We first aimed to determine whether PfCK2 β subunits play essential functions in parasite proliferation. CK2β proteins possess two conserved pairs of cysteine residues that form the base of a zinc finger required for homo- and heterodimerization 29 , a prerequisite for CK2 holoenzyme formation 30 . To generate plasmids able to disrupt pfck2β genes, an internal fragment of the coding sequence excluding those crucial motifs was amplified and cloned into the transfection vector pCAM-BSD, which confers resistance to blasticidin 31 . It was expected that integration of the constructs (pCAM-BSD-KOPfCK2β1 and pCAM-BSD-KOPfCK2β2) into their cognate genomic loci by single crossover homologous recombination would result in a pseudo-diploid configuration, in which both truncated copies would be unable to express a functional protein (Figure 1A). After two independent transfections of pCAM-BSD-KOPfCK2β1 into 3D7 parasites, integration was monitored by PCR in the blasticidin-resistant populations (Figure 1B; see Additional file 1, table A1, for a list of all primers used in this study), using primer combinations that allow discrimination between the episome, the wild-type locus, and the disrupted locus. Only the episome (lane 3) and the wild-type locus (lane 1) were detectable, with no sign of integration even after prolonged culture (15 weeks). This might be due either to the fact that pfck2β1 is essential to the completion of the asexual parasite cycle, or to the possible refractoriness of the locus to recombination.

To ascertain whether pfckβ1 can be disrupted if the subunit is provided through expression of an episomal copy of the gene, we constructed a complementation plasmid (pCHD-PfCK2β1) containing the full-length pfckβ1 gene under the control of the PfHsp86 promoter and a 3' UTR. In parallel with the transfection of the pCAM-BSD-KOPfCK2β1 plasmid alone, a population of parasites was transfected with both pCAM-BSD-KOPfCK2β1 and pCHD-PfCK2β1. PCR (Figure 1B) and Southern blot (Figure 1C) analyses showed that disruption of the targeted locus occurred only in this doubly transfected, doubly resistant parasite culture. Probing the membrane with a BSD probe (Figure 1C, left panel) confirmed that integration had occurred only in the parasites transfected with both the knockout and complementation plasmids. The pfckβ1 probe hybridized to the 11.1 kb band, corresponding to the wild-type locus, in the lane that contained DNA from untransfected parasites and from parasites transfected with the knockout plasmid alone (lanes 1 to 3). This band was undetectable in DNA from parasites transfected with both the knockout and complementation plasmid (lane 4), indicating that the gene can be disrupted only when an additional cassette coding for the PfCK2β1 subunit is provided to the parasites. There are multiple possibilities for recombination of the knockout and complementation plasmids before or after integration, which presumably accounts for the additional bands. The crucial observation is the disappearance of the wild-type-specific band only in the doubly transfected parasite culture. Similar results were obtained for PfCK2β2 (see Additional file 2, Figure A1), indicating that both regulatory subunits play crucial roles in parasite proliferation in the red blood cell.

To generate parasite lines expressing tagged PfCK2β1 or PfCK2β2 from the endogenous loci, we transfected wild-type parasites with 3'-HA-tagged constructs (pCAM-BSD-HA-pfck2β1 or pCAM-BSD-HA-pfck2β2), which were expected not to cause loss of function of the target protein upon integration into the cognate locus. Integration of the tagging construct and disappearance of the wild-type locus were readily achieved for both genes (see 2 file 5, Figure A2, Figure A3). To determine the subcellular localization of HA-tagged PfCK2α 19 , PfCK2β1 or PfCK2β2 subunits, we examined these transgenic lines in IFAs using antibodies against the HA epitope and against Exp2, a parasitophorous vacuole marker 32 (data for localization of all three subunits are presented in Figure 2A-C). In line with the pleiotropic nature of CK2 in eukaryotic cells, a signal for all three subunits was detectable in both the cytoplasmic and nuclear compartments of the parasite. The location of HA-tagged PfCK2α and PfCK2β2 in rings, trophozoites, and schizonts was further examined by immuno-electron microscopy, which confirmed presence of the protein in both cytoplasm and nucleus (Figure 3). Furthermore, bands of the expected sizes of all three subunits were clearly visible on anti-HA western blots in both cytoplasmic and nuclear fractions of unsynchronized cultures of the three HA-tagged parasite lines (Figure 2D). Antibodies against aldolase, a cytoplasmic marker, and against histone H3, a nuclear marker, were used to probe the same western-blot membrane to ascertain the purity of the subcellular fractions (Figure 2E). In conclusion, our data suggest that PfCK2 subunits are expressed throughout erythrocytic schizogony, and are present in various subcellular locations.

