Influential factor contributing to the isoform-specific
inhibition by ATP of human mitochondrial
NAD(P)
+
-dependent malic enzyme
Functional roles of the nucleotide binding site Lys346
Ju-Yi Hsieh
1
, Guang-Yaw Liu
2
and Hui-Chih Hung
1,3
1 Department of Life Sciences, National Chung-Hsing University, Taichung, Taiwan
2 Institute of Immunology, Chung-Shan Medical University, Taichung, Taiwan
3 Institute of Bioinformatics, National Chung-Hsing University, Taichung, Taiwan
Malic enzymes are a family of oxidative decarboxy-
lases that catalyze the oxidation of L-malate to pyru-
vate with accompanying reduction of NAD(P)
+
to
NAD(P)H and release of CO
2
[1–7]. The enzyme
requires a divalent metal ion (Mn
2+
or Mg
2+
) for
enzyme catalysis [8,9]. In addition, the metal ion is
Keywords
allosteric regulation; ATP inhibition; cofactor
specificity; cooperativity; mutagenesis
Correspondence
H.-C. Hung, Department of Life Sciences
and Institute of Bioinformatics, National
Chung-Hsing University, 250 Kuo-Kuang
Road, Taichung, 40227 Taiwan
Fax: +886 4 22851856
Tel: +886 4 22840416 (ext. 615)
E-mail: hchung@dragon.nchu.edu.tw
(Received 21 June 2008, revised 1 September
2008, accepted 4 September 2008)
doi:10.1111/j.1742-4658.2008.06668.x
Human mitochondrial NAD(P)
+
-dependent malic enzyme (m-NAD-ME)
is a malic enzyme isoform with dual cofactor specificity, ATP inhibition
and substrate cooperativity. The determinant of ATP inhibition in malic
enzyme isoforms has not yet been identified. Sequence alignment of nucleo-
tide-binding sites of ME isoforms revealed that Lys346 is conserved
uniquely in m-NAD-ME. In other ME isoforms, this residue is serine. As
the inhibitory effect of ATP is more pronounced on m-NAD-ME than on
other ME isoforms, we have examined the possible role of Lys346 by
replacing it to alanine, serine or arginine. Our kinetic data indicate that the
K346S mutant enzyme displays a shift in its cofactor preference from
NAD
+
to NADP
+
upon increasing k
cat,NADP
and decreasing K
m,NADP
.
Furthermore, the cooperative binding of malate becomes less significant in
human m-NAD-ME after mutation of Lys346. The hvalue for the wild-
type is close to 2, but those of the K346 mutants are approximately 1.5.
The K346 mutants can also be activated by fumarate and the cooperative
effect can be abolished by fumarate, suggesting that the allosteric property
is retained in these mutants. Our data strongly suggest that Lys346 in
human m-NAD-ME is required for ATP inhibition. Mutation of Lys346 to
Ser or Ala causes the enzyme to be much less sensitive to ATP, similar to
cytosolic NADP-dependent malic enzyme. Substitution of Lys to Arg did
not change the isoform-specific inhibition of the enzyme by ATP. The
inhibition constants of ATP are increased for K346S and K346A, but are
similar to those of the wild-type for K346R, suggesting that the positive
charge rather than group specificity is required for binding affinity of ATP.
Thus, ATP inhibition is proposed to be determined by the electrostatic
potential involving the positive charge on the side chain of Lys346.
Abbreviations
c-NADP-ME, cytosolic NADP
+
-dependent malic enzyme; k
cat,NAD,
catalytic constant using NAD
+
as the cofactor; k
cat,NADP,
catalytic constant
using NADP
+
as the cofactor; K
i,ATP(NAD),
inhibition constant of ATP for NAD
+
;K
i,ATP(NADP),
inhibition constant of ATP for NADP
+
; ME, malic
enzyme; m-NAD-ME, mitochondrial NAD(P)
+
-dependent malic enzyme; m-NADP-ME, mitochondrial NADP
+
-dependent malic enzyme.
