|Home | About | Journals | Submit | Contact Us | Français|
CK2 is a ubiquitous but enigmatic kinase. The difficulty in assigning a role to CK2 centers on the fact that, to date, no biologically relevant modulator of its function has been identified. One common theme revolves around a constellation of known substrates involved in growth control, compatible with its concentration in the nucleus and nucleolus. We had previously described the identification of two catalytic subunits of CK2 in Trypanosoma brucei and characterized one of them. Here we report the characterization of the second catalytic subunit, CK2α’, and the identification and characterization of the regulatory subunit CK2β. All three subunits are primarily localized to the nucleolus in T. brucei. We also show that CK2β interacts with the nucleolar protein NOG1, adding to the interaction map which previously linked CK2α to the nucleolar protein NOPP44/46, which in turn associates with the rRNA binding protein p37. CK2 activity has four distinctive features: near equal affinity for GTP and ATP, heparin sensitivity, and stimulation by polyamines and polybasic peptides. Sequence comparison shows that the parasite orthologues have mutations in residues previously mapped as important in specifying affinity for GTP and stimulation by both polyamines and polybasic peptides. Studies of the enzymatic activity of the T. brucei CK2s show that both the affinity for GTP and stimulation by polyamines have been lost and only the features of heparin inhibition and stimulation by polybasic peptides are conserved.
The eukaryotic protein kinase CK2, originally called casein kinase II, was one of the earliest identified protein kinases. The kinase primarily consists of two subunits, a catalytic subunit (CK2α or CK2α’) and a regulatory subunit (CK2β). The typical stoichiometry is two catalytic subunits (homomeric or heteromeric) coupled with two regulatory subunits . Under differing salt conditions, other forms have been observed [2,3]. In addition, there are a number of reports where the subunits function independently of each other, often pairing with other molecules [4–6]. CK2 is distantly related to the CMGC group of protein kinases, particularly the cyclin-dependent kinases (CDKs) . Unlike the CDKs, the activity of the catalytic subunit does not require the presence of the regulatory subunit.
Though extensively studied, the actual role of CK2 has remained elusive. CK2 appears to be essential in every organism tested thus far . Genetic analysis of CK2 function in yeast indicates a role in cell-cycle progression and cell morphology [9,10]. The enzyme is highly pleiotropic, with over 300 known substrates . The largest classes of substrates are involved in either gene expression or protein synthesis, in agreement with the genetic analysis in yeast. The localization of CK2 is consistent with its known substrates. Though dispersed throughout the cell, the protein shows highest abundance in the nucleus, particularly the nucleolus [12–14], which is the site of ribosome biogenesis and the location of many known CK2 substrates.
Confounding the determination of the role of CK2 is its apparently constitutive activation, contrasting with most protein kinases which cycle between an active and inactive state (see  and references therein). Regulation of CK2 activity appears to be primarily at the level of protein abundance. Even the regulatory subunit CK2β does not appear to directly regulate the activity of the enzyme, but rather to modify its interactions with other molecules. While no in vivo effectors of CK2 are known, a number of in vitro effectors have been identified. These effectors include heparin, which inhibits CK2 activity, and polyamines like spermine, which increase activity [16,17]. Polybasic peptides can modulate the ability of CK2 to phosphorylate specific substrates [18–20], which could be relevant in vivo. Another hallmark of CK2 activity is its ability to utilize GTP as well as ATP as a phosphate donor . The physiological significance of this property is not known.
The protozoan parasite Trypanosoma brucei, the causative agent for African trypanosomiasis or sleeping sickness, has a complex life cycle involving growth and development in insect and mammalian hosts. The procyclic form (which resides in the insect midgut) and mammalian slender bloodform can be readily cultured in vitro. The genome of the organism has recently been elucidated , as has the kinome . Two catalytic subunits of CK2 were identified . One of the subunits, CK2α (CK2A1, systematic name Tb09.211.4890) was previously characterized . Unlike most other CK2s, the purified T. brucei CK2α subunit, as well as CK2α immunoprecipitated from the parasites, have only weak affinity for GTP. Here we report the characterization of the second catalytic subunit as well as the identification and characterization of the regulatory subunit CK2β.
