pCR2.1-TOPO, pETDuet-1, pGEX4T3, and pGADT7 plasmids were purchased from Invitrogen, Novagen, Life Sciences, and Clontech respectively. pCAM-BSD and pCAM-BSD-HA plasmids were kind gifts from Dr. C. Doerig (Monash University, Melbourne, Australia).

Monoclonal anti-HA, anti-penta His, anti-GST, anti-H3, and anti-HA peroxidase antibodies were purchased from Roche, Qiagen, Sigma-Aldrich, Millipore, and Abcam respectively. Anti-actin1 and anti-SOD1 antibodies were used as previously described (Daher et al., 2006, 2010).

Putative eIF2β, eIF2γ, and eIF5 sequences were searched using BLASTp on sequences available in PlasmoDB databases. Protein sequences (human and P. falciparum) were aligned using the ClustalW program.

The phylogenetic analysis of eukaryotic initiation factor eIF2β was performed using 76 amino acid sequences of eIF2β from 23 apicomplexans, 5 mammals, 1 amphibian, 8 fish, 9 plants, 8 arthropods, 12 archaea, 1 yeast (fungi), 2 amoebozoa, 1 cercozoa, 1 foraminifera, and 5 excavata. Before phylogenetic analysis, sequences were submitted to the protein families database (Pfam, Finn et al., 2014) and conserved domains (CDD, Marchler-Bauer et al., 2014) for family and domain assignment, respectively. All sequences included in further analysis belonged to the eIF5/eIF2B family (pfam01873), and contained the characteristic structural domain eIF2β (Accession: PRK03988).

The sequences were aligned with MAFFT (v7), configured for the highest accuracy (Katoh and Standley, 2013). After alignment, ambiguous regions were removed with Gblocks (v0.91b) (Castresana, 2000). The final alignment contained 125 gap free amino acid positions. All evolutionary and phylogenetic analysis were performed in the Molecular Evolutionary Genetics Analysis (MEGA, v6) software (Tamura et al., 2013). The best-fit model of sequence evolution was selected based on Bayesian Information Criterion (BIC) and Akaike Information Criterion, corrected (AICc). The JTT (Jones et al., 1992) model with a proportion of Gamma distributed (G) and invariant (I) sites which had the lowest values of BIC and AICc was chosen as best-fitting model for the actual data. The evolutionary history was inferred using the Maximum Likelihood method based on JTT model and assuming a proportion of Gamma distributed (with shape parameter α = 1.19) and invariant sites (= 0.02). Initial trees for the heuristic search (Nearest-Neighbor-Interchange: NNI) were obtained by applying Neighbor-Joining algorithms and the topology with higher log likelihood value was selected. The reliability of the internal branches was assessed using the Bootstrap methods with 1000 replicates. Graphical representation and editing of the phylogenetic trees were performed with MEGA. The accession numbers of the sequences are provided in the phylogenetic tree (Figure 2 and Table S2).

The Robetta server (http://robetta.bakerlab.org/) was used for tertiary modeling. The best structure was ranked using the RESPROX (Berjanskii et al., 2012) qualifying server. Finally, the structure was prepared using the Schrodinger's Maestro (Schrödinger: Schrödinger maestro Package In: maestro, version 99. New York: LLC; 2014) package Protein Preparation Wizard. The Protein Databank (PDB) crystal structure of Pyrococcus furiosus (PDB: 2DCU, chain B) was also prepared and used for a structural alignment of the two initiation factors (implemented in Maestro). P. furiosus (PDB: 2DCU, chain B) was chosen since this was the closest structural homolog to P. falciparum eIF2β according to the Dali server (Holm and Rosenström, 2010).

P. falciparum 3D7 clone was grown according to Trager et al. (Trager and Jensen, 1976), with slight modification (Fréville et al., 2012). Parasites were synchronized by a double sorbitol treatment as previously described (Vernes et al., 1984).

To isolate total DNA or proteins, parasitized erythrocytes were lysed by saponin (Umlas and Fallon, 1971) and pelleted. Soluble proteins extracts were prepared according to Fréville et al. (2013).

Genomic DNA (gDNA) was extracted using the KAPA Express Extract kit (KAPABioSystem) as described in the manufacturer's protocol.

