Báo cáo khoa học: Identification of domains involved in the allosteric regulation of cytosolic Arabidopsis thaliana NADP-malic enzymes ppt

Identification of domains involved in the allosteric
regulation of cytosolic Arabidopsis thaliana NADP-malic
enzymes
Mariel C. Gerrard Wheeler
1
, Cintia L. Arias
1
, Vero
´
nica G. Maurino
2
, Carlos S. Andreo
1
and
Marı
´
a F. Drincovich
1
1 Centro de Estudios Fotosinte
´
ticos y Bioquı
´
micos, Universidad Nacional de Rosario, Argentina
2 Botanisches Institut, Universita
¨
tzuKo
¨
ln, Cologne, Germany
Introduction
Malic enzymes (MEs) catalyse the reversible oxidative
decarboxylation of l-malate to pyruvate, CO

2
and
NAD(P)H in the presence of a divalent cation [1]. This
enzyme is widely distributed in nature, as the substrates
and products of the reaction participate in different met-
abolic pathways. In plants, photosynthetic and nonpho-
tosynthetic NADP-dependent isoenzymes (NADP-ME;
EC 1.1.1.40) have been found in both plastids and
cytosol [2]. Moreover, the co-expression of different
NADP-ME isoenzymes has been observed in the same
cell, and even in the same subcellular compartment [3].
The elucidation of the biological role of the different
NADP-ME isoenzymes, apart from being involved in
C
4
photosynthesis or crassulacean acid metabolism, will
Keywords
allosteric; Arabidopsis thaliana; isoenzymes;
NADP-malic enzyme; regulation
Correspondence
M. F. Drincovich, Centro de Estudios
Fotosinte
´
ticos y Bioquı
´
micos (CEFOBI),
Universidad Nacional de Rosario, Suipacha
531, 2000 Rosario, Argentina
Fax: 54 341 4370044
Tel: 54 341 4371955

E-mail:
(Received 19 June 2009, revised 28 July
2009, accepted 4 August 2009)
doi:10.1111/j.1742-4658.2009.07258.x
The Arabidopsis thaliana genome contains four genes encoding NADP-
malic enzymes (NADP-ME1–4). Two isoenzymes, NADP-ME2 and
NADP-ME3, which are shown to be located in the cytosol, share a
remarkably high degree of identity (90%). However, they display different
expression patterns and show distinct kinetic properties, especially with
regard to their regulation by effectors, in both the forward (malate oxida-
tive decarboxylation) and reverse (pyruvate reductive carboxylation) reac-
tions. In order to identify the domains in the primary structure that could
be responsible for the regulatory differences, four chimeras between these
isoenzymes were constructed and analysed. All chimeric versions exhibited
the same native structures as the parental proteins. Analysis of the chime-
ras constructed indicated that the region from amino acid residue 303 to
the C-terminal end of NADP-ME2 is critical for fumarate activation.
However, the region flanked by amino acid residues 303 and 500 of
NADP-ME3 is involved in the pH-dependent inhibition by high malate
concentration. Furthermore, the N-terminal region of NADP-ME2 is
necessary for the activation by succinate of the reverse reaction. Overall,
the results show that NADP-ME2 and NADP-ME3 are able to distinguish
and interact differently with similar C
4
acids as a result of minimal struc-
tural differences. Therefore, although the active sites of NADP-ME2 and
NADP-ME3 are highly conserved, both isoenzymes acquire different allo-
steric sites, leading to the creation of proteins with unique regulatory mech-
anisms, probably best suited to the specific organ and developmental
pattern of expression of each isoenzyme.

Abbreviations
GFP, green fluorescent protein; ME, malic enzyme.
FEBS Journal 276 (2009) 5665–5677 ª 2009 The Authors Journal compilation ª 2009 FEBS 5665
require further effort, as the gene family of this protein
is more complex than expected [4].
The Arabidopsis thaliana genome contains four
NADP-ME genes [3,4]. One gene encodes a plastidic
enzyme (NADP-ME4 [3]), but the other three isoenzymes
do not possess predictable organellar sorting sequences
and thus are thought to be located in the cytosol
(NADP-ME1–3). Previous studies have indicated dif-
ferential expression patterns for each isoenzyme [3].
In this regard, although NADP-ME2 and NADP-
ME4 are constitutively expressed in mature organs,
NADP-ME1 is restricted to secondary roots and
NADP-ME3 to trichomes and pollen [3]. Although
the four isoenzymes share a high degree of identity
(75–90%), the recombinant enzymes show distinct
structural and kinetic properties [3,5]. Specifically,
the isoenzymes behave differently in terms of regula-
tion by metabolic effectors, NADP-ME2 being the
most highly regulated, especially by activation [5].
In particular, NADP-ME2 and NADP-ME3 share
90% identity (Fig. 1), are encoded in the same
chromosome and belong to the cytosolic dicot group
in a phylogenetic tree constructed with plant NADP-
ME sequences [3]. In the malate oxidative decar-
boxylation reaction, although NADP-ME2 is highly
activated by aspartate, fumarate and succinate,
NADP-ME3 is inhibited by fumarate with no modifi-

cation of the enzymatic activity in the presence of
aspartate and succinate [5]. Furthermore, although suc-
cinate and fumarate show strong activation of the
NADP-ME2 pyruvate reductive decarboxylation reac-
tion (up to 400%), these metabolites act as inhibitors
of the NADP-ME3 reverse reaction [5]. Two NADP-
ME2 amino-terminal deletions previously analysed
indicated that some residues from this region are criti-
cal for aspartate and CoA activation [5]. However,
regions involved in the differential regulation by fuma-
rate and succinate could not be mapped by this
approach. Moreover, the mutation of R115 in NADP-
ME2 indicated that this amino acid residue is involved
in fumarate activation [5]. However, this residue is
conserved in NADP-ME3, indicating that other amino
Fig. 1. Sequence alignment of A. thaliana
NADP-ME2 and NADP-ME3. Regions of the
primary structure of each isoenzyme that
are involved in fumarate activation (in
yellow), CoA activation (underlined) and
malate inhibition (in green) of the forward
reaction are highlighted. In addition, the
regions involved in succinate activation of
the reverse reaction are highlighted in light
blue. Nonconserved residues between the
two sequences are shown in bold. ‘*’, iden-
tical residues; ‘:’, conserved substitution; ‘.’,
semiconserved substitution.
Regulation of A. thaliana NADP-malic enzyme activity M. C. Gerrard Wheeler et al.
5666 FEBS Journal 276 (2009) 5665–5677 ª 2009 The Authors Journal compilation ª 2009 FEBS

