Evolutionary and mechanistic insights into substrate and product accommodation of CTP:phosphocholine cytidylyltransferase from Plasmodium falciparum

Article abstract of DOI:10.1111/febs.12282

The enzyme CTP:phosphocholine cytidylyltransferase (CCT) is essential in the lipid biosynthesis of Plasmodia (Haemosporida), presenting a promising antimalarial target. Here, we identified two independent gene duplication events of CCT within Apicomplexa and characterized a truncated construct of Plasmodium falciparum CCT that forms a dimer resembling the molecular architecture of CCT enzymes from other sources. Based on biophysical and enzyme kinetics methods, our data show that the CDP-choline product of the CCT enzymatic reaction binds to the enzyme considerably stronger than either substrate (CTP or choline phosphate). Interestingly, in the presence of Mg²⁺, considered to be a cofactor of the enzyme, the binding of the CTP substrate is attenuated by a factor of 5. The weaker binding of CTP:Mg²⁺, similarly to the related enzyme family of aminoacyl tRNA synthetases, suggests that, with lack of Mg²⁺, positively charged side chain(s) of CCT may contribute to CTP accommodation. Thermodynamic investigations by isothermal titration calorimetry and fluorescent spectroscopy studies indicate that accommodation of the choline phosphate moiety in the CCT active site is different when it appears on its own as one of the substrates or when it is linked to the CDP-choline product. A tryptophan residue within the active site is identified as a useful internal fluorescence sensor of enzyme-ligand binding. Results indicate that the catalytic mechanism of Plasmodium falciparum CCT may involve conformational changes affecting the choline subsite of the enzyme. Database Model data are available in the Protein Model DataBase (PMDB) under the accession number PM0078718 (PfCCT(528-795)) and PM0078719 (PfCCT MΔK) Structured digital abstract PfCCT MΔK and PfCCT MΔK bind by mass spectrometry studies of complexes (View interaction) PfCCT MΔK and PfCCT MΔK bind by comigration in gel electrophoresis (View interaction) PfCCT MΔK and PfCCT MΔK bind by molecular sieving (View interaction) CTP:phosphocholine cytidylyltransferase (CCT) is essential in lipid biosynthesis of Plasmodia, presenting a promising antimalarial target. We identified two independent gene duplication events of CCT within Apicomplexa and characterized a dimeric truncated construct of Plasmodium falciparum CCT. Catalytic mechanism of PfCCT may involve conformational changes affecting the choline subsite of the enzyme whereas Mg²⁺ cofactor is dispensable for CTP substrate binding.

Full text of DOI:10.1111/febs.12282

Evolutionary and mechanistic insights into substrate and product accommodation of
CTP:phosphocholine cytidylyltransferase from Plasmodium falciparum
1
*
1
2
2
3
Gergely N. Nagy , Lívia Marton , Balázs Krámos , Julianna Oláh , Ágnes Révész , Károly
3
4
5
5,6
Vékey , Frédéric Delsuc , Éva Hunyadi-Gulyás , Katalin F. Medzihradszky , Marina
7
7
7
1,8*
Lavigne , Henri Vial , Rachel Cerdan , Beáta G. Vértessy
1
Institute of Enzymology, Research Centre for Natural Sciences, Hungarian Academy of
Sciences, 29 Karolina street 1113 Budapest, Hungary.
2
Department of Inorganic and Analytical Chemistry, Budapest University of Technology and
Economics, 4 Szent Gellért tér, 1111 Budapest, Hungary.
3
Institute of Organic Chemistry, Research Centre for Natural Sciences, Hungarian Academy
of Sciences, Budapest, 1025, Pusztaszeri ut 59-67, Hungary
4
Institut des Sciences de l'Evolution, UMR5554-CNRS-IRD, Université Montpellier 2, Place
Eugène Bataillon 7, 34095 Montpellier, France.
5
Laboratory of Proteomics Research, Biological Research Centre of Hungarian Academy of
Sciences, Temesvári krt. 62., H-6726 Szeged, Hungary.
6
Mass Spectrometry Facility, Department of Pharmaceutical Chemistry, University of
California San Francisco, San Francisco, CA 94158-2517, USA.
7
Laboratory Dynamique des Interactions Membranaires Normales et Pathologiques,
Université Montpellier 2, Centre National de la Recherche Scientifique, UMR 5235, Place
Eugène Batallion 7, 34095 Montpellier, France.
8
Department of Applied Biotechnology and Food Science, Budapest University of
Technology and Economics, 4 Szent Gellért tér, 1111 Budapest, Hungary.
Article Type: Original Article
This article has been accepted for publication and undergone full peer review but has not
been through the copyediting, typesetting, pagination and proofreading process, which may
lead to differences between this version and the Version of Record. Please cite this article as
doi: 10.1111/febs.12282
This article is protected by copyright. All rights reserved.

*
To whom correspondence should be addressed (gnagy@enzim.hu, vertessy@enzim.hu).
Abstract
The enzyme CTP:phosphocholine cytidylyltransferase (CCT) is essential in lipid biosynthesis
of Plasmodia (Haemosporida), presenting a promising antimalarial target. Here, we identified
two independent gene duplication events of CCT within Apicomplexa and characterized a
truncated construct of Plasmodium falciparum CCT that forms a dimer resembling the
molecular architecture of CCT enzymes from other sources. Based on biophysical and
enzyme kinetics methods, our data show that the CDP-choline product of the CCT enzymatic
reaction binds to the enzyme considerably stronger as compared to either substrate (CTP or
2
+
choline phosphate). Interestingly, in the presence of the Mg , considered to be a cofactor of
the enzyme, the binding of the CTP substrate is attenuated by a factor of 5. The weaker
2
+
binding of CTP:Mg , similarly to the related enzyme family of aminoacyl tRNA synthetases,
2
+
suggests that in lack of Mg , positively charged side chain(s) of CCT may contribute to CTP
accommodation. Thermodynamic investigations by isothermal titration calorimetry and
fluorescent spectroscopy studies indicate that accommodation of the choline phosphate
moiety in the CCT active site is different when it appears on its own as one of the substrates
or when it is linked to the CDP-choline product. A tryptophan residue within the active site is
identified as a useful internal fluorescence sensor of enzyme-ligand binding. Results argue
that the catalytic mechanism of PfCCT may involve conformational changes affecting the
choline subsite of the enzyme.
Key words
CTP:phosphocholine cytidylyltransferase, malaria, lipid biosynthesis, thermodynamics of
ligand binding, gene duplication
Abbreviations
PfCCT, Plasmodium falciparum CTP:phosphocholine cytidylyltransferase
Δ720-737
MΔK a truncated construct of PfCCT (PfCCT (528-795)
95 with the lack of the lysine-rich segment 720-737
encompassing residues 528-
7
CTP cytidine triphosphate
ChoP choline-phosphate
This article is protected by copyright. All rights reserved.

CDPCho CDP-choline
PP pyrophosphate,
i
K
M
Michaelis constant
cat turnover number
inhibition constant
k
K
i
PC phosphatidylcholine
ITC isothermal titration calorimetry
ML maximum likelihood
MESG 2-amino-6-mercapto-7-methylpurine ribonucleoside
Database: model data are available in the Protein Model DataBase (PMDB) under the
accession number PM0078718 (PfCCT (528-795) and PM0078719 (PfCCT MΔK).
Introduction
Malaria continues to be a major world health problem affecting hundred million people
annually, mostly in African sub-Saharan countries. The disease is caused by the infection and
destruction of red blood cells by protozoan parasites belonging to the genus Plasmodium,
Plasmodium falciparum and Plasmodium vivax being the most widespread [1, 2]. There is
thus an urgent need to identify new drug targets and develop new pharmacophores with
unique structures and mode of action [3, 4].
Successful growth and multiplication of the parasite requires temporally controlled metabolic
programs leading to the duplication of the structural components. The main Plasmodium
membrane constituents are glycerophospholipids, produced by de novo pathways using
precursors that are actively transported from the host cytoplasm [5, 6]. Phosphatidylcholine
(
PC) is the major phospholipid in P. falciparum membranes [7, 8] and its synthesis relies on
numerous pathways, seldom found together in a single organism [5, 6]. Choline-mimicking
compounds arrest parasite growth with potent in vivo antimalarial properties [9]. The clinical
candidate albitiazolium [10] that specifically inhibits de novo phosphatidylcholine
biosynthesis is structurally unrelated to existing antimalarial agents and is currently being
developed in human phase 2 clinical trials for parenteral cures of severe malaria [11].
This article is protected by copyright. All rights reserved.