The coding regions of PfeIF2β and PfeIF2γ were obtained by RT-PCR using the primers described in Table S1 and subcloned in E. coli expression vectors pETDuet-1 and pGEX4T3 which allows the expression of proteins fused with a 6-His- or GST-tag respectively. For the expression of PfeIF5 or PfeIF2β in Xenopus oocytes, they were amplified with p12–p13 and p14–p15 respectively (Table S1) and subcloned in pGADT7 vector allowing the production of capped RNA (cRNA) using the T7 promoter. cRNA was obtained as previously described (Fréville et al., 2014) and used for the expression of HA tagged proteins in oocytes.

For mutant constructs of PfeIF2β, PCR-based site-directed mutagenesis (Qbiogene) was used. The pETDuet-PfeIF2β plasmid was used as template with primers p4–p5 and p6–p7 (Table S1) to obtain PfeIF2β-²⁹AGEAKA³⁴ and PfeIF2β-¹⁰³KAAA¹⁰⁶ respectively. For the PfeIF2β-²⁹AGEAKA³⁴/¹⁰³KAAA¹⁰⁶ mutant construct, it was obtained by PCR using the primers p6–p7 and the PfeIF2β-²⁹AGEAKA³⁴ plasmid as template. All the open reading frames (ORFs) as well as the mutation points were checked by sequencing.

Expression and purification of PfPP1 was previously described (Fréville et al., 2013). The expression of wild PfeIF2β recombinant protein and the mutated versions were carried out in the E. coli BL21 strain as previously described (Fréville et al., 2013) with slight modifications. Briefly, inductions were carried out at 30°C for 3 h in the presence of 50 μM ZnCl₂ and tagged-proteins purifications were done under non-denaturing conditions as described by manufacturers' protocol using Ni-NTA agarose beads (Qiagen). The purity, checked by 15% SDS-PAGE followed by SimplyBlue™ safe staining (Invitrogen), was >90%. Recombinant PfeIF2β-6His protein was subjeicted to peptide mass fingerprint by MALDI-TOF mass spectrometry to confirm its identity.

For the GST tagged proteins, the expression was induced at 37°C for 3 h in the presence of 0.5 mM IPTG and 1 mM MnCl2 for PfPP1-GST or 25 mM MgCl2 and 50 μM ZnCl2 for the expression of PfeIf2β-GST and PfeIf2γ-GST. Proteins were purified according to the manufacturers' instructions using glutathione Sepharose beads (Sigma). For GST pull down experiments, purified recombinant proteins were bound to glutathione-Sepharose beads overnight at 4°C and washed with a buffer containing 20 mM Tris-HCl pH7.4, 500 mM NaCl, 50 μM ZnCl_2, and 0.1% Triton X-100 before use.

The antisera anti-PfeIF2β was raised according to the protocol previously described (Fréville et al., 2013).

Two μg of PfeIF2β-6His recombinant protein (wild-type or mutated) were incubated with PfPP1-GST, PfeIF2γ-GST or GST bound to gluthatione-Sepharose beads, and 25 μg of BSA in binding buffer (20 mM Tris-HCl pH7.4, 500 mM NaCl, 20 mM Hepes, 0.2 mM EDTA, 0.1% Triton X-100, 1 mM DTT, protease cocktail inhibitor and 1 mM MnCl₂, 50 μM ZnCl₂ or 25 mM MgCl₂ according to uses proteins) for 1–2 h at 4°C on wheel. After five washes with binding buffer, proteins were eluted in loading buffer, separated on SDS-PAGE and blotted to nitrocellulose. Blots were revealed with anti-His or anti-GST mAb antibodies. Horseradish peroxidase labeled anti-mouse (1:50,000) was used as secondary antibodies followed by chemiluminescence detection (DURA, Pierce).

Preparation of Xenopus oocytes and microinjection experiments were performed as previously described (Vicogne et al., 2004). Briefly, in each assay, 20 oocytes removed from at least two or three different animals were microinjected with PfeIF2β-6His (wild-type or mutated) recombinant protein. Progesterone was used as a positive control for oocyte maturation. GVBD was detected by the appearance of a white spot at the apex of the animal pole after 15 h. In order to carry out immunoprecipitation, extracts from 20 oocytes removed from at least two or three animals were prepared 15 min after the microinjection of PfeIF2β (wild-type or mutated) as previously described (Vandomme et al., 2014).

