Identifying interaction motifs in CK2β – a ubiquitous kinase regulatory subunit

Identifying interaction motifs inCK2b – a ubiquitous kinaseregulatory subunitVictor Martin Bolanos-Garcia1, Juan Fernandez-Recio1,2, Jorge E. Allende3 andTom L. Blundell1

1 Department of Biochemistry, University of Cambridge, 80 Tennis Court Road, Cambridge CB2 1GA, UK2 Institut de Recerca Biomedica (IRB-PCB), Parc Cientific de Barcelona, 08028 Barcelona, Spain3 Instituto de Ciencias Biomedicas, Facultad de Medicina, Universidad de Chile, Santiago 650499, Republic of Chile

Opinion TRENDS in Biochemical Sciences Vol.31 No.12

Glossary

Acidic loop: The region of CK2b encompassing residues 55–64 that is

characterized for acting as the binding site for various substrates and that

adopts more than one conformation.

Destruction box: A conserved motif of nine amino acids (RxxLxxxxN) required

for the programmed proteolytic degradation of proteins such as mitotic cyclins

and substrates of the anaphase promoting complex.

Non-obligate protein–protein complexes: Complexes with dissociation

constants in the micromolar range that are formed by proteins that can also

exist in the free form at physiological concentration. Typically, these

complexes involve smaller contact areas between protomers and more planar

and polar interfaces on average than obligate complexes.

Nuclear localization sequence: Amino acid sequence required for the transport

of nuclear proteins to the nucleus via the nuclear pore complex. A typical

nuclear localization sequence consists of a short stretch of 4–8 amino acid

residues containing several lysine and arginine residues. Another typical

arrangement, known as bipartite nuclear localization sequence, contains two

Casein kinase 2 (CK2) is probably the most ubiquitousserine/threonine kinase found in eukaryotes: it phos-phorylates >300 cellular proteins, ranging from transcrip-tion factors to proteins involved in chromatin structureand cell division. CK2 is a heterotetrameric enzyme thatinduces neoplastic growth when overexpressed. The bsubunit of CK2 (CK2b) functions as the regulator of thecatalytic CK2a and CK2a0 subunits, enhancing their sta-bility, activity and specificity. However, CK2b also func-tions as a multisubstrate docking platform for severalother binding partners. Here, we discuss the organizationand roles of interaction motifs of CK2b, postulate newprotein-interaction sites and map these to the knowninteraction motifs, and show how the resulting complex-ity of interactions mediated by CK2 gives rise to theversatile functions of this pleiotropic protein kinase.

IntroductionNative casein kinase 2 (CK2) is a constitutively activeenzyme that is regulated not by second messengers butrather through protein–protein interactions and changesin its concentration, extent of phosphorylation, state ofoligomerization and localization [1,2]. Most of the 300 ormore substrates of CK2 reported so far correspond toproteins that participate in cell signalling [1]. The CK2holoenzyme forms a heterotetrameric complex consistingof two catalytic (CK2a and CK2a0) and two regulatory(CK2b) subunits (i.e. a a2b2 complex). The two CK2b

subunits associate to form a stable dimer, and this dimer-ization can take place in the absence of the catalyticsubunits. Moreover, assembly of CK2 complexes contain-ing the catalytic subunits requires the prior formation ofdimers of the regulatory CK2b subunit [3].

In contrast to the activity of regulatory subunits of otherkinases such as cAMP- and cyclin-dependent proteinkinase, CK2b does not switch on or off the intrinsic activityof the catalytic subunits [4]. Although CK2b seems tointeract directly with >40 different proteins [5,6], theimportance in living cells of some of these interactionsremains to be confirmed. The interacting partners ofCK2b include other protein kinases such as A-Raf,

Corresponding author: Bolanos-Garcia, V.M. ([email protected]).Available online 3 November 2006.

www.sciencedirect.com 0968-0004/$ – see front matter � 2006 Elsevier Ltd. All rights reserve

Chk1, Chk2, PKC-z, Mos and p90rsk [7–12]. The fact thatthese interactions often compete with CK2a bindingindicates that they might arise through recognition byCK2b of motifs that are common to several other proteinkinases in addition to CK2a. Many interaction partners ofCK2b are potential substrates of CK2, such as the putativetumour suppressor protein Doc1, the mitochondrial trans-lational initiation factor 2, the propionyl CoA carboxylase b

subunit, the Fas-associated protein FAF1, fibroblastgrowth factor 2, p53, p21WAF1 and p27KIP1 [12–17].

On the basis of this large number of interactingpartners, it is not surprising that CK2b also has functionsin addition to regulating the CK2 holoenzyme. Indeed,CK2b has an inhibitory effect on some substrates [18], ithas been detected in locations where catalytic subunitsare apparently absent [19], and in normal cells it issynthesized in higher concentrations than are the cata-lytic CK2a and CK2a0 subunits [20]. Furthermore, theconcentration of CK2b in tumour cells differs from that innormal tissue [21], and in mice the functional loss ofCK2b is lethal [22].