FEBS Journal 275 (2008) 5383–5392 ª2008 The Authors Journal compilation ª2008 FEBS 5383
required for enzyme folding and stability [10]. These
enzymes exist universally in nature, with sequence con-
servation and overall structural similarity among vari-
ous species [11–18]. Based on their cofactor specificity,
mammalian malic enzymes have been divided into three
isoforms: cytosolic NADP
+
-dependent (c-NADP-ME)
[19–22], mitochondrial NADP
+
-dependent (m-NADP-
ME) [23–25], and mitochondrial NAD(P)
+
-dependent
(m-NAD-ME) [6,26,27]. m-NAD-ME has dual cofactor
specificity in that it can employ both NAD
+
and
NADP
+
as cofactors, but it favors NAD
+
physiologi-
cally [6,28]. Human m-NAD-ME, through the produc-
tion of NADH and pyruvate, may be involved in
glutaminolysis in rapid growing tissues and tumors
[26,27,29–31].
m-NAD-ME is a unique isoform that has a multi-
faceted regulatory system for control of its catalytic
activity [4,6,27,28,30,32,33]. It displays substrate coop-
erativity in l-malate, and the enzyme can be allosteri-
cally activated by fumarate [16,30,33]. Previous studies
have suggested that ATP may work through an alloste-
ric mechanism to inhibit enzyme activity [16,30,34–36].
These allosteric properties of m-NAD-ME imply a
particular role in the malate and glutamine oxidation
pathways in tumor mitochondria [26,27,30,31,37].
However, site-directed mutagenesis and detailed kinetic
studies have demonstrated that ATP inhibition may
take place in the active site rather than the allosteric
site [36,38].
Several crystal structures of malic enzymes from var-
ious species have been determined, revealing that the
enzyme is a homotetramer comprising a dimer of
dimers quaternary structure. When the substrate
malate is bound, the enzyme is more compact at the
active site. These current structures establish malic
enzyme as a new group of oxidative decarboxylases
with a discrete backbone structure [11,12,14,15,
17,18,39]. Human m-NAD-ME and its substrate,
cofactor, metal ions, regulator and transition-state
analogue inhibitors have been described [13–16]. In the
structure of human m-NAD-ME, each monomer has a
separate active site that can bind one molecule of
NAD
+
or ATP. In addition, another molecule of
NAD
+
or ATP resides at an exo site in the tetramer
interface of each subunit (Fig. 1A). In the dimer inter-
face, an allosteric site is found to bind to the allosteric
activator fumarate [16]. Structural information on
pigeon c-NADP-ME and Ascaris suum m-NAD-ME is
also available [11,12,18], revealing that there is no
additional exo site in these two enzymes. Nevertheless,
the Ascaris suum NAD-malic enzyme can also be acti-
vated by fumarate, which binds to a separate allosteric
site [40].
Mammalian malic enzymes have distinct cofactor
specificities among their isoforms. Previous studies
have suggested that Lys362 in pigeon c-NADP-ME
[41] and Gln362 in human m-NAD-ME [28] have a
remarkable effect on the cofactor binding and speci-
ficity for these enzyme isoforms. Our recent data
clearly indicate that the Q362K mutant of human
m-NAD-ME is a non-allosteric, non-cooperative and
NADP
+
-specific enzyme, just like NADP-ME [28].
Additionally, Q362K is more sensitive to ATP, and the
inhibition constant is smaller than that of wild-type
A
B
Fig. 1. Surface representation of the crystal structure of the human
mitochondrial malic enzyme and sequence alignments of malic
enzymes from various sources. (A) Crystal structure of the enzyme
in complex with ATP, L-malate, Mn
2+
and fumarate (Protein Data
Bank code 1PJ4). The active site and exo site regions are shown
apart, with an ATP molecule, shown by the ball-and-stick method.
Fumarate is colored yellow and Mn
2+
is colored red. This image
was produced using PYMOL (DeLano Scientific LLC, San Carlos,
CA). (B) Sequence alignments of 14 malic enzymes with amino acid
sequences around the NAD
+
-binding region in the active site. The
amino acid sequences of malic enzymes were searched by BLAST
[43], and the alignments were generated using CLUSTAL W [44]. The
amino acid residues highlighted in color are residue 346 in the
NAD
+
-binding region.
Human mitochondrial NAD(P)
+
-dependent malic enzyme J.-Y. Hsieh et al.
5384 FEBS Journal 275 (2008) 5383–5392 ª2008 The Authors Journal compilation ª2008 FEBS
[28]. This unexpected result might be due to an addi-
tional positive charge introduced into the nucleotide
binding site. Sequence alignments of the nucleotide-
binding region among malic enzymes have revealed
that, in addition to Gln362, Lys346 is entirely con-
served among the NAD
+
-dependent malic enzymes,
but in NADP
+
-dependent malic enzymes, this lysine is
replaced by serine (Fig. 1B). As the inhibitory effect of
ATP is more pronounced on m-NAD-ME than on
other ME isoforms, we have investigated the possi-
ble role of Lys346 on the cofactor specificity, ATP
inhibition and malate cooperativity of the human
m-NAD(P)-dependent malic enzyme.