Xenopus oocytes were also used in order to test the interaction of PfeIF2β with its partners PfeIF2γ and PfeIF5. cRNA corresponding to HA-PfeIF2β or HA-PfeIF5 proteins was microinjected followed by the microinjection of His-PfeIF2γ or His-PfeIF2β recombinant proteins respectively. Protein extractions, immunoprecipitation and immunoblots experiments were carried out as previously described (Vandomme et al., 2014).

To examine the interaction of PfeIF2β with XePP1, the experiments were performed as previously described (Fréville et al., 2013).

For western blots, 10 μg/ lane of P. falciparum soluble proteins from asynchronous cultures was separated on a 4–20% SDS-PAGE and blotted onto nitrocellulose. For the detection of PfeIF2β, blots were probed with primary mouse anti-PfeIF2β serum (1:1000 in PBS milk 5%).

The detection of native PfeIF2β in total proteins extracted from asynchronous cultures of parasites was carried out by using PfPP1-6His beads. After pre-clearing on Ni-NTA agarose beads, 2 mg of total protein extracts were incubated at 4°C overnight with PfPP1-6His beads. After washing steps, proteins were eluted with SDS-PAGE loading buffer, separated by SDS-PAGE and blotted to nitrocellulose. Blots were revealed with pre-immune serum, anti-PfeIF2β serum, or anti-His mAb antibodies. Horseradish peroxidase labeled anti-mouse (1:50,000) was used as secondary antibodies followed by chemiluminescence detection (DURA, Pierce).

The PfeIF2β disruption plasmid (pCAM-PfeIF2β) was generated by the insertion of a PCR product corresponding to a 5′ portion from the PfeIF2β sequence (846 bp) into the pCAM-BSD vector which contains a cassette conferring resistance to blasticidin. The insert was obtained using 3D7 genomic DNA as template and the primers p18–p19 (with PstI and BamHI sites respectively, Table S1). Attempts to check the accessibility of the PfeIF2β locus were performed by transfecting wild 3D7 parasites with 3′ tagging constructs. To this end, the 3′ end of the PfeIF2β sequence (845 bp, omitting the stop codon) was amplified by PCR using 3D7 genomic DNA and the primers p16–p17 (containing PstI and BamHI restriction sites respectively, Table S1). The 3′ tagging plasmids were generated by inserting the PCR product into the pCAM-BSD-hemagglutinin (HA). Ring stage 3D7 parasites were transfected with 100 μg of plasmid DNA by electroporation, according to Sidhu et al. (2005). To select transformed parasites, 48 h after transfection, blasticidin antibiotic (Invivogen) was added to a final concentration 2.5 μg/ml. Resistant parasites appeared after 4–6 weeks and were maintained under drug selection.

To verify the presence of correct constructs in transfected parasites, plasmid rescue experiments were carried out. Genomic DNA extracts (KAPA Express Extract) from wild-type or transfected parasites were used to transform DH5α E. coli cells (Invitrogen). Plasmid DNA was purified from bacterial clones and digested with restriction enzymes (PstI and BamHI).

Genotypes of Pfeif2β Knock-Out parasites were analyzed by PCR on genomic DNA using the primers p24 (derived from the 5′ non-translated region and absent in the construct) and p30 (Table S1) specific for the pCAM-BSD vector. Genotypes of Pfeif2β Knock-In parasites were analyzed using primers p24–p31 (reverse primer corresponding to HA tag, Table S1).

Five milliliters of unsynchronized blood cultures of parasites P. falciparum 3D7 or PfeIF2β-HA recombinant strain (5% parasitemia) were centrifuged (700 g, 5 min), and the pellet was fixed 24 h with 10% neutral formalin and paraffin embedded. Morphological assessment was obtained by examining sections (4 μm) stained with hematoxylin-eosin-safran.

Immunofluorescence assays (IFA) were done to immunolocalize PfeIF2β-HA tag recombinant proteins. Sections of P. falciparum eIF2β-HA tag recombinant strain were incubated for 1 h at 37°C with an anti-HA tag (biotine) rabbit polyclonal antibody (1:100 dilution; Abcam). After washing in PBS, sections were incubated with streptavidine-Alexa fluor 488-labeled (1:200; Invitrogen) added with 4′,6-diamidino-2-phenylindole dihydrochloride (DAPI, 0.2 μg/ml; Invitrogen) for 1 h at 37°C. After washing, slides were mounted with an anti-fade mounting medium (Bio-Rad) and analyzed using a Zeiss LSM880 confocal microscope (Zeiss).

