Cross talk between protein kinase CK2 and eukaryotic ... - Springer Link

0 downloads 0 Views 943KB Size Report
of eIF2 with eIF2B (which catalyses the GDP to GTP ex- change) [12], as well as with eIF5 (which acts as a GTPase activating protein) [13]. Interestingly, both ...
cgSpringer

Molecular and Cellular Biochemistry 274: 53–61, 2005.

2005

Cross talk between protein kinase CK2 and eukaryotic translation initiation factor eIF2β subunit The carboxyl-terminal domain of eIF2β binds to specific sequences of CK2α which affect its activity Franc Llorens,1 Stefania Sarno,2 Eduard Sarr´o,1 Anna Duarri,1 Nerea Roher,1 Flavio Meggio,2 Maria Plana,1 Lorenzo A. Pinna2 and Emilio Itarte1 1

Departament de Bioqu´ımica i Biologia Molecular, Unitat de Bioqu´ımica de Ci`encies, Universitat Aut`onoma de Barcelona, Edifici Cs, Campus de Bellaterra, 08193 Bellaterra (Barcelona), Spain; 2 Dipartimento di Chimica Biologica, Universit`a di Padova, Padova, Italy

Abstract The β-subunit of eukaryotic translation initiation factor eIF2 is a substrate and a partner for protein kinase CK2. Surface plasmon resonance analysis shows that the truncated form corresponding to residues 138–333 of eIF2β (eIF2β-CT) interacts with CK2α as efficiently as full length eIF2β, whereas the form corresponding to residues 1–137, which contains the CK2 phosphorylation sites, (eIF2β-NT) does not bind. The use of different mutants and truncated forms of CK2α allowed us to map the basic segment K74–K83 at the beginning of helix αC and residues R191R195K198 in the p + 1 loop as the main determinants for the binding to eIF2β-CT of either the isolated CK2α subunit or the CK2 holoenzyme. The presence of eIF2β-CT stimulated the activity of CK2α towards the RRRAADSDDDDD peptide substrate; effect that was not observed with the CK2α K74-77A whose ability to bind to eIF2β-CT is severely impaired. Gel filtration analysis confirmed the ability of CK2α to form complexes with eIF2β-CT, and the contribution of the basic cluster in CK2α (K74–K77) in this association. (Mol Cell Biochem 274: 53–61, 2005) Key words: eIF2, enzymic regulation, protein kinase CK2, protein-protein interaction, protein phosphorylation, surface plasmon resonance

Introduction Protein kinase CK2 (also known as casein kinase II or casein kinase 2) is a constitutively active enzyme, ubiquitous and essential for life in eukaryotes [1, 2]. Protein kinase CK2 is known to phosphorylate more than 300 different substrates [1–3], which include many proteins involved in transcription,

RNA processing and translation. Moreover, there is a wide array of proteins able to interact with either the catalytic (CK2α and CK2α  ) or the regulatory (CK2β) subunits of CK2. Many of these proteins are able to serve as substrates for CK2 and a restricted number of them also exert a regulatory role on CK2 activity [2, 4]. These regulatory effects seem to vary depending on the form of the enzyme (either free catalytic

Address for offprints: E. Itarte, Departament de Bioqu´ımica i de Biolog´ıa Molecular, Unitat de Bioqu´ımica de Ci`encies, Universitat Aut`onoma de Barcelona, Edifici Cs, Campus de Bellaterra, 08193 Bellaterra (Barcelona), Spain (E-mail: [email protected])