Here, we describe the organization of the various motifsof CK2b that mediate protein–protein interactions anddiscuss the roles of this subunit in CK2-dependent andCK2-independent regulation. In addition, we proposenovel protein interaction sites in this subunit using

interdependent positively charged clusters separated by a linker region of 10–

12 amino acids.

d. doi:10.1016/j.tibs.2006.10.005

mailto:[email protected]

http://dx.doi.org/10.1016/j.tibs.2006.10.005

Box 1. ODA: a new predictor of protein interaction sites

The optimal docking area (ODA) method identifies protein surface

patches with propensity for protein–protein interactions on the basis

of desolvation energy [62]. The method calculates the effective

desolvation energy for burying a continuous surface patch from the

solvent according to atomic solvation parameters [23]. The docking

surface energy is calculated for patch residues that lie within a

sphere of radius d (where d = 1, 2, . . ., 20 A) from the centre of

coordinates of each residue side chain. The patch with lowest

desolvation energy value is called an ODA. Significant ODAs are

those with an optimal desolvation energy of less than �10.0 kcal/

mol, and can be graphically displayed on the surface of the proteins

by colouring each surface residue according to its ODA value

(typically from blue to red, indicating increasing propensities for a

docking site). The ability of the ODA method to identify protein

interaction sites has been validated on a set of three-dimensional

structures of non-homologous proteins involved in non-obligate

protein–protein interactions. Significant ODAs were found in half of

the test proteins, and >80% of the ODAs identified corresponded to

a known protein interaction site [62].

The ODA method has proved to be a fast and powerful tool for

structural analyses of protein–protein interactions. For example,

ODA analysis of multiprotein drug efflux pumps in bacteria has

identified important differences between the outer membrane

channels TolC (Escherichia coli) and VceC (Vibrio cholerae) with

implications for the proposed mode of interaction with the inner

membrane proton antiporters AcrAB and VceAB, respectively

[63,64]. In addition, ODA analysis of phosphoenolpyruvate carbox-

ylase (PEPC) has provided structural hints for a novel mechanism of

allosteric regulation and has helped to derive a conserved mode of

tetramer formation for archaeal, bacterial and eukaryote phosphoe-

nolpyruvate carboxylases [65]. More recently, ODA calculations

have provided rationales for the mechanism of interaction of

different polygalacturanases with their inhibitors – an interaction

that is important for plant innate immunity [66].

Opinion TRENDS in Biochemical Sciences Vol.31 No.12 655

bioinformatics tools that include estimation of the optimaldesolvation area (ODA; Box 1), residue conservation andphylogenetic trace analysis [23–27].

Organization of the protomer and the homodimerThe CK2b regulatory subunit is a compact, globularhomodimer that shows high conservation across species(Figure 1a). The N-terminal domain (residues 5–104) isglobular and organized as four a-helices (a1, a2, a3 anda4). Helices a1 (residues 9–14), a2 (residues 27–31) and a3(residues 46–54) wrap around a4 (residues 66–89), whichholds a protruding ‘acidic loop’ (residues 55–64; see Glos-sary). Helices a4 and a5 barely interact because they makean angle of 958 with each other. Instead, a5 shows moreinteractions with residues located in the zinc-finger region(Figure 1b), and it seems more appropriate to consider a5(residues 91–102) as part of this latter region.

Residues 105–147 form the core of the zinc-finger [28].Each Zn2+ ion is coordinated with tetrahedral symmetry byfour conserved cysteine residues: Cys109, Cys114, Cys137and Cys140. The crystal structure of the CK2b dimershows that the zinc-finger motifs of each protomer interactgreatly with each other and that the protein–protein inter-face thus formed involves hydrophobic and polar interac-tions. Although a small area is buried in the interface(540 A2 for each protomer), eight hydrogen bonds formedbetween interfacial residues contribute further to the sta-bility of the CK2b dimer. Helix a6 (residues 163–175)establishes contacts with the strand b1 of the zinc-fingermotif, which is a chief contributor to dimer stability, as the

www.sciencedirect.com

mutation of Cys109 and Cys114 to serine has shown [28].Furthermore, 14 out of 42 residues in the dimer interface ofthe zinc-finger motif are absolutely conserved and theremaining 28 are highly conserved, underlining therelevance of dimer formation for CK2 activity [29]. Theseobservations are consistent with reports that Lys139 has arole in homodimerization, but inconsistent with reportsthat Asn67 and Met52 might also be involved [30].

The C-terminal region (residues 178–205) contains alarge loop (residues 178–193) and helix a7 (residues 194–200). This region also contains the residues Thr213, whichis phosphorylated by the checkpoint kinase Chk1 [31], andSer209, which is phosphorylated in a cell-cycle-dependentmanner by p34cdc2 [32]. Helix a7 is located away fromhelices a1 to a6 of the same protomer; however, theC-terminal residues 190–205 contribute to the formationof the CK2b dimer [29] (Figure 1b). Deletion of the last 37C-terminal residues (CK2b mutant D179–215) results in afairly stable CK2b dimer, whereas the full-length CK2b

dimer on its own is prone to aggregation in vitro [33].

Interaction motifs of CK2b and their protein partnersThe functions of CK2b are mediated by several interactionsequence motifs (Figure 1a). When mapped onto the three-dimensional structure, these motifs can be most usefullyconsidered in the context of organization of the structureinto three main regions: the N-terminal, juxta-dimer inter-face and C-terminal regions (Figure 1b).