The cytoplasmic and nuclear extracts were prepared as previously described (Voss et al., 2002). Sorbitol synchronized parasites at ring-stage (10–15 hpi), trophozoite-stage (22–28 hpi), and schizont-stage (40–42 hpi) were incubated in lysis buffer containing 20 mM HEPES pH7.8, 10 mM KCl, 1 mM EDTA, 1 mM DTT, 1% Triton X-100, and protease inhibitor cocktail for 5 min on ice. After centrifugation at 2500 g, the supernatants corresponding to the cytoplasmic fractions were recovered. To extract the nuclei, the pellets were washed twice with the lysis buffer and were further incubated in extraction buffer containing 20 mM HEPES pH7.8, 800 mM KCl, 1 mM EDTA, 1 mM DTT, and protease inhibitor at 4°C on a rotator for 30 min. The nuclear extracts were centrifuged at 12,000 g for 30 min and the supernatants corresponding to the nuclear extracts were harvested. Ten micrograms were used in Western blot assays as described above. The band intensities were quantified using the Image quant TL8.1 software (GE Healthcare, Imager Las 4000) in which the band intensity of ring nuclear fraction. The results represent the mean of fold change between nuclear and cytoplasm fractions.

The eIF2β gene bioinformatically identified in the P. falciparum genome (PfeIF2β) (accession number PF3D7_1010600) encodes a sequence of 222 amino acids, which is 111 amino acids shorter than human eIF2β. To clearly identify the ORF of this gene, PCRs using internal primers of the coding sequence and primers (Table S1) derived from genomic DNA were performed on cDNA obtained from total RNA. This approach determined the start and stop codons with an ORF of 666 nucleotides (Figure S1). The deduced sequence of 222 amino acids, with an expected molecular mass of 25.3 kDa, showed an overall identity with human eIF2β of 47% (Figure 1A). This identity is mainly observed at the C-terminus (nearly 150 residues) which contains a conserved region also present in the eIF5 family. Closer examination of PfeIF2β, revealed the presence of one lysine block and one GTP binding motif while the human counterpart contains 3 lysine blocks, known to bind to mRNA, and 2 GTP binding motifs (Figure 1A). In addition, PfeIF2β possesses one canonical binding motif, RVxF (¹⁰³KVAW¹⁰⁶) as well as the FxxR/KxR/K (²⁹FGEKKK³⁴) motif, described to be involved in binding to PP1. The presence of four conserved cysteines in PfeIF2β (Figure 1A), known to be involved in the structural stability of the C-terminus in the presence of Zinc ion (Gutiérrez et al., 2002), should also be noted.

To further confirm the expression of eIF2β by P. falciparum at its expected size, an antiserum raised against the His-tagged recombinant protein (Figure S2) was used in immunoblots on a total extract of blood parasites from an asynchronous culture. As shown in Figure 1B (lane 2), the immunoblot showed a specific band at 25 kDa confirming the expected size of PfeIF2β and excluding a longer form of this protein. Proteomic analysis available at the PlasmoDB revealed that PfeIF2β is expressed during the intraerythrocytic development of P. falciparum (Pease et al., 2013).