54 subunit or holoenzyme) and the particular protein substrate. The stable interaction with protein partners located in specific organelles or subcellular compartments and the effects that these interactions might cause on CK2 activity is believed to contribute to the modulation of CK2 activity towards protein targets involved in specific functions within the cell [1–4]. Protein kinase CK2 is present in ribosomal-enriched cellular subfractions which contain several structural proteins and translation initiation factors that are substrates for CK2 [1, 3, 5]. This suggests that CK2 might be important for maintaining some components of the translation machinery in a functional conformation. Translation initiation factor eIF2, a key factor in protein synthesis initiation, is a well known substrate for CK2, which phosphorylates it on its β subunit [3, 6]. Protein synthesis initiation requires the participation of multiple protein components, which must form transient, specific complexes following a fixed and precise sequence, in a clock-work manner [7, 8]. eIF2 is believed to participate in various stages of translation initiation, what seems to be due to the presence in its structure of different domains, each one involved in specific functions. The conventional form of eIF2 is a trimer made of one copy of each of the three different subunits (eIF2α, eIF2β and eIF2γ ). Of these subunits, eIF2α is widely accepted as exerting a key regulatory role in translation initiation through phoshorylation on its Ser51 by different eIF2α kinases [7, 8]. The eIF2β subunit binds mRNA [9] and contributes, together with eIF2γ , to GTP and Met-tRNAi Met binding [10, 11]. In the assembled pre-initiation complexes eIF2β remains exposed to the environment, what allows its interaction with other translation factors. The eIF2β mediates the association of eIF2 with eIF2B (which catalyses the GDP to GTP exchange) [12], as well as with eIF5 (which acts as a GTPase activating protein) [13]. Interestingly, both eIF2B (eIF2Bε subunit) [14] and eIF5 [15] have been recently shown to be CK2 substrates ‘in vivo’. We have recently shown that besides serving as a substrate for CK2 holoenzyme, eIF2β binds to free CK2α, and inhibits its activity on calmodulin and β-casein [16]. The structural elements in eIF2β responsible for this binding seem to be located in the region encompassing its central and C-terminal domains, whereas the phosphorylation sites for CK2 holoenzyme are inserted in its N-terminal region. Interestingly, the central and C-terminal domains of eIF2β seem to be sufficient for the formation of the 43 S pre-initiation complex in eukaryotes [10], whereas the N-terminal domain would be involved in subsequent steps of translation initiation. In the present work we show that eIF2β interacts with specific regions of the catalytic subunit of CK2 (CK2α) and by doing that it affects its catalytic properties. This renders eIF2β a potential CK2α anchoring protein with functional consequences on its activity.

Materials and methods Protein expression and purification CK2α wild-type and mutated forms, CK2β, [17] and His6 eIF2β wild-type and truncated forms were expressed and purified as described previously [16].

SPR analysis BIAcore X system (BIAcore) was used to detect the interactions. CK2α or eIF2β-CT were covalently linked to a sensor chip CM5 (BIAcore) using amine coupling chemistry [16]. Proteins were injected over the surface with a flow rate of 10 µl/min in running buffer HBS (10 mM HEPES pH 7.4, 150 mM NaCl, 3 mM EDTA, 0.005% v/v surfactant P20) at 25 ◦ C. After injection HBS replaced the protein solutions in a continuous flow rate of 10 µl/min. All samples were run simultaneously over a flow cell containing a blank surface. Results were evaluated using the SPR kinetic evaluation software BIAevaluation 3.0 (BIAcore).

Far-western assays Human recombinant CK2α wild-type and mutated forms (75 pmol each) were resolved on SDS/PAGE 12% (w/v) gel, transferred on to PVDF membranes and blocked for 1 h with 5% dry milk in TPBS buffer (0.1% (v/v) Tween 20 phosphate-buffered saline (137 mM NaCl, 2.7 mM KCl, 10 mM Na2 HPO4 , 1.8 mM KH2 PO4 ). After overnight incubation with renaturalization buffer (50 mM Tris-HCl pH 7.5, 300 mM KCl, 0.1% Triton X-100), His6 -eIF2β-CT (20 µg/ml) or TPBS as a control were added for 1 h with gentle shaking. Membranes were washed three times with TPBS and developed with the anti-His antibody.

Pull-down assays His6 -eIF2β-CT expressed and purified as described previously [16] was mixed with native CK2 holoenzyme, purified from rat liver [18], for 1 h at 4 ◦ C on shaking. A slurry of NiNTA–agarose (1:1; 20 µl) was added and left for 1 h at 4 ◦ C under gentle shaking. Samples were centrifuged at 12000 g for 5 min at 4 ◦ C in an Eppendorf microfuge to sediment the His6 -eIF2β-CT bound proteins/Ni-NTA–agarose complexes. Beads were washed with PBS three times, and the proteins were eluted with SDS/PAGE sample buffer and subjected to SDS/PAGE, transferred onto PVDF membranes and probed using anti-CK2α or anti-CK2β antibodies.