De novo PC biosynthesis through the CDP-choline (CDPCho) dependent Kennedy pathway
is of crucial importance in P. berghei [12] and P. falciparum. Within this pathway, the
CTP:phosphocholine cytidylyltransferase (2.7.7.15) enzyme [13-16] that catalyses the
second step of the CDPCho pathway by condensing CTP to phosphocholine (ChoP) to obtain
the energized CDPCho is rate limiting and thus can be considered as a potential molecular
drug target.
The CCT genes are annotated in different Plasmodium genomes as unique copies.
Surprisingly, the plasmodial CCT genes encode two putatively duplicated CCT segments
(
(
(
abbreviated as CMP segments), each including a cytidylyltransferase catalytic domain
C1/C2), a membrane/lipid binding domain (M1/M2) and a putative phosphorylation domain
P1/P2). Between the two CMP segments, there is a long linker (240-278 aa), which is
partially conserved. Other known CCTs possess a single catalytic domain and also one copy
of membrane binding and phosphorylation domains, and are reported to form homodimers in
solution [17]. Hence, the two duplicated segments of plasmodial CCT proteins might
associate to form an intramolecular “pseudo-dimer”, although this suggestion has not yet
been investigated experimentally. It has been shown that a recombinant PfCCT protein
528-896
construct (PfCCT
), containing the second catalytic domain (C2) followed by the
membrane interaction domain (M2) is enzymatically active [18].
CCT expression is apparent in all life stages of Plasmodium [11, 19]. Microarray data
(
PlasmoDB) [20] show a transcription increase in late schizonts and significant transcription
in gametocytes and sporozoites, as well. Membrane interaction of an amphipathic segment of
the membrane binding domain (M) regulate CCT enzymatic activity while removal of the
membrane binding segment from the CCT protein results in a constitutively active enzyme
[
21]. For the rat CCT enzyme, an in vitro kinetic study suggested random mechanism for
M
binding and release of the substrates/products, with K values of 10 mM and 0.3 mM for
CTP and ChoP, respectively [22].
Although CCT activity is crucial for Plasmodia at their blood stage, the enzyme:substrate
binary complexes have not yet been investigated in detail. To date, no three-dimensional
structure of any CCT:substrate complex is available. Current structural concepts rely on the
three-dimensional CCT:CDPCho structure of the rat catalytic domain (PDB ID:3HL4) [23]
and structures of other cytidylyltransferase:nucleotide complexes (ECT:CMP and GCT:CTP,
This article is protected by copyright. All rights reserved.

PDB IDs: 3ELB and 1COZ, respectively [24]). It is also not known whether binary
CCT:substrate complexes already adopt catalytically competent conformation or such
conformation is reached only in the ternary complex of the enzyme with both substrates.
In the present study, we have investigated the evolutionary origin of the two domains and the
catalytic mechanism and ligand binding properties of P. falciparum CCT. We reconstructed
the most likely evolutionary scenario that led to the duplication of the CMP segments in CCT
within Ampicomplexa. We then designed and characterized a catalytic domain construct of
PfCCT using the C2 catalytic domain, and showed that this construct is enzymatically
functional and forms a dimer, similarly to other CCTs. We constructed a three-dimensional
structural model of this enzyme and used a continuous spectrophotometric method to monitor
its steady-state kinetic behaviour. Binding of substrates and products was followed by
fluorescence spectroscopy and isothermal titration calorimetry (ITC). Relying on a Trp
fluorophore, located in the active site, we could characterise alterations in the binding modes
of substrate and product containing the choline moiety. Thermodynamic analysis revealed
2
+
that the presence of Mg , the cofactor of CCT catalysis, attenuates the binding of CTP.
Taken together, the present ligand binding and kinetic studies of the enzyme contribute to
understanding the catalytic mechanism of PfCCT.
Results
Phylogenetic analyses revealed two independent duplication events of CMP segments in
CCTs within Apicomplexa
The CCT catalytic domain (C) is widely conserved from plants to animals. Putative CCT
enzymes harbouring duplicated C domains are found only in three Apicomplexan parasites,
Plasmodium, Babesia and Theileria, within the class Aconoidasida. Sequence alignment
revealed that the duplicated C domains of plasmodial CCT are highly identical (Fig. 1). This
can be exemplified with PfCCT, showing 90% sequence identity and 97% similarity between
the N- and the C-terminal catalytic domains (C1 and C2, respectively). To reconstruct the
history of duplication events that led to the existence of duplicated CCT genes in
Apicomplexa, we performed a phylogenetic analysis. The alignment of the full CCT protein
sequences is presented in Fig. S1. It shows that the single CMP segment observed in Coccidia
(Cryptosporidium, Neospora, and Toxoplasma) is homologous to the N-terminal segment of
Aconoidasida (other genera). The maximum likelihood (ML) tree obtained for the 16 full
This article is protected by copyright. All rights reserved.

Apicomplexan CCT proteins is shown in Fig. 2A. The phylogenetic relationships among the
Apicomplexan species included are fully compatible with the tree recently inferred from
genome-scale data [25]. Indeed, Babesia microti also appears as a distinct lineage that
diverged early within Piroplasmida, rendering the genus Babesia paraphyletic. It is also worth
noting that Piroplasmida members are particularly fast evolving compared to all other
Apicomplexans.
The ML tree inferred for the 27 CMP protein segments of Apicomplexan CCTs is presented
in Fig. 2B. By reconstructing the CMP segment duplication history during the evolution of
Apicomplexa, this phylogeny allows testing the hypothesis of a single ancestral duplication
of CMP segment in the CCT gene predating the separation between Piroplasmida (Babesia
and Theileria) and Haemosporida (Plasmodium) [5]. Under this hypothetical scenario, N-
terminal and C-terminal segments are predicted to form distinct monophyletic subtrees whose
internal relationships should exactly mirror the species tree of Fig. 2A. The ML tree in Fig.
2
B clearly shows that it is not the case. Indeed, the sequences of the duplicated C-terminal
segments fall into two distinct monophyletic groups, each one being closely related to a
cluster of N-terminal sequences within both Piroplasmida and Haemosporida. The most likely
scenario to explain such a phylogenetic pattern is that two independent domain duplication
events occurred in the last common ancestor of Piroplasmida on one side, and Haemosporida
on the other side. Moreover, the large difference observed in sequence divergence between
the N-terminal and C-terminal segments in Piroplasmida and Haemosporida suggests that
these two independent duplications have occurred at different points in time. Indeed, the high
degree of sequence divergence observed among the duplicated segments in Babesia and
Theileria species might be explained by a duplication event early in the evolutionary history
of Piroplasmida. Conversely, the high similarity of duplicated CMP segments observed
between Plasmodium species may be the result of more recent duplication event in the
Haemosporida lineage. Based on this evolutionary scenario resulting in two very similar
catalytic domains of PfCCT and due to the fact that full-length PfCCT could not be expressed
and purified for biochemical analysis up to date, we decided to study one of the PfCCT
catalytic domains as a model.
This article is protected by copyright. All rights reserved.

Homology model and characterization of PfCCT catalytic domain
A homology model of both catalytic domains of PfCCT was built based on the published
structure of the homodimer catalytic domain of rat CCT (see Supplementary Material for
details) [23]. Analysis of these models shows that neither of the very few residues that differ
in C1 and C2 is bordering the conserved active site of the enzyme. The high degree of
sequence identity between the two catalytic domains of PfCCT likely allows one single
domain to dimerize as in other studied CCTs. We engineered a truncation construct
528-795
(
PfCCT
binding M2 domain and is therefore expected to be constitutively active. A lysine-rich region
residues 720-737, termed as K-loop in Fig. 3A), constituting a species-specific insertion in
) that contains the C2 catalytic domain but does not contain the membrane-
(
Plasmodia sequences, was removed to aid purification of the recombinant construct (see
Materials and Methods for details). Comparing homology models either containing or lacking
this lysine-rich segment revealed no structural differences affecting the folding of the
catalytic domain (compare models in wheat and light blue colours in Fig. 3A). The construct
Δ720-737
PfCCT (528-795)
lacking the lysine-rich segment (abbreviated as MΔK) was
expressed in E. coli. Removal of the lysine-rich segment improved the thermal stability of the
m
PfCCT construct as judged by an approx. 9 C increase in melting temperature (T ), measured
by differential scanning fluorimetry (Fig. S2).
We used in silico analysis of structural disorder (Fig. 3B) to assess further structural
characteristics. Within the full length CCT sequence, there are numerous highly
flexible/disordered segments in the linker region connecting the two duplicated segments.
The lysine-rich segment is present in both C1 and C2 domains and is predicted to possess
high disorder. The region encompassing residues 540-610, present within our engineered
Δ720-737
PfCCT (528-795)
(MΔK) construct, is also predicted to possess high
flexibility/disorder.
Gel filtration of PfCCT MΔK under native conditions resulted in a single peak, at an elution
volume corresponding to a molecular mass of approximately 268 kDa (Fig. S3), which is
unexpectedly high (for comparison, the molecular mass of the PfCCT MΔK polypeptide
calculated from the amino acid sequence is 30.9 kDa). Denaturing SDS PAGE analysis
revealed that this peak contains one polypeptide component for which the apparent molecular
mass based on the SDS PAGE analysis is estimated to be 37 kDa (Fig. 4A), again
This article is protected by copyright. All rights reserved.