All sequences included in the phylogenetic analysis belonged to the eIF-5/eIF-2B family (pfam01873), and contained the characteristic structural domain eIF-2B (Accession: PRK03988). The evolutionary history of 76 eIF2β amino acid sequences from apicomplexan, mammals, amphibian, fish, plants, arthropods, archaea, fungi, amoebozoa, cercozoa, foraminifera, and excavata was inferred using the maximum likelihood method based on the JTT+G+I model (see Section Materials and Methods). The tree topology with the highest log likelihood is shown in Figure 2A. The eIF2β of all independent taxa included in the analysis formed monophyletic clades, except for archaea that was fractioned in two clusters. The topology of tree reflects well the current classification of Eukaryotes (Adl et al., 2012). For example, the eIF2β sequences from Apicomplexan (Alveolata), Cercozoa and Foraminifera clustered together. Members of Metazoa (Opisthokonta) formed a monophyletic clade, but the fungi Schizosaccharomyces pombe (Opisthokonta) clustered together with Plantae (Archaeplastida) and no with the Metazoan as expected (Figure 2A). However, the molecular structure of S. pombe eIF2β resembles that of Metazoan and no Plantae. A remarkable difference between Opisthokonta and Archaeplastida eIF2β is that while members of Opisthokonta present an N-terminus with two lysine blocks, members of Plantae lack this domain (Figure 2B). The N-terminus is also absent in Apicomplexa, Amoebozoa, Cercozoa, and Archaea. Consequently, these groups of organisms do not present the additional two lysine blocks found in the N-terminus of Opisthokonta. It is noteworthy that although Reticulomyxa filose (Foraminifera) and Trypanosoma spp. (Excavata) present an N-terminus, it is not similar to that found in Opisthokonta (Figure 2B). From this phylogenetic tree, the most parsimonious explanation for the evolution of the N-terminal extension is that it evolved independently in Opisthokonta, Excavata, and Foraminifera probably and may have different functions within each of these groups. In agreement with this, the eIF2β from Excavata and Foraminifera do not present the typical lysine blocks found in all the other Eukaryotes. Remarkably, despite the fact that Apicomplexa eIF2β lack the N-terminus, it presents the two main PP1 binding domains found in Mammals and Amphibians (FxxR/KxR/K and R/KxV/IxF/W), which suggests a conserved function (Figure 2B).

Using the Multiple Mapping Method (http://www.fiserlab.org/servers/M4T), we observed that the putative tertiary structure of P. falciparum eIF2β (60–201 residues) was similar to the architecture of aIF2β of P. furiosus (PDB: 2DCU, chain B) with an αββααββαββ topology where the last two β sheets are conserved for zinc ion binding (Figure 3A). The helix-turn-helix domain interacts with the γ-domain of initiation factor (Gutiérrez et al., 2004; Sokabe et al., 2006) and Figure 3B shows that its position and conformation are conserved. Since the entire sequence of the β-domain was used, the modeled P. falciparum eIF2β depicts the additional ~30 residues extending both termini.

Since eIF2β is known to play a role in the initiation of translation and had been reported to be a component of the translational machinery by interacting with eIF2γ and eIF5, we next examined the ability of PfeIF2β to interact with the corresponding initiation factors present in P. falciparum. Both PfeIF2γ (PF3D7_1410600) and PfeIF5 (PF3D7_1206700) were obtained and sequenced. Deduced amino acid sequence analyses showed that both proteins exhibited conserved regions known to be involved in the interaction with eIF2β (Figures S3A,B). The expression in E. coli of recombinant tagged protein was successful for PfeIF2γ but failed for PfeIF5, which could be related to the lack of expression due to the presence of amino acid repeats and/or to toxicity for bacteria. We then tested by a pull-down assay whether PfeIF2β interacts with PfeIF2γ. Results presented in Figure 4A (lane 3) indicated that PfeIF2β was indeed able to bind GST-PfeIF2γ. The direct interaction was supported by the absence of eIF2β binding to GST alone (Figure 4A, lane 2). In an independent approach, we further confirmed this PfeIF2β-PfeIF2γ interaction using Xenopus oocytes model where the recombinant cRNA of PfeIF2β (expressing a HA-tagged protein) and His-tagged recombinant protein PfeIF2γ were micro-injected. Co-immunoprecipitation with either anti-His or anti-HA antibodies followed by immnunoblotting analyses clearly showed an interaction between PfeIF2β and PfeIF2γ (Figure 4B, lane 3). The use of anti-His or anti-HA antibodies on extracts from singly micro-injected oocytes (Figure 4B, lanes 1 and 2) confirmed the specificity of the interaction.

To further examine the capacity of PfeIF2β to interact with PfeIF5, the recombinant cRNA of PfeIF5 (expressing a HA-tagged protein) and His-tagged recombinant protein PfeIF2β were micro-injected into Xenopus oocytes. The co-immunoprecipitation/immunoblot assays clearly revealed an interaction between PfeIF2β and PfeIF5 (Figure 4C, lane 3). Taken together, these experiments indicate that the proteins produced are correctly folded and that PfeIF2β binds to proteins of the translation initiation eIF complex.