The work described uses either the procyclic T. brucei strain 29–13 or the single marker bloodform line, which are derivatives of the 427 strain . Both lines express T7 RNA polymerase and the tetracycline (Tet) repressor allowing for Tet-regulated expression of transfected genes. Procyclic cells were grown in SDM-79 (JRH Biosciences) supplemented with 15% fetal calf serum containing 15 μg/ml G418 and 50 μg/ml hygromycin to maintain the T7 RNA polymerase and Tet-repressor genes. Bloodform cells were grown in HMI-9 with 2.5 μg/ml G418. Both procyclic and bloodform cells were transfected as described  with 2.5 μg/ml phleomycin and/or puromycin 1μg/ml added for selection of constructs. Expression from Tet-regulated constructs was induced with 1–2 μg/ml Tet.
All primers used for amplification, site-directed mutagenesis and real-time PCR are listed in Table 1. For in vitro expression of proteins, the coding regions of the genes were amplified and cloned into pRSETA. Plasmid pRSETA-CK2α was described earlier . The CK2α’ and CK2β coding regions were amplified from genomic DNA with the primers P-CK2α’1 and P-CK2α’2, and P-CK2β1 and P-CK2β2 respectively. PCR products for these reactions, as well as all others described, were initially cloned into pGEM-T easy (Promega) and sequenced. Inserts were excised with BamHI and XhoI and cloned into pRSETA.
All of the plasmids used for expression in T. brucei employed Tet-regulated T7 promoters. To express myc-tagged proteins in T. brucei, the CK2A1 gene in pLEW-CK2α-myc (a pLEW79 derivative) was replaced with ORFs encoding either CK2α’ (CK2A2) or CK2β (CK2B). CK2A2 was amplified with the primers P-CK2α’2 and P-CK2α’3. After cloning into pGEM-T easy, the gene was released by digestion with HindIII and XhoI and ligated into a similarly digested vector. The CK2B gene was excised from the pGEM-T easy plasmid with EcoRV and XbaI. The target vector was initially digested with HindIII and the site filled in with Klenow DNA polymerase. Next, the plasmid was digested with XbaI and the CK2β fragment inserted. For co-expression of CK2β-myc with other tagged genes, the phleomycin resistance gene was excised with SpeI and SalI and replaced with the puromycin resistance gene from pLew79(PAC)GFP+3′MCS (Saveria, Kessler, and Parsons, submitted).
The plasmid pLEW79-TAP  was modified to form pLEW-MHTAP to facilitate cloning and separate the calmodulin binding domain of the TAP tag (which also includes an IgG binding domain from protein A) from the protein of interest. The primers MHTAP S and MHTAP AS were used in site-directed mutagenesis using the Quick-change protocol (Stratagene). The modified plasmid contains a small polylinker (AvrII-HindIII-XhoI-BamHI) in lieu of the luciferase reporter, plus a single myc tag and eight histidine codons upstream of the TAP tag.
To insert the CK2 genes into pLEW-MHTAP, CK2A1 was amplified from genomic DNA with the primers P-CK2α1 and P-CK2α2, and the fragment was then digested with HindIII and BamHI. CK2A2 was amplified with the primers P-CK2α’2 and P-CK2α’3 and the fragment digested with XbaI. Finally, the plasmid pRSETA-CK2β was digested with BglII and BamHI. The fragments were then separately ligated into appropriately digested pLEW-MHTAP.
To construct the plasmid pB42AD-CK2β used for the two-hybrid screen, CK2β was excised from the pGEM clone used to generate pRSET-CK2β with EcoRI and XhoI and ligated into similarly digested pB42AD . The Saccharomyces cerevisiae NOG1 gene was amplified with the primers ScNOG1 f1 and r1. The PCR product was re-amplified with primers ScNOG f2 and r2. The PCR product was then transformed along with digested pOBD2 as described .
For RNAi, gene fragments were inserted into the vector pZJM, which contains head-to-head, Tet-regulated T7 promoters allowing for the expression of double-stranded RNA . Primers P-CK2α3 and P-CK2α4 were used to amplify a portion of CK2A1. This PCR product, as well as the CK2α’ PCR product described above, were digested with XhoI and HindIII prior to ligation into XhoI-HindIII digested pZJM. To simultaneously repress both CK2α and CK2α’, the 5′ end of CK2A2 was amplified with the primers P-CK2α’4 and P-CK2α’5 and the product digested with XhoI. The fragment was ligated into the XhoI digested pZJM-CK2α plasmid created above to create pZJM-CK2αα’.
For generation of antiserum to CK2α, pRSET-CK2α was transformed into E. coli BL21. CK2α was found predominantly in inclusion bodies. This fraction was separated by SDS-PAGE. The region containing CK2α was cut out and used to immunize rabbits as described . Other primary antibodies and antisera used include E7 anti-β-tubulin monoclonal antibody (mAb) (Developmental Studies Hybridoma Bank, University of Iowa), anti-myc 9E10 mAb (Covance), rabbit anti-HRP (Sigma), and rabbit anti-NOG1 .
