplemented with 0.1 mM CaCl2 or
2.5 mM MgCl2 and 2 mM ThDP. Each experiment was conducted at
least three times.
2.5. Transketolase activity
The enzyme’s catalytic activity in the transferase reaction was
measured spectrophotometrically at 340 nm in the reaction mixture
with 0.5–1 μg/ml TK, 0.1 mM CaCl2, 0.1 mM ThDP, 7 mM sodium ar-
senate, 3.2 mM dithiothreitol, 0.37 mM NAD+, 3 U/ml glyceraldehyde
3-phosphate dehydrogenase, 3.2 mg/ml potassium salt of a phospho-
pentose mixture, 50 mM glycylglycine, pH 7.6 [29].
2.6. Preparation of holotransketolase and semiholotransketolase
For the preparation of holoTK, apoTK was incubated 5 min with
0.1 mM CaCl2 or 2.5 mM MgCl2 and 0.1 mM ThDP. For the preparation
of holoTK without free cations, holoTK was passed through sephadex G-
50, equilibrated with 50 mM glycylglycine, pH 7.6, then 0.1 mM ThDP
was added.
For the preparation of semiholoTK1 (the enzyme in which ThDP is
bound only with one active center and which has 50% activity com-
pared to the activity of holoTK), 0.6–5.0 mg/ml apoTK was incubated
40 min with 0.1 mM CaCl2 and ThDP concentration, equimolar TK
concentration.
SemiholoTK2 is the enzyme where only the active center with low
affinity to the coenzyme is functioning. To obtain such a preparation, it
is necessary to block active center 1 with high affinity to ThDP, using an
inactive analog of the coenzyme, oxythiamine diphosphate (its affinity
to TK active centers is several times higher than the affinity of native
ThDP) [18]. Then the TK with blocked active center 1 was supple-
mented with 0.1 mM ThDP. For preparation of semiholoTK 2,
0.6–5.0 mg/ml apoTK was incubated 40 min with 0.1 mM CaCl2 and
oxythiamine diphosphate concentration, equimolar TK concentration,
then 0.1 mM ThDP was added. It has 50% activity compared to the
activity of holoTK.
Ca2+ binding to TK was also studied by isothermal titration ca-
lorimetry with VP-ITC instrument (MicroCal LLC, East Northampton,
MA, USA) using a 1.4-mL cell. All experiments were carried out at 25 °C
in 50 mM HEPES, passed through Chelex, pH 7.6. Titration experiments
were performed by making successive 10-μl injections of 2 mM CaCl2
solution into the buffer solution with 12–15 μM TK. The interval be-
tween injections was 8–15 min. The data analysis was carried out with
the MICROCAL ORIGIN 7.0 software using the one-set of centers model
[31].
2.9. Circular dichroism absorbance
Circular dichroism (CD) absorbance was recorded at 320 nm in ki-
netic regime on an IBM PS 1- controlled Mark V dichrograph (Jobin
Ivon, France). Spectra were recorded in a 1-cm quartz cuvette at a 5-s
averaging time. The medium was composed of 5.4 μM (0.86 mg/ml) TK,
45 μM ThDP and 0.1 mM CaCl2 (or 2.5 mM MgCl2), 50 mM glycylgly-
cine, pH 7.6.
2.10. Determination of Km for X5P and HPA
2.7. Removal of calcium from apoTK
The enzyme’s catalytic activity in the one-substrate reaction was
measured spectrophotometrically with X5P at 340 nm in the reaction
mixture with 5 μg/ml TK, 0.1 mM CaCl2, 0.1 mM ThDP, 0.003-0.1 mM
X5P, 50 mM glycylglycine, pH 7.6 [7], and with HPA at 230 nm in the
reaction mixture with 50 μg/ml TK, 0.1 mM CaCl2, 0.1 mM ThDP, 0.03-
0.6 mM HPA, 10 mM glycylglycine, pH 7.6 [32]. The enzyme’s catalytic
activity in the two-substrate reaction was measured with HPA in the
reaction mixture with 5 μg/ml TK, 0.1 mM CaCl2, 0.1 mM ThDP,
0.06–1 mM HPA, 1.5 mM R5P, 10 mM glycylglycine, pH 7.6 [32]. In the
case of semiholoTK the concentration of ThDP was the equimolar
concentration of TK. The reaction was started by the addition of holoTK
or semiholoTK. HPA concentration was measured with lactate dehy-
drogenase and NADH [33]. When Mg2+ (2.5 mM) was used instead of
Ca2+ from apoTK was removed by the procedure [16]. Calcium
removal was performed only in differential scanning calorimetry (DSC)
experiments. 1 mg/ml TK in 50 mM potassium phosphate buffer, pH 7.6
was incubated with 10 mM EGTA for 1 h, then dialyzed against 250
volumes of 20 mM potassium phosphate buffer, pH 7.6 in the presence
of 16 mM EDTA for two days at 4 °C with one change of buffer. The
enzyme was then passed through Sephadex G-50, equilibrated with a
metal-sterile 50 mM glycylglycine buffer obtained by passing the buffer
through Chelex 100 resin. The calcium content was controlled by direct
measurement of calcium with the aid of TMX in 20 mM HEPES, pH 7.6
[30]. The concentration of TMX was 0.1 mM. As a standard, a calcium
solution with an initial concentration of 0.5 mM was used. Its
124

O.N. Solovjeva et al.