considerably higher that the calculated molecular mass of 30.9 kDa. SDS-PAGE and gel
filtration data therefore indicate anomalous migration of the protein in gel electrophoresis and
possible formation of higher oligomers under native conditions. The presence of flexible
segments as suggested by the disorder predictors (Fig. 3B) may explain this behaviour and
also suggests that estimation of the oligomerization status of the native molecular mass of this
protein based on size exclusion chromatography is not reliable, since deviations from the
ideally expected hydrodynamic behaviour of globular proteins are expected to greatly perturb
elution characteristics.
To investigate the oligomerization state of PfCCT MΔK, we therefore performed electrospray
ionization mass spectrometry (MS), which has been shown to provide valuable data on
protein-protein complexation [26]. The electrospray mass spectrum of PfCCT MΔK under
native conditions (Fig. 4B) reveals two abundant ion series in the mass range m/z 1900-4000.
In the upper mass range (m/z 3000-4000), peaks corresponding to a dimer structure can be
observed. The charge state distribution is rather narrow, encompassing mainly the z = +18 to
+
20 ions, indicating that the native protein conformation survives the transfer from solution
to the mass spectrometer [27]. The molecular mass of the dimer was determined to be 61,840
200 Da (c.f. calculated dimer molecular mass: 61.8 kDa). At the lower mass range (m/z
900-3000), the ion series corresponding to the monomer form is present. The charge
distribution is centered on the +13 ions, and the mass calculated from the distribution is
0,870 ±200 Da, in very good agreement with the calculated molecular mass. Based on the
±
1
3
integrated peak areas, a monomer: dimer ratio of 10:13 was estimated. Signals due to higher
oligomers - trimer, tetramer etc. - are practically not observable, setting an upper limit of 3 %
for their relative abundance on the basis of the signal to noise ratio. The preference for dimer
formation suggests that the dimer structure reflects the association existing in solution, i.e., it
is not due to non-specific aggregation upon ionization. Dimerization of the PfCCT MΔK
construct is also suggested by the 3D structural model and is in agreement with the oligomer
properties of other CCT enzymes.
Random order kinetic mechanism shown by steady-state analysis
The enzymatic properties of the PfCCT MΔK construct were determined by using a
continuous spectrophotometric assay [28] that was adopted for other pyrophosphate
producing enzymes [29, 30]. Substrate titrations with CTP and ChoP were performed while
This article is protected by copyright. All rights reserved.

the other substrate was added at saturating concentrations. Plot of the initial rate as a function
of CTP concentration provided a saturation curve (see Fig. 5A), which fits to the
-1
Michaelis-Menten equation providing the apparent kinetic parameters 1.45 ± 0.05 s for kcat
and 168 ± 17 μM for KM,CTP (Table 1). In the ChoP titration, we observed substrate inhibition
at [ChoP] >5 mM range. The ChoP titration data were fitted with a substrate inhibition
equation (shown in Materials and Methods) providing an apparent KM,ChoP of 1.8 mM (Fig.
5
B). As a consequence, in CTP titration experiment, ChoP was used at a sub-inhibitory
concentration of 5 mM. Neither of the CTP and ChoP substrate saturation data showed
significant cooperative fashion, as judged by the Akaike’s Information Criterion (AIC) test,
comparing Michaelis and Hill fitted models. Deletion of the lysine-rich segment did not
modify enzyme activity (data not shown).
In order to assess binding order of the substrates CTP and ChoP, we performed a kinetic
analysis based on CDPCho product inhibition. Initial rate of CCT reaction was measured in
the function of either substrate while the concentration of the other substrate and CDPCho
product was kept constant (Fig. 5C-D). We found that the inhibitory pattern of CDPCho
showed a major competitive character either with respect to both CTP and ChoP substrate,
which can be interpreted within the frames of a random Bi-Bi mechanism[31]. The apparent
inhibitory constant K of CDPCho is estimated from the fitting to be 65 μM.
i
Comparison of the heat effects of substrate and product binding at the choline subsite
Since isothermal titration calorimetry (ITC) is an especially useful technique for quantitative
characterization of biological interactions, we used ITC to obtain thermodynamic parameters
of binding of substrates and products to PfCCT MΔK. Titration calorimetry experiments were
2
+
performed in the same buffer as used in the activity assay, including 5 mM Mg (Fig. 6 A-D,
Table 2).
For both CTP and CDPCho titration of MΔK (Fig. 6A and D), the best fit to the experimental
data was obtained by fitting a one-site independent binding model. This, together with the
observed stoichiometry close to the ratio 1:1, is consistent with kinetic titration data (Table
2
). The binding of these ligands to the enzyme did not show any significant cooperativity
under our experimental conditions. In the case of the CDPCho product, the dissociation
constant K of 89 μM is in good agreement with the inhibitory constant K of 65 μM,
d
i
This article is protected by copyright. All rights reserved.

determined in the steady-state experiments. The CTP equilibrium ligand binding experiment
reports similar affinity (K =294 μM) as the value of the respective Michaelis constant (K
d
M
=
168 μM). Both CDPCho and CTP ligands exhibited negative binding enthalpy and entropy
values, which indicate that binding of the ligands are enthalpically favoured and entropically
disfavoured. The latter observation for CTP binding is consistent with previous NMR data of
CTP binding to GCT, a related cytidylyltransferase enzyme [32]. Nevertheless, CDPCho
binding to the enzyme occur with more than twofold higher enthalpy as compared to CTP
binding, shedding light on the potentially different interactions with these two ligands. These
could be due to either different accommodation of the pyrophosphate group of CTP and β
phosphate group of CDPCho and/or contribution of the interactions of trimethylammonium
moiety of CDPCho at the choline binding subsite in the active site of CCT.
Titration of the enzyme with PP
calorimetric signals may be interpreted by estimating that the dissociation constant of the
CCT:PP complex is in the mM range. On the other hand, we could not measure any
i
product yielded only minor heat effects (Fig. 6C). These
i
significant calorimetry of apo MΔK enzyme for ChoP titration (up to 2 mM, already above
the K , determined in the kinetric assay) (Fig. 6B). This result indicates that either ChoP
M
binding to the apo enzyme does not occur under these conditions or the enthalpic character of
this binding is negligible when ChoP is present only on its own.
In order to reveal any indirect effect of ChoP binding, we applied displacement ligand
titration method, that has been shown to be effective in detecting ligand binding events with
low mM affinities [33]. Immediately following ChoP titrations, the MΔK solutions
containing ChoP were further titrated with CDPCho to assess an equilibrium ligand
displacement titration of ChoP to CDPCho. These subsequent product titrations yielded
thermodynamic parameters very similar to those obtained from apo MΔK CDPCho titrations
in the absence of ChoP (Fig. 6I and J, Table 2). Hence, results did not show any detectable
binding of ChoP to be determined suggesting that under our experimental conditions, ChoP
binding to the apo enzyme cannot be observed, arguing for a lower limit of the dissociation
M
constant Kd,ChoP>5 mM (to be compared with a K of 1.8 mM).
This article is protected by copyright. All rights reserved.

2
+
Mg attenuates enzyme affinity towards CTP
2
+
2+
Mg is reported to be required for CCT catalysis [18, 34, 35], therefore Mg is a generally
used additive in CCT assay conditions. However, no data are available on its effect on ligand
binding to the enzyme. We therefore carried out a set of ITC experiments in the absence of
2+
added Mg (Fig. 6E, F, G, H). Titration of the enzyme with CDPCho reported similar
2
+
thermodynamic parameters in the absence and presence of Mg cofactor (with dissociation
constants 44 μM and 89 μM, Fig. 6D, H, respectively). CTP affinity, on the contrary, seems
to be strongly dependent on the presence of the metal cofactor. Addition of 0.5 mM EDTA
2
+
resulted in significantly higher affinity of the enzyme towards CTP, than in presence of Mg
(
(
the observed dissociation constant being 59.5 μM as compared to 294 μM, respectively)
Fig. 6A and E, Table 2). ChoP titration and PP titration provided practically the same results
i
2
+
in the absence as well as in the presence of Mg (Fig. 6B, C, G, H).
Intrinsic fluorescent signal reports on differences between ChoP and CDPCho bound
states of the enzyme
In order to gain information on the microenvironment of the active site, we investigated the
change in the intrinsic Trp fluorescence signal. PfCCT MΔK has three tryptophan residues:
W682, W692 and W743 (see Fig. 1). W682 is buried at the dimer interface while the other
two residues are more solvent-accessible; W743 is part of the αL-helix, located on the
construct surface (see Fig. 3A), while W692 establishes a cation-π interaction with the
trimethylammonium moiety of CDPCho product at the active site (Fig. 3A, Fig. 7) [23]. To
identify the tryptophan residue that is responsible for the observed signal changes, we created
two constructs by site directed mutagenesis, performing a W/H exchange at position 743, and
W743H
a W/Y exchange at position 692, respectively. The MΔK
showed similar ligand binding
W692Y
(
Fig 6K, Table 2) and catalytic properties, whereas the MΔK
mutant, the interaction
with CDPCho is still present, but becomes somewhat perturbed (Fig 6L, Table 2, K
d
increases threefold).
Addition of CDPCho led to an increase in the intensity of the intrinsic Trp fluorescence
signal of PfCCT MΔK (Fig. 8A). Moreover, the fluorescence emission maximum shifted
from 347.7 to 344.2 nm. Plotting the emission maximum shifts (λmax) against CDPCho
concentration resulted in a saturation curve, indicating saturation of enzyme active sites with
This article is protected by copyright. All rights reserved.