55 Gel filtration assays 15 µg of purified CK2α wild type or CK2α K74-77A mutant, either alone or pre-incubated in the presence of eIF2β-CT protein at 1:2 molar ratio, were loaded in a Superdex 75 (Amersham Biosciences) gel filtration column. The column was equilibrated with a buffer containing 50 mM Tris-HCl pH 7.5, 7 mM β-mercaptoethanol and 0.3 M NaCl and the protein elution was analyzed by OD monitor. Phosphorylation assays Protein kinase CK2 activity was assayed as described previously [14] using 0.1 mM of specific peptide (RRRAADSDDDDD) and 125 µM [γ -32 P]ATP as substrates (specific activity 500 cpm/pmol) in the presence or absence of His6 eIF2β-CT at a 1:1 molar ratio with CK2α.

Results Deciphering the determinants in CK2α involved in its physical interaction with eIF2β We have recently shown that a truncated form of human eIF2β encompassing its central and carboxyl-terminal domains (His6 -eIF2β-CT) was able to interact with the catalytic subunit of CK2 (CK2α) in far-western assays [16], whereas the truncated form corresponding to its N-terminal region (residues 1–137) (His6 -eIF2β-NT) did not bind. The kinet-

ics of the interactions have now been determined by surface plasmon resonance analysis (SPR). The data showed that His6 -eIF2β-CT binding to CK2α was indistinguishable from that detected with the full length His6 -eIF2β polypeptide (Fig. 1). In contrast, no binding was detected with His6 -eIF2β-NT, what confirmed that the N-terminal region, which harbours the CK2 phosphorylation sites, has no contribution to the stable association between eIF2β and CK2α. Moreover, the data on SPR indicated that the N-terminal region of eIF2β did not influence the binding between its C-terminal region and CK2α. Therefore, the His6 -eIF2β-CT form was used in all of the subsequent experiments. Mutational analysis of CK2α unravelled the presence in its sequence of certain amino acid stretches or of specific residues which are crucial for its function [17]. Recombinant human CK2α forms mutated in these regions have now been used to map the structural elements involved in the interaction of free CK2α to eIF2β-CT. The 61 residues region at the carboxyl-terminal part in CK2α are known to confer this subunit specific functions not exerted by CK2α  , which lacks this region. Deletion of the 56 residues in the carboxyl-terminal part of CK2α (CK2α 336–391) had only a small effect on its binding to eIF2β-CT (Fig. 2A). The first 30 residues at the amino-terminal part of CK2α had little contribution to the binding since the CK2α 2–30 truncated form was able to bind to eIF2β-CT although, in this case the kinetics of the association was somehow slower than with CK2α. The amino-terminal segment is involved in the constitutive activation of the free CK2α by making contacts with the p + 1 loop [19]. Interestingly, mutation

Fig. 1. SPR analysis of the interaction of the catalytic subunit of protein kinase CK2 (CK2α) with His6 -eIF2β. Representative sensorgrams obtained by injection of 35 µl of a 100 nM solution of His6 -eIF2β-wt, His6 -eIF2β-CT (138–333) or His6 -eIF2β-NT (1–137) at a flow rate of 10 µl min over a sensor with immobilised CK2α. The response obtained with a control sensor surface (without protein immobilised) was substracted from each sensorgram. The response difference is measured in RU.