The presence of two PP1 putative binding motifs in the PfeIF2β gene product led us to examine the ability of the endogenous eIF2β expressed by P. falciparum to bind to PfPP1. Pull-down experiments were carried out using recombinant His-PfPP1 retained on Ni-NTA agarose beads and a soluble extract of blood parasites. Eluted proteins with loading buffer were analyzed for the presence of endogenous PfeIF2β. As expected, immunoblot with anti-His mAb antibody showed the presence of His-PfPP1 (Figure 5A, lane 1). When the eluted proteins were immunoblotted with a polyclonal antibody raised against His-PfeIF2β, we observed a band at 25 kDa, corresponding to the expected size of eIF2β. An additional band was detected at 35 kDa which might correspond to His-PfPP1 as the polyclonal used could contain antibodies against the His-tagged PfeIF2β used for immunization (Figure 5A, lane 3). Although these data indicate that endogenous PfeIF2β can bind to PfPP1, it could not exclude an indirect interaction. In order to examine whether PfPP1 directly and physically binds PfeIF2β, we performed GST-pulldown experiments. Figure 5B clearly shows that GST-PfPP1 binds recombinant His-PfeIF2β (lane 3) while GST alone did not pull down PfeIF2β (lane 2).

To further clarify the molecular basis of interaction of PfeIF2β with PfPP1, we tried to identify the contribution of the putative RVxF and FxxR/KxR/K binding motifs present in PfeIF2β. To this end different versions of mutated PfeIF2β containing either single or combined mutations of putative binding motifs (Figure 5C) were used in GST-PP1 pulldown experiments. Bacterially expressed proteins were purified and incubated with GST-PfPP1 in vitro. As expected, wild-type PfeIF2β protein bound PfPP1 (Figure 5D, lane 4). PfeIF2β proteins with a single mutation in one binding motif (PfeIF2β ¹⁰³KAAA¹⁰⁶ or PfeIF2β ²⁹AGEAKA³⁴) retained their ability to bind PfPP1 (Figure 5D, lanes 6 and 8 respectively). However, PfeIF2β protein with combined mutations in the two motifs (PfeIF2β ²⁹AGEAKA³⁴/¹⁰³KAAA¹⁰⁶) was defective in PfPP1-binding (Figure 5D, lane 10). GST alone did not pull down wild-type PfeIF2β or any mutated protein used (Figure 5D, lanes 3, 5, 7, and 9). These data indicate that the RVxF and FxxR/KxR/K motifs are the central binding motifs of PfeIF2β to PfPP1 and that either native motif is sufficient to mediate binding.

To evaluate the impact of PfeIF2β on the activity of PfPP1 and in the absence of Plasmodium specific substrate to this enzyme, nonspecific substrates such as pNPP or phosphopeptide (K-R-p-T-I-R-R) were tested. Although PfPP1 was able to dephosphorylate these substrates, no effect was observed when PfeIF2β was added at different concentrations (not shown). We therefore turned our attention to the Xenopus oocyte model in which the micro-injection of phosphatases regulators could regulate the G2/M transition assessed by the appearance of Germinal Vesicle Break Down (GVBD) (Daher et al., 2007a; Fréville et al., 2013; Vandomme et al., 2014). In this context, we reasoned that the GVBD could serve as a surrogate marker to evaluate the functional role of the complex eIF2β-PP1. Using this model, we first showed that a micro-injection of 60 ng of recombinant PfeIF2β protein was able to induce GVBD in all micro-injected oocytes (Figure 6A). To further confirm these results, wild-type and different single or double mutated PfeIF2β proteins were used. Results depicted in Figure 6A showed that all protein versions did induce GVBD except the PfeIF2β double mutated in RVxF and FxxR/KxR/K motifs. To explain the effect of PfeIF2β in Xenopus oocytes, we checked whether PfeIF2β could bind to XePP1. Immunoblot analysis of eluates from either anti-His or anti-XePP1 immunoprecipitations revealed the co-immunoprecipitation of PfeIF2β and XePP1 (Figure 6B, lane 3). The single mutation of either RVxF or FxxR/KxR/K motifs did not affect the binding of PfeIF2β to XePP1 (Figure 6B, lanes 4 and 5) while the double mutation of these two motifs did abolish the binding capacity of PfeIF2β (Figure 6B, lane 6). Collectively, these data suggest that PfeIF2β rather blocked the PP1 activity in oocytes and one of the two binding motifs is sufficient for PfeIF2β to fulfill to its function.