To gain insight into the functions of PfCK2 through identification of the proteins with which the holoenzyme interacts, we subjected protein extracts from the parasite line expressing HA-tagged PfCK2β1 to immunoprecipitation using the anti-HA antibody; the same analysis was performed on extracts from wild-type parasites as a negative control. The immunoprecipitates were separated by SDS-PAGE, and a western blotting was performed using anti-PfCK2β1 to verify the immunoprecipitation efficiency (see Additional file 2, Figure A4A). A parallel gel was loaded and slices excised (see Additional file 2, Figure A4B) for mass spectrometry analysis.

Potential interactors (see Additional file 3, Table A2) were identified based on the number of unique peptides recovered, and on the unweighted spectrum count of the number of detected peptides. We considered proteins as potential hits if they were present in the immunoprecipitate from the HA-tagged line but were absent (or present in trace amounts only) in those from the wild-type parasites. The same experiment was performed in parallel with the parasite line expressing HA-tagged PfCK2β2 (see Additional file 4, Table A3, for the list of recovered proteins), although in this case we were unable to perform western-blot verification of the immunoprecipitate because of the poor quality of the anti- PfCK2β2 antibody.

PfCK2α was pulled down with both HA-PfCK2β1 and HA-PfCK2β2, in line with the previously reported in vitro interaction between the catalytic subunit and both regulatory subunits 19 . Interestingly, PfCK2β1 and PfCK2β2 were present in similar amounts in both immunoprecipitates, suggesting that complexes containing both regulatory subunits may be relatively abundant.

We examined the distribution of all other potential interactors among the various parasite metabolic processes described in the Malaria Parasite Metabolic Pathways (MPMP) website (curated by H. Ginsburg; http://sites.huji.ac.il/malaria/) (Figure 4, top panel). Although some pathways were not represented at all in the immunoprecipitates, or had only a very small number of entries, others displayed many hits in both lists.

A different distribution pattern was seen when the same protocol, using the same anti-HA antibody, was applied to parasites expressing other HA-tagged kinases, such as PfCK1 (data not shown); this result lends support to the specificity of the pull-down approach. For both PfCK2β subunits, the highest numbers of potential interactors were found in the processes of 1) post-translational modifications, and 2) mitosis and chromosome separation. As we were aiming to understand the nuclear function of PfCK2, we focused further investigations on the latter process.

Each metabolic process covers several specific pathways. Of the various pathways grouped under the 'mitosis and chromosome separation' process, some contained very few or no hits (Figure 4, bottom panel). By contrast, 12 proteins belonging to the 'nucleosome assembly and regulation pathway' (see Additional file 2, Figure A5) were found in the PfCK2β immunoprecipitates (Table 1). These included both P. falciparum nucleosome assembly proteins PfNapL (PFL0185c) and PfNapS (PFI0930c). Co-immunoprecipitation experiments coupled to western-blot analysis established that HA-tagged PCK2α can also pull down both PfNaps (see Additional file 2, Figure A6). Naps are negatively charged proteins that act as histone chaperones and function as histone depositors, allowing ordered formation of chromatin at definite sites. PfNapL and PfNapS proteins interact in vitro with P. falciparum histones H2A, H2B, H3, and H4, and are in vitro substrates for human CK2 33 34 . A number of histones and their variants, and several proteins thought to be involved in virulence-gene regulation through association with sub-telomeric chromatin at the nuclear periphery, were also identified in the β subunit precipitates. These included two of the four Alba P. falciparum proteins that have recently been identified as perinuclear DNA-RNA-binding proteins and thought to play a role in gene regulation in asexual stages 35 ; the prototype Alba protein was described as an archeal chromatin protein 36 .