For immunoprecipitation of proteins generated in vitro, a coupled transcription and translation reaction (TnT, Promega) was used. Templates for protein production included the genes, which were cloned into the E. coli expression vector pRSETA (which generates His-tagged proteins), and PCR products containing the untagged ORFs plus a T7 promoter. The latter were generated from the cloned genes using the T7 containing 5′ primer CK2α T7 5′ or CK2β T7 5′, matched with the corresponding 3′ primers CK2α 3′ TnT or CK2β TnT 3′. Radiolabeled proteins were generated by adding [35S]-methionine to the reactions. For immunoprecipitation, Protein A magnetic beads (Dynal) were pre-loaded with anti-CK2α antiserum at 1:250 and incubated with in vitro translated proteins. Beads were washed three times with SK-lysis buffer (50 mM Tris pH7.5, 150 mM NaCl, 2 mM EGTA, 1% NP-40, 0.25% deoxycholate) and samples were analyzed by SDS-PAGE and phosphorimaging.
For immunoprecipitation from T. brucei, cells containing inducible expression cassettes were treated with Tet for 16–24 hours and sample lysates prepared as described . Immunoprecipitations were performed as described above, using 108 cells per reaction.
Immunoblots followed the protocol described  except those imaged by infrared (IR) detection using the Odyssey imaging system (Licor). For infrared imaging, blots were blocked in Odyssey blocking buffer. Blots were subsequently probed with either rabbit anti-CK2α antiserum (1:5000) or mouse anti-β-tubulin mAb (1:100,000). Primary antibodies were detected with goat anti-rabbit Ig conjugated with AlexaFluor 680 (50 ng/ml) or goat anti-mouse Ig conjugated with IRDye 800 (25 ng/ml) and imaged on the Odyssey.
Cells were fixed and stained for immunofluorescence as described . Antibodies used were 9E10 anti-myc 5 μg/ml, rabbit anti-HRP (Sigma) 1:200, and rabbit anti-NOG1 1:200 . Primary antibodies were visualized with FITC-conjugated goat anti-mouse Ig or Texas Red conjugated goat anti-rabbit Ig (Southern Biotechnology).
Proteins were purified from two liters of cells after overnight induction with Tet. Cells were collected by centrifugation and washed with PBS containing 10 mM glucose. Cells were then lysed in IPP 150 [33,34] containing 0.1% BSA and the protease inhibitors PMSF, leupeptin, aprotinin, E-64, pepstatin, and TLCK at concentrations described . Tagged protein complexes were then purified as described . Briefly, this entails affinity purification on human Ig-sepharose, cleavage of the protein A tag with TEV protease, and subsequent binding and release from calmodulin coupled to agarose column. The resulting proteins have 63 additional amino acids at their C-termini as compared to the native proteins. Preliminary heparin inhibition experiments indicated that additional kinase activities were present in the CK2α-TAP (but not the CK2α’-TAP or the CK2β-TAP) preparations. To remove these activities, the salt concentration was increased from 150 mM NaCl to 1M NaCl in the wash buffer for the Ig-sepharose affinity step (for this protein only). Although NaCl inhibits CK2α, control experiments showed that this inhibition is reversible (data not shown).
Protein samples were separated by SDS-PAGE on 8–16% gradient gels (Cambrex). Gels were silver stained using SilverQuest (Invitrogen) following manufacturer’s protocol and protein bands excised. Excised bands were processed and digested with trypsin as described  and analyzed by LC-MS/MS using an LCQ deca XP. Collision spectra were analyzed using Sequest and compared against the T. brucei database (www.genedb.org).
Two different yeast two-hybrid systems were used. To analyze the interaction between the T. brucei proteins the system described by Estojak et al  was used. To analyze the interaction between the S. cerevisiae proteins the system described by Uetz et al was used .
Kinase reactions were performed as described  using TAP-purified protein, with the addition of casein (5 μg per reaction) or bovine calmodulin (5 μg per reaction) as substrates. Reactions were separated by SDS-PAGE (10% acrylamide) and labeled proteins detected by phosphorimaging. Signals were quantified using ImageQuant (Molecular Dynamics).