MolecularCatalysis466(2019)122–129
Ca2+, then Ca2+ was not removed from the first active center.
The Km values for X5P and HPA were determined by fitting the
curve of the experimental dependence of the reaction rate on the sub-
strate concentration. Each point on the experimental curve is the initial
reaction rate when adding different concentrations of the substrate.
Kinetic parameters were determined by fitting the dependence of TK
reaction rate on the concentration of one of the substrates at a constant
concentration of the other substrate. Experimental data were analyzed
using the Michaelis equation
M and 1.4 × 10-7 M, Kd for Mg2+ is 1.4 × 10-7 M and 2.0 × 10-6 M for
the first and second active centers of TK, respectively (Table 2, N 1, 2).
Thus, in the two active centers of TK, negative cooperativity is observed
in the binding of both Ca2+ and Mg2+. The affinity of Mg2+ for apoTK
is an order of magnitude lower than that of Ca2+ for both active cen-
ters.
K
d for ThDP was 3.6 × 10−8 M and 3.8 × 10-7 M in the presence of
Ca2+ and 2.9 × 10-7 M and 2.0 × 10-6 M in the presence of Mg2+, for
the first and second active centers of TK, respectively (Table 2, N 3, 4
and [15,19]. When Ca2+ was in only one active center (the second
active center did not contain cation), then Kd for ThDP was 4.5 × 10-6
M, i.e. in the absence of a cation in the second active center, the affinity
of the first active center of TK for ThDP is reduced by 2 orders of
magnitude (Table 2, N5).
v = V · [S] / ([S] + Km)
and the equation
v = (V2 + V1 · Km2 / [S]) / (1 + Km2 / [S] + Km1 · Km2 / [S]2)
Using the CD method, it was shown that when ThDP binds with
apoTK, a negative absorption band is formed at 320 nm [34]. There is
interdependence between dichroic absorption value and the catalytic
activity of TK [35]. At high concentrations of ThDP (usually
80–100 μM) ThDP binding occurs rapidly. If the concentration of the
coenzyme is low, complete binding of ThDP occurs within 30–40 min,
since a lag phase precedes the steady state which is attributed to the
low rate of the interaction between the coenzyme and the apoenzyme
[15]. We determined the ThDP binding by the CD absorption mea-
surement in the kinetic regime at a middle concentration of ThDP of
45 μM. In the presence of Mg2+ in both active centers, ThDP has not
only a lower affinity for TK than in the presence of Ca2+ (Table 2), but
also binds more slowly (compare Fig. 2, N 1, 2). Almost complete
binding of ThDP in the presence of Ca2+ occurs in 2 min, and in the
presence of Mg2+ in 5 min. In the absence of Ca2+ in one of the two
active centers, ThDP is fully bound to the first active center in 5 min, as
in the presence of Mg2+. Accordingly, a decrease in dichroic absorption
is half the maximum because the binding of ThDP occurs only in the
first active center containing Ca2+ (Fig. 2, N 3). Half-binding time for
ThDP is 5 s in the presence of Ca2+ and 40 s in the presence of Mg2+
and without cations (in the latter case Ca2+ and consequently ThDP are
bound in one of the two active centers).
which corresponds to the following tentative reaction scheme:
K
1
K 2
m
m
E ↔ ES1 ↔ ES2 → V2
↓
V1
where E designates TK; S, one of the substrates (X5P or R5P); ES1, the
enzyme harboring the substrate in one of the active centers; and ES2,
the enzyme, in which the second active center is also engaged by the
substrate.
In order to determine the Hill coefficient, the experimental data
were fitted to the Hill equation
v = (Vmax · Sn) / (Km + Sn)
The package “Gnuplot” was used for curve fitting.