We tested whether recombinant PfCK2α could phosphorylate selected chromatin-associated proteins from the potential PfCK2β interactor list. Kinase assays were performed with glutatione S-transferase (GST)-tagged PfCK2α, using a GST-PfCK2α K72M dead mutant as a negative control (Figure 5). PfCK2α autophosphorylation generated a band of ~60 kDa in size, but the dead mutant did not autophosphorylate, and was unable to phosphorylate exogenous substrates, as previously reported 19 . When subjected to a kinase assay using PfCK2α, both recombinant PfNaps were phosphorylated, with PfNapL (55 kDa) giving a very strong signal (after as short an exposure time as 30 minutes), and PfNapS (37 kDa) appearing as a much less efficient substrate (even after a longer exposure time of 2 hours) (Figure 5A). No phosphorylation of the Nap proteins was seen, either with the PfCK2 kinase-dead-mutant control, nor with the unrelated PK PfPK6 37 38 , despite the ability of the latter enzyme to autophosphorylate and to phosphorylate myelin basic proteins in our assay conditions (Figure 5D).

We next examined the ability of histones to act as substrates for recombinant GST-PfCK2α. Recently, several phosphorylation sites in plasmodial histones have been identified ( 39 ; Dastidar et al., unpublished data), and because several histones were recovered in the PfCK2β precipitates (Table 1), we thought that PfCK2α might be one of the kinases that can phosphorylate histones. When purified parasite native histones were subjected to a GST-PfCK2α kinase assay, preferential phosphorylation of histone H2B (13 kDa) was seen, whereas the dead-mutant kinase failed to phosphorylate any of the histones (Figure 5B). We also tested in PfCK2α kinase assays all four P. falciparum Alba members expressed as GST fusion proteins 35 , as well as the nuclear protein PfSir2, which is implicated in establishment of heterochromatin in sub-telomeric region and hence represses var gene expression 40 41 42 . PfAlba3, PfAlba4, and Sir2 were not phosphorylated by GST-PfCK2α (Figure 5E). By contrast, PfAlba1 (55 kDa) and PfAlba2 (52 kDa) gave signals on the autoradiographs after a kinase assay was performed with GST-PfCK2α, whereas none of the substrates subjected to the GST-PfCK2α dead mutant was phosphorylated (Figure 5C).

Fragments from central region of the PfCK2β1 and PfCK2β2 ORFs were amplified by PCR and inserted between the BamHI and NotI sites of the pCAM-BSD plasmid 31 , which contains the Aspergillus terreus blasticidin-S-deaminase gene, whose gene product confers resistance to blasticidin (see Additional file 1, table A1, for cloning primers).

A plasmid for in vivo episomal expression of PfCK2β1 and PfCK2β2 subunits were constructed as follows. The full-length coding sequences were first inserted between the BglII and NotI sites of plasmid pHGB 56 , and then transferred into plasmid pCHD-1/2 56 by a clonase reaction (Gateway system, Invitrogen) according to the manufacturer's instructions. Plasmid pCHD-1/2 includes a cassette encoding human dihydrofolate reductase, conferring resistance to the antifolate drug WR99210. Parasites that were transfected with both a KO plasmid and a complementation plasmid were cultured under double-drug selection (see below).

The 3' end of the PfCK2β2 coding sequence (573 bp, omitting the stop codon) was amplified by PCR using primers incorporating PstI and BamHI restriction sites and inserted between the same sites of the pCAM-BSD-HA plasmid 57 . For pfCK2β1, the 3' end of the gene (540 bp) was amplified, then the PCR product was treated with the restriction enzymes PstI and BglII and inserted between the PstI and BamHI sites of pCAM-BSD-HA.

Four P. falciparum strains (3D7A 58 and three HA-tagged strains of PfCK2 subunits: PfCK2αHA, PfCK2β1HA and PfCK2β2HA) 19 were grown in human erythrocytes as described previously 59 . The medium for the HA-tagged lines was supplemented with blasticidin (2.5 μg/ml; VWR International Ltd., West Chester, PA, USA). Cultures of the P. falciparum clone 3D7A were maintained at 37°C in RPMI 1640 medium (Gibco) supplemented with 25 mmol/l sodium bicarbonate, 2 mmol/l L-glutamine, 300 mmol/l hypoxanthine, 10 μg/ml gentamicin, and lipid-rich bovine serum albumin (BSA) (AlbuMAX II; Sigma-Aldrich, St Louis, MO, USA). Cultures were seeded at 5% hematocrit volume and maintained at a parasitemia of 1 to 10% with daily changes of medium. The incubator was flushed with a gas mixture containing 5% CO₂. For transfection, asexual blood-stage parasites were synchronized by sorbitol treatment 60 so as to result in the majority being ring-stage parasites. Forty-eight hours later, the ring-stage parasites were transfected by electroporation with 100 μg of purified plasmid DNA in buffer (Cytomix; Gibco-BRL, Grand Island, New York, NY, USA) as described previously 61 . Blasticidin (2.5 μg/ml) was added to the culture medium to select for transformed parasites. Parasites under double selection had 5 nmol/l WR99210 added to the culture medium in addition to blasticidin. Parasites were maintained in this supplemented medium from 2 days post-transfection.