56

Fig. 2. Interaction of the C-terminal domain of His6 -eIF2β with CK2α wild type, mutated or truncated forms. Representative sensorgrams obtained by the injection of 35 µl of a 100 nM solution of (A) CK2α wt, CK2 2–30 and CK2α 336–391 or (B) CK2α wt, CK2α K74-77A, CK2α K79R80K83A, CK2α R191R195K198A and R278K279R280A, at a flow rate of 10 µl min over a sensor with immobilised His6 -eIF2β-CT. The response obtained with a control sensor surface (without protein immobilised) was substracted from each sensorgram. The response difference is measured in RU. (C) Interaction of the Cterminal domain of eIF2β with CK2α deleted and truncated forms by Far-western assay. CK2α wt, CK2 2–30, CK2α 336–391, CK2α K74-77A and CK2α K79R80K83A, (75 pmol each) were resolved on SDS/PAGE (12% gels) and transferred on to PVDF membranes. Ponceau Red staining was used to confirm transfer efficiency (results not shown). After blocking and renaturalization, the membranes were incubated either in the absence or in the presence of His6 -eIF2β-CT (20 µg/ml), washed and developed against anti-His antibody.

57 to alanine of the basic amino acids in the activation segment (R191R195K198A) is highly detrimental to the binding to eIF2β-CT (Fig. 2B). CK2α also contains a basic stretch (residues 74 to 83) which participates in the protein substrate recognition by interacting with acidic determinants in the phosphorylatable sequence. This basic stretch is composed by a cluster of four lysines (K74-77) followed by a sequence that contains two more lysines (K79 and K83) and one arginine (R80), and each one of these two parts are important for CK2α activity [17]. The sensorgrams obtained with the CK2α K74-77A and the CK2α K79R80K83A mutants demonstrated that these residues are also important for the binding to eIF2β-CT (Fig. 2B). The CK2α amino acid sequence also contains a triad of basic residues (R278K279R280) located just ahead of helix αJ [19, 20]. Mutation at these residues was found to cause a slight increase in the activity of the free CK2α and CK2 holoenzyme [20]. However, the sensorgrams obtained with the CK2α R278K279R280A mutant indicate that it plays little role in its binding to eIF2β-CT. The influence of some of these CK2α mutations on its ability to bind to eIF2β-CT was also tested by far-western analysis. This technique confirmed that deletions at either the amino- or the carboxyl-termini of CK2α did not prevent

its binding to eIF2β-CT, whereas mutations at either one of the two parts of the basic stretch was strongly deleterious for this binding (Fig. 2C).

Binding of CK2α to CK2β in the reconstituted CK2 holoenzyme did not mask the regions in CK2α involved in its binding to eIF2β-CT We have previously observed that free His6 -CK2β was also able to bind to eIF2β although less efficiently than His6 CK2α [16]. Moreover, the reported crystal structure of the CK2 holoenzyme shows that the K74-77 and the K79R80K83 regions in human CK2α are not involved in their association to the CK2β dimer to form the tetramer [21]. Furthermore, previous studies showed that the CK2α K74-77A and CK2α K79R80K83A mutants retain the ability to associate to CK2β [17]. Thus, we decided to explore the possibility that these regions in CK2α are also involved in the binding of eIF2β-CT to the CK2 tetramer. The sensorgrams obtained with the CK2, purified as holoenzyme, as a ligand and His6 eIF2β-CT immobilised on the chip showed a profile typical of a high binding efficiency (Fig. 3A). The ability to bind was significantly reduced when the CK2α K74-77A mutant

Fig. 3. Interaction of the C-terminal domain of His6 -eIF2β with CK2 holoenzyme. (A) Representative sensorgrams of SPR analysis obtained by injecting 35 µl of a 100 nM solution of CK2 holoenzyme composed of CK2β wild type and either CK2α wild type (CK2αβ), or the CK2α K74-77A (CK2αK74-77Aβ), or CK2α R191R195K198A (CK2α R191R195K198Aβ) mutants, at a flow rate of 10 µl min over a sensor with immobilised His6 -eIF2β-CT. (B) Equimolar amounts of His6 -eIF2β-CT were mixed with native CK2 holoenzyme purified from rat liver. After interaction a slurry of Ni-NTA–agarose was added to recover His6 -tagged protein. After washing the beads, the supernatant and pellets were analysed by Western-blot against CK2α and CK2β subunits.