In this paper, we provide detailed kinetic evidence to
describe the determinants of isoform-specific ATP inhi-
bition for human m-NAD-ME. Site-directed mutagen-
esis was used to characterize the role of Lys346 in the
enzyme.
Results
Characterization of human wild-type and K346
mutant m-NAD-ME
The kinetic parameters for the recombinant ME
employing NAD
+
or NADP
+
as the cofactor were
determined with or without fumarate (Table 1). With
fumarate, the K
m,NAD
or K
m,NADP
values for wild-type
and K346 m-NAD-ME were decreased, and the k
cat
values for these enzymes were increased, indicating
that the allosteric property was retained in the mutant
enzymes. The K
m,NAD
values for the K346A and
K346S enzymes were increased slightly compared with
the wild-type enzyme, but the K
m,NAD
value was
decreased slightly for the K346R enzyme. The
K
m,NADP
value for K346A was increased slightly, while
those for both K346S and K346R were decreased. Fur-
thermore, the K
m,NADP
values for the wild-type and
K346A enzymes were greater than their K
m,NAD
val-
ues; however, in the K346S and K346R enzymes, the
K
m,NADP
values were slightly smaller than the K
m,NAD
values. The c-NADP-ME, which is NADP-specific,
had a very low K
m,NADP
value compared with its
K
m,NAD
: the K
m,NADP
⁄K
m,NAD
ratio was approxi-
mately 10 000. Among these mutants, the K346S
enzyme had the smallest K
m,NADP
⁄K
m,NAD
ratio, indi-
cating that this mutant has a decreased apparent affin-
ity towards NAD
+
but an increased apparent affinity
towards NADP
+
. Fumarate can decrease the K
m
val-
ues for these enzymes, but seemed to be less efficient
in decreasing the K
m,NADP
values. The decrease in
K
m,NAD
values of these enzymes by fumarate was
approximately 1.3–3.1-fold, while the decrease in
K
m,NADP
values was approximately 1.2–1.5-fold.
The k
cat,NAD
values of the K346 mutants showed a
1.5-fold decrease compared with the wild-type enzyme.
In contrast, the k
cat,NADP
values for both the K346S
and K346R enzymes were superior to those of wild-
type. The K346A enzyme, similar to wild-type, dis-
played poor catalytic efficiency when NADP
+
was
utilized as the cofactor. Furthermore, the k
cat
values
for the wild-type and K346 mutant enzymes using
NADP
+
as the cofactor were much less than those
using NAD
+
as the cofactor. For the K346S enzyme,
the k
cat,NADP
value was increased to a certain extent
similar to k
cat,NAD
, and was found to be approximately
five times greater than that of the wild-type enzyme.
Furthermore, the k
cat,NADP
of m-NAD-ME was only
8% relative to c-NADP-ME, but increased to approxi-
mately 40% for K346S. These results indicate that the
K346S enzyme shows an increased catalytic turnover
rate when utilizing NADP
+
as the cofactor. Neverthe-
less, the k
cat,NADP
⁄K
m,NADP
ratio of c-NADP-ME
was over 400-fold larger than that of K346S, indi-
cating that the K346S enzyme retains dual cofactor
specificity.
Table 1. Kinetic parameters for wild-type and K346 human m-NAD-ME1. ), no fumarate added; +, with 3 mMfumarate added.