Parasite cultures were lysed in 0.15% saponin. The parasite pellets were resuspended in cold phosphate-buffered saline, and treated with 150 μg/ml proteinase K and 2% SDS at 55°C for 2 to 3 hours. Total DNA was extracted in phenol:chloroform:isoamyl alcohol (25:24:1) and precipitated in ethanol and 0.3 mol/l sodium acetate. DNA was digested with ClaI and EcoRI for PfCK2β1, and with ClaI and NcoI for PfCK2β2.

Cytoplasmic and nuclear extracts were prepared as described previously 41 62 . Using standard procedures. SDS-PAGE was performed and DNA was transferred to nitrocellulose membranes, which were probed using either commercially available antibodies against HA-tag (ab9110; 1:1000 dilution), histone H3 (ab1791; 1:2500 dilution), P. falciparum aldolase (ab38905; 1:1000) (all Abcam, Cambridge, MA, USA) or subunit-specific antibodies generated in rabbits against the PfCK2β1-derived peptide DSNKDLQDSKSDKS (1:500 dilution) or the PfCK2β2-derived peptide DEINRDSEEMYKNK (1:1000 dilution (both BioGenes GmBH, Berlin, Germany) After probing with appropriate secondary antibodies conjugated to horseradish peroxidase (GE Healthcare Life Sciences, Princeton, NJ, USA), the membranes were developed with a commercial substrate (Super Signal West Pico Chemiluminescent Substrate; Thermo Fisher Scientific Inc., Rockford, IL, USA) in accordance with the manufacturer's recommendations.

IFAs were performed on synchronized stages (ring, trophozoite, and schizont) of the three HA-tagged strains of the PfCK2 subunits as described previously 56 . The fixed cells were incubated with primary antibodies anti-HA (ab9110; 1:2000 dilution) and anti-Exp2 32 (1:800 dilution) (both Abcam). After incubating with appropriate secondary antibodies, the cells were examined under a microscope (Nikon Instruments, Tokyo, Japan).

Infected erythrocytes were fixed in 1% glutaraldehyde in RPMI-HEPES buffer for 1 hour at 4°C. After washing, polymerization in agar (Type IX; Sigma-Aldrich), and dehydration with ethanol, the samples were transferred into embedding resin (LR-White; London Resin Company Ltd, Reading, Berkshire, UK), and polymerized for 12 hours at 4°C, and cut. Ultrathin sections of the sample were collected and mounted on Cu/Pd grids. Free aldehyde groups were blocked by incubating the samples for 5 minutes in 50 mmol/l NH₄Cl. Sections were blocked in phosphate-buffered saline containing 5% (v/v) normal goat serum, 1% (w/v) BSA and 0.01% (v/v) Tween-80. The washed grids were incubated for 2 hours in antiserum diluted in phosphate-buffered saline containing 1% (v/v) normal goat serum, 1% (w/v) BSA and 0.01% (v/v) Tween-80 and anti-HA antibody (ab9110; Abcam). Samples were washed and incubated for 45 minutes with 12-nm gold-labeled goat anti-rabbit IgG (111-205-144; Jackson Immunoresearch Laboratories, Inc., West Grove, PA, USA). Washed sections were stained for 15 minutes with aqueous 4% uranyl acetate followed by staining for 2 minutes with 1% lead citrate. Sections were analyzed using an electron microscope (JEM-1200EX; JEOL Ltd, Tokyo, Japan) at 120 kV.