58 Table 1. Summary of proteins checked for interaction between CK2 and His6 -eIF2β Analyte on chip

Analyte on solution

Interaction

CK2α CK2α CK2α eIF2β-CT eIF2β-CT eIF2β-CT eIF2β-CT

eIF2β-Wt eIF2β-NT eIF2β-CT αWt α2–30 α336–391 a K74-77A

+++ − +++ +++ +++ +++ +

eIF2β-CT eIF2β-CT

αK79R80K83A αR191R195K198A

+ +

eIF2β-CT eIF2β-CT eIF2β-CT eIF2β-CT

αR278K279R280A holo wt holo K74-77A holo R191R195K198A

+++ +++ ++ +

Fig. 4. Effect of the C-terminal domain of His6 -eIF2β on the activity of CK2α wild type and the CK2α K74-77A mutant. The activity of CK2α wild type and CK2α K74-77A mutant was tested using the synthetic peptide RRRAADSDDDDD in the presence or absence of His6 -eIF2β-CT at a 1:1 molar ratio to CK2α.

The symbols indicate interaction (+) or no interaction and the strength of the binding.

was used to reconstitute the CK2 holoenzyme, effect that was even more dramatic with the CK2α R191R195K198A mutant. These data indicate that these two basic stretches are also involved in the binding of the CK2 holoenzyme to eIF2β. A summary of the results on the interaction between eIF2β and CK2α, either as free subunit or incorporated into the CK2 holoenzyme, obtained in this work using the different mutants is shown in Table 1. The ability of eIF2β-CT to bind to native CK2 holoenzyme, purified from rat liver, was also tested by pull-down experiments. As observed in Fig. 3B, pre-incubation of rat liver CK2 with His6 -eIF2β-CT prior to the addition of Ni-NTA-agarose led to the total recovery of CK2α and CK2β in the pellet, whereas only small amounts of these CK2 subunits were detected in this fraction in the control samples.

to His6 -eIF2β-CT affected the catalytic activity of CK2α. Unexpectedly, His6 -eIF2β-CT did not inhibit but stimulated CK2α activity on this peptide (Fig. 4). To explore if this effect required the direct association between His6 -CK2α and His6 eIF2β-CT we made use of the CK2α K74-77A mutant which has little affinity for His6 -eIF2β-CT. This mutant has a basal activity lower than that of wild-type CK2α but is strongly activated by its binding to CK2β [18]. In contrast to this, the presence of His6 -eIF2β-CT has only a minor effect on its activity (Fig. 4). The ability of His6 -eIF2β-CT to form stable complexes with CK2α was then tested by gel filtration experiments (Fig. 5). Pre-incubation of CK2α with His6 -eIF2β-CT lead to the appearance of a second peak that eluted well ahead of control free CK2α. This peak was almost negligible when His6 -CK2α K74-77A was used instead of CK2α, what agrees with the much weaker association between CK2α K74-77A and His6 -eIF2β-CT. His6 -eIF2β-CT, in the absence of CK2α, was eluted after 13 ml (data not shown).

Association to His6 -eIF2β-CT increases the catalytic activity of His6 -CK2α on the specific substrate peptide

Discussion

We have shown that high molar ratios of His6 -eIF2β-CT inhibited the activity of free His6 -CK2α on calmodulin and βcasein [16]. The use of specific peptides to assay CK2α activity is considered a more reliable measure of its potential catalytic efficiency since it avoids possible constrains to the correct positioning of the phosphorylatable sequence at the catalytic site due to interactions between CK2 and other regions in the protein substrate. We have now made use of the specific peptide RRRAADSDDDDD to determine if the binding

Interaction of protein kinase CK2 with other cellular proteins has emerged as a potentially important mechanism to control its activity on specific cellular processes [1–4]. Many of the CK2 interacting partners also serve as protein kinase CK2 substrates, but, in some cases, a closer look to the residues involved in their association has shown that the regions containing the phosphorylation sites do not overlap with those involved in binding. The data obtained in the present study confirmed that this is specially evident with