acid residues are responsible for the differential regula-
tion by fumarate. In this work, NADP-ME2 (Uniprot
Accession Number Q9LYG3) and NADP-ME3
(Uniprot Accession Number Q9XGZ0) are experimen-
tally shown to be located in the cytosol; moreover, the
relationship between the primary structure and differ-
ences in regulation was investigated by the character-
ization of complementary chimeras between the two
isoenzymes. The segments swapped in the construction
of the chimeras allowed us to evaluate the eight non-
conserved amino acid residues between NADP-ME2
and NADP-ME3. These amino acid residues are sepa-
rated into two regions: three are located in the first
segment swapped and five in the second (Fig. 1). Using
this approach, specific segments of the primary struc-
ture responsible for regulatory differences were identi-
fied, indicating that minimal structural changes are
responsible for the distinct behaviour of these two
highly similar NADP-ME isoenzymes.
Results
Subcellular localization of A. thaliana NADP-ME2
and NADP-ME3
In order to determine the subcellular localization of
A. thaliana NADP-ME2 and NADP-ME3, the full-
length cDNA of each isoenzyme was fused in frame to
the green fluorescent protein (GFP) coding sequence,
and the localization of the fluorescence was assayed by
transient expression in A. thaliana cell cultures. Figure 2
clearly shows that NADP-ME2 and NADP-ME3 are
both homogeneously distributed in the cytosol. A con-

trol assay with the GFP coding region shows the locali-
zation of free GFP in the cytosol and nucleus (Fig. 2).
Structural characterization of chimeric
NADP-MEs
In order to examine the sequence domains responsible
for regulatory differences between the cytosolic isoen-
zymes NADP-ME2 and NADP-ME3, four chimeric
proteins (named ME2.3, ME2.3¢, ME3.2 and ME3.2¢;
Fig. 3) were successfully expressed in E. coli and puri-
fied to homogeneity. To determine whether the chime-
ric proteins display any structural changes in relation
to the parental proteins, CD spectra for all chimeric
and parental enzymes were compared. In all cases, the
CD spectra obtained after corrections for protein con-
centration were very similar (data not shown), indicat-
ing that there was no significant loss of secondary
structure in the chimeric proteins.
Monomeric molecular masses of 65 kDa were
determined by SDS-PAGE for all chimeras (data not
Fig. 2. Subcellular localization of A. thaliana NADP-ME2 and NADP-ME3. Transient expression of 35S::NADP-ME2::GFP, 35S::NADP-
ME3::GFP and 35S::GFP in A. thaliana protoplasts. Bright field images with the superimposed GFP fluorescence images shown at the top.
Fluorescence distribution is shown at the bottom. The scale bar represents 12 lm.
M. C. Gerrard Wheeler et al. Regulation of A. thaliana NADP-malic enzyme activity
FEBS Journal 276 (2009) 5665–5677 ª 2009 The Authors Journal compilation ª 2009 FEBS 5667
shown), which are in accordance with those
obtained for the parental recombinant proteins [3].
Native electrophoresis of the purified proteins indi-
cated that the parental and chimeric proteins
presented almost identical electrophoretic mobility
(data not shown and [3]). Moreover, the parental

and chimeric proteins presented highly similar native
molecular masses by gel filtration chromatography,
Fig. 3. Chimeric NADP-MEs constructed and analysed in the present work. The conserved restriction sites EcoRV and BclI at positions 910
and 1500, respectively, of the cDNA of parental enzymes (NADP-ME2 and NADP-ME3) were used to construct the complementary chimeric
enzymes ME2.3, ME2.3¢, ME3.2 and ME3.2¢. These sites correspond to positions 303 and 500, respectively, in the protein sequence of
NADP-ME2 and NADP-ME3. The recombinant NADP-MEs that are activated by fumarate or CoA or inhibited by high malate concentration
at pH 7.0 (for the forward reaction) or that are activated by succinate (for the reverse reaction) are indicated on the right by ‘4’. Regions of
the primary structure of each parental NADP-ME that are involved in fumarate, CoA and succinate activation and malate inhibition are
indicated.
Table 1. Properties of parental and chimeric NADP-MEs. The indicated values are the average of at least three different measure-
ments ± SD. For k
cat
calculations, a 65 kDa monomeric molecular mass was used for all isoenzymes. Some values for parental NADP-ME2
and NADP-ME3, obtained previously [3], are included for comparison. (k
catD
⁄ k
catC
, k
cat decarboxylation
⁄ k
cat carboxylation
; M, monomeric molecular
mass; N, native molecular mass; NI, no inhibition was observed.).
NADP-ME2 NADP-ME3 ME2.3 ME2.3¢ ME3.2 ME3.2¢
Malate oxidative decarboxylation, pH 7.5
k
cat
(s
)1
) 324 ± 29 268 ± 24 148 ± 13 156 ± 16 45 ± 5 94 ± 13

c
K
mNADP
(lM) 72 ± 7 6.5 ± 0.6 31 ± 5 55 ± 9 23 ± 1 31 ± 8
c
K
m
L
-malate
(mM) 3.3 ± 0.4 0.8 ± 0.1 1.1 ± 0.1 4.8 ± 0.5 2.3 ± 0.4 1.7 ± 0.1
c
Malate oxidative decarboxylation; pH 7.0
K
r
(mM) NI 0.6 ± 0.1 9.6 ± 0.2 NI NI 1.1 ± 0.1
F NI 0.1 ± 0.05 0.7 ± 0.1 NI NI 0.3 ± 0.1
Pyruvate reductive carboxylation; pH 7.0
k
cat
(s
)1
) 75 ± 3 237 ± 11 164 ± 16 19 ± 2 27 ± 3 181 ± 19
K
mpyruvate
(mM) 0.5 ± 0.05 5.0 ± 0.2 7.9 ± 0.3 2.1 ± 0.2 2.4 ± 0.3 31 ± 4.2
Relation forward ⁄ reverse reaction
k
catD
⁄ k
catC