CDPCho (Fig. 8B). The dissociation constant of PfCCT MΔK with CDPCho complex, was
found to be 19.2 ± 2.4 μM, which is in adequate agreement with the value of 44 ± 2 μM,
determined by isothermal titration calorimetry (Table 2). Titration of PfCCT MΔK with ChoP
substrate is not accompanied with either increase of fluorescence intensity or blue shift of
W743H
fluorescence emission maximum (Fig. 8B). For the MΔK
mutant, titration with
CDPCho resulted in similar intensity increase and shift of emission maximum as compared to
the wild type, with an observed binding constant 32.4 ± 5.8 μM (Fig. 8B). In contrast, no
W692Y
such fluorescence signal change was observed upon titration of MΔK
with CDPCho
(
Fig. 8B). We conclude that the change of fluorescent emission spectral parameters induced
by binding of CDPCho to the enzyme are due to the W692 residue, present within the active
site.
Discussion
Here we report a detailed ligand binding and kinetic study of a truncated construct of P.
falciparum CCT (PfCCT MΔK) that contains a single catalytic domain but lacks the
membrane binding domain. To evaluate compatibility of the construct with the full-length
PfCCT sequence, we investigated the conservation pattern of the duplicated CMP segments
in Apicomplexa. Results of the phylogenetic analyses indicate two independent ancestral
gene duplication events in Haemosporida (Plasmodia) and Piroplasmida, the former having
occurred more recently. This indicates reduced possibility for the functional divergence of C1
and C2 domains. Widespread sequence conservation and the rat CCT crystal structure [23]
enabled the creation of a homology model of the investigated PfCCT construct, containing
the C2 domain. These considerations together with the present results indicating dimerization
and steady-state kinetic characteristics of PfCCT MΔK verify that this construct constitutes a
reliable catalytically functional model for the physiologically expected C1/C2 pseudo-
heterodimer catalytic domain assembly of full length PfCCT.
The observed competitive inhibition pattern of CDPCho versus both substrates indicates that
CDPCho can bind to the same enzyme conformation as either substrate. Interpretation of our
product inhibition measurements in light of the above considerations suggests random
binding mechanism for PfCCT MΔK, in agreement with previous data for rat CCT [22]. In
such kinetic mechanism, either substrate (CTP or ChoP) can bind to the apo enzyme on its
This article is protected by copyright. All rights reserved.

own. However, in our ligand binding experiments based on either fluorescence spectroscopy
or isothermal titration microcalorimetry, no binding of ChoP to the apoenzyme could be
observed, in contrast to the suggested kinetic mechanism. It has to be emphasized that both
techniques could be used with success to follow binding of CDPCho to the apoenzyme, and
although fluorescence spectroscopy is an indirect method, isothermal titration
microcalorimetry is considered to be a generally applicable direct technique to evaluate
molecular interactions. To reconcile the data from the kinetic and ligand binding experiments,
it is important to realize that kinetic studies have their own limitations, as well. Accordingly,
it was shown that an observed kinetic pattern for random mechanism may not identify the
relevant set of elementary steps [36].
The fluorimetric titrations in our study rely on the signal of W692, located in the choline
subsite of the active site. The homology model reveals that this residue possibly monitors
CDPCho binding to the active site by juxtaposing the trimethylammonium moiety of
CDPCho to the indole ring of W692 (Fig. 7). The lack of any fluorescent signal upon ChoP
addition is likely due to a different microenvironment of the choline subsite. ITC experiments
also including displacement ligand titrations reported considerable heat effects for CDPCho
binding, whereas no such heat effects were found for ChoP, strengthening the suggestion that
the interaction of the ChoP moiety with the choline binding subsite in the enzyme depends on
whether ChoP is present on its own or within CDPCho.
2
+
Mg was reported to be a co-factor of cytidylyltransferases [35], however, we found that
2
+
CTP and CDPCho bind to the CCT enzyme even in the absence of Mg (cf. Fig. 6C, F). This
correlates well with successful co-crystallization of enzymes belonging to the
2+
cytidylyltransferase family with CDPCho and CTP, in the absence of Mg [23, 24].
2+
Moreover, we also observed that addition of Mg significantly decreased binding affinity of
CCT towards CTP. In the absence of the divalent cation, binding of the nucleotide needs to
be accomplished by adequate charge compensation of the phosphate chain, presumably
performed by positively charged and polar residues. Considering the sequence conservation
within the catalytic domain of CCTs as well as the structural models, residues R755 and
K663 as well as T761 and T762 [37] may be proposed to fulfil such coordinating roles for the
2
+
phosphate chain. We propose that Mg may be important for efficient catalysis but not for
nucleotide binding for CCT enzymes belonging to the HxGH enzyme superfamily [38]. Lack
2
+
of requirement for Mg during nucleotide binding to the enzyme may also present a possible
This article is protected by copyright. All rights reserved.

mode of action in the HxGH enzyme superfamily. Such a situation was experimentally
verified for a class I aminoacyl tRNA synthetase that binds its cognate nucleotide ATP with
2
+
considerably higher affinity in the absence, than in presence of Mg [39].
In our study, we delineated the possible evolutionary steps leading towards Plasmodial CCT
proteins, identified a useful fluorophore within the active site of CCT, and provided insights
into kinetic and ligand binding characteristics of a PfCCT construct that can be expressed and
purified for detailed studies. Based on these results, further research directions are suggested
to focus on a detailed in-depth kinetic study of the elementary steps of CCT catalysis, and on
generating experimental three-dimensional structures to provide a full coherent knowledge of
the mechanism of action of this important enzyme.
Materials and Methods
Chemicals
Restriction enzymes, T4 DNA ligase, DNA polymerases and DNA purification kit were
obtained from Machenary-Nagel. CTP, CDPCho, Sypro Orange, inorganic pyrophosphatase,
purine nucleoside phosphorylase and antibiotics were purchased from Sigma-Aldrich (St.
Luis, MO, USA). ChoP was from TCI Europe N.V.
IPTG was obtained from Fisher Scientific GmbH, Germany. MESG (2-amino-6-mercapto-7-
methylpurine riboside) was obtained from Berry and Associates (Dexter, MI). Ni-NTA was
from Quiagen, protease inhibitor cocktail tablets were purchased from Roche. All other
chemicals were of analytical grade of the highest purity available.
Phylogenetic analyses
A first dataset containing the amino acid sequences of CCT proteins was assembled from
GenBank for 16 apicomplexan taxa including species belonging to the genera
Crytosporidium, Neospora, Toxoplasma, Babesia, Theileria, and Plasmodium. These full
length protein sequences were aligned using the probabilistic consistency-based program
ProbCons [40] with default parameters on the Phylogeny.fr portal [41]. This resulted in an
alignment with 1122 amino acid sites. In order to reconstruct the history of domain
duplication within these proteins, a second data set was constructed by extracting the regions
of the full alignment corresponding to the N-terminal and C-terminal CMP segments, leading
This article is protected by copyright. All rights reserved.

to 16 and 11 sequences, respectively. Indeed, only members of the genera Babesia, Theileria,
and Plasmodium possess a duplicated C-terminal domain, whereas Crytosporidium,
Toxoplasma and Neospora exhibit a CCT gene with a single CMP segment. The amino acid
sequences of the 27 extracted domains were realigned using the phylogeny-aware PRANK
program on the webPRANK server [42] with default parameters. This resulted in a final
alignment with 362 sites.
Ambiguously aligned sites were subsequently removed from both datasets using the GBlocks
program [43] with default relaxed parameters. This led to final alignments with 456 and 217
amino acid sites for the full gene and domain data sets, respectively. Maximum-likelihood
phylogenetic reconstruction was conducted on these two data sets using PHYML 3.0 [44]
with NNI branch swapping on a BIONJ starting tree. For both datasets, the LG+G4 model of
protein sequence evolution was selected as the best fitting model by ProtTest 3 [45]. The
statistical robustness of phylogenetic inference was assessed by performing 100 bootstrap
replicates using the same heuristic search strategy as for the initial analyses.
Homology modelling
Homodimer homology models have been built for constructs PfCCT (528-795) and PfCCT
MΔK using the catalytic domain of the rat CCT which included the CDPCho product (PDB
ID: 3HL4) [23]. Both homology models were built using the same computational procedure.
The alignment previously published in the literature [23] was slightly modified in the
neighbourhood of the lysine-rich segment. The identity of the alignment was 42.21%. For
homology modelling only residues 581 to 775 were considered, as in the rat CCT X-ray
structure no template was found for the starting and end sequences of the construct. The
MODELLER 9v8 software [46] was used to create 100 homology models for both construct
using the same alignment. The low energy conformations with low objective function values
were visually inspected and one with appropriately coordinated CDPCho ligand was selected
for each construct. The selected structures were optimised with CHARMM27 force field [47]
and CHARMM software [48] using the GBSW implicit solvent model. The topology file and
non-standard parameters added for the CDPCho ligand can be found in the Supplementary
Material.
The selected homology models were validated using the PROCHECK [49], WHAT_CHECK
[
50]and ERRAT [51] programs (see the Supplementary Material for further details, Fig. S4,
This article is protected by copyright. All rights reserved.