Northern blots were prepared and hybridized with RNA probes as described . For real-time PCR analysis, RNA was isolated from cells using TRIzol (Invitrogen) and treated with DNase using DNAfree (Ambion) according to manufacturer’s protocol. RNA was converted into cDNA using MultiScribe reverse transcriptase (Applied Biosystems). Primers for real-time PCR (900 nM) were CK2α-RT F and R, CK2α’-RT F and R, CK2β-RT F and R and α-tub F and R (α-tubulin). For each real-time reaction, the cDNA generated from 25 ng of RNA was amplified and detected using SYBR Green PCR master mix (Applied Biosystems). Each sample was analyzed in triplicate, normalized against α-tubulin, and referenced against a standard curve generated using genomic DNA.
We previously reported the characterization of a 41 kDa T. brucei CK2α catalytic subunit encoded by the CK2A1 gene and identified CK2A2 (systematic name Tb927.2.2430), which encodes a distinct, 43 kDa catalytic subunit, CK2α’ . As previously noted, both proteins contain the sequence features characteristic of CK2 catalytic subunits, including a conserved tyrosine in region I, a valine in subdomain II, a basic region downstream of subdomain II, and a DWG motif (in lieu of DFG) in subdomain VII. The basic region has been shown to be involved in nuclear localization and substrate recognition of other CK2αs. Reciprocal BLAST analysis indicates that the previously described Leishmania major CK2α gene (LmjF35.1730)  has the highest homology to TbCK2A1, while the L. major ORF LmjF02.0360 has the highest homology to TbCK2A2. In Trypanosoma cruzi a single CK2α gene was identified (Tc00.1047053510761.60), while two CK2α’ genes were found (Tc00.1047053508741.320, Tc00.1047053503513.10). The existence of two genes in the T. cruzi genome sequence is not surprising since the strain used for the genome project is a hybrid between two distinct genotypes with significant allelic divergence. In the catalytic subunits there is relatively little sequence in addition to the kinase domains but the CK2α’ proteins have somewhat longer amino terminal extensions (62 in CK2α’ versus 42 amino acids in CK2α in the case of T. brucei). Within the catalytic domain, TbCK2α and TbCK2α’ respectively have ~ 60% and 50% identity to both human catalytic subunits.
Although most organisms have two types (α and α’) of catalytic subunits, usually only one regulatory subunit is found per organism. Querying the genome sequences (www.genedb.org) yielded a single intact CK2β gene in each species (Tb11.01.2590, LmjF36.5090; Tc00.1047053503757.40). Interestingly, T. cruzi also possesses a pseudogene, which contains an internal stop codon. The alignment of the three predicted trypanosomatid CK2β proteins with the human, Caenorhabditis elegans, and Saccharomyces cerevisiae proteins is shown in Figure 1. In pairwise comparisons the trypanosomatid sequences showed over 50% identity amongst each other, and 30–35% sequence identity with human, S. cerevisiae, S. pombe, C. elegans, and Arabidopsis. All three trypanosomatid proteins contain the characteristic zinc finger motif found in other CK2β proteins (overlined in black in Figure 1). CK2β proteins are typically phosphorylated close to the amino terminus (residues 2 and 3 of the human protein) by the CK2 catalytic subunit . Only the T. brucei protein has a consensus CK2 phosphorylation site near its amino terminus (first asterisk, Figure 1).
An antiserum was raised to recombinant CK2α. This antiserum detected three proteins on immunoblots, the largest and smallest of which were reliably detected. The smallest species comigrated with in vitro translated CK2α at approximately 38 kDa (Figure 2A). Only the comigrating species was depleted when the level of CK2α mRNA was reduced by RNAi (Figure 2B). Similarly, it was immunoprecipitated by the antiserum, whereas the other molecules were not (data not shown). Hence the smallest detected band on immunoblots is CK2α.
To verify that the CK2β identified was a true homologue, we tested its ability to interact with CK2α (Figure 2C). After expression in a coupled transcription and translation system, CK2α and 35S-CK2β proteins were mixed. As seen in Figure 2C, anti-CK2α antiserum co-precipitated the labeled CK2β when CK2α was present, but not in its absence. We observed in these studies that T. brucei CK2β migrated aberrantly, with the untagged version observed at approximately 44 kDa, as compared to the predicted 34 kDa. The reason for this unusual migration is not clear, although we note the presence of both highly acidic and basic regions in the C-terminal extension unique to this species. In similar experiments using CK2α’ and CK2β, we were unable to detect any interaction (data not shown). In contrast to CK2α, in vitro translated CK2α’ was not enzymatically active (data not shown), suggesting that it may be misfolded. We were unable to test the interaction in the yeast two-hybrid system since both CK2α’ and CK2β were strong transcriptional activators.