3. Results
3.1. The affinity of the active centers of transketolase for cations and ThDP
Native holoenzyme from baker’s yeast contains only Ca2+ (two
calcium ions per dimeric molecule of enzyme, i.e. one Ca2+ for each
active center) [11]. When removing cations from the medium, Ca2+
remains bound only in one active center of apoTK (Table 1, N1). If
ThDP is added to the apoenzyme in the presence of Ca2+, and then
excess Ca2+ is removed from the medium, then 2 calcium ions per di-
meric molecule of holoTK are determined (Table 1, N2).
Thus, not only the binding of ThDP in the first active center affects
the affinity of ThDP to the second one [15,19], but also the binding of
Ca2+ in the second active center affects the affinity of ThDP to the first
one, decreasing it by two orders of magnitude (Table 2, N 5). Accord-
ingly, in the absence of Ca2+ in the second (conditionally) subunit,
ThDP binds by its diphosphate with the first subunit, and by its ami-
nopyrimidine ring with the second subunit, devoid of Ca2+. So, the
absence of Ca2+ changes the structure of the second part of the active
center so that the affinity of ThDP for the enzyme decreases by 2 orders
of magnitude.
If ThDP is added to the apoenzyme that contains Ca2+ in one active
center, in the presence of Mg2+ in the medium, and then excess Mg2+ is
removed, then one calcium ion is determined per dimeric holoTK mo-
lecule (Table 1, N3). In both cases the holoTK activity is fully de-
termined without the addition of cofactors to the reaction mixture.
Thus, Ca2+ is not replaced by Mg2+ in the first active center, i.e. when
the "magnesium" enzyme is formed without preliminary removing of
Ca2+, in fact, it contains Ca2+ in the one active center and Mg2+ in the
other. This was the reason why positive cooperative interaction instead
of negative in ThDP binding in the presence of Mg2+ was determined
[15].
Using DSC, we showed that in the absence of Ca2+, the overall
structure of the subunits changes and, apparently, the structure changes
also in the region of active centers.
3.2. Differential scanning calorimetry of transketolase
About the difference in the affinity of active centers for Ca2+
binding was known only that it is easily removed from the second active
center and is tightly bound to the first active center of apoTK [11].
We determined the affinity of cations for apoTK after removal of
Ca2+ from two active centers (Table 1, N4). Kd for Ca2+ is 1.2 × 10−8
The effect of the cofactors on the thermal stability of the TK mole-
cule was studied by DSC. Thermal denaturation in all experiments was
irreversible, what was shown by repeated warming up (data not
shown). In Fig. 3A are shown the calorimetric curves of apoTK after
Table 1
Calcium content in the forms of transketolase.
№
TK form
Ca2+, mol/mol TK
Ca2+, mol/mol TK 11]
1
2
3
4
apoTK + Ca2+→ sephadex
0.95
1.95
0.90
0.3
0.05
0.05
0.05
0.9
1.7
apoTK + Ca2+ + ThDP → sephadex
apoTK + Mg2+ + ThDP → sephadex
apoTK after EDTA
0.02
0.3
125

O.N. Solovjeva et al.
MolecularCatalysis466(2019)122–129
Table 2
Binding constants for cofactors of transketolase for the first (upper line) and second (lower line) active centers.
N
1
2
3
4
5
Kd, M
Our data
Data of [15]
Data of [19]
for Ca2+
1.2
1.4
1.4
2.0
3.6
3.8
2.9
2.0
4.5
no
0.1 x 10−8
0.1 x 10−7
0.1 x 10−7
0.1 x 10−6
0.3 x 10−8
0.4 x 10−7
0.4 x 10−7
0.2 x 10−6
0.1 × 10−6
for Mg2+
for ThDP ( + 0.1 mM Ca2+
)
8.0 x 10−8
5.0 x 10−7
4.0 x 10−7
5.7 x 10−6
3.2 x 10−8
2.5 x 10−7
for ThDP ( + 2.5 mM Mg2+
)
positive cooperativity (wrong)
for ThDP (no cations added)
So, the thermal stability of holoTK is higher than apoTK, both in the
absence of cations and in the presence of Ca2+ (compare curves 1 and
3; 2 and 5). The thermal stability of both apoTK and holoTK increases
with Ca2+ binding (compare curves 1 and 2; 3 and 5). Mg2+ does not
affect the thermal stability of the holoTK (compare curves 3 and 4).
