Our reverse-genetics data strongly suggest that both PfCK2β subunits are essential for parasite viability, indicating that they have non-redundant functions in the parasite. This was unexpected given the dispensability of the β subunits for viability in S. cerevisiae
(which has two CK2β subunits that can be knocked out individually or simultaneously [14
]) and ib Schizosaccharomyces pombe
(which had a single β subunit) [43
]. However, the single β subunit present in C. elegans
] and in mice [20
] is essential. The different interactomes for the CK2β subunits in S. cerevisiae
] are consistent with functional specialization of each β subunit. The immunoprecipitates obtained with PfCK2β1 and PfCK2β2, despite displaying significant overlap in the identity of recovered proteins, also contained hits that were found in only one of the two immunoprecipitates (see Additional file 5
, Table A4). This would be consistent with the non-redundancy in function suggested by reverse genetics; however, we do not have sufficient data to draw any firm conclusion in this respect. The dataset presented in the supplementary material (see Additional file 3
, Table A2; see Additional file 4
, Table A3) is from one of three experiments. Although several hits obtained in the first dataset were confirmed in two further experiments, fewer proteins were recovered in the latter, pre-empting statistical analysis, More work is needed to confirm that individual proteins are truly absent from one of the other PfCK2β interactome. However, the data (see Additional file 3
, Table A2; see Additional file 4
, Table A3) nevertheless enabled us to propose assignment of specific functions to PfCK2.
Using recombinant proteins in pull-down experiments, we previously found that PfCK2α is able to interact with both β subunits in vitro
]. MS analysis of immunoprecipitated subunits (this study) indicates that these interactions are also found in parasite extracts. CK2 is known to have several functions in various organisms [11
], and the glimpse into the interactomes of the P. falciparum
CK2 regulatory subunits provided by our study is consistent with their involvement in a number of cellular processes. Implementing the same protocol with different HA-tagged kinases such as PfCK1 (data not shown) yielded protein repertoires that were clearly distinct from those described here for the PfCK2β subunits, lending support to the specificity of the approach. Potential PfCK2 interactors were found in 17 out of 18 metabolic processes described on the MPMP website. Many processes displayed only a very small number of hits; whether or not these represent genuine interactors will require in-depth investigations of each individual case.
By contrast, other processes and pathways exhibited a large proportion of their proteins recovered in the immunoprecipitates, suggesting that a number of distinct functional complexes had been pulled down. A striking example is the 'nucleosome assembly and regulation' pathway, many components of which were found in both PfCK2β1 and PfCK2β2 immunoprecipitates (Table ). This included both PfNAPs, many histones, and several proteins involved in chromatin remodelling (ISWI, CHD1, high mobility group protein, and structure-specific recognition protein). Interestingly, all recovered proteins were members of the so-called 'activating factors' in this pathway, whereas none of the 'inactivating factors' were present in the immunoprecipitates (see Additional file 2
, Figure A5), which again is consistent with the specificity of the approach. Some of these interactions were validated by coupled co-immunoprecipitation/western-blot analysis (see Additional file 2
, Figure A6) and in vitro
kinase assays (Figure ).
Both PfNap proteins were phosphorylated by recombinant PfCK2α. In yeast, phosphorylation of Nap1by one of the two catalytic subunits of CK2 is implicated in cell cycle progression, as cells expressing CK2 phosphorylation site Nap1 mutants display a lengthened S-phase and a shortened G1 phase, together resulting in a delay in mitosis. Interestingly, these mutants retained the wild-type ability to interact with histones [45
]. It has been proposed that phosphorylation of the PfNaps by human recombinant CK2 may affect their affinity for histones [34
]. PfCK2 might also act on PfNaps in vivo
, thus affecting their function, and in accordance with this hypothesis, recent global phosphoproteomic analyses of parasite extracts [46
] found both PfNaps be phosphorylated on acidic peptides predicted to be CK2 substrates by the NetPhos tool (Table ).