We further tested interactions in vivo. When T. brucei stable transfectants were induced with Tet to express CK2β-myc, anti-myc mAb co-precipitated CK2α, as revealed by immunoblotting (Figure 2D). The experiments were extended by examining TAP-tagged proteins. CK2B, CK2A1 and CK2A2 were cloned into the Tet-regulated TAP tag vector pLEW-MHTAP and procyclic form transfectants containing integrated plasmids were isolated. Following induction, the TAP-tagged proteins, as well as any proteins associated with them, were recovered using the standard TAP procedure of affinity chromatography on immunoglobulin and calmodulin matrices.
Equivalent kinase activities (see below) of the different preparations were separated by SDS-PAGE and silver stained (Figure 3A). Each sample showed a prominent band whose migration corresponded to ~8 kDa larger than the unmodified proteins (8 kDa of the tag remains after proteolytic removal of the protein A domain), although the tagged CK2α’ was relatively less abundant than the other tagged proteins. In the CK2β-TAP preparation, a second major band was seen, as well as a number of fainter bands. The two major bands and the 44 kDa region were excised, digested with trypsin and processed for protein mass spectrometry. Eight tryptic fragments of CK2β and a single tryptic fragment of the TAP tag were identified in the 52 kDa band, showing it to contain the purified tagged CK2β protein. Eleven tryptic fragments of CK2α were identified in the 40 kDa band. Finally, a single tryptic fragment of CK2β was identified in the 44 kDa region (the same size as in vitro translated CK2 β), as was a single tryptic fragment of NOPP44/46, a nucleolar protein which has previously been shown to interact with and be phosphorylated by CK2α . No peptides corresponding to CK2α’ were identified in any of the processed bands.
The CK2α-TAP and CK2α’-TAP preparations showed different staining patterns. Neither showed a stoichiometric association with CK2β. Unlike CK2β-TAP, where two bands with equal intensity were detected on the silver stain, surprisingly only one prominent band was detected for CK2α-TAP. To determine if the TAP tag on CK2α had disrupted the interaction with CK2β, we co-expressed both CK2α-TAP and CK2β-myc in T. brucei procyclic forms. TAP pulldowns showed the presence of little CK2β-myc (Figure 3B), although the protein was clearly expressed as shown by immunoprecipitation with anti-myc mAb (please note, the presence of CK2α-TAP following immunoprecipitation with anti-myc mAb is uninformative since the TAP tag interacts directly with the Ig reagents). Since Figure 2D shows that CK2β-myc can interact with endogenous CK2α, these data suggest that the C-terminal TAP on the catalytic subunits interferes with their interaction with CK2β. The crystal structure of both the human and maize holoenzymes do not provide any into this finding, since the carboxyl terminus of CK2α is distal to its site of interaction with CK2β [39,40].
CK2α’-TAP expressed very poorly (as did a CK2α’-myc protein, not shown), so the concentration of the protein kinase eluted in affinity purification was also low. Hence, at least 20-fold more cell equivalents of the CK2α’ preparation were used for gel analysis. Contaminants were correspondingly increased, as seen in the stained gel. The 50 kDa band (asterisk) reacted with anti-myc mAb (myc is contained as an epitope tag upstream of the cleaved TAP tag), showing that it was indeed CK2α’. In this gel it is difficult to know whether there is any co-purifying CK2β.
The CK2α and CK2α’ preparations enriched by the TAP purification procedure provided a convenient source of CK2α catalytic subunits largely in the absence of the regulatory subunit. Preparations of CK2β-TAP protein allowed for characterization of the CK2α2β2 holoenzyme, although trace amounts of CK2α’ could theoretically be present. We will refer to these preparations as CK2α, CK2α’, and holoenzyme, keeping in mind that the composition of the latter is predominantly CK2α2β2. Initial assays using casein as a substrate indicated that all of the purified proteins were active. The preparations had over a thousand-fold higher protein kinase activity than mock purifications from untransfected procyclic forms (data not shown). In addition to the phosphorylated casein, additional proteins were also phosphorylated and detected in Figure 4A. The identity of most of these proteins is unknown, although the 52 kDa species seen only in the holoenzyme assay likely corresponds to CK2β-TAP, which was prominent on Coomassie blue-stained gels (see Fig. 3A) and which contains a CK2 phosphorylation site. CK2α (and the holoenzyme, which contains CK2 α) phosphorylated two electrophoretically distinct subunits of casein, while CK2α’ predominantly phosphorylated only the larger subunit. The basis for this difference was not determined.