N-terminal region

The N-terminal region contains two main CK2b

autophosphorylation sites, Ser2 and Ser3 [34], which arehighly conserved between species. On the basis of residueconservation, it seems likely that Ser4 also constitutesanother important autophosphorylation site. Indeed, ithas been shown that CK2 phosphorylates adjacent serineresidues [35,36]. Autophosphorylation of CK2 on CK2b

seems to take place through an ‘intra-oligomeric’ mechan-ism in which the two CK2b subunits of a protomer arephosphorylated by the catalytic subunits of another adja-cent protomer [37]. Moreover, additional phosphorylationmechanisms seem to exist in vivo, because phosphorylationof free CK2b by the p34cdc2/cyclinB kinase, but not phosphor-ylation of CK2b associated with the CK2 holoenzyme, hasbeen observed [38].

The N-terminal region of CK2b contains 20 of the total28 glutamic acid and aspartic acid residues present inthe primary structure. All acidic residues contribute asurface area of 3600 A2, which represents >40% of thetotal surface area of the CK2b protomer [28]. This regionpresents docking sites for protein substrates involved incell localization and activation, translation control andprotein degradation [1,39]. The segment Tyr46–Leu54 ofCK2b located in helix a3 contains a putative ‘destructionbox’ resembling that of cyclin B [40]. CK2b shows othersimilarities to cyclin proteins, including ligand specificity,modulation of CK2 catalytic activity, the formation ofcomplexes and cell-cycle-dependent activity. The destruc-tion box is close to Ser2 and Ser3, and replacing these tworesidues significantly decreases CK2b degradation, indi-cating that autophosphorylation is also involved in regu-

Figure 1. Structural features of CK2b. (a) Aligned sequences of CK2b from different organisms, showing overlap of some of the interaction motifs. The motifs are coloured

as follows: main phosphorylation sites (residues 2–4), lime green; Nopp140 interaction site (5–20), orange; destruction box (46–54), magenta; acidic loop (55–64), red; dimer

interface residues (143–148), dark blue; residues involved in interaction with the cell-cycle regulators p21 and p53 (106–116, 124, 134, 141, 145–149, 152), grey; C-terminal

region including the CK2a-binding residues (187–192), A-Raf- and Mos-binding residues, pale pink (187–205). The CK2b residues involved in interaction with L41 are boxed

in brown. Residue Met195 of human CK2b, which corresponds to the site of the only single nucleotide polymorphism identified so far, is shown in pink. The potential new

interaction motifs described here are shown in green. The secondary structure elements indicated are taken from PDB accession code 1JWH (residues 3–205) [29].

(b) Ribbon representations of the CK2b dimer. The two views are perpendicular to the two-fold axis (left) and along it (right). In both representations, the N-terminal, zinc-

finger and C-terminal regions are shown for one protomer (left). The Zn2+ ion of the zinc-finger is shown as a blue sphere. The seven a-helices of this protomer are indicated

and the interaction motifs are coloured as in (a). Figure generated with Pymol [61].

656 Opinion TRENDS in Biochemical Sciences Vol.31 No.12

lating the stability of this subunit [41]. It is possible thatthe autophosphorylation of Ser2 and Ser3 affects theaccessibility of substrates that bind to the destructionbox. The first 20 amino acid residues of CK2b are alsoinvolved in interaction with Nopp140, a protein that bindsthe ‘nuclear localization sequence’ and shuttles between


the cytoplasm and the nucleus [42]. Thus, it is likely thatthe order in which substrate interactions take place withthese two overlapping motifs has considerable importancein the regulation of CK2 functions.

Another protein that specifically binds to the N-terminalregion of CK2b is L41. This highly basic, ribosomal protein


stimulates the phosphorylation of DNA topoisomerase IIaas a result of its direct association with CK2b [43]. Asequence analysis of point mutants in this CK2b regionhas revealed that Asp26, Met52 and Met78 are crucial forinteraction with L41 [43]. Residues Asp26 and Met52 alsoform part of the overlapping destruction box and acidic-loopmotifs. The N-terminal region of CK2b also contains thebinding site for the cytoplasmic domain of CD5, a trans-membrane protein expressed on T cells and the B1 subset ofB cells that modulates activation mediated by the antigenreceptor. Indeed, this CK2b–CD5 interaction has led to theidentification of anovel pathway for controllingactivation inT and B cells [44].

The electron density of the acidic loop is not well definedin two CK2b homodimer crystal structures that have beenreported recently (PDB accession code 1RQF [33] and1JWH [28]), indicating that this region is very flexible.Similar clusters of acidic amino acids are typicallyobserved in CK2 substrates, and it is speculated thatresidues 55–64 (DLEPDEELED) of CK2b are autoinhibi-tory sequences, reminiscent of those present in otherkinases [45]. This region of CK2b participates in down-regulation of CK2 [46] and contains the binding motif forpolybasic peptides that trigger CK2 activity towards cal-modulin [47]. Mutations that neutralize negative chargesin this segment raise the basal activity of the holoenzyme,indicating that changes in the size and shape of this regiongreatly affect CK2 function [48]. Initial reports suggestedthat region 55–64 of CK2b contains the binding motif forpolyamines such as spermidine [49], whereas other reportsidentified different spermidine-interacting residues suchas Tyr80 [50]. More recent studies have shown that sper-midine does not interact with CK2b in living cells [51].