4.3 1.1 0.9 8.3 1.7 0.5
Structural properties
M (kDa)
a
65 65 65 65 65 65
N (kDa)
b
243 249 228 255 234 241
a
Determined by SDS-PAGE.
b
Determined by gel filtration chromatography.
c
At pH 8.0, as inhibition by high malate concentration was
observed at values lower than pH 8.0.
Regulation of A. thaliana NADP-malic enzyme activity M. C. Gerrard Wheeler et al.
5668 FEBS Journal 276 (2009) 5665–5677 ª 2009 The Authors Journal compilation ª 2009 FEBS
which were consistent with tetrameric oligomeric
states (Table 1).
Kinetic characterization of chimeric NADP-MEs in
the oxidative decarboxylation direction
Previous results have indicated that NADP-ME2 dis-
plays higher decarboxylation activity at a lower pH
(optimum pH 6.8) than NADP-ME3 (optimum pH
7.7) [5]. Nevertheless, as at pH 7.5 both isoenzymes
retained 95% of the maximal activity (data not
shown), this pH was chosen for the comparative char-
acterization of the kinetic and regulatory properties of
the chimeric proteins (Table 1). As ME3.2¢ was inhib-
ited by high malate concentration at pH values lower

than pH 8.0 (data not shown), the kinetic analysis of
this chimera was performed at pH 8.0 (Table 1).
All chimeras showed less specific activity than the
parental isoenzymes (Table 1). The k
cat
values obtained
were between 48% and 14% of the value obtained for
NADP-ME2, the parental enzyme with the highest
specific activity (Table 1). The K
mNADP
and K
ml-malate
values of the chimeric proteins were of the same order
of magnitude as those of NADP-ME2 (Table 1). These
results indicate that, despite the differences detected,
the binding sites for the substrates were integral in the
chimeric proteins.
Regulatory properties of the chimeric NADP-MEs
in the oxidative decarboxylation direction
Several compounds were tested as possible effectors of
the enzymatic activity of each chimeric NADP-ME in
the direction of the oxidative decarboxylation of
l-malate (Fig. 4), and compared with the results
obtained with the parental isoenzymes [5]. With the
exception of acetyl-CoA and CoA, the effectors were
tested at two concentrations, 0.5 and 2 mm, which are
referred to as low and high concentrations, respectively.
As in the case of NADP-ME2 and NADP-ME3,
oxaloacetate and ATP were the strongest inhibitors of
the enzymatic activity of all chimeric proteins (Fig. 4).

ME2.3¢ and ME3.2 were inhibited only by high ATP
concentration, whereas all other chimeras were inhib-
ited by both high and low ATP concentrations, proba-
bly because of the higher affinity for this inhibitor.
Glucose-6-phosphate also inhibited the enzymatic
activity of all proteins (Fig. 4), although oxaloacetate
and ATP were the strongest inhibitors.
The effects of acetyl-CoA and CoA on the enzy-
matic activity were tested at 20 lm (Fig. 4). Acetyl-
CoA did not modify significantly the activity of the
chimeras. However, CoA maintained its status as an
activator in the chimeric proteins ME2.3¢ and ME3.2,
whereas no modification of ME2.3 and ME3.2¢ activi-
ties were observed in the presence of this compound
(Fig. 4).
Furthermore, both succinate and aspartate were able
to activate all chimeric enzymes, although activation
by succinate was observed only at high concentration
in the case of ME2.3 and ME3.2¢ (Fig. 4). By contrast,
fumarate activated only ME3.2 and inhibited ME3.2¢,
but only at high concentration in the latter case
(Fig. 4).
Kinetic characterization of chimeric NADP-MEs in
the oxidative decarboxylation direction at pH 7.0
The effect of the substrate l-malate on the forward
reaction of NADP-ME2 and NADP-ME3 was analy-
sed at pH 7.0. The kinetic measurements showed that
NADP-ME3 was partially inhibited by high concentra-
tions of l-malate at this pH (Fig. 5). The kinetic data
obtained for NADP-ME3 at pH 7.0 fitted to an equa-

tion in which two different sites for malate, one cata-
lytic and one allosteric, are considered (see Materials
and methods [6]). When occupied, the allosteric site
decreases the activity of the enzyme, rendering a par-
tial inhibition that is characterized by a K
r
value of 0.6
and u value of 0.1 (Table 1). The inhibition of NADP-
ME3 by high malate concentration was pH dependent,
as no inhibition was observed at pH 7.5 (Table 1 and
[3]). However, NADP-ME2 was not inhibited by high
malate concentration at pH 7.0 (Fig. 5).
In order to identify sequence segments in the pri-
mary structure of NADP-ME2 and NADP-ME3
responsible for the differential behaviour at pH 7.0,
the chimeric proteins were analysed at this pH. The
results indicated that only ME2.3 and ME3.2¢ were
inhibited by high malate concentration. The data
obtained for these enzymes at pH 7.0 fitted well to the
equation that considers that the enzyme binds malate
at two different sites, one catalytic and the other allo-
steric, as in the case of NADP-ME3 (data not shown).
The K
r
values obtained (Table 1) indicate that the
malate allosteric site of ME2.3 and ME3.2¢ displays
lower affinity than that of the parental enzyme
NADP-ME3. In turn, the higher u parameters for the
chimeras indicate a smaller decrease in the catalytic
activity when the allosteric site is occupied by the

inhibitor, in comparison with the parental enzyme
NADP-ME3 (Table 1). In the case of ME3.2 ¢, the inhi-
bition by high malate concentration was also observed
at pH 7.5 (not shown), but not at pH 8.0, which was
used for the kinetic characterization of the chimera
(Table 1).
M. C. Gerrard Wheeler et al. Regulation of A. thaliana NADP-malic enzyme activity
FEBS Journal 276 (2009) 5665–5677 ª 2009 The Authors Journal compilation ª 2009 FEBS 5669
Reversibility of the reaction catalysed by chimeric
NADP-MEs
The four chimeric NADP-MEs were tested for their
capability to catalyse the reverse reaction: the pyru-
vate reductive carboxylation. As the carboxylation
reaction catalysed by the parental isoenzymes showed
an optimum at pH 7.0 [5], this pH value was used
for kinetic analysis of the chimeras. All chimeric pro-
teins showed less specific activity than NADP-ME3
(Table 1). ME2.3 and ME3.2¢ are the chimeras with
the highest k
cat
values for the reverse reaction, with
Fig. 4. Regulatory properties of the chimeric NADP-ME isoenzymes in the oxidative decarboxylation direction. NADP-ME forward activity
was measured for each isoenzyme at pH 7.5 in the absence or presence of 0.5 or 2 m
M of each effector [indicated as Succinate 0.5 or 2;
Fumarate 0.5 or 2; Asp (aspartate) 0.5 or 2; OAA (oxaloacetate) 0.5 or 2; ATP 0.5 or 2; Glucose 6P (glucose-6-phosphate) 0.5 or 2] or 20 l
M
of CoA or acetyl-CoA. The results are presented as the percentage of activity in the presence of the effectors in relation to the activity mea-
sured in the absence of the metabolites for each of the respective enzyme constructs. The assays were performed at least in triplicate and
the error bars indicate SD. Significant inhibition (as indicated in Materials and methods): dark grey and single-hatched bars. Significant activa-
tion (as indicated in Materials and methods): light grey and double-hatched bars. The results for parental NADP-ME2 and NADP-ME3,