Fig. S5, Table S1 and Table S2). The two homology models exhibited very similar
characteristics and differed significantly only in the presence/absence of the Plasmodia
lineage-specific, lysine-rich segment. The models indicated that this insert lay reasonably far
away from the protein domain that harbours the active site of the constructs, thus it is
unlikely to have a major influence on the binding of the CDPCho ligand. Based on this and
on the results of disorder predictions the homology model of MΔK protein was chosen for
modelling the active site of PfCCT with ligands.
Cloning and mutagenesis
The PfCCT cDNA sequence (PF3D7_1316600) was codon-optimized for expression in E.
coli (GenScript). The DNA fragment corresponding to residues 528 to 795 of PfCCT was
cloned into the expression vector pET15b (Novagen) using the NheI/BamHI sites for
production of N-terminal 6×His-tagged protein. Site-directed mutagenesis to produce MΔK
(
528-795, delta 720-737) construct of P. falciparum CCT, lacking the lysine-rich,
Plasmodium specific segment (720-737) and its W692Y and W743H point mutants were
performed by the Quickchange method (Agilent) (mutagenesis primers shown in Table S3).
All constructs were verified by DNA sequencing.
Protein expression and purification
The recombinant His-tagged PfCCT MΔK (528-795, delta 720-737) protein was expressed
using the BL21 (DE3) Rosetta E. coli expression system. 0.5 l of LB medium was inoculated
with Rosetta cells from a 5 ml overnight culture and grown at 37°C until OD600 reached 0.4.
After 30 min cooling at 20°C, protein expression was induced by 0.5 mM isopropyl
thiogalactoside for 20 h at 20°C. The cells were harvested by centrifugation at 4000g for 20
min and stored at -80°C. Thawed cells were resuspended in 30 ml of 20 mM Hepes pH 7.5,
1
00 mM NaCl, 1 mM PMSF, 1% Triton X100, 3 μg/ml RNase, 3 μg/ml DNase, EDTA
containing protease inhibitor cocktail tablet (Roche), 1 mM benzamidine and 10 mM
β-mercaptoethanol. The solution was sonicated, centrifuged, and then the supernatant was
2
+
supplemented with 1.5 mM Mg to form chelate complex with EDTA present and was
applied onto a pre-equilibrated benchtop nickel-nitrilotriacetic acid-agarose affinity-
chromatography column (Qiagen). The protein was eluted with Hepes 20 mM, NaCl 100 mM
(
pH 7.5) (buffer A) containing 250 mM imidazole. The eluted samples were dialysed
This article is protected by copyright. All rights reserved.

overnight in buffer A and the next day the sample was concentrated and further purified on a
HiLoad XK-16 Superdex 200 column (GE Healthcare) previously equilibrated with buffer A.
The purified sample was shown to be over 90% homogeneous based on SDS-page analysis,
concentrated to 10 mg/ml by using ultrafiltration membrane (Sartorius), then supplemented
2
with 10% glycerol, and finally was flash-frozen in liquid N in aliquots and stored at -80°C.
SDS-PAGE analysis of purification procedure is shown in Fig. 4A. Protein concentration is
in monomers throughout the whole article. It was determined spectrophotometrically from the
-1
-1
absorbance at 280 nm by using an extinction coefficient 31400 M cm value as calculated
on the basis of amino acid composition by using ProtParam server [52] and a molecular
weight of 30898 for the His-tagged monomer (throughout the present study, molar
concentrations of the CCT protein construct refer to the monomer). Absorbance
measurements were carried out using a Nanodrop spectrophotometer. Protein samples were
identified by mass spectrometric analysis of the tryptic digest on a LCQ-Fleet Ion Trap
LC/MS.
Differential scanning fluorimetry
Solutions of 5x Sypro Orange and 0.5 mg/ml protein were prepared in final volume of 25 µl
in 96-well PCR plate wells. The plates were sealed with adhesive PCR film (Thermo
Scientific), and heated in a Stratagene Mx3000P real-time PCR instrument (Agilent) from 25
to 85°C in increments of 0.5°C. Three parallels were set up for each condition. To obtain the
m
midpoint of protein melting transition defined as a melting temperature T , a Boltzmann
model (Eq.1) was used to fit the fluorescence data using OriginPro 8 (OriginLab Corp,
Northampton, MA, USA).
F − F
1
0
F = F +
(Eq.1)
T
0
(T −T)/dT
m
1
+ e
is the fluorescence intensity at T temperature, T
are the pretransitional and posttransitional fluorescence intensities, respectively. Data points
after the fluorescence intensity maximum F were excluded from fitting.
where F
T
m 0 1
, melting temperature, F and F
1
This article is protected by copyright. All rights reserved.

Mass Spectometry
In the mass spectrometric study of protein complexes, a commercial Waters QTOF Premier
instrument equipped with electrospray ionization source was used in positive ion mode. Mass
spectra were obtained under native conditions: namely, the ions were generated from aqueous
2
4 3
0 mM NH HCO buffer solution (pH = 7.8) containing the PfCCT MΔK protein at 40 μM
monomer concentration. These conditions allow transfer of the native protein complex
present in the solution into the gas phase. The capillary voltage was 2800 V, the sampling
cone voltage was 128 V and the temperature of the source was kept at 90 °C. Mass spectra
were recorded in the mass range of 1500-8000 m/z.
Fluorescence spectroscopy
Emission curves were recorded on a Jobin Yvon SpeX Fluoromax-3 spectrofluorometer using
1
0 mm path length 20°C thermostated cuvettes with excitation at 295 nm (slit 1 nm),
emission between 320 and 400 nm (slit 5 nm). 3 μM of protein in buffer A was titrated with
addition of 1-2 μl aliquots from concentrated stock ligand solutions. After 3 minutes
incubation, three emission scans were recorded, and then averaged. Emission maximum was
determined by fitting the upper 20% fluorescence intensity of the emission curves with Gauss
distribution equation Eq.2:
2
(
λ − λ
)
max
2
−
2ω
y = y + A⋅e
(Eq. 2)
0
where λ, wavelength in nm, λmax, calculated emission peak maximum in nm.
The dissociation constant of ligand binding was calculated by fitting to the calculated
emission shift data with a quadratic equation (Eq.3) (as previously [53-55]:
2
0.5
y = A ⎡(c + K + x) − (c + K + x) − 4cx) ⎤ / 2c (Eq. 3)
⎣
⎦
where
x
, ligand concentration (
μM);
y
, shift of emission of ligand addition (nm); A, emission
shift observed at saturation; c, enzyme concentration (μM); K, dissociation constant (μM).
This article is protected by copyright. All rights reserved.

Steady-state kinetic assay
A continuous coupled enzyme activity assay was employed based on [29]. The assay was
performed in 10 mm path length 20°C thermostated cuvettes in a dual wavelength Specord
2
00 spectrophotometer, with a reaction volume of 600
concentrations were set such that CCT catalysis was the rate limiting step; [CCT M
.16 M, [inorganic pyrophosphatase] = 2.08 U/ml, [purine nucleoside phosphorylase] = 2.08
μ
l. Enzyme and substrate
ΔK] =
0
μ
2
+
U/ml, [Mg ] = 5 mM, [MESG] = 0.33 mM, and variable concentrations of phosphocholine
and CTP substrates in buffer A. The reaction was started by a final quick addition of CCT
M
Δ
K to the reaction mixture. Initial velocity was determined from the slope of the first 10%
of progress curve. For CTP substrate titrations, CTP concentration was varied between 12
M and 1.2 mM while ChoP concentration was kept at 5 mM. For ChoP substrate titrations,
μ
ChoP concentration was varied between 0.2 mM and 20 mM while CTP concentration was
kept at 1 mM. Kinetic data were fitted with Eq. (4, 5) (Michaelis-Menten equation and
substrate inhibition equation, respectively):
Vmax ∗[S]
v =
v =
(Eq. 4)
(Eq. 5)
K +[S]
M
Vmax
⎛
⎜
⎝
KM [S]⎞
1
+
+
⎟
⎠
[S]
Ki
-1
In the above equations, v is the reaction rate (s ), vmax is the maximum velocity of the
-1
-1
reaction (s
μ
M ), [S] is the concentration of the varied substrate (
i
μM) and K describes the
binding of a substrate molecule to the enzyme (
the rate of the reaction.
μM), with the result being a half decrease in
Product inhibition kinetic measurements
Product inhibition studies were performed using the above assay with CDPCho as inhibitor
present in 0 µM, 40 µM and 100 µM concentrations. Substrate concentration was varied in
the presence of fixed saturating concentrations of the other one. Inhibitory constant (K
i
) was
This article is protected by copyright. All rights reserved.