Juxta-dimer interface region

In CK2b, residues Asp105, Arg111, Glu115, Lys134,Asp142 and Lys147 are scattered at the edge of the zinc-fingermotif in the juxta-dimer interface region. Side chainsof some of these conserved residues are exposed to thesolvent and therefore can accommodate multiple proteinligands [28,33]. Proteins that directly interact with thejuxta-dimer interface region of CK2b include the keyregulators of cell division p53 [15] and p21WAF1 [16,33].p21WAF1, a potent inhibitor of cyclin-dependent kinases(CDKs), belongs to thekinase inhibitorprotein (KIP) family.Inhibition of CDKs induces cell-cycle arrest in G phase.

p21WAF1 and,more recently, p27KIP1 have been reportedto bind specifically to the regulatory CK2b subunit but notto the catalytic CK2a subunits [17]. In a study usingdeletion mutants, it has been shown that the regionbetween residues 72 and 149 of CK2b is necessary forp53 binding [15]. The spatial disposition of docking sitesand the twofold, non-crystallographic symmetry of thedimer might suggest that more than one molecule ofthese cell-cycle regulators can bind at the same time. Inthe crystal structure of a complex formed between ap21WAF1 peptide (residues 45–65) and CK2b that we andour co-workers [33] recently reported, however, only onemolecule of peptide interacts with the CK2b dimer (PDBaccession code 1RQF). The residues of CK2b involved inthis interaction with the p21WAF1 peptide are Tyr113 and


Glu115 from one protomer, and Leu124, Glu130, Ala131,Lys134, Thr145 and His152 from the other [33]. Whetheror not p27KIP1, a cell-cycle inhibitor closely related top21WAF1, binds to this region of CK2b, remains to beinvestigated. Interestingly, residues 123–127 (GLSDI) cor-respond to the epitope for the only monoclonal antibody toCK2b described so far [52].

C-terminal region

The C-terminal region of CK2b (residues 178–205) isinvolved in formation of the CK2b homodimers, but it alsohas other functions (Figure 1b). The crystal structure of theCK2 holoenzyme shows that this region, specifically resi-dues 187–192, is responsible for the direct interaction ofCK2b with the catalytic CK2a and CK2a0 subunits [29].The crystal structure of the CK2 holoenzyme (a2b2) showsthat the average size of the CDK2a/CDKb interface is�832 A

´ 2, which is within the range commonly observedin ‘non-obligate protein–protein complexes’ [53].

Because the dissociation constant (Kd) of the interactionbetween the CK2a and CK2b subunits of CK2 is very high(Kd = 5.4 nM), it was initially suggested that neither sub-unit would dissociate from the assembled holoenzyme.Although the molecular mechanisms that control the dis-sociation of CK2a and CK2b have not been described, thepossibility that CK2a and CK2b can in fact associate anddissociate is supported by recent microscopic observations,in which CK2a and CK2b seem to exist independently invivo [54]. Moreover, the asymmetric expression of theCK2b subunit in different cell compartments might haveimportant implications in CK2-dependent and CK2-inde-pendent regulation (Box 2).

The C-terminal region of CK2b also mediatesassociations with the human cell-cycle checkpoint kinaseChk1 and the Mos and A-Raf protein kinases, and theseassociations apparently involve some of the residues thatinteract with CK2a [6,55]. CK2b seems to recognize resi-dues neighbouring the glycine-rich conserved sequence 1 ofprotein kinases. The interaction between CK2b and Chk1constitutes a novel mechanism of Chk1 activation andleads to an increase in the Cdc25C-phosphorylating activ-ity of Chk1 [31], underlining the important role that CK2b

has in cell-cycle regulation. In contrast its activation ofChk1, the interaction of CK2b with Mos, a mitogen-acti-vated protein kinase required for the coordinated activa-tion of Cdc2 and the G2–meiosis I transition [56],constitutes an example of the dual roles of CK2b becauseit negatively regulates progesterone-induced maturation[57]. The interaction of CK2b (residues 194–200) with A-Raf leads to abnormal activation of the latter. It remains tobe determined whether the enhanced A-Raf activated isassociated with solid human tumours and proliferatingtissue formation. Remarkably, only A-Raf, and not B-Rafor C-Raf-1, interacts with CK2b [58]. This direct, isoform-specific activation of Raf by CK2b links two differentprotein kinases that were initially thought to functionseparately in vivo.

Predicting potential interaction sitesIn addition to direct structure determination, our group[23,24] has also used bioinformatics approaches that

Box 2. Does CK2b undergo CK2-independent regulation?

In 1998, Allende and Allende [67] suggested that CK2b might be a

‘wild-card’ regulatory subunit that could regulate several kinases.

Indeed, the asymmetric expression of CK2b might constitute an

important in vivo regulatory mechanism of CK2 – a notion that is

supported by recent experimental evidence. For example, the Ste

protein, which competes with CK2b for binding to the catalytic CK2a

subunit, gives rise to the formation of oligomers that have activity

and/or regulatory properties altered to such an extent that the

addition of polylysine no longer stimulates CK2 [68]. Furthermore,

fluorescence imaging of living cells has demonstrated the asymmetric

concentration and independent movement of CK2a and CK2b

subunits and has facilitated the identification of slow- and fast-

moving free subunits [54,69].