K
m,NAD
(mM)
K
m,NADP
(mM)
K
m,NADP
⁄
K
m,NAD
k
cat,NAD
(s
)1
)
k
cat,NADP
(s
)1
)
k
cat,NADP
⁄
k
cat,NAD
k
cat,NAD
⁄
K
m,NAD
(s
)1
ÆmM
)1
)
k
cat,NADP
⁄
K
m,NADP
(s
)1
ÆmM
)1
)
m-NAD-ME )0.62 ± 0.06 1.76 ± 0.40 2.84 126.31 ± 10.5 9.04 ± 0.5 0.07 203.73 5.14
+ 0.26 ± 0.08 1.27 ± 0.31 4.88 156.80 ± 8.6 23.06 ± 1.3 0.15 603.07 18.16
K346A m-NAD-ME )1.16 ± 0.24 2.44 ± 0.32 2.10 90.51 ± 9.8 3.27 ± 0.2 0.04 78.02 1.34
+ 0.37 ± 0.07 1.99 ± 0.27 5.38 115.24 ± 11.6 3.69 ± 0.9 0.03 311.46 1.85
K346S m-NAD-ME )1.02 ± 0.17 0.80 ± 0.06 0.78 71.39 ± 6.1 42.63 ± 1.3 0.60 69.99 53.29
+ 0.48 ± 0.07 0.53 ± 0.09 1.10 78.79 ± 3.5 48.00 ± 2.1 0.61 164.15 90.57
K346R m-NAD-ME )0.24 ± 0.03 0.50 ± 0.02 2.08 79.57 ± 4.9 20.58 ± 1.5 0.26 331.54 41.16
+ 0.18 ± 0.02 0.33 ± 0.05 1.83 85.06 ± 7.4 26.07 ± 0.9 0.31 472.56 79
c-NADP-ME )18.6 ± 4.56 5.3 ·10
)3
±1.4 ·10
)3
2.8 ·10
)4
34.18 ± 3.8 113.21 ± 6.4 3.31 1.84 21 360.38
J.-Y. Hsieh et al. Human mitochondrial NAD(P)
+
-dependent malic enzyme
FEBS Journal 275 (2008) 5383–5392 ª2008 The Authors Journal compilation ª2008 FEBS 5385
Inhibitory effect of ATP on human wild-type and
K346 mutant m-NAD-ME
ATP can effectively inhibit the catalytic activity of
human m-NAD-ME, but this effect is not obvious for
the c-NADP-ME isoform (Fig. 2, open and closed cir-
cles, respectively). Although c-NADP-ME was slightly
inhibited, the inhibitory effect of ATP was less pro-
nounced with NADP
+
as the cofactor (Fig. 2B, closed
circles). In contrast, ATP inhibition was more pro-
minent in m-NAD-ME with NADP
+
as the cofactor
(Fig. 2B, open circles). The differential inhibitory effect
of ATP might be due to the protective effect on
m-NAD-ME and c-NADP-ME of their isoform-specific
cofactors. Furthermore, the inhibitory effect of ATP on
the K346 mutant enzymes was less pronounced than
that on wild-type m-NAD-ME, except for K346R.
K346R is the mutant that is most susceptible to ATP
inhibition; the residual enzyme activities of wild-type
and K346R were at the same level either with NAD
+
or NADP
+
as the cofactor (Fig. 2, closed squares),
indicating that conserving the positive charge on this
residue is crucial for ATP inhibition. The K346S and
K346A enzymes were less sensitive to ATP (Fig. 2,
open and closed triangles, respectively). However, with
NADP
+
as the cofactor, ATP inhibition appears
more obvious for the wild-type and K346 mutant
m-NAD-ME (Fig. 2B).
ATP acts as an active-site inhibitor of m-NAD-ME
following a competitive mechanism for NAD
+
and
malate [36,38]. Our data indicate that the K346 mutants
were also inhibited by ATP. Inhibition experiments were
used to determine the binding constant of ATP for these
mutant enzymes. For the wild-type m-NAD-ME, the
inhibition constants of ATP for NAD
+
(K
i,ATP(NAD)
)
and NADP
+
(K
i,ATP(NADP)
) were 0.30 and 0.33 mm,
respectively (Table 2). For c-NADP-ME, the values of
K
i,ATP(NAD)
and K
i,ATP(NADP)
were 4.77 and 5.91 mm,
respectively (Table 2), which are approximately 16-fold
greater than those of m-NAD-ME. This observation
coincided with the ATP effect experiments (Fig. 2),
demonstrating the low affinity of this c-NADP-ME
isoform for ATP. The inhibition constants of ATP for
K346 mutant enzymes, however, were quite different.