calculated by global fit of the CTP titration at saturated ChoP concentration curves with Eq. 6
describing full competitive inhibition
v = Vmax /[1+(K
M
/[S])(1+[I]/K
i
)]
(Eq. 6)
ITC measurements
Calorimetric measurements were performed on a MicroCal-ITC 200 titration calorimeter (GE
Healthcare) at 20°C. The protein samples were dialysed against buffer A, including 10 mM
2+
β
-mercaptoethanol or 0.5 mM TCEP as reducing agent and 5 mM Mg or 0.5 mM EDTA, as
indicated. Ligands were freshly dissolved into dialysis buffer and subsequently, pH was
adjusted, prior to ITC measurements.
CTP and CDPCho concentrations were determined using molar extinction coefficients 9000
-1
-1
-1
-1
M cm at 271 nm, pH 7.0 and 5000 M cm at 260 nm, respectively. In the experimental
setup, the cell of the instrument was filled with protein and the syringe with the respective
ligand. Each titration typically included 27 steps of injection with 1.5
μl of ligand per
injection spaced 180 s apart from each other, with the injection syringe rotating at 500 rpm.
In case of displacement ligand titrations, the enzyme previously titrated with ChoP, reaching
a final concentration of 2-5 mM was immediately further titrated with CDPCho. The data
were analysed using MicroCal ORIGIN software, following the directions of the
manufacturer, in order to calculate the thermodynamic parameters: dissociation constant (K
d
),
stoichiometry (N), enthalpy ( H) and entropy ( S). The heats of dilution obtained from
Δ
Δ
titration of the same ligand solutions into buffer were subtracted from the titration data before
the data was fitted to the appropriate model.
Acknowledgements
This work was supported by the French “Agence Nationale de la Recherche” (ANR-09-
BLAN-0397) (ML, HV, RC), the European Community (FP7/EviMalar Network of
Excellence No. 242095) (HV, RC), ANR-NKTH ADD-MAL (Agence Nationale de la
Recherche - Nemzeti Kutatasi es Fejlesztesi Hivatal, "Antimalarial drug discovery and
development of new in vitro assays for the optimization of antimalarial therapy") (BGV,
GNN, LM) and the New Széchenyi Plan (TÁMOP-4.2.2/B-10/1-2010-0009) (JO, BK, BGV,
This article is protected by copyright. All rights reserved.

GNN). JO acknowledges receipt of an EU Marie Curie ERG Fellowship (Project
Oestrometab"). This work has been supported by the Hungarian Scientific Research fund
grant No. OTKA NK83857) (ÁR, KV), and NK84008 (BGV)). GNN was supported by the
"
(
Zsuzsa Szabó Foundation. This study represents the contribution ISEM 2013-031 of the
Institut des Sciences de l’Evolution de Montpellier.
References
1
.
WHO (2011) World Malaria Report 2011. In. World Health Organization, Geneva,
Switzerland.
2
.
Murray CJL, Rosenfeld LC, Lim SS, Andrews KG, Foreman KJ, Haring D, Fullman N,
Naghavi M, Lozano R & Lopez AD (2012) Global malaria mortality between 1980 and 2010: a
systematic analysis. The Lancet 379, 413-431.
3
.
Fairhurst RM, Nayyar GM, Breman JG, Hallett R, Vennerstrom JL, Duong S, Ringwald
P, Wellems TE, Plowe CV & Dondorp AM (2012) Artemisinin-Resistant Malaria: Research
Challenges, Opportunities, and Public Health Implications. Am J Trop Med Hyg 87, 231-241, doi:
1
4
0.4269/ajtmh.2012.12-0025.
.
Dondorp AM, Fairhurst RM, Slutsker L, Macarthur JR, Breman JG, Guerin PJ, Wellems
TE, Ringwald P, Newman RD & Plowe CV (2011) The threat of artemisinin-resistant malaria. N
Engl J Med 365, 1073-1075, doi: 10.1056/NEJMp1108322.
5
.
Dechamps S, Shastri S, Wengelnik K & Vial HJ (2010) Glycerophospholipid acquisition
in Plasmodium - a puzzling assembly of biosynthetic pathways. International journal for parasitology
4
0, 1347-1365, doi: 10.1016/j.ijpara.2010.05.008.
Vial H & Mamoun C (2005) Plasmodium Lipids: Metabolism and Function. In Molecular
6
.
Approach to Malaria (Sherman I, Ed, ed, pp. 327-352. ASM Press, Washington DC.
7
3
8
.
Vial HJ & Ancelin ML (1992) Malarial lipids. An overview. Subcellular Biochemistry 18, 259-
06.
.
Vial HJ, Eldin P, Tielens AG & van Hellemond JJ (2003) Phospholipids in parasitic
protozoa. Mol Biochem Parasitol 126, 143-154.
9
.
Wengelnik K, Vidal V, Ancelin ML, Cathiard AM, Morgat JL, Kocken CH, Calas M,
Herrera S, Thomas AW & Vial HJ (2002) A class of potent antimalarials and their specific
accumulation in infected erythrocytes. Science 295, 1311-1314.
1
0.
Vial HJ, Wein S, Farenc C, Kocken C, Nicolas O, Ancelin ML, Bressolle F, Thomas A &
Calas M (2004) Prodrugs of bisthiazolium salts are orally potent antimalarials. Proceedings of the
National Academy of Sciences, USA 101, 15458-15463.
1
1.
Wein S, Maynadier M, Bordat Y, Perez J, Maheshwari S, Bette-Bobillo P, Tran Van Ba C,
Penarete-Vargas D, Fraisse L, Cerdan R, et al. (2012) Transport and pharmacodynamics of
albitazolium,
a
candidate antimalarial drug. Br
J
Pharmacol, doi: 10.1111/j.1476-
5
1
381.2012.01966.x.
2.
Dechamps S, Wengelnik K, Berry-Sterkers L, Cerdan R, Vial HJ & Gannoun-Zaki L
(2010) The Kennedy phospholipid biosynthesis pathways are refractory to genetic disruption in
Plasmodium berghei and therefore appear essential in blood stages. Molecular and biochemical
parasitology 173, 69-80, doi: 10.1016/j.molbiopara.2010.05.006.
1
3.
Jackowski S (1994) Coordination of membrane phospholipid synthesis with the cell cycle.
J Biol Chem 269, 3858-3867.
1
3
4.
Kent C (1995) Eukaryotic phospholipid biosynthesis. Annual Review of Biochemistry 64,
15-343, doi: 10.1146/annurev.bi.64.070195.001531.
This article is protected by copyright. All rights reserved.

1
5.
Kent C (2005) Regulatory enzymes of phosphatidylcholine biosynthesis: a personal
perspective. Biochim Biophys Acta 1733, 53-66, doi: S1388-1981(04)00223-9 [pii]
1
1
0.1016/j.bbalip.2004.12.008.
6. Fagone P & Jackowski S (2012) Phosphatidylcholine and the CDP-choline cycle. Biochim
Biophys Acta, doi: S1388-1981(12)00206-5 [pii]
1
1
0.1016/j.bbalip.2012.09.009.
7.
Taneva S, Dennis MK, Ding Z, Smith JL & Cornell RB (2008) Contribution of each
membrane binding domain of the CTP:phosphocholine cytidylyltransferase-alpha dimer to its
activation, membrane binding, and membrane cross-bridging. J Biol Chem 283, 28137-28148, doi:
1
0.1074/jbc.M802595200
M802595200 [pii].
8.
1
Yeo HJ, Larvor MP, Ancelin ML & Vial HJ (1997) Plasmodium falciparum
CTP:phosphocholine cytidylyltransferase expressed in Escherichia coli: purification,
characterization and lipid regulation. Biochem J 324 ( Pt 3), 903-910.
1
9.
Yeo HJ, Sri Widada J, Mercereau-Puijalon O & Vial HJ (1995) Molecular cloning of
CTP:phosphocholine cytidylyltransferase from Plasmodium falciparum. Eur J Biochem 233, 62-72.
0. Llinas M, Bozdech Z, Wong ED, Adai AT & DeRisi JL (2006) Comparative whole
2
genome transcriptome analysis of three Plasmodium falciparum strains. Nucleic Acids Res 34,
1
1
2
166-1173, doi: 34/4/1166 [pii]
0.1093/nar/gkj517.
1.
Friesen JA, Campbell HA & Kent C (1999) Enzymatic and cellular characterization of a
catalytic fragment of CTP:phosphocholine cytidylyltransferase alpha. J Biol Chem 274, 13384-
1
2
3389.
2.
Veitch DP, Gilham D & Cornell RB (1998) The role of histidine residues in the HXGH
site of CTP:phosphocholine cytidylyltransferase in CTP binding and catalysis. Eur J Biochem 255,
2
27-234.
3.
2
Lee J, Johnson J, Ding Z, Paetzel M & Cornell RB (2009) Crystal structure of a
mammalian CTP: phosphocholine cytidylyltransferase catalytic domain reveals novel active site
residues within a highly conserved nucleotidyltransferase fold. J Biol Chem 284, 33535-33548, doi:
M109.053363 [pii]
1
2
0.1074/jbc.M109.053363.
4.
Weber CH, Park YS, Sanker S, Kent C & Ludwig ML (1999) A prototypical
cytidylyltransferase: CTP:glycerol-3-phosphate cytidylyltransferase from bacillus subtilis. Structure
7
, 1113-1124, doi: st7906 [pii].
2
5.
Cornillot E, Hadj-Kaddour K, Dassouli A, Noel B, Ranwez V, Vacherie B, Augagneur Y,
Bres V, Duclos A, Randazzo S, et al. (2012) Sequencing of the smallest Apicomplexan genome
from the human pathogen Babesia microti. Nucleic Acids Res 40, 9102-9114, doi:
1
0.1093/nar/gks700
gks700 [pii].
6. Loo JA (2000) Electrospray ionization mass spectrometry: a technology for studying
2
noncovalent macromolecular complexes. Int J Mass Spectrom 200, 175-186, doi: Doi
1
2
0.1016/S1387-3806(00)00298-0.
7. Ashcroft AE (2005) Recent developments in electrospray ionisation mass spectrometry:
noncovalently bound protein complexes. Nat Prod Rep 22, 452-464, doi: Doi 10.1039/B417724j.
2
8.
Webb MR (1992) A continuous spectrophotometric assay for inorganic phosphate and
for measuring phosphate release kinetics in biological systems. Proc Natl Acad Sci U S A 89,
4
2
884-4887.
9. Lloyd AJ, Thomann HU, Ibba M & Soll D (1995) A broadly applicable continuous
spectrophotometric assay for measuring aminoacyl-tRNA synthetase activity. Nucleic Acids Res 23,
2
886-2892, doi: 5b0116 [pii].
This article is protected by copyright. All rights reserved.