Lastly, in Arabidopsis thaliana, a genome-scale search of CK2

subunit expression has revealed that the CK2b subunit is more

diversely distributed than the catalytic CK2a subunits [70]. Moreover,

this study has shown that some CK2b isoforms are present in the

nucleus and the cytosol, whereas others are located exclusively in one

of these compartments [70]. The higher concentrations of the CK2b

subunit, with respect to those of the catalytic CK2a and CK2a0

subunits, in a specific subcellular compartment imply the existence of

both CK2-dependent and CK2-independent responses (Figure I). In

this scheme, the high concentration of CK2b recruits the catalytic

CK2a and CK2a0 subunits to form the CK2 holoenzyme (Figure Ia).

Partners that interact with CK2b when assembled as part of the CK2

holoenzyme result in a CK2-dependent cellular response. The

remaining pool of free CK2b is available to interact with protein

substrates in a CK2-independent manner (Figure Ib,c). Moreover, the

establishment of these possible regulatory mechanisms might

depend not only on the relative concentration of free CK2b and its

substrates in a particular subcellular compartment but also on their

phosphorylation and/or oligomerization state. One of the outcomes of

CK2b binding to CK2a is that the catalytic activity can be exported out

of the cell to function as an ectokinase [71]. Thus, it might be possible

that binding to the CK2b dimer serves as an export mechanism for

some substrates.

Figure I. Effect of dimeric CK2b concentration on CK2-dependent and CK2-independent responses. Shown is the potential role of dimeric CK2b in cell regulation when,

in a particular subcellular compartment, the concentration of the CK2b dimer is much higher than that of the catalytic CK2a and CK2a0 subunits. (a) This asymmetric

concentration might act as the driving force that leads to a greater degree of interactions of the CK2 holoenzyme with substrates that specifically or preferentially

interact with CK2b, thus conferring specificity to CK2. (b) The pool of free CK2b available might bind other kinases such as Ras-A and Mos to modify their catalytic

activity in a CK2-independent fashion. (c) When the binding to another target protein (such as p53, p21, p27) does not involve the C terminus of CK2b (178–205), then

binding occurs irrespective of whether the CK2b is bound to CK2a or another protein kinase.


facilitate the identification of potential interaction sites.For example, the ODA method (Box 1) indicates that themost favourable docking region in CK2b is positionedaround the binding site for CK2a and CK2a, and thatresidues YGL (i.e. residues 80–82 in a4), HARY (residues84–87 in a4) and DGA (residues 155–157 in the loop regionlocated between b3 and a6) might constitute substrateinteraction sites not previously described (Figure 2). Thepotential interaction sites are highly conserved among


species and define two novel interaction motifs(Figure 2). Computer programs that predict interfaceresidues on the basis of the evolutionary relationshipsamong sequence homologues, such as ConfSurf [27], andthose that characterize functional regions according tophysicochemical and structural constrains in additionto evolutionary conservation, such as ProMate [25] andCrescendo [26], also predict that residues YGL, HARY andDGA will lie at the interface.

Figure 2. Comparison of surface residues of CK2b and Cyclin A identified by the ODA method. The absolute ODA scale, 0–40, is shown on the right; the colours, ranging

from light blue to red, represent increasing propensities for a binding site. (a) Interaction of CK2b with CK2a. To aid comparison with the CyclinA–CDK2 complex, the

interaction of only one CK2b protomer with the catalytic CK2a subunit (cyan) is shown. (b) CyclinA–CDK2 complex. Cyclin A presents one interaction region (the XRL motif)

that is involved in CDK2 binding. Note the scattered distribution and far larger number of potential docking sites in CK2b than in Cyclin A. (c) Predicted interaction sites in

one CK2b protomer. The catalytic CK2a subunit is shown (cyan) to aid visualization of these sites in CK2b.


A comparison of the ODAs of CK2b–CK2a and CyclinA–CDK2 complexes (Figure 2) shows that the number ofODAs is higher in CK2b than in Cyclin A – a feature thatis consistent with the far larger number of CK2b ligands.The CK2a and CK2a0 catalytic subunits and the CDK2 arekinases closely related in the phylogenetic tree, sharing33% identity [59]. CDK2 activity is regulated by its directinteraction with Cyclin A [60]. Figure 2 also shows that, incontrast to the wide distribution of docking sites in CK2b,the main Cyclin A substrate-binding region is unambigu-ously located in a surface-exposed hydrophobic site com-monly referred as the ‘RXL recognition motif’.

Although the characterization in vitro and in vivo of thepotential protein interaction sites of CK2b remains to beaccomplished, the effectiveness of this approach for iden-tifying interaction sites has been demonstrated in varioussystems (Box 1).

Concluding remarksThe CK2b subunit of the CK2 holoenzyme showsremarkable plasticity. The organization of motifs inCK2b constitutes the structural basis of specificity forthe astonishing number of CK2 substrates identified sofar and is the hallmark of the function of CK2b as amultisubstrate docking platform. In addition to the reg-ulation of CK2 activity and stability imposed by thevarious interaction motifs in CK2b, the asymmetric


expression of free subpopulations of this subunit indifferent cell compartments might drive its directassociation with specific substrates and might contributeto modulation of substrate turnover. We suggest that suchassociations are likely to imply the existence of both CK2-dependent and CK2-independent interactions.