One of the key hallmarks of the CK2 kinase is its sensitivity to heparin . We have previously demonstrated that TbCK2α purified from E. coli was inhibited by heparin at nanomolar concentrations . The TAP-purified CK2α and holoenzyme were also inhibited by nanomolar concentrations of heparin with just 4% and 3% percent activity respectively at 50 nM heparin. CK2α’ was slightly more resistant to heparin inhibition with 31% activity at 50 nM heparin. That over 94% of the activity showed a simple inhibition curve demonstrated that, although not purified to homogeneity, there was only a trace contamination with other kinase activities in the preparations. An additional inhibitor of CK2 activity is 5,6-dichloro-1-β-D-ribofuranosylbenzimidazole (DRB). We assayed the effect of DRB on isolated CK2α’ and holoenzyme. These had IC50s of 31 μM and 110 μM respectively (not shown), in comparison to an in vitro IC50 of 6μM for CK2 isolated from HeLa cells .
Polyamines generally have a stimulatory effect on CK2 holoenzyme activity in vitro, which is mediated through the CK2β subunit . We tested the effect of the polyamine spermine on the T. brucei holoenzyme activity. At levels showing the highest activity in other systems, no significant stimulatory effect was seen (Figure 4C).
Polybasic peptides, such as poly-L-lysine, stimulate the ability of CK2 holoenzyme to phosphorylate certain proteins, the prototype of which is calmodulin. We tested whether the T. brucei holoenzyme retained this property, using bovine calmodulin as substrate. Poly-L-lysine yielded a 1000-fold stimulation phosphorylation of calmodulin, but had no detectable effect on phosphorylation of the 52 kDa species thought to be CK2β-TAP (Figure 4D).
One difference between the mammalian holoenzyme and its isolated catalytic subunits is sensitivity to high salt . The activity of CK2α subunits decrease with increasing salt while the activity of either the α2β2 or α’2β2 holoenzymes increases three- to four-fold at 50–200 mM NaCl, before dropping at higher concentrations. Like mammalian catalytic subunits, T. brucei CK2α and CK2α’ were both sensitive to salt, losing over 50% of their activity at 100 mM and 200 mM NaCl respectively (Figure 4E). In contrast to the mammalian CK2 holoenzymes, the T. brucei holoenzyme showed no salt activation, but rather lost 50% of its activity at 50 mM NaCl (Figure 4E, triangles).
Unlike other protein kinases, CK2 can typically use GTP as well as ATP as a phosphate donor. We have previously reported that T. brucei CK2α has only a very low affinity for GTP . This property is conveyed by the catalytic subunit  and, by extension, is exhibited by the corresponding holoenzyme. Excess unlabeled nucleotides were added to the reactions containing CK2α’ to determine if they could compete with the labeled ATP (Figure 4F). Among the four ribonucleotides, only ATP could compete for binding. The deoxyribonucleotide dATP also strongly competed with ATP.
We had previously shown that CK2α concentrates in the nucleolus with diffuse nuclear and cytoplasmic staining. We tested the localization of CK2α’-TAP and CK2β-myc. The former was detected by virtue of the binding of the tag to the Fc portion of IgG, while the latter was detected with anti-myc mAb. Unfortunately, the presence of the protein A portion of TAP tag complicates co-localization using antibodies. However, using DAPI as a nuclear stain, a significant proportion of the CK2α’-TAP can be seen to be nuclear (Figure 5A). It appeared to be concentrated in the region of the nucleus which stains poorly with the DNA dye DAPI, i.e., the nucleolus. However, as expected for a pleiotropic protein kinase, there was additional diffuse cytoplasmic staining (Figure 5A). Similarly, CK2β-myc showed general cytoplasmic staining with a single region of more concentrated staining (Figure 5B). This region of staining was localized to the nucleolus as seen by the overlay with staining for the nucleolar marker NOG1.
In earlier work, we demonstrated that CK2α interacts with the nucleolar protein NOPP44/46, which also interacts with the nucleolar GTP binding protein NOG1. The studies presented in Figs. 2 and and33 show that CK2α and CK2β interact with one another. We therefore decided to test whether CK2β interacted with NOG1, using a yeast two-hybrid approach. Indeed the two proteins were found to interact (Figure 6A). The interaction was also detected for the yeast orthologues of these proteins, originally in a genome-wide screen using ScNOG1, and then confirmed by testing the specific ORFs (Figure 6B). We confirmed the interaction in T. brucei by co-immunoprecipitation analysis in which the association of CK2β-myc and NOG1 was tested. CK2β-myc was expressed in T. brucei procyclic forms in the Tet inducible system. We have previously shown that NOG1 is sensitive to proteolysis during extended incubations, such as those required for immunoprecipitations, leading to the generation of multiple immunoreactive species . As shown in Figure 6C, NOG1 was detected in anti-myc immunoprecipitates only when the expression of CK2β-myc was induced. As expected, it was detected in all anti-NOG1 immunoprecipitates, whether or not CK2β-myc was induced. In the reciprocal experiment, anti-NOG1 coprecipitates contained CK2β-myc.