With the isothermal titration calorimetry, we showed that the
binding of both Ca2+ and Mg2+ to apoTK occurs without a change of
heat capacity (data not shown), so it was impossible to determine if
Ca2+ is bound not only to the active centers. However, in the presence
of excess Ca2+ in the medium, the affinity of the active centers of
holoTK to substrates in transferase reaction varies [14,37].
3.3. The affinity of the active centers of transketolase to substrates
Fig. 2. CD kinetic measurement of binding of 5.4 μM apoTK with 45 μM ThDP
in the presence of 0.1 mM Ca2+ (1), 2.5 mM Mg2+ (2) and without cations in
the medium (Ca2+ is in the one active center of TK) (3).
Previously, we showed that in a two-substrate (transferase) reaction
in the absence of free cations in the medium, as well as in the presence
of Mg2+, both active centers of TK are equivalent in their affinity for
X5P. In the presence of Ca2+, this equivalence is disrupted and negative
cooperativity is observed. The activity in both semiholoTK (with only
the first center active and with only the second center active) was the
same, but they differ in the Km value (Table 3) [14,38]. Similar results
were obtained for HPA (Table 3). The difference in the binding of two
substrates is that the value of Hill coefficient for X5P was 0.67
and for X5P – 1.78 0.26.
Table 3. Kinetic parameters of TK in two-substrate reaction with
X5P [14,38] and HPA (own data). The concentration of second sub-
strate R5P was 1.5 mM (saturating)
We performed similar measurements for a one-substrate reaction,
using X5P and HPA as substrates (Table 4). When the substrate binds in
the first active center of holoTK, a burst of activity is observed, in
comparison to semiholoTK (Fig. 4, curves 1, 2).
Previously [14,38], we used the following formula to fit the de-
pendence of the initial reaction rate on substrates concentrations for
two-substrate reaction V(S):
removal of 85% of calcium from the TK molecule (curve 1) and after the
addition of 0.1 mM Ca2+ (curve 2). Ca2+ increases the thermal stability
of apoTK from 50.7 °C to 54.4 °C. The temperature maximum of the
enzyme in the presence of Ca2+ does not exceed the temperature
maximum of 15% of the subunits containing Ca2+ only in the active
center (the second peak on curve 1) and in the absence of Ca2+ in the
medium. So, only Ca2+ of active centers, not excess Ca2+, changes the
overall structure of the apoenzyme. Accordingly, only Ca2+ bound in
the active centers changes the whole structure of the apoenzyme. The
absence of an additional contribution of excess of Ca2+ or Mg2+ in the
medium by the magnitude of the temperature maximum of apoTK has
also been shown in [35].
Fig. 3B shows the data of thermal denaturation of holoTK in the
absence of cations in medium (accordingly, cations are only in the ac-
tive centers) (curve 3) and when Mg2+ (curve 4) or Ca2+ (curve 5) are
added to the medium. Maximum of thermal stability of holoTK is
61.6 °C in the absence of cations, 61.8 °C in the presence of Mg2+ and
66.9 °C in the presence of Ca2+. So, excess only Ca2+, not Mg2+
changes the structure of holoTK molecule. Apparently, for this reason,
X-ray structure analysis of TK was carried out only in the presence of
Ca2+. In the presence of Mg2+ it was not possible to obtain crystals of
TK from S. cerevisiae [8,36].
0.02,
v = (V2 + V1 · Km2 / [S]) / (1 + Km2 / [S] + Km1 · Km2 / [S]2)
which corresponds to the following tentative reaction scheme:
Fig. 3. Temperature dependences of excess heat capacity (ΔCp) for
apoTK (A) and holoTK (B) in 50 mM glycylglycine, pH 7.6. (1)
apoTK after treatment with EGTA (0.3 Ca2+, mol/mol remains) in
the absence of added Ca2+; (2) apoTK with addition of 0.1 mM
Ca2+; (3) holoTK formed with an excess of Ca2+ and ThDP and
passed through a Sephadex column to remove Ca2+ from the
medium, then 2 mM ThDP was added; (4) to (3) was added
2.5 mM Mg2+; (5) to (3) was added 0.1 mM Ca2+. TK con-
centration was 0.5 mg/ml. Scan rate 1 °C/min.
126

O.N. Solovjeva et al.
MolecularCatalysis466(2019)122–129
Table 3
Kinetic parameters of transketolase in two-substrate reaction with X5P [14,38] and HPA (own data). The concentration of second substrate R5P was 1.5 mM
(saturating).