The pace of progress in the development ofbioinformatics tools has facilitated the identification ofpotential protein–protein interaction sites, includingthose that might constitute the target of new therapeuticagents. On the basis of calculations of surface residueswith normalized interface propensity, evolutionary con-servation among homologous sequences and ODAs, itseems likely that additional residues in CK2b will beidentified to interact directly with as yet unidentifiedprotein substrates. Because the experimental verifica-tion of sites predicted to be involved in protein–proteininteractions is lagging behind, the development of high-throughput methods to characterize such sites isurgently needed.

It is our hope that the ideas discussed here willstimulate further studies on the biological and structuralcharacterization of CK2 and its regulatory subunit.

References1 Meggio, F. and Pinna, L.A. (2003) One-thousand-and-one substrates of

protein kinase CK2? FASEB J. 17, 349–368


2 Niefind, K. and Issinger, O.G. (2005) Primary and secondaryinteractions between CK2a and CK2b lead to ring-like structures inthe crystals of the CK2 holoenzyme. Mol. Cell. Biochem. 274, 3–14

3 Graham, K.C. and Litchfield, D.W. (2000) The regulatory b subunit ofprotein kinase CK2 mediates formation of tetrameric CK2 complexes.J. Biol. Chem. 275, 5003–5010

4 Sarno, S. et al. (1999) Cooperative modulation of protein kinase CK2by separate domains of its regulatory b-subunit. Biochemistry 39,2324–2329

5 Filhol, O. et al. (2004) Protein kinase CK2: a new view of an oldmolecular complex. EMBO Rep. 5, 351–355

6 Guerra, B. and Issinger, O.G. (1999) Protein kinase CK2 and its role incellular proliferation, development and pathology. Electrophoresis 20,391–408

7 Boldyreff, B. and Issinger, O-G. (1997) A-Raf kinase is a newinteracting partner of protein kinase CK2 b subunit. FEBS Lett.403, 197–199

8 Guerra, B. et al. (2003) Modulation of human checkpoint kinaseChk1 by the regulatory b-subunit of protein kinase CK2. Oncogene22, 4933–4942

9 Bjorling-Poulsen, M. et al. (2005) The ‘regulatory’ b-subunit of proteinkinase CK2 negatively influences p53-mediated allosteric effects onChk2 activation. Oncogene 24, 6194–6200

10 Bren, G.D. et al. (2000) PKC-z-associated CK2 participates in theturnover of free IkBa. J. Mol. Biol. 297, 1245–1258

11 Kusk, M. et al. (1999) Interactions of protein kinase CK2b subunitwithin the holoenzyme and with other proteins. Mol. Cell. Biochem.191, 51–58

12 Grein, S. et al. (1999) Searching interaction partners of proteinkinase CK2b subunit by two-hybrid screening. Mol. Cell. Biochem.191, 105–109

13 Olsen, B. et al. (2003) Protein kinase CK2 phosphorylates theFas-associated factor FAF1 in vivo and influences its transport intothe nucleus. FEBS Lett. 546, 218–222

14 Bonnet, H. et al. (1996) Fibroblast growth factor-2 binds to theregulatory b subunit of CK2 and directly stimulates CK2 activitytoward nucleolin. J. Biol. Chem. 271, 24781–24787

15 Appel, K. et al. (1995) Mapping of the interaction sites of the growthsuppressor protein p53 with the regulatory b-subunit of protein kinaseCK2. Oncogene 11, 1971–1978

16 Romero-Oliva, F. and Allende, J.E. (2001) Protein p21WAF1/CIP1 isphosphorylated by protein kinase CK2 in vitro and interacts withthe amino terminal end of the CK2 b subunit. J. Cell. Biochem. 81,445–452

17 Tapia, J.C. et al. (2004) Cell cycle regulatory protein p27KIP1 is asubstrate and interacts with the protein kinase CK2. J. Cell.Biochem. 91, 865–879

18 Schuster, N. et al. (1999) Regulation of p53 mediated transactivationby the b-subunit of protein kinase CK2. FEBS Lett. 447, 160–166

19 Guerra, B. et al. (1999) Protein kinase CK2: evidence for a proteinkinase CK2b subunit fraction, devoid of the catalytic CK2a subunit, inmouse brain and testicles. FEBS Lett. 462, 353–357

20 Korn, I. et al. (2001) The activity of CK2 in the extracts of COS-7 cellstransfected with wild type and mutant subunits of protein kinase CK2.Mol. Cell. Biochem. 227, 37–44

21 Stalter, G. et al. (1994) Asymmetric expression of protein kinase CK2subunits in human kidney tumors. Biochem. Biophys. Res. Commun.202, 141–147

22 Buchou, T. et al. (2003) Disruption of the regulatory b subunit ofprotein kinase CK2 in mice leads to a cell-autonomous defect andearly embryonic lethality. Mol. Cell. Biol. 23, 908–915

23 Fernandez-Recio, J. et al. (2004) Identification of protein–proteininteraction sites from docking energy landscapes. J. Mol. Biol. 335,843–865

24 Fernandez-Recio, J. et al. (2005) Optimal docking area: a new methodfor predicting protein–protein interaction sites. Proteins 58, 134–143