T. brucei bloodform parasites proliferate more than two-fold faster than procyclic forms in vitro. When the abundances of the relevant transcripts from both stages were quantified by real-time PCR, only CK2α mRNA appeared to increase and that increase was modest (average of 67%, Figure 7A). When the abundances of the mRNAs were compared to genomic equivalents, the mRNA abundance of CK2β was approximately two-fold higher than that of the other two genes. Quantitation of the transcript abundance in slender bloodforms, non-dividing stumpy bloodforms, and procyclic forms from the pleiomorphic strain, TREU667, gave similar results (data not shown). The abundance of CK2α protein, as revealed by anti-CK2α antibodies, was comparable between procyclic 29–13 cells and single marker bloodform cells when normalized to β-tubulin, which is constitutively expressed (Figure. 7B). Attempts to raise anti-peptide antisera to CK2α’ yielded only low titer antibodies with poor specificity.
CK2 has been shown to be an essential kinase in most if not all organisms. We investigated its role in T. brucei by reducing its expression via RNA interference (RNAi). Regions of either CK2A1 (nt 1–480) or CK2A2 (nt 641–1171) were cloned into the RNAi vector pZJM for Tet-regulated expression of double-stranded RNA. Following isolation of stable procyclic form transfectants, RNAi was induced. Surprisingly, the doubling time increased only about 20% (data not shown). We next tested the possibility that there is redundancy between the two kinases by targeting both transcripts simultaneously for RNAi by expressing double-stranded RNA encompassing regions of both genes (nt 1–480 of CK2A1 plus nt 35–521 of CKA2). An additive rather than synergistic effect was seen, with a generation time 40% longer than that of uninduced cells (Figure 8A). The endogenous full-length RNAs at 1.46 kb (CK2α) and 3.29 kb (CK2α’) were reduced by ~75% at day 2 through day 7 (Figure 8B) following induction. It is clear that the maximal growth rate of T. brucei requires CK2 function. Since we saw only a 75% reduction in the mRNA abundances, it remains possible that CK2 activity is essential. Similar findings were seen when double RNAi was performed using bloodstream forms (data not shown). The cell-permeant CK2 inhibitor DRB affected the growth of both stages with 50% inhibition at the highest concentration tested, 200 μM (data not shown), although off-target effects cannot be excluded. Other than slowed population growth, no phenotype was observed for depletion of CK2 activity by RNAi or DRB treatment.
The nucleolus, in addition to being the site of ribosome biogenesis, also plays a role in other pathways like gene silencing, cell proliferation, and tumor suppression. Intriguingly, a role for CK2 has been implicated in all of these functions. CK2 is known to accumulate in the nucleus  and interacts with several nucleolar proteins [24,43,44]. Here we show that the α’ and β subunits of CK2, like the α subunit, are concentrated in the nucleus and nucleolus in T. brucei. It has been demonstrated previously that tagged yeast CK2α1 is able to pull down the GTP-binding protein NOG1, as well as several other proteins in the pre-ribosome , and hence likely associates with the 66S precursor particle. In this report we demonstrate that CK2β interacts with NOG1 of both T. brucei and S. cerevisiae, suggesting that the interaction of NOG1 and CK2β may be one means of associating CK2 with the pre-ribosomal particle. The web of interactions in trypanosomes is extended by our previous work showing that CK2α interacts with the nucleolar protein NOPP44/46 , which in turn interacts with NOG1 . NOPP44/46 also interacts with p34/37 , which binds 5S rRNA . It is likely that several of these interactions are transient, perhaps even sequential, rather than all of the proteins residing together in a stable particle. For example, most of NOG1 resides in the 66S pre-ribosome, whereas NOPP44/46  and CK2 (data not shown) do not. Little is known about CK2 substrates in trypanosomes. In addition to NOPP44/46 , tubulin also appears to be a substrate of trypanosomatid CK2 .