Substrate
X5P
Enzyme form
Cofactor
Km1 (μM)
Km2 (μM)
V1 (U/mg)
V2 (U/mg)
HoloTK
HoloTK
SemiholoTK 1
SemiholoTK 2
HoloTK
HoloTK
HoloTK
SemiholoTK 1
SemiholoTK 2
HoloTK
Ca2+
no
25
75
6
9
773
75
300
9
15
13
3
4
69
68
17
2
Ca2+
Ca2+
Mg2+
Ca2+
no
115
7
38
71
3
5
11
76
4
1
71
5
HPA
205
580
580
30
25
30
475
580
6
25
2.52
2.1
0.29
0.1
4.08
4.2
0.1
0.3
Ca2+
Ca2+
Mg2+
280
600
40
30
2.1
4.2
0.15
0.2
600
30
Km1 Km2
primary and the tertiary structure [6–8]. Kinetic differences of active
centers are formed during the process of catalysis. The reason for this is
that the active centers are located between the subunits close to each
other, which determines their interaction [8]. Changes in protein
structure as a result of the binding of various ligands can be determined
by the DSC method. The higher the maximum of the thermal transition,
the more ordered the structure of the molecule. For TK, we showed that
different values of the maximum of the thermal transition (in ascending
order) are determined (1) for apoTK in the absence of cations in the
active centers, (2) for apoTK in the presence of Ca2+, (3) for holoTK
formed in the presence of Ca2+ with the subsequent removal of cations
from the medium and (4) for holoTK in the presence of Ca2+ in the
medium. A holoenzyme formed in the presence of Mg2+ and with Mg2+
in the medium has the same maximum as the holoenzyme in the ab-
E ↔ ES1 ↔ ES2 → V2
However, the sharp decrease of the enzyme activity at ˜10 μM for
X5P and at ˜100 μM for HPA in the presented here data cannot be de-
scribed by the above formula. We propose the following mechanism of
this sharp decrease in activity. At lower substrate concentration a
substrate molecule bound in one of the two centers of the dimer is
transformed into the product and releases before the enzyme binds a
second molecule of the substrate in the second center. In this way, only
one of the two centers is occupied. At some critical concentration of
substrate, the enzyme binds a second molecule of the substrate before
the release of the product from the first substrate molecule. Such an
occupation of both centers results in a decrease of the enzyme activity.
Thus, the increase of substrate concentration above some critical value
results in the transformation of a practically whole amount of enzyme
into a less active form. If this is true, then in such a situation, the whole
interval of applied S can be divided into the two parts. When the sub-
strate concentration is lower than its critical value, only one center is
occupied, and the enzyme is in the active form. If the substrate con-
centration is higher than the critical value, both centers are occupied,
and the enzyme is in a less active form. Since in any of these situations
the enzyme is practically homogeneous with regards to its activity, the
two parts of the measured V(S) can be analyzed independently to de-
termine the characteristics of the enzyme reaction. To this end, we used
the Michaelis equation. Fig. 5 show the best fit of the Michaelis equa-
tion to these two parts of the measured curve for X5P (curves A, B) and
HPA (curves C, D).
sence of cations in the medium. Accordingly, in the presence of Ca2+
,
but not Mg2+ in the medium, the thermostability of holoTK increases.
Therefore, it can be assumed that in holoTK Ca2+ binds outside the
active centers. Previously it was shown only that TK can bind up to 9
molecules of ThDP [41].
In Fig. 3, curve 1, Ca2+ is removed from 85% protein and is absent
in the medium. In this case, the maximum of the thermal transition is
almost equal to the maximum for the apoenzyme in the presence of an
excess of Ca2+ (curve 2). Consequently, the change in the overall
structure of the apoenzyme is due to the presence of Ca2+ only in the
active centers. After binding of Ca2+ with the first active center, the
affinity of the second active center to Ca2+ decreases by an order of
magnitude (Table 2, N1). In other words, the altered structure of one
subunit (containing Ca2+ in one, conditionally first, active center)
changes the structure of the second active center, as a result of which
the affinity of Ca2+ to the second active center decreases. Similar re-
sults were obtained for the binding of Mg2+, but the affinity of both the
first and second active centers for Mg2+ was an order of magnitude
lower than for Ca2+ (Table 2, N2).
The fact that the Michaelis equation fits both parts of the measured
V(S) indicates the correctness of our suggestion about the homogeneity
of the enzyme in both analyzed here situations. It was previously shown
that active centers of TK work at flip-flop mechanism [39,40].