25 Neuvirth, H. et al. (2004) ProMate: a structure based predictionprogram to identify the location of protein–protein binding sites.J. Mol. Biol. 338, 181–199

26 Chelliah, V. et al. (2004) Distinguishing structural and functionalrestraints in evolution in order to identify interaction sites. J. Mol.Biol. 342, 1487–1504


27 Glaser, F. et al. (2003) ConSurf: identification of functional regions inproteins by surface-mapping of phylogenetic information.Bioinformatics 19, 163–164

28 Chantalat, L. et al. (1999) Crystal structure of the human proteinkinase CK2 regulatory subunit reveals its zinc finger-mediateddimerization. EMBO J. 18, 2930–2940

29 Niefind, K. et al. (2001) Crystal structure of human protein kinase CK2:insights into basic properties of the CK2 holoenzyme. EMBO J. 20,5320–5331

30 Canton, D.A. et al. (2001) Assembly of protein kinase CK2:investigation of complex formation between catalytic and regulatorysubunits using a zinc finger-deficientmutant of CK2b.Biochem. J. 358,87–94

31 Kristensen, L.P. et al. (2004) Phosphorylation of the regulatoryb-subunit of protein kinase CK2 by checkpoint kinase Chk1:identification of the in vitro CK2b phosphorylation site. FEBS Lett.569, 217–223

32 Litchfield, D.W. et al. (1995) The protein kinase from mitotic humancells that phosphorylates Ser-209 on the casein kinase II b-subunit isp34cdc2. Biochim. Biophys. Acta 1269, 69–78

33 Bertrand, L. et al. (2004) Structure of the regulatory subunit of CK2 inpresence of a p21WAF1 peptide demonstrates flexibility of the acidicloop. Acta Crystallogr. D60, 1698–1704

34 Litchfield, D.W. et al. (1991) Phosphorylation of the b subunit of caseinkinase II in human A431 cells. Identification of theautophosphorylation site and a site phosphorylated by p34cdc2.J. Biol. Chem. 266, 20380–20389

35 Pinna, L.A. (1990) Casein kinase 2: an ‘eminence grise’ in cellularregulation? Biochim. Biophys. Acta 1054, 267–284

36 Luscher, B. et al. (1990) Myb DNA binding inhibited byphosphorylation at a site deleted during oncogenic activation.Nature 344, 517–522

37 Pagano, M.A. et al. (2005) Autophosphorylation at the regulatory b

subunit reflects the supramolecular organization of protein kinaseCK2. Mol. Cell. Biochem. 274, 23–29

38 Meggio, F. et al. (1995) Phosphorylation and activation of proteinkinase CK2 by p34cdc2 are independent events. Eur. J. Biochem.230, 1025–1031

39 Bibby, A.C. and Litchfield, D.W. (2005) The multiple personalities ofthe regulatory subunit of protein kinase CK2: CK2 dependent and CK2independent roles reveal a secret identity for CK2b. Int. J. Biol. Sci. 1,67–79

40 Allende, J.E. and Allende, C.C. (1995) Protein kinases. 4. Proteinkinase CK2: an enzyme with multiple substrates and a puzzlingregulation. FASEB J. 9, 313–323

41 Zhang, C. et al. (2002) Phosphorylation regulates the stability of theregulatory CK2b subunit. Oncogene 21, 3754–3764

42 Li, D. et al. (1997) Specific interaction between casein kinase 2 and thenucleolar protein Nopp140. J. Biol. Chem. 272, 3773–3779

43 Ahn, B.H. et al. (2003) Identification of mutations in protein kinaseCKIIb subunit that affect its binding to ribosomal protein L41 andhomodimerization. J. Biochem. Mol. Biol. 36, 344–348

44 Raman, C. et al. (1998) Regulation of casein kinase 2 by directinteraction with cell surface receptor CD5. J. Biol. Chem. 273,19183–19189

45 Soderling, T.R. (1990) Protein kinases. Regulation by autoinhibitorydomains. J. Biol. Chem. 265, 1823–1826

46 Boldyreff, B. et al. (1993) Reconstitution of normal and hyperactivatedforms of casein kinase-2 by variably mutated b-subunits. Biochemistry32, 12672–12677

47 Meggio, F. et al. (1994) Casein kinase 2 down-regulation and activationby polybasic peptides are mediated by acidic residues in the 55–64region of the b-subunit. A study with calmodulin as phosphorylatablesubstrate. Biochemistry 33, 4336–4342

48 Leroy, D. et al. (1999) Dissecting subdomains involved in multiplefunctions of the CK2 b subunit. Mol. Cell. Biochem. 191, 43–50

49 Leroy, D. et al. (1995) Direct identification of a polyamine bindingdomain on the regulatory subunit of the protein kinase casein kinase 2by photoaffinity labeling. J. Biol. Chem. 270, 17400–17406

50 Leroy, D. et al. (1997) Binding of polyamines to an autonomous domainof the regulatory subunit of protein kinase CK2 induces aconformational change in the holoenzyme. A proposed role for thekinase stimulation. J. Biol. Chem. 272, 20820–20827


51 Lawson, K. et al. (2006) A novel protein kinase CK2 substrate indicatesCK2 is not directly stimulated by polyamines in vivo. Biochemistry 45,1499–1510