Table 2 Phosphopeptides identified in parasite extracts in two published phosphoproteome studies [46,47].a
We found that PfCK2α preferentially phosphorylated histone H2B in vitro
in a mixture of native histones extracted from the parasite. Histone post-translational modifications such as methylation and acetylation play key roles in the control of expression of virulence genes (for example, the var
family) in malaria parasites [48
]) Similarly, histone phosphorylation is known to play an important regulatory role in other eukaryotes [49
], and in yeast, CK2 phosphorylates histone H4 at serine 1 as part of a DNA damage-signaling pathway [50
]. Although previous studies have identified few phosphorylation marks in P. falciparum
histones, we recently characterized several novel phosphorylation marks on the parasite's histones (Dastidar et al.
, unpublished). Whether PfCK2 is responsible for phosphorylation of any of these sites in vivo
, and the functional consequences of these phosphorylation events, remain to be investigated. Nevertheless, taken together, our data strongly suggest that PfCK2-dependent phosphorylation may play a role in chromatin dynamics.
We also tested other known nuclear proteins with a potential role in gene regulation to see if they could be substrates for the kinase. Of all the proteins tested, recombinant PfAlba1 and PfAlba2 were found to be substrates for the PfCK2α in vitro
(Figure ), and PfAlba1 was found to be phosphorylated in parasite extracts on several residues [46
], including one in an acidic stretch consistent with CK2 substrate preferences (Table ). By contrast, recombinant PfAlba3, PfAlba4, and PfSir2 were not phosphorylated by the kinase (Figure ). PfAlba1 and PfAlba2 belong to the same phylogenetic cluster, which is distinct from that containing PfAlba3 and PfAlba4 ([51
]), consistent with their different ease if phosphorylation by PfCK2. Moreover, the former two PfAlbas possess an arginine- and glycine-rich RGG domain in addition to the Alba domain. It has been shown in S. cerevisiae
that phosphorylation of the RGG domains affects protein localization and transport between the nucleus and cytoplasm [53
]. Given that PfAlba1 and PfAlba2 concentrate at the perinuclear space in ring-stage parasites, but expand to the cytoplasm of trophozoites and schizonts [51
], it is possible that phosphorylation by PfCK2 may contribute to the regulation of their nucleocytoplasmic transport.
Several proteins involved in mitosis and chromosome separation and in DNA replication and DNA damage pathways (for example, several replication factor C subunits, ATP-dependent DNA helicase, DNA ligase 1) were also pulled down. Implication of CK2 in the DNA damage response is well documented in other systems [12
], and a DNA repair pathway has been characterized in P. falciparum
], of which the DNA polymerase and DNA ligase components can be retrieved in the CK2β immunoprecipitates (see Additional file 3
, Table A2; see Additional file 4
, Table A3; see Additional file 5
, Table A4). Because DNA repair is markedly reduced in the highly drug-resistant W2 isolate, which displays a higher rate of drug-resistance acquisition than do drug-sensitive strains [55
], it might be of interest to compare CK2 activity in the W2 versus drug-sensitive clones.
Although many of the proteins identified in the immunoprecipitates were of nuclear origin, a number of potential interactors were associated with other cellular compartments. Some of the recovered proteins were putative or established Maurer's cleft proteins (for example, Maurer's cleft 2 transmembrane protein; Pfmc-2TM), proteins involved in invasion and motility processes (for example, glutamate-riche protein, merozoite surface protein 2 precursor, RhopH3, and other rhoptry-associated proteins) or in cell-cell interaction processes (for example, mature parasite-infected erythrocyte surface antigen, PfEMP2 and rifins). Likewise, several proteins and enzymes involved in redox metabolism were recovered (thioredoxin, thioredoxin-related protein, thioredoxin peroxidase, glutathione peroxidase, glutathione S-transferase, ribonucleotide reductase). This diversity is consistent with our IFA, western blot, and immuno-EM data localizing all the three subunits to both the nuclear and cytoplasmic compartments. Although these results require confirmation by other approaches, and despite the fact that we cannot exclude false positives in our interactome list, they are consistent with an involvement of PfCK2 kinase in varied cellular functions in the parasite.
The availability of parasite lines expressing individual tagged PKs for the almost entire P. falciparum
] now allows implementation of a kinome-wide investigation of kinase function based on the interactomics approach used here for PfCK2. Preliminary data obtained with other PKs suggest that distinct functions can therefore be assigned to different enzymes on the basis of specifically pulled-down protein repertoires.