In holoTK, unlike apoTK, the overall structure of the enzyme
(maximum of thermal transition) differs in the presence of Ca2+ in the
medium and when cations are removed from the medium (Fig. 3, curves
3,5). Adding Mg2+ to the medium (when Ca2+ is present in the active
4. Discussion
The subunits of the dimeric TK molecule are identical in both the
Table 4
Kinetic parameters of transketolase in one-substrate reaction.
Substrate
X5P
Enzyme form
Cofactor
Km1 (μM)
Km2 (μM)
V1 (U/mg)
V2 (U/mg)
HoloTK
HoloTK
SemiholoTK 1
SemiholoTK 2
HoloTK
HoloTK
HoloTK
SemiholoTK 1
SemiholoTK 2
HoloTK
Ca2+
no
0.39
25.3
22.4
0.02
0.03
1.2
24.34
25.3
1.69
0.03
0.20
0.16
0.003
0.02
0.32
0.32
0.01
0.01
Ca2+
Ca2+
Mg2+
Ca2+
no
23.5
21
1.4
1.2
0.15
0.30
0.02
0.02
21
1.2
HPA
21
115
24
1,5
4
2.0
110
115
5
4
0.032
0.019
0.002
0.001
0.044
0.042
0.002
0.002
Ca2+
Ca2+
Mg2+
25
2.0
0.018
0.32
0.002
0.03
600
127

O.N. Solovjeva et al.
MolecularCatalysis466(2019)122–129
Fig. 4. Dependence of holoTK (1) and semiholoTK (2) activity in a
one-substrate reaction in the presence of Ca2+ on the concentra-
tion of X5P (A) and HPA (B). The reaction mixture contained: A –
7 mM sodium arsenate, 0.37 mM NAD+, 3 U/ml glyceraldehyde 3-
phosphate dehydrogenase, 3.2 mM dithiothreitol, 0.1 mM CaCl2,
0.1 mM ThDP, 5 μg/ml TK, 50 mM glycylglycine, pH 7.6. The
reaction was initiated by the addition of 0.003-0.1 mM X5P; B –
0.1 mM CaCl2, 0.1 mM ThDP, 50 μg/ml TK, 10 mM glycylglycine,
pH 7.6. The reaction was initiated by the addition of 0.03-0.6 mM
HPA. In the reaction with both substrates, semiholoTK contained
ThDP concentration, equimolar TK concentration.
centers) does not shift the maximum of the holoTK thermal transition
(Fig. 3, curve 4).
values determined for X5P and HPA in one-substrate and two-substrate
reaction. In the absence of cations in the medium or in the presence of
Mg2, a single Km value is determined (Table 3,4). This is in direct
agreement with the DSC data on the increase in thermostability of
holoTK in the presence of Ca2. Apparently, this occurs as a result of the
binding of Ca2+ outside the active centers of TK.
The presence of an acceptor substrate (R5P) reduces the affinity of
holoTK for the donor substrate (X5P and HPA) compared with its affi-
nity in the one-substrate reaction. This can be explained by the com-
petitive relationship of donor and acceptor substrates, which was
shown earlier [42].
An interesting fact is the sharp drop in activity in the one-substrate
reaction, which we explain by changing the flip-flop mechanism. A si-
milar sharp decrease in activity was observed for ketoglutarate dehy-
drogenase from pigeon breast muscle [43]. In the plots of the depen-
dence of the enzyme activity on the concentration of ketoglutarate,
hyperbola is transformed into a curve with a maximum and minimum
after the reduction of the 2 SH-groups of the enzyme (Fig. 6).
Each of the two active centers of TK is formed by two subunits. Each
ThDP molecule is bound by a pyrophosphate tail through a cation with
the first subunit and an aminopyrimidine ring with the second subunit
[8] (Fig. 1). In the absence of a cation, ThDP binding does not occur
[13]. A two-step ThDP binding mechanism was previously shown. Kd
for the primary fast binding step of ThDP through PP-domain and the
forward rate constants for the secondary slow binding step of ThDP
through the aminopyrimidine ring are similar for two active centers.
The observed cooperativity in the binding of ThDP to TK is asso-
ciated with the backward reaction of the secondary binding of ThDP,
which indicates the conformational instability of one of the active
centers in the holodimer [17]. If Ca2+ is bound only in the first active
center (there are no cations in the medium), then the affinity of ThDP to
this active center decreases by two orders of magnitude (Table2, N5), as
compared to the variant when Ca2+ is bound in both active centers. At
the same time, the binding time of ThDP with TK increases (Fig.2, curve
3). It can be assumed that in the absence of Ca2+ in the second active
center, the parameters of primary ThDP binding to Ca2+ of the first
active center in the first subunit do not change, and the affinity of its
aminopyrimidine ring changes to the amino acids of the first active
center located in the second subunit. Accordingly, Ca2+, which binds to
the second subunit in the second active center, changes the structure of
the first active center in the same subunit.