52 Nastainczyk, W. et al. (1995) Isolation and characterization of amonoclonal anti-protein kinase CK2 b-subunit antibody of the IgGclass for the direct detection of CK2 b-subunit in tissue cultures ofvariousmammalianspeciesandhumantumors.Hybridoma14, 335–339

53 Jones, S. and Thornton, J.M. (1996) Principles of protein–proteininteractions. Proc. Natl. Acad. Sci. U. S. A. 93, 13–20

54 Filhol, O. et al. (2003) Live-cell fluorescence imaging reveals thedynamics of protein kinase CK2 individual subunits. Mol. Cell. Biol.23, 975–987

55 Chen, M. et al. (1997) The casein kinase II b subunit binds to Mos andinhibits Mos activity. Mol. Cell. Biol. 17, 1904–1912

56 Frank-Vaillant, M. et al. (2001) Interplay between Cdc2 kinase and thec-Mos/MAPK pathway between metaphase I and metaphase II inXenopus oocytes. Dev. Biol. 231, 279–288

57 Lieberman, S.L. and Ruderman, J.V. (2004) CK2 b, which inhibits Mosfunction, binds to a discrete domain in the N-terminus of Mos. Dev.Biol. 268, 271–279

58 Hagemann, C. et al. (1997) The regulatory subunit of protein kinaseCK2 is a specific A-Raf activator. FEBS Lett. 403, 200–202

59 Litchfield, D.W. (2003) Protein kinase CK2: structure, regulation androle in cellular decisions of life and death. Biochem. J. 369, 1–15

60 Rosenblatt, J. et al. (1992) Human cyclin-dependent kinase 2 isactivated during the S and G2 phases of the cell cycle andassociates with cyclin A. Proc. Natl. Acad. Sci. U. S. A. 89, 2824–2828

61 DeLano, W.L. (2002) The PyMOL Molecular Graphics System. DeLanoScientific, San Carlos, CA, USA

Endeav

The quarterly magazine

philosophy o

You can access Ende

ScienceDirect, where you

editorial comment and a c

illustrated articles on th

Featuri

Death by hypnosis: an 189its European reverberat

NASA and the search for life iLinnaeus’ herbarium cabinet: a piece of furn

Civilising missions, natural history and British indFreudian snaps

Coming

Making sense of modernity’s maladies: economies of healEngineering fame: Isambard K

Disputed discovery: vivisection and experim‘‘A finer and fairer future’’: commodifying wage earn

‘But man can do his duty’: Charles Darwin

And much, mu

Endeavour is available on Science


62 Fernandez-Recio, J. et al. (2005) Optimal docking area: a newmethod for predicting protein–protein interaction sites. Proteins 58,134–143

63 Fernandez-Recio, J. et al. (2004) A model of a transmembrane drug-efflux pump from Gram-negative bacteria. FEBS Lett. 578, 5–9

64 Federici, L. et al. (2005) The crystal structure of the outer membraneprotein VceC from the bacterial pathogen Vibrio cholerae at 1.8 Aresolution. J. Biol. Chem. 280, 15307–15314

65 Matsumura, H. et al. (2006) A novel mechanism of allosteric regulationof archaeal phosphoenolpyruvate carboxylase: a combined approach tostructure-based alignment and model assessment. Protein Eng. Des.Sel. 19, 409–419

66 Sicilia, F. et al. (2005) The polygalacturonase-inhibiting protein PGIP2of Phaseolus vulgaris has evolved a mixed mode of inhibition ofendopolygalacturonase PG1 of Botrytis cinerea. Plant Physiol. 139,1380–1388

67 Allende, J.E. and Allende, C.C. (1998) Promiscuous subunitinteractions: a possible mechanism for the regulation of proteinkinase CK2. J. Cell. Biochem. (Suppl. 30/31), 129–136

68 Romero-Oliva, F. et al. (2003) Dual effect of lysine-rich polypeptides onthe activity of protein kinase CK2. J. Cell. Biochem. 89, 348–355

69 Theis-Febvre, N. et al. (2005) Highlighting protein kinase CK2movement in living cells. Mol. Cell. Biochem. 274, 15–22

70 Salinas, P. et al. (2006) An extensive survey of CK2 a and b subunits inArabidopsis: multiple isoforms exhibit differential subcellularlocalization. Plant Cell Physiol. 47, 1295–1308

71 Rodriguez, F. et al. (2005) Protein kinase CK2 holoenzyme producedectopically in human cells can be exported to the external side of thecellular membrane. Proc. Natl. Acad. Sci. U. S. A. 102, 4718–4723

our

for the history and

f science.

avour online on

’ll find book reviews,

ollection of beautifully

e history of science.

ng:

4 Hungarian case andions by E. Lafferton

n the universe by S.J. Dick

iture and its function by S. Muller-Wille

ustry: Livingstone in the Zambezi by L. Dritsas

by P. Fara

soon:

th and disease in the industrial revolution by M. Brown

ingdom Brunel by P. Fara

ent in the 19th century by C. Berkowitz

ers in American pulp science fiction by E. Drown

’s Christian belief by J. van der Heide

ch more. . .

Direct, www.sciencedirect.com

Identifying interaction motifs in CK2β – a ubiquitous kinase regulatory subunit

Documents

Transcript of Identifying interaction motifs in CK2β – a ubiquitous kinase regulatory subunit