obtained previously [5], are included for comparison.
Regulation of A. thaliana NADP-malic enzyme activity M. C. Gerrard Wheeler et al.
5670 FEBS Journal 276 (2009) 5665–5677 ª 2009 The Authors Journal compilation ª 2009 FEBS
values even higher than that of NADP-ME2
(Table 1). These proteins were able to catalyse the
reductive carboxylation reaction at higher rates than
the oxidative decarboxylation reaction (Table 1).
However, ME2.3 and ME3.2¢ display the lowest affin-
ity towards pyruvate in comparison with all the other
isoenzymes (Table 1).
Regulatory properties of the chimeric NADP-MEs
in the reductive carboxylation direction
The effect of several metabolites on the reductive
carboxylation reaction of the chimeric proteins was
analysed and compared with the results obtained with
the parental enzymes (Fig. 6).
l-Malate, one of the products of the reverse reac-
tion, was the strongest inhibitor of the enzymatic activ-
ity of all the chimeric versions (Fig. 6). Aspartate also
inhibited the reductive carboxylation of all chimeras
and NADP-ME2, but did not modify the enzymatic
activity of NADP-ME3 (Fig. 6).
With regard to succinate, this organic acid activated
the chimeric enzymes possessing the amino-terminal
region of NADP-ME2, ME2.3 and ME2.3¢ (Fig. 6).
By contrast, succinate did not modify the activity of
ME3.2 and ME3.2¢ (Fig. 6). In the case of fumarate,
all chimeras showed activation by this compound
(Fig. 6), as did NADP-ME2, whereas the parental iso-
enzyme NADP-ME3 was inhibited by both succinate

and fumarate [5].
Discussion
A. thaliana NADP-ME2 and NADP-ME3 are shown
to be located in the cytosol. The measurement of
enzymatic activity in the presence of several putative
metabolic effectors indicated distinct regulatory pat-
terns for both isoenzymes (Figs 4–6). In order to iden-
tify the key sequence regions associated with the more
relevant kinetic differences between these highly similar
isoenzymes, several chimeras of these proteins were
constructed and analysed. All the chimeric proteins
showed structural integrity by CD analysis and conser-
vation of the quaternary conformation (Table 1).
Thus, the absence of severe structural changes with
respect to the parental enzymes, and the fact that the
chimeras were functional proteins (Table 1), allowed
us to use them as a tool to compare regulatory pat-
terns and to evaluate the determinants of the primary
sequence associated with them.
Regulatory regions associated with fumarate
and CoA activation of the malate oxidative
decarboxylation reaction of NADP-ME2
Several compounds were tested as possible modifiers of
the malate oxidative decarboxylation reaction cataly-
sed by NADP-ME2 and NADP-ME3. Oxaloacetate,
ATP, glucose-6-phosphate and acetyl-CoA similarly
affected the activity of both native enzymes and chime-
ras (Fig. 4 [5]). However, succinate, fumarate, aspar-
tate and CoA produced differential effects on the
activity of NADP-ME2 and NADP-ME3, as well as

the different chimeras analysed (Fig. 4).
Fumarate activation was observed for the parental
NADP-ME2 and the chimeric enzyme ME3.2 (Fig. 4).
These proteins share a common region, which extends
from amino acid residue 303 to the C-terminal end of
NADP-ME2, suggesting that this sequence is associ-
ated with the activation mechanism by this compound
(Figs 1 and 3). Moreover, amino acid residues from
both segments swapped (from amino acid residue 303
Fig. 5. NADP-ME2 and NADP-ME3 forward activity as a function of malate concentration at pH 7.0. Free Mg
2+
and NADP concentrations
were kept constant at 10 and 1.0 m
M, respectively, in all cases. A typical result is shown from at least three independent determinations.
The data were fitted to the Michaelis–Menten equation for NADP-ME2 or to the model described in Materials and methods for NADP-ME3
(see equation in [6]), and are presented as the percentage of maximum activity. The absolute values corresponding to 100% of activity are
497 and 218 UÆmg
)1
for NADP-ME2 and NADP-ME3, respectively. Malate inhibition was not observed for either isoenzyme at pH 7.5
(Table 1 and [3]).
M. C. Gerrard Wheeler et al. Regulation of A. thaliana NADP-malic enzyme activity
FEBS Journal 276 (2009) 5665–5677 ª 2009 The Authors Journal compilation ª 2009 FEBS 5671
to 500 and from amino acid residue 500 to the car-
boxyl-terminal end of NADP-ME2; Figs 1 and 3) are
involved in this regulation, as the chimeras ME2.3¢
and ME 3.2¢, both bearing only one of these segments,
are not activated by fumarate (Fig. 3).
Like NADP-ME2, human mitochondrial NAD(P)-
ME is allosterically activated by fumarate [7]. In this
case, fumarate binds at the dimer interface, where