To explore the role of PfeIF2β in the P. falciparum blood stage life cycle, silencing its expression by disrupting the gene using the pCAM vector system was attempted. Blood ring stage parasites of the 3D7 strain were transfected with a pCAM-BSD-PfeIF2β construct containing a 5′ fragment derived from the genomic Pfeif2β sequence and the blasticidin resistance gene (Figure 7A). From three independent transfection experiments, the analysis of genomic DNA obtained from resistant stable parasites by PCR (from 2 months up to 6 months of culture under blasticidin pressure) with specific primers (Table S1) did not detect the presence of viable knock out parasites (Figure 7B, lane 6). The wild type Pfeif2β gene was still amplified in genomic DNA and the plasmid remained episomal even after prolonged culture (Figure 7B, lanes 4 and 5 respectively). At this stage, it could not be excluded that the absence of viable parasites with an interrupted Pfeif2β gene could be attributed to the lack of accessibility of Pfeif2β gene locus. To examine this hypothesis, we introduced a targeted modification in the locus without loss-of-function by transfecting ring stage parasites with a plasmid containing the 3′ end of the Pfeif2β coding region fused to the HA sequence (Figure 8A). Using a specific primer of Pfeif2β derived from the upstream region of the construct PfeIF2β-HA and a primer corresponding to the HA sequence (Table S1), genotype analysis by PCR showed the correct integration of Pfeif2β-HA into the locus (Figure 8B, lane 4) and indicates the accessibility of the Pfeif2β locus to genetic manipulations. The integrity of the HA-tagged PfeIF2β was confirmed by immunoblot using an anti-HA monoclonal antibody and protein extracts from Knock-in parasites (Figure 8C). In order to further explore the knock in parasites, we attempted to establish stable clonal lines by limiting dilution. Surprisingly, we were unable to isolate clones expressing PfeIF2β-HA which could be attributed to the low efficiency of transfection and recombination in P. falciparum. Nevertheless, the above data suggest that PfeIF2β is likely essential for the development of intraerythrocytic stages of P. falciparum.

In previous report, it has been demonstrated that eIF2β is exclusively present in cytoplasm of mammalian cells (Bohnsack et al., 2002). We first sought to investigate the localization of PfeIF2β in blood stage parasites by IFA. To this end, anti-HA-biotin antibody (for transfected parasites) or anti-eIF2β antisera and fixed/permeabilized thin smears of wild parental parasites on glass slides, fixed/permeabilized parasites in suspension or thin sections of fixed parasites were used. As shown in Figure 9A, only the use of anti-HA-biotin antibody combined with thin sections of fixed parasites resulted in staining of late transfected parasites (trophozoites, schizonts). The signal was mainly detectable within the parasite and seems to be restricted to the cytoplasm as no overlapping was observed with the nuclear staining. Unfortunately, all assays aiming to express GFP-tagged PfeIF2β either by episomal expression or by gene replacement failed so far. The prediction of classical nuclear localization signals in PfeIF2β (Nguyen Ba et al., 2009) incited us to further examine the parasite compartments that contain PfeIF2β using subcellular fractionation method. The use of cytoplasmic and nuclear markers confirmed the absence of protein cross contamination during fractionation (Figure 9B, upper and middle panels). Immunoblots depicted in Figure 9B (lower part) clearly confirmed the presence of PfeIF2β in the cytoplasm fraction obtained from asynchronous cultures. Intriguingly, PfeIF2β was also detectable in the nuclear fraction. The difference of data between immunoblots and IFA as to the presence of PfeIF2β in the nucleus could be attributed to a poor accessibility of the protein and/or its low abundance in this compartment. When extracts corresponding to about 10 μg of each fraction from synchronized parasite populations were tested, PfeIF2β was detected in cytoplasm and nuclear fractions obtained from ring, trophozoite, and schizont stages (Figure 9C, upper panel). The relative quantification of band intensities of PfeIF2β revealed about 2.5-fold higher in the nuclear extracts when compared to the cytoplasm fractions extracted from ring forms. However, about 11-fold and 10.5-fold increase were detected in the cytoplasm extracts when compared to nuclear extracts of trophozoites and schizonts respectively. When the anti-actin antibody was used, results showed that the actin abundance is comparable between nuclear fractions extracted from ring, trophozoite and schizont stages (Figure 9C, lower panel). Similar results were obtained with the actin detected in the cytoplasmic extracts. This supports the comparison of the level of PfeIF2β in the same compartment of each stage and further strengthens the differential distribution of PfeIF2β during the progression of blood stage parasites.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.