For the K346R enzyme, the inhibition pattern is similar
to that for wild-type m-NAD-ME (Fig. 2), and the
K
i,ATP
values for the wild-type and K346R were almost
identical (Table 2), indicating that this mutant did not
change its property towards ATP inhibition. In contrast,
the K
i,ATP
values for K346A and K346S were greater
Fig. 2. Inhibitory effect of ATP on human m-NAD-ME and c-NADP-
ME. ATP inhibition of the wild-type enzyme and Q346 mutant
m-NAD-ME was assayed using NAD
+
(A) or NADP
+
(B) as the
cofactor. The assay mixture contains 15 mMmalate, 10 mMMgCl
2
,
and 1 mMNAD
+
or NADP
+
in a 50 mMTris ⁄HCl (pH 7.4) buffer
system. The ATP concentration ranged from 0 to 3 mM. Closed
circles, c-NADP-ME; open circles, wild-type m-NAD-ME; closed
triangles, 346A m-NAD-ME; open triangles, 346S m-NAD-ME;
closed squares, 346R m-NAD-ME.
Table 2. Inhibition constants for wild-type and K346 human
m-NAD-ME.
K
i,ATP(NAD)
(mM)
K
i,ATP(NADP)
(mM)
G
(kcalÆmol)
)1
for
K
i,ATP(NAD)a
G
(kcalÆmol)
)1
for
K
i,ATP(NADP)a
m-NAD-ME 0.30 ± 0.04 0.33 ± 0.03 0.00 0.00
K346A
m-NAD-ME
1.74 ± 0.22 1.80 ± 0.15 1.06 1.02
K346S
m-NAD-ME
2.23 ± 0.12 2.23 ± 0.14 1.21 1.15
K346R
m-NAD-ME
0.40 ± 0.06 0.25 ± 0.03 0.17 )0.17
c-NADP-ME 4.77 ± 0.49 5.91 ± 1.06 1.67 1.74
a
Calculation of the Gvalue for the K346 mutant enzymes was
based on the equation: G=)RT ln[K
i(wild-type)
⁄K
i(mutant)
].
Human mitochondrial NAD(P)
+
-dependent malic enzyme J.-Y. Hsieh et al.
5386 FEBS Journal 275 (2008) 5383–5392 ª2008 The Authors Journal compilation ª2008 FEBS
than that of wild-type. For the K346S enzyme, the
K
i,ATP
value was sevenfold greater than for the wild-type
m-NAD-ME and was similar to that for c-NADP-ME,
suggesting that this mutant enzyme is less sensitive to
ATP, just like c-NADP-ME. In addition, the K
i,ATP
values with either NAD
+
or NADP
+
as the cofactor
were almost identical, suggesting that the binding affin-
ity of ATP with the enzyme was not associated with the
cofactor specificity.
Cooperative effect of malate on human wild-type
and K346 mutant m-NAD-ME
The initial velocity of the wild-type and K346 mutant
m-NAD-ME measured for various concentrations of
malate with either NAD
+
or NADP
+
as the cofactor
showed sigmoidal kinetics (Fig. 3, closed circles),
which implies cooperative binding of malate. However,
this cooperative effect was abolished by fumarate. The
sigmoidal curve changed to a hyperbolic one at a satu-
rated fumarate concentration of 3 mm, which can fully
activate m-NAD-ME activity (Fig. 3, closed triangles).
For the K346 mutant enzymes, the sigmoidal kinetics
at various malate concentrations without fumarate
were less obvious than for wild-type, suggesting that
Ser346 has an effect on malate cooperativity (Fig. 3,
closed circles). All of the K346 mutant enzymes dem-
onstrated lower cooperativity for malate binding, but
the enzyme activities were still activated by fumarate
and the cooperative effect was abolished by fumarate
Fig. 3. Cooperativity of L-malate for the
human mitochondrial malic enzyme. The
assay mixture contains 10 mMMgCl
2
,
1.0 mMNAD
+
or NADP
+
, and various
concentrations of malate in a 50 mM
Tris ⁄HCl (pH 7.4) buffer system with or
without fumarate. The fumarate concentra-
tions used in the experiments are 0 mM
(closed circles), 0.2 mM(open circles) and
3mM(closed triangles). Panels on the left
(A, C, E and G) show the titration curves
using NAD
+
as the cofactor, while those on
the right (B, D, F and H) show the titration
curves using NADP
+
as the cofactor. (A,B)
Wild-type enzyme; (C,D) K346A enzyme;
(E,F) K346S enzyme; (G,H) K346R enzyme.
J.-Y. Hsieh et al. Human mitochondrial NAD(P)
+
-dependent malic enzyme
FEBS Journal 275 (2008) 5383–5392 ª2008 The Authors Journal compilation ª2008 FEBS 5387