four amino acid residues are involved: R67, R91,
E59 and D102 [7–9]. Only two of these amino acid
Fig. 6. Regulatory properties of the chimeric NADP-ME isoenzymes in the reductive carboxylation direction. NADP-ME reverse activity was
measured for each isoenzyme at pH 7.0 in the absence or presence of 1, 7.5 or 15 m
M of each effector [indicated as L-malate, Succinate,
Fumarate and Asp (aspartate) 1, 7.5 and 15 m
M]. The results are presented as the percentage of activity in the presence of the effectors in
relation to the activity measured in the absence of the metabolites, for each of the respective enzyme constructs. The assays were
performed at least in triplicate, and error bars indicate SD. Significant inhibition (as indicated in Materials and methods): dark grey and single-
hatched bars. Significant activation (as indicated in Materials and methods): light grey and double-hatched bars. The results for parental
NADP-ME2 and NADP-ME3, obtained previously [5], are included for comparison.
Regulation of A. thaliana NADP-malic enzyme activity M. C. Gerrard Wheeler et al.
5672 FEBS Journal 276 (2009) 5665–5677 ª 2009 The Authors Journal compilation ª 2009 FEBS
residues (R91 and D102; homologous to R115 and
D126 of NADP-ME2, respectively) are conserved in
A. thaliana NADP-ME2 (Fig. 1 and [5]), suggesting
that the mechanism of activation should be different
between the two isoenzymes. Moreover, these two
amino acid residues are also conserved in NADP-
ME3 (Fig. 1), which is not activated by fumarate
(Fig. 4). Therefore, other amino acid residues differ-
ent from those proposed for the human isoenzyme
[7–9] are necessary to control the binding capacity
and fumarate response of NADP-ME2. Several
amino acid residues in this C-terminal region (Figs 1
and 3) are good candidates to be involved in this
activation, and their role remains to be determined
by mutational studies. Specifically, from the five non-
conserved amino acid residues in the suggested
domain involved in fumarate activation, the muta-

tions at positions 357 and⁄ or 360 could be involved
in the differential regulation, as well as the conserved
change at position 543 (Fig. 1).
The activation of NADP-ME2 by fumarate could be
relevant in vivo,asA. thaliana accumulates large
amounts of fumarate and malate during the day and
uses these organic acids as a way to transport carbon
to other organs, and as energy and carbon sources in
conditions of energy demand [10,11]. In this sense, the
activation by fumarate of NADP-ME2, which is
expressed in photosynthetic and nonphotosynthetic
organs of A. thaliana, may be linked to the higher
utilization of organic acids on energy demand by the
activation of this isoenzyme when the fumarate con-
centration increases. However, our data suggest that
NADP-ME3, which is restricted to pollen and tri-
chomes, is not linked to this organic acid utilization
and regulation.
By contrast, the region between amino acid residues
303 and 500 of NADP-ME2 may be associated with
CoA activation of the l-malate decarboxylation reac-
tion because only NADP-ME2, ME2.3¢and ME3.2
showed activation by this compound (Figs 1 and 3).
Previous studies have indicated that the deletion of 44
amino acid residues from the amino-terminal region of
NADP-ME2 provides an enzyme that is not activated
at all by CoA [5]. Thus, the activation by this metabo-
lite may require the participation of the region flanked
by amino acid residues 303 and 500 of NADP-ME2
interacting with residues from the amino-terminal

region.
Although NADP-ME3 is not activated by aspar-
tate and succinate (Fig. 4), it is surprising that the
activity of the four chimeras increases in the pres-
ence of both effectors, although to a different extent
(Fig. 4). In this way, it can be inferred that the acti-
vation by these metabolites is mediated by several
amino acid residues, not found in NADP-ME3, but
distributed in the different protein segments of
NADP-ME2 that were swapped by the construction
of the chimeras.
Regulatory region associated with the
pH-dependent malate inhibition of NADP-ME3
Malate inhibition of the forward reaction at pH 7.0
was observed only in the case of NADP-ME3, ME2.3
and ME3.2¢ (Fig. 3, Table 1). These proteins share a
common region, which extends from amino acid resi-
dues 303 to 500 of NADP-ME3, suggesting that this
sequence is associated with the mechanism of substrate
inhibition (Figs 1 and 3). The fact that the inhibition
by high substrate concentration was associated with a
limited region of the protein supports the hypothesis
of the existence of an allosteric site responsible for
such regulation, as in the case of maize photosynthetic
NADP-ME [6]. In agreement with this, the kinetic
data obtained for the enzymes that were inhibited by
l-malate fitted very well to the equation that considers
an allosteric inhibitor binding site for malate (Fig. 4
[6]). Moreover, as the inhibition by malate is pH
dependent (Table 1), it is concluded that the amino

acid residue(s) involved in this allosteric regulation
may change the protonation state between pH 7.0 and
pH 7.5, leading to the loss of inhibition at higher pH.
In the particular case of ME3.2¢, the loss of malate
inhibition is observed at higher pH (pH 8.0, Table 1).
It is thus possible that in this chimera a change in the
pKa value of the amino acid residue(s) involved in
malate inhibition may occur as a result of interaction
with different amino acid residues in the allosteric site.
The amino acid residue changes at positions 357, 420
and ⁄ or 481 between NADP-ME2 and NADP-ME3 are
good candidates for involvement in the pH-dependent
regulation by malate, as they involve changes from
noncharged amino acid residues in NADP-ME2 to
positive amino acid residues, depending on pH, in
NADP-ME3 (Fig. 1).
The inhibition by excess l-malate was marked as a
pH-dependent characteristic of MEs implicated in C
4
photosynthesis [6]. This in vivo regulatory mechanism
was suggested to occur through the pH change induced
in illuminated chloroplasts, and ensures that NADP-
ME is fully active only at pH 8.0, when carbon fixa-
tion is in progress. Thus, the pH-dependent malate
inhibition of NADP-ME3 was unexpected, as it is a
cytosolic isoenzyme not implicated in photosynthesis.
In this way, this isoenzyme from a C
3
species displays
the feature of malate inhibition at pH 7.0 associated

M. C. Gerrard Wheeler et al. Regulation of A. thaliana NADP-malic enzyme activity
FEBS Journal 276 (2009) 5665–5677 ª 2009 The Authors Journal compilation ª 2009 FEBS 5673
with C
4
photosynthesis, probably as an evolutionary
ancestor of C
4
NADP-ME. Similarly, a nonphoto-
synthetic recombinant NADP-ME from tobacco also
showed partial inhibition by l-malate [12]. Further
studies should be conducted to reveal whether the
pH-dependent regulation of a nonphotosynthetic
isoenzyme may be relevant in vivo, especially with
regard to the localization of NADP-ME3 in the cyto-
sol of pollen and trichome cells.
Regulatory region associated with succinate
activation of the pyruvate reductive
carboxylation reaction of NADP-ME2
The activation of the pyruvate reductive carboxylation
reaction by succinate was only observed in the case of
NADP-ME2 and the chimeras ME2.3 and ME2.3¢
(Fig. 6). Thus, a regulatory region associated with acti-
vation by this metabolite could be defined, which com-
prises the first 303 amino-terminal amino acid residues
of NADP-ME2 (Figs 1 and 3). However, the three
nonconserved amino acid changes between NADP-
ME2 and NADP-ME3 are located in the amino-termi-
nal region (Fig. 1), which is not involved in succinate
activation, as NADP-ME2 lacking the first 44 amino
acid residues is still activated by succinate [5]. Thus,