Parasite cultures were lysed in 0.15% saponin and pellets were stored at -80°C. Proteins were extracted on ice using mammalian protein extraction reagent (M-PER) (Profound™ Mammalian HA-tag IP/Co-IP application; Thermo Fisher Scientific) with protease inhibitors (1:25; Complete Protease Inhibitor Cocktail; Roche Diagnostics, Basel, Switzerland) and benzonase nuclease (1:500; Novagen, Darmstadt, Germany). The lysates were cleared by centrifugation at 11,180 g for 15 minutes at 4°C, and the concentration of proteins measured by the Bradford assay. Proteins were loaded on a spin column with a pre-inserted frit, preventing bead loss between washes, and 6 μl (50% v/v_ of anti-HA crosslinked onto sepharose beads were added (both included in the kit used for protein extraction; Thermo Fisher Scientific). The columns were then transferred to a rocker platform and incubated overnight. The flow-through was collected by centrifugation at 300 g for 10 seconds at room temperature. The columns were washed three times using the Tris-buffered saline (TBS) provided in the kit, followed by a short burst of centrifugation (300 g for 10 seconds). Finally, any proteins bound to the beads were eluted using boiled 4× Laemmli buffer, and recovered by centrifugation (15,600 g for 3 minutes). Aliquots of protein extracts and flow-through liquid were taken before and after immunoprecipitation to monitor the yield.

Samples were separated by SDS-PAGE on 10% polyacrylamide gel, which was then stained with Coomassie blue (Biosafe; Bio-Rad Laboratories, Inc., Hercules, CA, USA), and later placed in 10% ethanol and 1% acetic acid. Each lane was removed from the gel and cut into eight slices, then the proteins were digested in the gel using trypsin As follows. Briefly, the samples were reduced in 10 mmol/l dithioerythritol (DTE) and alkylated in 55 mmol/l iodoacetamide (IAA), the gel pieces were dried. The samples were incubated with 12.5 ng/ml trypsin overnight at 37°C. The tryptic peptides were extracted from the gel slices, dried, and resuspended in 2% acetonitrile (ACN):0.1% formic acid (FA) for liquid chromatography/tandem mass spectrometry (LC-MS-MS) analysis

A mass spectrometer (LTQ Orbitrap XL (Thermo Fischer Scientific) equipped with an ultraperformance LC (UPLC) system (NanoAcquity; Waters Ltd, Elstree, Hertfordhshire, UK) was used. Peptides were trapped in a custom-made precolumn (Magic C18 AQ stationary phase, 5 μm diameter, 200Å pore, 0.1 × 20 mm, Michrom Bioresource) and separated in a custom-made main column (Magic C18 AQ, 3 μm diameter, 100Å pore, 0.75 × 150 mm), using a run of 53 minutes and a gradient of H2O:ACN:FA 98%:2%:0.1% (solvent A) and ACN:H2O:FA 98%:1.9%:0.1% (solvent B). The gradient was run at a flow rate of 250 nl/min as follows: 100% A for 3 min, 30% B within 36 min, 47% B within 14 min, 90% B within 5 min held for 5 min and 100% A for 17 min. The MS/MS was operated in an information-dependent mode, in which each full MS analysis was followed by 10 MS/MS acquisitions, during which the most abundant peptide were selected for collision-induced dissociation (CID) to generate tandem mass spectra. The normalized collision energies were set to 35% for CID.

Data search was performed using Mascot software (version 2.3; Matrix Science Inc., Boston, MA, USA) and Proteome Discoverer (v.1.1; Thermo Fisher Scientific), and the sequences searched against a concatenated database consisting of non-redundant Plasmodium database (PlasmoDB, version 6.4; http://plasmodb.org) and the reversed-sequence version of the same database and common contaminant proteins. Finally, results were imported into Scaffold (version 3_00_8 Proteome Software; http://proteomesoftware.com/) for validation of protein identification, normalization, and comparison of spectral count. Peptide identifications were accepted if they could be established at a probability of greater than 95% as determined by the ProteinProphet algorithm 63 . Protein identifications were accepted if they were assigned at least two unique, validated peptides, and could be established with at least 99% probability as determined by the ProteinProphet algorithm 64 .

The GST-PfCK2α and GST-PfCK2α-dead K72M mutant fusion proteins 19 were purified from E. coli (BL21 Gold strain) as described previously 65 . PfCK2α kinase activity was determined by measuring ³³P or ³²P incorporation using γ[³³P]-ATP or γ[³²P]-ATP. All the substrates (Alba1, Alba2, Alba3, Alba4, PfSir2; kindly provided by C. Scheidig-Benatar), and extracted P. falciparum histones were treated with active PfCK2α and dead mutant as described previously 65 . Samples were run on a 16% gel, which was dried and exposed to film for autoradiography.