59

Fig. 5. Gel filtration analysis of the association between His6 -eIF2β-CT and CK2α wild type and the CK2α K74-77A mutant. The association of CK2α wild type and CK2α K74-77A mutant with His6 -eIF2β-CT was analyzed using a Superdex 75 chromatography as described in the experimental section. Elution profiles of CK2α subunit in the absence (A and B) and in the presence of a double molar amount of His6 -eIF2β-CT (C and D) are shown.

eIF2β since the determinants involved in the high affinity, stable binding to CK2α are located in its C-terminal region, whereas its N-terminal region, which harbours the CK2 phosphorylation sites, has no contribution to this binding [16]. The use of different CK2α mutants has allowed us to evidence that the basic stretch (K74–K83) and the basic residues in the p + 1 loop at the end of the activation segment of CK2α (R191R195K198A) are necessary for its binding to eIF2β-CT, whereas the basic residues R278K279R280, and the amino- and carboxyl-terminal regions have little influence on this binding. Moreover, it is worth to note that neither the mutants of the basic stretch nor the one of the activation segment had lost completely their ability to bind eIF2β-CT. Thus, it is likely that the tight CK2α/eIF2β-CT binding is not restricted to a single region but it requires the cooperation of both segments in CK2α. The binding site for hsp90 has also been mapped at the region encompassing residues 62 to 83 in CK2α [22], a region suspected to be also involved in the binding to nucleolin [23]. In a similar vein, the first N-terminal 84 residues in CK2α are known as involved in its binding to ATF1 [24], although the specific residues responsible for this binding are unknown. Moreover, we have previously observed that the lysine cluster K74–K77 in the basic stretch in CK2α is specially required for its binding to grp94 [25]. Within the basic stretch, the lysine cluster (K74–K77) is implicated in the inhibitory effect

of heparin, whereas both the K74-77 quartet and the 79– 83 segment appear to be involved in substrate recognition [17, 20]. However, the lysine cluster can not be considered as a general recognition site for the CK2α partners, since HS1, a partner and a substrate for CK2α [26, 27], was found to interact with the CK2α mutant K74-77A as efficiently as with CK2α (wt). Reconstitution of the CK2 tetramer did not block the association of CK2α to eIF2β-CT. This behoviour is similar to that reported previously for bovine prion protein bPrP [28], but differs from that observed with HS1 and grp94, which did not bind to the holoenzyme [25, 26]. These differences would no rely on impediments to the accessibility of the lysine cluster (K74-77) in CK2α once incorporated into the CK2 holoenzyme, since the reported crystal structure has shown that it does not participate in the binding to the CK2β dimer and remains exposed to the solvent [21]. On the other hand, our results indicate that the lysine cluster is critical for the binding to eIF2β-CT of free CK2α but somehow less important for the CK2 holoenzyme, whereas the basic residues in the p+1 loop (R191R195K198A) are equally necessary for the binding of either the free subunit or the holoenzyme. However, it must be kept in mind that when considering the binding of some partners to CK2 holoenzyme, the contribution of binding sites provided by the CK2β subunit can not be disregarded. In this context, we have previously observed that free CK2β was also able to bind to eIF2β although less efficiently than CK2α