some of the semiconserved or conserved amino acid
residue changes may be involved in succinate activa-
tion. Good candidates are the changes in charge at
positions 253 and ⁄ or 295 (Fig. 1).
By contrast, fumarate was able to activate the
reverse reaction catalysed by NADP-ME2 (Fig. 6 [5]).
However, although NADP-ME3 is not activated by
this compound at all, it is surprising that the activity
of the four chimeras is increased by fumarate (Fig. 6).
In this regard, ME3.2¢, the chimera that shares the
minimum number of amino acid residues with NADP-
ME2, is less activated by this compound than the
other compounds (Fig. 6). In this way, amino acid res-
idues of different segments of NADP-ME2 swapped in
the construction of the chimeras are involved in the
fumarate activation of the pyruvate reductive carboxyl-
ation reaction. However, the different degree of fuma-
rate activation shown by the chimeras (e.g. 150% in
the case of ME3.2 and 671% in the case of ME2.3¢
with 7.5 mm of fumarate) may indicate that some
regions are more critical than others in the activation
by this allosteric modulator. This hypothesis should be
tested by site-directed mutagenesis of candidate amino
acid residues from the different regions and by estima-
tion of the kinetic parameters of fumarate activation.
Aspartate inhibits NADP-ME2 and all the chimeras
analysed, but is unable to decrease NADP-ME3 activ-
ity, although tested at high concentration (Fig. 6).
These results are consistent with an allosteric type of
inhibition, in which amino acid residues from the dif-

ferent segments used to construct the chimeras are
involved. However, as some chimeras are inhibited
only at high aspartate concentrations (e.g. ME3.2¢ is
not inhibited at all at 1 mm aspartate), some segments
of the primary structure seem to be more critical than
others (Fig. 6). This hypothesis can be tested by site-
directed mutagenesis of candidate amino acid residues
from the different regions, and by estimation of the
affinity towards aspartate as inhibitor.
The forward and reverse reactions are distinctly
regulated by effectors which are associated with
different protein determinants
Finally, the results obtained suggest that the regulation
of the forward and reverse NADP-ME activities is
mediated by different protein regions (Figs 4 and 6).
In this sense, the dual effect of aspartate in the
NADP-ME2 reaction, which activates the decarboxyl-
ation and inhibits the carboxylation reaction, can be
explained through a conformational change in the
enzyme induced by the substrate malate, which can be
important for exposing an aspartate-activating binding
site [5].
By contrast, succinate and fumarate strongly
increased the activity of NADP-ME2 in both direc-
tions of the reaction (Figs 4 and 6). However, despite
the structural similarity between these two organic
acids, the kinetic results indicated that the activation
by these compounds was mediated by different bind-
ing sites. This conclusion is supported by previous
studies, which showed that R115 mutation of NADP-

ME2 abolished the activating effect of fumarate, but
did not modify the activity measured in the presence
of succinate [5]. In turn, our experimental data clearly
indicate that, for each separate metabolite, the regula-
tion of the direct reaction is mediated by different
sites than in the reverse reaction (Fig. 3). Again,
conformational changes induced in the protein by
l-malate or pyruvate could provide an explanation
for these observations. A future challenge will be to
determine the three-dimensional structure of a plant
NADP-ME in the presence and absence of the sub-
strates to analyse the conformational changes that are
induced by the binding of the substrate, which may
influence the observed allosteric regulation of each
isoenzyme. As suggested recently, a knowledge of the
allosteric interactions could be very useful in protein
design inhibition or activation to influence protein
function as required [13].
Regulation of A. thaliana NADP-malic enzyme activity M. C. Gerrard Wheeler et al.
5674 FEBS Journal 276 (2009) 5665–5677 ª 2009 The Authors Journal compilation ª 2009 FEBS
Concluding remarks
Although A. thaliana NADP-ME2 and NADP-ME3
share a high identity, they display distinct regulatory
properties. Thus, the activities of these two isoen-
zymes, which colocalize in the cytosol, are modulated
by different effectors. In this work, the regions
involved in some of these differential regulations were
mapped in the primary structure of each isoenzyme.
Succinate, fumarate, malate and aspartate are all
involved in the modulation of NADP-ME activity.

Nevertheless, although these compounds are structur-
ally similar, NADP-ME2 and NADP-ME3 are able to
distinguish and interact differently with these C
4
acids
as a result of minimal differences in the primary pro-
tein structure (Fig. 1). The precise identification of
these differences will be important for the enzymatic
engineering of NADP-ME aimed at creating enzymes
able to fulfil desired reactions responding to specific
effectors.
Materials and methods
Heterologous expression and purification of the
recombinant enzymes
The pET32 vectors containing the full-length cDNA
sequences of NADP-ME2 and NADP-ME3, pET-ME2 and
pET-ME3 [3], were used to express each NADP-ME fused
in-frame to a histidine tag to facilitate the purification of
the expressed fusion protein by a nickel-containing His-
Bind column (Novagen, Gibbstown, New York, USA). The
induction and isolation of the proteins were performed as
described previously [14,15]. The fusion proteins were
digested with enterokinase as described in [3], and the pro-
teins were further purified using an affinity Affi Gel Blue
column (BioRad, Hercules, CA, USA) and analysed by
SDS-PAGE. The purified enzymes were concentrated on
Centricon YM-50 (Millipore, Billerica, MA, USA) and
stored at )80 °Cin50mm Tris ⁄ HCl pH 8.0, 10 mm MgCl
2
and 50% (v ⁄ v) glycerol for further studies.