Despite the use of multiple methodologies, we were unable determine whether CK2α’ interacts with CK2β. However, one conclusion that can be made based on the TAP purification is that unlike CK2α, CK2α’ is not present in high quantities in the CK2β complexes. This result could be due to a weak interaction or a much lower abundance of CK2α’ protein as compared to CK2α in procyclic forms (although the abundances of the cognate mRNAs are similar). Our consistent difficulty in obtaining significant expression of tagged CK2α’ even though CK2α was expressed abundantly using the same promoter, and 5′ and 3′ UTRs, suggests that expression of this protein may be under tight control.
As mentioned earlier, the distinctive features of CK2 activity include its near equal affinity for both GTP and ATP, modulation by polyamines, activation by salt, sensitivity to polybasic peptides, and inhibition by heparin. The T. brucei CK2 isoforms share only the last two traits. Effects of polyamines, salt, and polybasic peptides have been shown to be mediated by a region of CK2β which has multiple acidic residues (underlined in Figure 1). Photoaffinity labeling studies using a spermine analog demonstrated covalent linkage to a residue within aa 72–78 of Drosophila CK2β , which contains two acidic residues conserved in TbCK2β. When three nearby acidic residues were mutated to alanine, the Drosophila enzyme showed higher basal activity and less stimulation by spermine . These three acidic residues (60, 61, 63 in human and Drosophila, filled circles on Figure 1) are not conserved in T. brucei. In fact, within the “acidic” region, two basic residues are found in the trypanosomatid CK2βs. Hence the lack of stimulation of CK2 activity by polyamines fits well with data in other systems. The biological implications are unclear since CK2 activity does not appear to be directly modulated by polyamines in vivo .
In other organisms, the interaction of CK2α with CK2β results in a decrease in enzymatic activity unless salt is added. This finding has been hypothesized to be due to tertiary interactions between the α2β2 tetramers (deduced from secondary contacts seen in the crystal structure), such that two acidic residues of the β subunits contact the active site via ionic interactions . Addition of salt would overcome this interaction. Since the trypanosomatid CK2β proteins lack these specific acidic residues (see Figure 1, arrows), the holoenzyme would not be expected to show activation by salt. Indeed we saw no evidence for salt activation.
In other systems, the ability of poly-L-lysine to stimulate phosphorylation is specific to certain substrates, of which calmodulin is the prototype. Mutation of acidic residues 55, 57, or the triad 59–61 strongly reduced poly-L-lysine stimulation of human CK2 . Since the T. brucei enzyme lacks acidic residues at aa 59–61 (open circles), we did not expect poly-L-lysine to strongly activate CK2 phosphorylation of calmodulin. However, strong activation towards this substrate was observed, comparable to that seen with wild-type human enzyme by the above authors. Calmodulin is a physiological substrate of CK2 , and its CK2 phosphorylation sites (T79, S81, S101) are maintained in T. brucei. CK2 activation by polybasic peptides is not restricted to poly-L-lysine but can be achieved with proteins such as histones [18–20], suggesting that this conserved property may be physiologically meaningful.
The heparin sensitivity of T. brucei CK2 is comparable to that of other CK2s. The final hallmark of CK2, the ability to use GTP as a phosphate donor, is also largely absent in the T. brucei enzymes. As reported earlier, one of the catalytic subunits (α) has only a weak affinity for GTP; while here we show evidence that the second subunit (α’) has no detectable affinity for GTP. Asn118 in subdomain V has previously been shown to be important in allowing both ATP and GTP utilization in Xenopus CK2α since its deletion or mutation to Asp reduced the ability to utilize GTP [55,56]. While TbCK2α’ possesses a Pro at this site (a substitution that was not tested), TbCK2α has a tolerated substitution (Ala). Recent analysis of human CK2α has shown that mutation of Val66 to Ala (the residue found in 87% of human protein kinases) and Met163 to Leu (the residue found in 78% of human protein kinases) abrogates the ability to utilize GTP . Val66 is conserved in both trypanosome isoforms, but Met163 is replaced by Ala in CK2α and Ile in CK2α’. It seems likely that the inability of the T. brucei isoforms to use GTP results from a constellation of differences that combine to alter the purine binding site. Taken together, our results show that T. brucei CK2 possesses a number of unusual properties, many of which correlate with structure-function predictions from other systems.
The E7 anti-beta tubulin mAb generated by Michael Klymkowsky was obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the NICHD and maintained by The University of Iowa, Department of Biological Sciences, Iowa City, IA 52242.
The authors thank Joe Stout for technical assistance. They additionally would like to thank Paul Englund and Mark Drew for the gift of pZJM, Achim Schnaufer for pLEW-TAP, and George Cross and Elizabeth Wirtz for both 29–13 and single marker cells. This work was supported in part by NIH R01 AI31077. DB received support from NIH F32 AI10637.