60 [16]. A weak binding to CK2β has also been reported for bPrP [28], whereas HS1 [26] and grp94 [25] did not bind. Whether CK2β cooperates with CK2α in the formation of complexes between eIF2β-CT and CK2 holoenzyme remains to be explored. The potential importance of the binding sites between CK2α and its partners is enlightened by the variety of effects they cause on its activity. Binding to bPrP activates calmodulin phosphorylation by CK2α [28] whereas binding to eIF2β-CT did not stimulate but rather decreased this phosphorylation [16]. Interestingly, binding to either bPrP [28] or to eIF2β-CT (this report) increased CK2α activity on the model substrate peptide, a fact that had also been observed previously with grp94 [25]. On the other hand, binding to HS1 decreased calmodulin phosphorylation by CK2α without altering its activity on the model peptide substrate [26]. As indicated above, grp94 and eIF2β-CT bind to the basic stretch in CK2α whereas mutation to alanines of the K74-77 cluster does not affect CK2α binding to HS1 [26]. The basic stretch makes contacts with the amino-terminal segment what stabilizes the active conformation of helix αC [19]. The amino-terminal segment also binds to the activation segment, and the sum of these contacts is responsible for keeping free CK2α in an open, constitutively active conformation. The basic residues of the p + 1 loop (R191R195K198A), which corresponds to the C-terminal region of the activation segment, are also involved in the binding to eIF2β-CT. Thus, it is tempting to speculate that the simultaneous binding of eIF2β-CT to the basic stretch and to the activation segment might help to stabilize the open conformation of the catalytic cleft favouring the phosphorylation of the substrates. On the other hand, the presence of bound eIF2β-CT may prevent the binding of larger protein substrates by steric reasons, leaving the catalytic site accessible only to short peptide substrates. This could explain why binding to eIF2β-CT increases its activity on the peptide substrate (this report) while having a slight inhibitory effect on that of calmodulin and β-casein [16]. The N-terminal region of eIF2β is responsible for the contacts with the C-terminal region of eIF2Bε (the catalytic subunit of the GEF-like factor eIF2B) during the GDP/GTP exchange step and with the C-terminal region of eIF5 (a GAPlike protein), when assembled into the multifactor complex [13, 29]. The C-terminal regions of eIF5 and eIF2Bε contain CK2 phosphorylation sites [14, 15], and mutation of the CK2 targeted serines to alanines in eIF2Bε abolishes its binding to eIF2β [15]. These data suggest that eIF2β might act as a CK2 anchoring protein for either free CK2α subunit or CK2 holoenzyme, what might serve to locate it in the close vicinity of its phosphorylation targets, and in this way help to maintain some components of the translation machinery in a functional conformation.

Acknowledgments This work was supported by grants SAF2002-03239 from MCYT, 2001SGR00199 from DGR (GENCAT), and the Spanish-Italian Integrated Action (HI2002-0022).

References 1. Pinna LA: Protein kinase CK2: a challenge to canons. J Cell Sci 115: 3873–3878, 2002 2. Litchfield DW: Protein kinase CK2: structure, regulation and role in cellular decisions of life and death. Biochem J 369: 1–15, 2003 3. Meggio F, Pinna LA: One-thousand-and-one substrates of protein kinase CK2? FASEB J 17: 349–368, 2003 4. Guerra B, Issinger OG: Protein kinase CK2 and its role in cellular proliferation, development and pathology. Electrophoresis 20: 391–408, 1999 5. Issinger OG: Phosphorylation of acidic ribosomal proteins from rabbit reticulocytes by a ribosome-associated casein kinase. Biochim Biophys Acta 477: 185–189, 1977 6. Tuazon PT, Merrick WC, Traugh JA: Site-specific phosphorylation of initiation factor 2 by three cyclic nucleotide-independent protein kinases. J Biol Chem 255: 10954–10958, 1980 7. Kimball SR: Eukaryotic initiation factor eIF2. Int. J. Biochem Cell Biol. 31: 25–29, 1999 8. Sonenberg N, Dever TE: Eukaryotic translation initiation factors and regulators. Curr Opinion Struct Biol 13: 56–63, 2003 9. Laurino JP, Thompson GM, Pacheco E, Castilho BA: The β subunit of eukaryotic initiation factor 2 binds mRNA through the lysine repeats and a region comprising the C2-C2 motif. Mol Cell Biol 19: 173–181, 1999 10. Thompson GM, Pacheco E, Melo EO, Castilho BA: Conserved sequences in the β subunit of archeal and eukaryal translation initiation factor 2 (eIF2), absent from eIF5, mediate interaction with eIF2γ . Biochem J 347: 703–709, 2000 11. Nika J, Rippel S, Hannig EM: Biochemical analysis of the eIF2βγ complex reveals a structural function for eIF2α in catalyzed nucleotide exchange. J Biol Chem 276: 1051–1056, 2001 12. Kimball SR, Heinzinger NK, Horetsky RL, Jefferson LS: Identification of interprotein interactions between the subunits of eukaryotic initiation factors eIF2 and eIF2B. J Biol Chem 273: 3039–3044, 1998 13. Das S, Maiti T, Das K, Maitra U: Specific interaction of eukaryotic translation initiation factor 5 (eIF5) with the β subunit of eIF2. J Biol Chem 272: 31712–31718, 1997 14. Wang X, Paulin FEM, Campbell LE, Gomez E, O’Brien K, Morrice N, Proud CG: Eukaryotic initiation factor 2B: identification of multiple phosphorylation sites in the ε-subunit and their functions ‘in vivo’. EMBO J 20: 4349–4359, 2001 15. Majumdar R, Bandyopadhyay A, Deng H, Maitra U: Phosphorylation of mammalian translation initiation factor 5 (eIF5) ‘in vitro’ and ‘in vivo’. Nucleic Acids Res 30: 1154–1162, 2002 16. Llorens F, Roher N, Miro FA, Sarno S, Ruiz FX, Meggio F, Plana M, Pinna LA, Itarte E: Eukaryotic translation-initiation factor eIF2beta binds to protein kinase CK2: effects on CK2alpha activity. Biochem J 375: 623–631, 2003 17. Sarno S, Vaglio P, Marin O, Issinger OG, Ruffato K, Pinna LA: Mutational analysis of residues implicated in the interaction between protein kinase CK2 and peptide substrates. Biochemistry 36: 11717–11724, 1997