Construction of chimeric NADP-ME sequences
NADP-ME chimeric proteins were generated by inter-
changing cDNA segments obtained using two conserved
restriction sites: EcoRV at position 910 and BclI at posi-
tion 1500 of the cDNA of NADP-ME2 and NADP-ME3
(Fig. 3). For this purpose, the plasmids pET-ME2 and
pET-ME3 [3] were treated with the corresponding restric-
tion endonucleases, and the fragments obtained were puri-
fied and recombined to obtain chimeric sequences in the
expression vector pET32. The plasmids constructed were
named as follows: pET-ME2.3, pET-ME2.3¢, pET-ME3.2
and pET-ME3.2¢ (Fig. 3). The chimeric constructs were
sequenced to verify the correct swapping of the fragments
and to ensure that no mistakes were introduced during
the subcloning procedures. The chimeric proteins were
expressed in BL21(DE3) cells and purified as described for
the parental enzymes.
CD measurements
CD spectra were obtained on a Jasco J-810 spectropolarim-
eter (Jasco, Easton, MD, USA) using a cell with a path
length of 0.1 cm and averaging 10 repetitive scans between
250 and 200 nm. Typically, 30 l g of each protein in 20 mm
NaCl ⁄ P
i
pH 7.5 were used for the assay. The mean amino
acid residue ellipticity was calculated as described in [14].
NADP-ME activity assays and protein
concentration measurement
The oxidative decarboxylation of l-malate (forward reac-
tion) was assayed spectrophotometrically using a standard

reaction mixture containing 50 mm Tris ⁄ HCl pH 7.5,
10 mm MgCl
2
,1mm NADP and 30 mml-malate in a
final volume of 0.5 mL. The reductive carboxylation of
pyruvate (reverse reaction) was measured in an assay
medium containing 50 mm Mops ⁄ KOH pH 7.0, 10 mm
MgCl
2
, 0.2 mm NADPH, 10 mm NaHCO
3
and 50 mm
pyruvate in a final volume of 0.5 mL. The linearity of
the reaction was monitored to detect any CO
2
loss during
the assay. One unit (U) is defined as the amount of
enzyme that catalyses the formation or consumption of
1 lmol of NAD(P)H per minute under the specified con-
ditions (e
340 nm
= 6.22 mm
)1
Æcm
)1
).
Initial velocity studies were performed by varying the
concentration of one of the substrates around its K
m
value, whilst keeping the other substrate concentrations

at saturating levels. All kinetic parameters were calculated
at least by triplicate determinations, and adjusted to
nonlinear regression using free concentrations of all
substrates [14].
When testing different compounds as possible inhibitors
or activators, NADP-ME forward activity was measured at
pH 7.5 in the absence or presence of 0.5 and 2 mm of succi-
nate, fumarate, aspartate, oxaloacetate, ATP or glucose-6-
phosphate or 20 lm of CoA or acetyl-CoA, whilst keeping
NADP and l-malate concentrations at the K
m
values for
each isoenzyme (Table 1). In the particular case of ME3.2¢,
a subsaturating and noninhibitory concentration of
l-malate (2.0 mm) was used. NADP-ME reverse activity
was measured at pH 7.0 in the absence or presence of 1.0,
7.5 or 15 mml-malate, succinate, fumarate or aspartate,
keeping the pyruvate concentration at the K
m
value for
each isoenzyme (Table 1). The results are presented as
the percentage of activity in the presence of the effector in
relation to the activity measured in the absence of the
M. C. Gerrard Wheeler et al. Regulation of A. thaliana NADP-malic enzyme activity
FEBS Journal 276 (2009) 5665–5677 ª 2009 The Authors Journal compilation ª 2009 FEBS 5675
metabolite. Significant activation or inhibition was deter-
mined after analysis of the activity measurements using
one-way analysis of variance (ANOVA). Minimum
significant differences were calculated by the Bonferroni,
Holm–Sidak, Dunett and Duncan tests (a = 0.05) using

the Sigma Stat Package. Only when the activity measure-
ments in the presence of a given metabolite were
significantly different from the activity measured in its
absence (P < 0.05) was activation or inhibition indicated
as significant in Figs 4 and 6. Moreover, inhibitions lower
than 80% and activations higher than 140% relative to the
activity in the absence of the metabolites are indicated.
When analysing the inhibition of NADP-ME by high
malate concentration at pH 7.0, the fitted model was one in
which the free enzyme is capable of binding malate at two
different sites, one catalytic (with a dissociation constant
K
c
) and the other allosteric (with a dissociation constant
K
r
). Thus, two different catalytic constants are postulated,
namely k
1
and k
2
, to indicate product formation when the
allosteric site is empty or occupied, respectively, u being
the ratio between k
2
and k
1
[6]. When no inhibition by high
substrate concentration was found, the kinetic data
obtained were fitted to a typical Michaelis–Menten equa-

tion. All activity assays were carried out at 30 °Cina
Helios b spectrophotometer (Unicam, Cambridge, UK).
The protein concentration was determined by the method
in [16].
Gel electrophoresis
SDS-PAGE was performed in 10% (w ⁄ v) polyacrylamide
gels according to [17]. Native PAGE was performed using
6% (w ⁄ v) polyacrylamide gels. Electrophoresis was run at
150 V at 10 °C. Gels were assayed for NADP-ME activity
by incubation in a solution containing 50 mm Tris ⁄ HCl pH
7.5, 30 mml-malate, 10 mm MgCl
2
,1mm NADP,
35 lgÆmL
)1
nitroblue tetrazolium and 0.85 lgÆmL
)1
phen-
azine methosulfate at 30 ° C.
Gel filtration chromatography
Molecular masses of recombinant native and chimeric pro-
teins of NADP-ME2 and NADP-ME3 were evaluated by
gel filtration chromatography on an A
¨
KTA purifier system
(GE Healthcare, UK) using a Tricorn Superdex 200 10 ⁄ 300
GL column (GE Healthcare, London, UK). The column
was equilibrated with 20 mm Tris ⁄ HCl at pH 8.0 in the
presence or absence of 100 mm NaCl, and calibrated using
the following molecular mass standards: carbonic anhy-

drase, 29 kDa; bovine serum albumin, 66 kDa; alcohol
dehydrogenase, 150 kDa; b-amylase, 200 kDa; apoferritin,
443 kDa; thyroglobulin, 669 kDa (Sigma-Aldrich, St Louis,
MO, USA). The sample and standards were applied sepa-
rately in a final volume of 100 lL at a constant flow rate of
0.5 mLÆmin
)1
.
Construction of NADP-ME::GFP fusions
The 35S::NADP-ME2::GFP and 35S::NADP-ME3::GFP
fusions were constructed using the AtNADP-ME2 and At-
NADP-ME3 full-length coding sequences fused to the GFP
coding sequence and flanked by the cauliflower mosaic virus
(CaMV) 35S promoter and terminator sequences. Full-length
NADP-ME2 and NADP-ME3 cDNAs were amplified
from A. thaliana flowers using the pair of primers ME2GFP-
F(5¢-CACCATGGGAAGTACTCCGACTGAT-3¢) and
ME2GFP-R (5¢-AGCCCTGTGTACAGAAACTACCGT-3¢)
and ME3GFP-F (5¢-CACCATGGGCACCAATCAGACT
CAG-3¢) and ME3GFP-R (5¢-AGTCCTGTCTACAGAA
ACTTCCGT-3¢), respectively. The fragments obtained were
cloned into pENTR ⁄ D-TOPO (Invitrogen, Carlsbad, CA,
USA). The GFP fusion constructs were made by subclon-
ing the coding sequence into pGWB5 (for C-terminal
GFP), a gateway-compatible binary vector designed for
35S promoter-driven expression of GFP fusion proteins
(kindly provided by T. Nakagawa, Shimane University,
Izumo, Japan). Cloning using gateway vector was per-
formed using the reagents and protocols from Invitrogen.
Transformation of A. thaliana cells and