3
0.
Varga B, Barabas O, Takacs E, Nagy N, Nagy P & Vertessy BG (2008) Active site of
mycobacterial dUTPase: structural characteristics and a built-in sensor. Biochem Biophys Res
Commun 373, 8-13, doi: S0006-291X(08)01015-2 [pii]
1
0.1016/j.bbrc.2008.05.130.
1.
3
Fromm H (1975) Initial rate enzyme kinetics. 1st Edition edn. Springer-Verlag, Berlin
Heidelberg New York.
3
2.
Stevens SY, Sanker S, Kent C & Zuiderweg ER (2001) Delineation of the allosteric
mechanism of a cytidylyltransferase exhibiting negative cooperativity. Nat Struct Biol 8, 947-952,
doi: 10.1038/nsb1101-947
nsb1101-947 [pii].
3
3.
Christensen T, Gooden DM, Kung JE & Toone EJ (2003) Additivity and the physical
basis of multivalency effects: a thermodynamic investigation of the calcium EDTA interaction. J
Am Chem Soc 125, 7357-7366, doi: 10.1021/ja021240c.
3
4.
Mages F, Rey C, Fonlupt P & Pacheco H (1988) Kinetic and biochemical properties of
CTP:choline-phosphate cytidylyltransferase from the rat brain. Eur J Biochem 178, 367-372.
3
5.
Tilley DM, Evans CR, Larson TM, Edwards KA & Friesen JA (2008) Identification and
characterization of the nuclear isoform of Drosophila melanogaster CTP:phosphocholine
cytidylyltransferase. Biochemistry 47, 11838-11846, doi: 10.1021/bi801161s.
3
6.
Frieden C (1976) On the kinetic distinction of ordered and random bireactant enzyme
systems. Biochem Biophys Res Commun 68, 914-917, doi: 0006-291X(76)91232-8 [pii].
3
7.
Huang HK, Taneva SG, Lee J, Silva LP, Schriemer DC & Cornell RB (2012) The
Membrane-Binding Domain of an Amphitropic Enzyme Suppresses Catalysis by Contact with
an Amphipathic Helix Flanking Its Active Site. J Mol Biol, doi: S0022-2836(12)00917-5 [pii]
1
3
0.1016/j.jmb.2012.12.003.
8.
Bork P, Holm L, Koonin EV & Sander C (1995) The cytidylyltransferase superfamily:
identification of the nucleotide-binding site and fold prediction. Proteins 22, 259-266, doi:
1
3
0.1002/prot.340220306.
9.
Kapustina M, Weinreb V, Li L, Kuhlman B & Carter CW, Jr. (2007) A conformational
transition state accompanies tryptophan activation by B. stearothermophilus tryptophanyl-tRNA
synthetase. Structure 15, 1272-1284, doi: S0969-2126(07)00324-3 [pii]
1
4
0.1016/j.str.2007.08.010.
0.
Do CB, Mahabhashyam MS, Brudno M & Batzoglou S (2005) ProbCons: Probabilistic
consistency-based multiple sequence alignment. Genome Res 15, 330-340, doi: 15/2/330 [pii]
1
4
0.1101/gr.2821705.
1.
Dereeper A, Guignon V, Blanc G, Audic S, Buffet S, Chevenet F, Dufayard JF, Guindon
S, Lefort V, Lescot M, et al. (2008) Phylogeny.fr: robust phylogenetic analysis for the non-
specialist. Nucleic Acids Res 36, W465-469, doi: 10.1093/nar/gkn180
gkn180 [pii].
4
2.
Loytynoja A & Goldman N (2010) webPRANK: a phylogeny-aware multiple sequence
aligner with interactive alignment browser. BMC Bioinformatics 11, 579, doi: 10.1186/1471-2105-
1
1
4
1-579
471-2105-11-579 [pii].
3.
Castresana J (2000) Selection of conserved blocks from multiple alignments for their use
in phylogenetic analysis. Mol Biol Evol 17, 540-552.
4
4.
Guindon S, Delsuc F, Dufayard JF & Gascuel O (2009) Estimating maximum likelihood
phylogenies with PhyML. Methods Mol Biol 537, 113-137, doi: 10.1007/978-1-59745-251-9_6.
4
5.
Darriba D, Taboada GL, Doallo R & Posada D (2011) ProtTest 3: fast selection of best-
fit models of protein evolution. Bioinformatics 27, 1164-1165, doi: 10.1093/bioinformatics/btr088
btr088 [pii].
This article is protected by copyright. All rights reserved.

4
6.
Eswar N, Webb B, Marti-Renom MA, Madhusudhan MS, Eramian D, Shen MY, Pieper
U & Sali A (2006) Comparative protein structure modeling using Modeller. Curr Protoc
Bioinformatics Chapter 5, Unit 5 6, doi: 10.1002/0471250953.bi0506s15.
4
7.
MacKerell AD, Jr., Banavali N & Foloppe N (2000) Development and current status of
the CHARMM force field for nucleic acids. Biopolymers 56, 257-265, doi: 10.1002/1097-
0
1
4
282(2000)56:4<257::AID-BIP10029>3.0.CO;2-W [pii]
0.1002/1097-0282(2000)56:4<257::AID-BIP10029>3.0.CO;2-W.
8.
Brooks BR, Brooks CL, 3rd, Mackerell AD, Jr., Nilsson L, Petrella RJ, Roux B, Won Y,
Archontis G, Bartels C, Boresch S, et al. (2009) CHARMM: the biomolecular simulation
program. J Comput Chem 30, 1545-1614, doi: 10.1002/jcc.21287.
4
9.
Laskowski RA, Rullmannn JA, MacArthur MW, Kaptein R & Thornton JM (1996)
AQUA and PROCHECK-NMR: programs for checking the quality of protein structures solved
by NMR. J Biomol NMR 8, 477-486.
5
3
5
0.
Hooft RW, Vriend G, Sander C & Abola EE (1996) Errors in protein structures. Nature
81, 272, doi: 10.1038/381272a0.
1. Colovos C & Yeates TO (1993) Verification of protein structures: patterns of nonbonded
atomic interactions. Protein Sci 2, 1511-1519, doi: 10.1002/pro.5560020916.
5
2.
Artimo P, Jonnalagedda M, Arnold K, Baratin D, Csardi G, de Castro E, Duvaud S,
Flegel V, Fortier A, Gasteiger E, et al. (2012) ExPASy: SIB bioinformatics resource portal.
Nucleic Acids Res 40, W597-603, doi: gks400 [pii]
1
5
0.1093/nar/gks400.
3.
Pecsi I, Szabo JE, Adams SD, Simon I, Sellers JR, Vertessy BG & Toth J (2011)
Nucleotide pyrophosphatase employs a P-loop-like motif to enhance catalytic power and
NDP/NTP discrimination. Proc Natl Acad Sci U S A 108, 14437-14442, doi: 1013872108 [pii]
1
5
0.1073/pnas.1013872108.
4.
Pecsi I, Leveles I, Harmat V, Vertessy BG & Toth J (2010) Aromatic stacking between
nucleobase and enzyme promotes phosphate ester hydrolysis in dUTPase. Nucleic Acids Res 38,
7
1
5
179-7186, doi: gkq584 [pii]
0.1093/nar/gkq584.
5.
Varga B, Barabas O, Kovari J, Toth J, Hunyadi-Gulyas E, Klement E, Medzihradszky
KF, Tolgyesi F, Fidy J & Vertessy BG (2007) Active site closure facilitates juxtaposition of
reactant atoms for initiation of catalysis by human dUTPase. FEBS Lett 581, 4783-4788, doi:
S0014-5793(07)00971-4 [pii]
1
5
0.1016/j.febslet.2007.09.005.
6.
Finn RD, Mistry J, Tate J, Coggill P, Heger A, Pollington JE, Gavin OL, Gunasekaran P,
Ceric G, Forslund K, et al. (2010) The Pfam protein families database. Nucleic Acids Res 38,
D211-222, doi: gkp985 [pii]
1
5
0.1093/nar/gkp985.
7.
Yang ZR, Thomson R, McNeil P & Esnouf RM (2005) RONN: the bio-basis function
neural network technique applied to the detection of natively disordered regions in proteins.
Bioinformatics 21, 3369-3376, doi: bti534 [pii]
1
0.1093/bioinformatics/bti534.
5
8.
Dosztanyi Z, Csizmok V, Tompa P & Simon I (2005) The pairwise energy content
estimated from amino acid composition discriminates between folded and intrinsically
unstructured proteins. J Mol Biol 347, 827-839, doi: S0022-2836(05)00129-4 [pii]
1
5
0.1016/j.jmb.2005.01.071.
9.
Ward JJ, McGuffin LJ, Bryson K, Buxton BF & Jones DT (2004) The DISOPRED
server for the prediction of protein disorder. Bioinformatics 20, 2138-2139, doi:
1
0.1093/bioinformatics/bth195
bth195 [pii].
This article is protected by copyright. All rights reserved.