61 18. Gil C, Plana M, Riera M, Itarte E: Rat liver pp49, a protein that forms complexes with protein kinase CK2, is composed of the β and the γ subunits of translation initiation factor eIF-2. Biochem Biophys Res Commun 225: 1052–1057, 1996 19. Niefind K, Guerra B, Pinna LA, Issinger OG, Schomburg D: Crystal structure of the catalytic subunit of protein kinase CK2 from Zea mays at 2.1 A resolution. EMBO J 17: 2451–2462, 1998 20. Sarno S, Vaglio P, Meggio F, Issinger OG, Pinna LA: Protein kinase CK2 mutants defective in substrate recognition. Purification and kinetic analysis. J Biol Chem 271: 10595–10601, 1996 21. Niefind K, Guerra B, Ermakowa I, Issinger OG: Crystal structure of human protein kinase CK2: insights into basic properties of the CK2 holoenzyme. EMBO J 20: 5320–5331, 2001 22. Miyata Y, Yahara I: Interaction between casein kinase II and the 90-kDa stress protein, HSP90. Biochemistry 34: 8123–8129, 1995 23. Li D, Dobrowolska G, Krebs EG: The physical association of casein kinase 2 with nucleolin. J Biol Chem 271: 15662–15668, 1996 24. Yamaguchi Y, Wada T, Suzuki F, Takagi T, Hasegawa J, Handa H: Casein kinase II interacts with the bZIP domains of several transcription

factors. Nucleic Acids Res 26, 3854–3861, 1998 25. Roher N, Sarno S, Mir´o F, Ruzzene M, Llorens F, Meggio F, Itarte E, Pinna LA, Plana M: The carboxy-terminal domain of Grp94 binds to protein kinase CK2α but not to holoenzyme. FEBS Lett 505: 42–46, 2001 26. Ruzzene M, Brunati AM, Sarno S, Donella-Deana A, Pinna LA: Hematopoietic lineage cell specific protein 1 associates with and down-regulates protein kinase CK2. FEBS Lett 461: 32–36, 1999 27. Ruzzene M, Brunati AM, Sarno S, Marin O, Donella-Deana A, Pinna LA: Ser/Thr phosphorylation of hematopoietic specific protein 1 (HS1). Eur J Biochem 267: 3065–3072, 2000 28. Meggio F, Negro A, Sarno S, Ruzzene M, Bertoli A, Sorgato MA, Pinna L: Bovine prion protein as a modulator of protein kinase CK2. Biochem J 352: 191–196, 2000 29. Asano K, Krishnamoorthy T, Phan L, Pavitt GD, Hinnebusch AG: Conserved bipartite motifs in yeast eIF5 and eIF2Bepsilon, GTPaseactivating and GDP-GTP exchange factors in translation initiation, mediate binding to their common substrate eIF2. EMBO J 18: 1673–1688, 1999