5. Conclusions
Cooperativity of the active centers of the TK from baker's yeast is
formed by the binding of a cation, and not ThDP, as previously thought.
As a result of the binding of Ca2+ or Mg2+ in one of the two active
centers of identical subunits (conditionally the first), the affinity of the
second active center to the cation decreases by an order of magnitude.
This decrease in affinity was previously shown for ThDP in the presence
The presence of Ca2+ in the medium also affects the kinetic para-
meters of substrate binding. Only in the presence of Ca2+ are two Km
Fig. 5. Fitting of the measured curve using the Michaelis equation, for one-substrate reaction with X5P of the first (A) and the second (B) parts and with HPA of the
first (C) and the second (D) parts.
128

O.N. Solovjeva et al.
MolecularCatalysis466(2019)122–129
[2] R.G. Duggleby, Acc. Chem. Res. 39 (2006) 550–557.
[3] J.J.L. Wang, P.R. Martin, C.K. Singleton, Biochim. Biophys. Acta Protein Struct.
Mol. Enzymol. 1341 (1997) 165–172.
[4] R.A.W. Frank, F.J. Leeper, B.F. Luisi, Cell. Mol. Life Sci. 64 (2007) 892–905.
[5] M. Sundström, Y. Lindqvist, G. Schneider, U. Hellman, H. Ronne, J. Biol. Chem. 268
(1993) 24346–24352.
[6] L.E. Meshalkina, G.A. Kochetov, Biochim. Biophys. Acta Enzymol. 51 (1979)
218–223.
[7] G.A. Kochetov, L.E. Meshalkina, R.A. Usmanov, Biochem. Biophys. Res. Commun.
69 (1976) 839–843.
[8] Y. Lindqvist, G. Schneider, U. Ermler, M. Sundström, EMBO J. 11 (1992)
2373–2379.
[9] F. Dyda, W. Furey, S. Swaminathan, M. Sax, B. Farrenkopf, F. Jordan, Biochemistry
32 (1993) 6165–6170.
[10] Y.A. Muller, Y. Lindqvist, W. Furey, G.E. Schulz, F. Jordan, G. Schneider, Structure
1 (1993) 95–103.
[11] G.A. Kochetov, P.P. Philippov, Biochem. Biophys. Res. Commun. 38 (1970)
930–933.
[12] M.V. Kovina, V.A. Selivanov, N.V. Kochevova, G.A. Kochetov, Biochemistry Mosc.
63 (1998) 988–995.
Fig. 6. The plots of the ketoglutarate dehydrogenase activity versus ketoglu-
tarate concentration before (A) and after (B) the reduction by dihydrolipoate
[43].
of both Ca2+ and Mg2+. In the absence of Ca2+ in the second active
center, the affinity of ThDP for the first one decreases by two orders of
magnitude. In other words, the absence of Ca2+ in the active center in
the region of the PP domain of one subunit changes the structure of the
active center in the region of the PYR domain of the second subunit.
The absence of Ca2+ in the second active center reduces by an order of
magnitude the binding rate of ThDP with apoTK (as shown by CD).
In the presence of Mg2+, the ThDP binding rate is an order of
magnitude lower than in the presence of Ca2+. Using DSC, it is shown
that Ca2+ and Mg2+ also change the overall structure of apoTK. The
overall structure of holoTK changes only in the presence of an excess of
Ca2+ in the medium, but not Mg2+. The nonequivalence of active
centers with respect to two donor substrates in a one-substrate reaction
in the presence of Ca2+ was shown. For both substrates, the affinity of
the second active center is lower compared with the affinity of the first.
In addition, the activity of the first active center in the holoTK is sig-
nificantly higher than in the semiholoTK and at the functioning of both
centers of the holoTK. In other words, the alternate work of active
centers in the dimer (flip-flop) reduces their activity. In addition, the
activity of the first active center in the holoTK is significantly higher
than in semiholoTK and with the operation of both centers of the
holoTK.
[13] G.A. Kochetov, Ann N.Y.Acad Sci. 378 (1982) 306–311.
[14] O.N. Solovjeva, I.A. Sevostyanova, V.A. Yurshev, V.A. Selivanov, G.A. Kochetov,
Protein J. 31 (2012) 137–140.
[15] R.M. Egan, H.Z. Sable, J. Biol. Chem. 256 (1981) 4877–4883.