microscopic analysis
Transient transformation of A. thaliana cells was conducted
as described by Koroleva et al. [18] using the hypervirulent
SV0 Agrobacterium strain (LBA4404.pBBR1MCS
virGN54D) carrying the NADP-ME2::GFP or NADP-
ME3::GFP plasmid. Detection of GFP was achieved using
a GFP filter (excitation, 460–500 nm; emission, 510–
560 nm) and a Nikon eclipse E800 microscope coupled to a
digital camera (ky-F1030, JVC, Yokohama, Japan).
Acknowledgements
MFD, CSA and MGW are members of the Researcher
Career of CONICET and CLA is a fellow of the same
institution. This work was supported by ANPCyT
(PICT N° 32233), CONICET and the Deutsche Fors-
chungsgemeinschaft.
References
1 Chang GG & Tong L (2003) Structure and function of
malic enzymes, a new class of oxidative decarboxylases.
Biochemistry 42, 12721–12733.
2 Drincovich MF, Casati P & Andreo CS (2001) NADP-
malic enzyme from plants: a ubiquitous enzyme
involved in different metabolic pathways. FEBS Lett
490, 1–6.
3 Gerrard Wheeler MC, Tronconi MA, Drincovich MF,
Andreo CS, Flu
¨
gge UI & Maurino VG (2005) A
Regulation of A. thaliana NADP-malic enzyme activity M. C. Gerrard Wheeler et al.
5676 FEBS Journal 276 (2009) 5665–5677 ª 2009 The Authors Journal compilation ª 2009 FEBS
comprehensive analysis of the NADP-malic enzyme

gene family of Arabidopsis thaliana. Plant Physiol 139,
39–51.
4 Maurino VG, Gerrard Wheeler MC, Andreo CS &
Drincovich MF (2009) Redundancy is sometimes seen
only by the uncritical: does Arabidopsis need six malic
enzyme isoforms? Plant Sci 176, 715–721.
5 Gerrard Wheeler MC, Arias CL, Tronconi MA, Mauri-
no VG, Andreo CS & Drincovich MF (2008) Arabidop-
sis thaliana NADP-malic enzyme isoforms: high degree
of identity but clearly distinct properties. Plant Mol Biol
67, 231–242.
6 Detarsio E, A
´
lvarez CE, Saigo M, Andreo CS & Drin-
covich MF (2007) Identification of domains involved in
tetramerization and malate inhibition of maize C
4
NADP-malic enzyme. J Biol Chem 282, 6053–6060.
7 Yang Z, Lanks CW & Tong L (2002) Molecular mecha-
nism for the regulation of human mitochondrial
NAD(P)
+
-dependent malic enzyme by ATP and fuma-
rate. Structure 10, 951–960.
8 Hung HC, Kuo MW, Chang GG & Liu GY (2005)
Characterization of the functional role of allosteric site
residue Asp
102
in the regulatory mechanism of human
mitochondrial NAD(P)

+
-dependent malate dehydroge-
nase (malic enzyme). Biochem J 392, 39–45.
9 Hsieh JY, Chiang YH, Chang KY & Hung HC (2009)
Functional role of fumarate site Glu59 involved in allo-
steric regulation and subunit–subunit interaction of
human mitochondrial NAD(P)
+
-dependent malic
enzyme. FEBS J 276, 983–994.
10 Fahnenstich H, Saigo M, Niessen M, Drincovich MF,
Zanor MI, Fernie A, Andreo CS, Flu
¨
gge UI &
Maurino VG (2007) Low levels of malate and fumarate
cause accelerated senescence during extended darkness
in Arabidopsis thaliana overexpressing maize C
4
NADP-
malic enzyme. Plant Physiol 145, 640–655.
11 Chia DW, Yoder TJ, Reiter WD & Gibson SI (2000)
Fumaric acid: an overlooked form of fixed carbon in
Arabidopsis and other plant species. Planta 211, 743–
751.
12 Mu
¨
ller GL, Drincovich MF, Andreo CS & Lara MV
(2008) Nicotiana tabacum NADP-malic enzyme: cloning,
characterization and analysis of biological role. Plant
Cell Physiol 43, 469–480.

13 Laskowski RA, Gerick F & Thornton JM (2009) The
structural basis of allosteric regulation in proteins.
FEBS Lett 583 , 1692–1698.
14 Detarsio E, Gerrard Wheeler MC, Campos Bermu´ dez
VA, Andreo CS & Drincovich MF (2003) Maize C
4
NADP-malic enzyme. Expression in Escherichia coli
and characterization of site-directed mutants at the
putative nucleotide-binding sites. J Biol Chem 278,
13757–13764.
15 Saigo M, Bologna FP, Maurino VG, Detarsio E, And-
reo CS & Drincovich MF (2004) Maize recombinant
non-C
4
NADP-malic enzyme: a novel dimeric malic
enzyme with high specific activity. Plant Mol Biol 55,
99–107.
16 Sedmak JJ & Grossberg SE (1977) A rapid, sensitive,
and versatile assay for protein using Coomassie brilliant
blue G 250. Anal Biochem 79, 544–552.
17 Laemmli UK (1970) Cleavage of structural proteins
during the assembly of the head of bacteriophage T4.
Nature 227, 680–685.
18 Koroleva OA, Tomlinson ML, Leader D, Shaw P &
Doonan JH (2005) High-throughput protein localization
in Arabidopsis using Agrobacterium-mediated transient
expression of GFP-ORF fusions. Plant J 41, 162–174.
M. C. Gerrard Wheeler et al. Regulation of A. thaliana NADP-malic enzyme activity
FEBS Journal 276 (2009) 5665–5677 ª 2009 The Authors Journal compilation ª 2009 FEBS 5677