Supplementary Material
Supplementary Material to this article contains
1
2
3
4
5
Details of construction and validation of homology models
Supplementary Figures S1, S2, S3, S4, S5
Supplementary Tables S1, S2, S3
Topology file and non-standard parameters added for the CDPCho ligand
Supplementary references
Tables for the manuscript:
„
Evolutionary and mechanistic insights into substrate and product accommodation of
CTP:phosphocholine cytidylyltransferase from Plasmodium falciparum”
1
*
1
2
2
3
Gergely N. Nagy , Lívia Marton , Balázs Krámos , Julianna Oláh , Ágnes Révész , Károly
3
4
5
5,6
Vékey , Frédéric Delsuc , Éva Hunyadi-Gulyás , Katalin F. Medzihradszky , Marina
7
7
7
1,8*
Lavigne , Henri Vial , Rachel Cerdan , Beáta G. Vértessy
a
Table 1 Kinetic parameters of PfCCT catalysis
CTP titration
ChoP titration
V
max
k
cat
K
M,CTP
k
cat
K
M,ChoP
K
I,ChoP
b
-1
-1
(
U /mg)
(s )
(μM)
(s )
(mM)
(mM)
2
810
1.45 ± 0.05 168 ± 17
1.2 ± 0.4
1.8 ± 1.1
10.5 ± 7.5
a
Kinetic parameters were obtained as described in the Materials and Methods.
b
One enzyme unit (U) is the number of nanomols of CDPCho produced per minute.
This article is protected by copyright. All rights reserved.

Table 2 Thermodynamic parameters of ligand binding to PfCCT MΔK constructs determined by direct and displacement titrations at
0°C
2
PfCCT construct
K
d
ΔH
ΔS
2
+
Mg
N
-1
-1 -1
(
μM)
0.88 ± 0.03 88.5 ± 7.2 -24.0 ± 1.0
0.62 ± 0.01 294 ± 38 -9.6 ± 1.7
(kcal mol ) (cal mol K )
MΔK
MΔK
MΔK
CDPCho
CTP
5 mM
5 mM
-63.3 ± 19.0
-16.6 ± 3.6
CDPCho (5 mM
5
mM
0.85 ± 0.01 88.5 ± 3.4 -23.2 ± 0.4
-60.5 ± 2.5
a
ChoP present)
MΔK
MΔK
MΔK
CDPCho
CTP
none
none
0.57 ± 0.02 44.4 ± 3.3 -22.9 ± 1.0
0.96 ± 0.02 59.5 ± 1.8 -8.6 ± 0.3
-58.1 ± 5.0
-10.1 ± 0.5
CDPCho (2 mM
none
0.75 ± 0.01 42.9 ± 0.8 -20.7 ± 0.1
0.96 ± 0.01 67.7 ± 0.9 -13.5 ± 0.1
-50.7 ± 1.0
a
ChoP present)
W743H
MΔK
CDPCho
CDPCho
none
none
-27.0 ± 0.4
-16.8 ± 8.6
W692Y
MΔK
0.83 ±0.24
218 ± 50
-9.8 ± 4.5
a
displacement titrations
This article is protected by copyright. All rights reserved.

Fig. 1 Residue conservation pattern within duplicated CCT catalytic domains in
Apicomplexa
Alignment of cytidylyltransferase catalytic domains sequences, whose boundaries are
identified by the Pfam server [56]. RAT, Rattus norvegicus CCT (Uniprot entry: P19836),
(
residues 80-208); TOXGO, Toxoplasma gondii, (B6KSU6) (54-203); PLAF7, Plasmodium
falciparum 3D7 strain (Q8IEE9) (36-182) and (621-767); PLABA, Plasmodium berghei
Q4YBQ9), (29-183), (588-736); BABMI, Babesia microti (CCF73098) (16-160), (314-510);
BABBO, Babesia bovis (A7AU63) (10-158), (323-448); THEAN, Theileria annulata
Q4UFG5), (1-131), (267-394); THEPA, Theileria parva (Q4N412) (13-150), (286- 413),
(
(
This article is protected by copyright. All rights reserved.

respectively. The alignment was initially performed by ClustalW, and then manually
adjusted. Residues conserved among cytidylyltransferase superfamily (GCT, ECT, CCT
enzymes) are highlighted with white letters on dark grey box, residues conserved among CCT
enzymes are highlighted with a light grey box. The positions of the W residues within the
PfCCT sequences are shown by grey and black arrows, the latter indicating the conserved W
residue bordering the active site.
Fig. 2 Phylogenetic analyses of domain duplication within Apicomplexa parasites
A) Maximum-likelihood phylogenetic tree of the 16 full length CCT Apicomplexan proteins.
Numbers at nodes are bootstrap percentages with bullets indicating a value of more than 95.
Scale is in mean number of substitutions per site. Structures of the CCT gene with location of
CMP segments in the main Ampicoplexan groups are represented following Dechamps et al.
[
5].
B) Maximum-likelihood phylogenetic tree of the 27 CMP protein segments of the CCT gene
in Apicomplexa. Numbers at nodes are bootstrap percentages, bullets indicating a value of
more than 95. Scale is in mean number of substitutions per site. Nodes without numbers have
bootstrap percentages less than 70. D1 and D2 indicate the two independent duplication
events required to explain the inferred phylogenetic pattern.
This article is protected by copyright. All rights reserved.

Fig. 3 Structural model of the PfCCT MΔK construct
A) Superimposed homology model of the homodimer forms of PfCCT MΔK (528-795, delta
7
20-737) (shown in cartoon model, coloured blue-white) and PfCCT (528-795) (shown in
cartoon model, coloured wheat) constructs, containing the C2-domain are superimposed. The
model were built with the Modeller program using the rat CCT crystal structure (PDB ID:
3
HL4) as template. The CDPCho ligand is shown in sticks, using atomic colouring (carbon,
yellow; oxygen, red; phosphorus, orange; nitrogen, blue). The tryptophans W682, W692,
W743 are showed in sticks, coloured in blue.
B) Disorder profile of full length PfCCT. The plot shows the probability of disorder predicted
by RONN [57], IUPred [58], and Disopred [59]. C1 and C2 domains, coloured in green, were
classified as ordered by all three predictors. The lysine-rich segments (K, coloured in cyan)
within the ordered conserved catalytic domains are predicted to be highly flexible.
This article is protected by copyright. All rights reserved.

Fig. 4 PfCCT MΔK is predominantly present in homodimer form in solution
A) Purification steps of PfCCT M K followed by 12% sodium dodecyl sulphate-
Δ
polyacrylamide gel electrophoresis. Lane 1, molecular mass standards; lane 2, E. coli extract
;
lane 3, sample eluted from Ni affinity purification; lane 4, sample after gel filtration.
B) Mass spectrum of PfCCT MΔK protein under native electrospray conditions. M and D
denotes monomer and dimer signals, respectively; whereas numbers indicate the charge
states. The two insets enlarged graphs for the monomer and dimer ion series regions, with
determined molar masses 30,870 (± 200) Da and 61,840 (± 200) Da, respectively.
This article is protected by copyright. All rights reserved.