[16] L.E. Meshalkina, G.A. Kochetov, Dokl. Akad. Nauk SSSR 248 (1979) 1482–1486.
[17] M.V. Kovina, V.A. Selivanov, N.V. Kochevova, G.A. Kochetov, FEBS Lett. 418
(1997) 11–14.
[18] H. Heinrich, H. Steffen, P. Janser, O. Wiss, Eur. J. Biochem. 30 (1972) 533–541.
[19] G.A. Kochetov, N.K. Tikhomirova, P.P. Philippov, Biochem. Biophys. Res. Commun.
63 (1975) 924–930.
[20] G. Schenk, R.G. Duggleby, P.F. Nixon, Int. J. Biochem.Cell Biol. 30 (1998)
1297–1318.
[21] G.A. Kochetov, O.N. Solovjeva, Biochim. Biophys. Acta Proteins Proteomics 1844
(2014) 1608–1618.
[22] I.A. Bykova, O.N. Solovjeva, L.E. Meshalkina, M.V. Kovina, G.A. Kochetov,
Biochem. Biophys. Res. Commun. 280 (2001) 845–847.
[23] E. Fiedler, R. Golbik, G. Schneider, K. Tittmann, H. Neef, S. König, G. Hübner, J.
Biol. Chem. 276 (2001) 16051–16058.
[24] G.A. Kochetov, I.A. Sevostyanova, JUBMB Life 62 (2010) 797–802.
[25] C.J. Gubler, L.R. Johnson, J.H. Wittorf, Methods Enzymol 18 A (1970) 120–125.
[26] O.N. Solovjeva, G.A. Kochetov, J. Mol. Catal. B: Enzymatic 54 (2008) 90–92.
[27] O.N. Solovjeva, Biochemistry (Mosc.) 67 (2002) 667–671.
[28] C.P. Heinrich, K. Noack, O. Wiss, Biochem. Biophys. Res. Commun. 49 (1972)
1427–1432.
[29] G.A. Kochetov, Methods Enzymol 90 (1982) 209–223 [32] S.T. Ohnishi, Analyt.
Biochem. 85 (1978) 165-179.
[30] S.T. Ohnishi, Analyt. Biochem. 85 (1978) 165–179.
[31] M.W. Freyer, E.A. Lewis, Methods Cell Biol. 84 (2008) 79–113.
[32] G.A. Kochetov, P.P. Philippov, Analyt. Biochem. 48 (1972) 286–291.
[33] A. Kornberg, Methods Enzymol. 1 (1955) 441–443.
[34] G.A. Kochetov, R.A. Usmanov, V.P. Merzlov, FEBS Lett. 9 (1970) 265–266.
[35] O.A. Esakova, L.E. Meshalkina, G.A. Kochetov, Life Sci. 78 (2005) 8–13.
[36] M. Nikkola, Y. Lindqvist, G. Schneider, J. Mol. Biol. 238 (1994) 387–404.
[37] I.A. Sevostyanova, V.A. Yurshev, O.N. Solovjeva, S.V. Zabrodskaya, G.A. Kochetov,
Proteins 71 (2008) 541–545.
[38] I.A. Sevostyanova, V.A. Selivanov, V.A. Yurshev, O.N. Solovjeva, S.V. Zabrodskaya,
G.A. Kochetov, Biochemistry (Mosc.) 74 (2009) 972–976.
[39] M.V. Kovina, G.A. Kochetov, FEBS Lett. 440 (1998) 1–4.
[40] I. Sevostyanova, O. Solovjeva, V. Selivanov, G. Kochetov, Biochem. Biophys. Res.
Commun. 379 (2009) 851–854.
Acknowledgments
We thank A.M. Arutyunyan for providing the RDA computer pro-
gram and help in the interpretation of the results on circular dichroism.
This research did not receive any specific grant from funding
agencies in the public, commercial, or not-for-profit sectors.
References
[41] A.G. Datta, E. Racker, J. Biol.Chem. 236 (1961) 617–623.
[42] O.N. Solov’eva, L.E. Meshalkina, M.V. Kovina, V.A. Selivanov, I.A. Bykova,
G.A. Kochetov, Biochemistry Mosc. 65 (2000) 1202–1205.
[43] V.I. Bunik, O.A. Buneeva, V.S. Gomazkova, FEBS Lett. 269 (1990) 252–254.
[1] F.J. Leeper, D. Hawksley, S. Mann, C. Perez Melero, M.D.H. Wood, Biochem. Soc.
Trans. 33 (2005) 772–775.
129