Phosphorylation of initiation factor-2a is required for activation of internal translation initiation during cell differentiation Gabi Gerlitz 1 , Rosemary Jagus 2 and Orna Elroy-Stein 1 1 Cell Research and Immunology, George S. Wise Faculty of Life Sciences, Tel Aviv University, Israel; 2 Center of Marine Biotechnology, University of Maryland Biotechnology Institute, Baltimore, USA The long uORF-burdened 5¢UTRs of many genes encoding regulatory proteins involved in cell growth and differenti- ation contain internal ribosomal entry site (IRES) elements. In a previous study we showed that utilization of the weak IRES of platelet-derived growth factor (PDGF2) is activated during megakaryocytic differentiation. The establishment of permissive conditions for IRES-mediated translation during differentiation has been confirmed by our demonstration of the enhanced activity of vascular endothelial growth factor, c-Myc and encephalomyocarditis virus IRES elements under these conditions, although their mRNAs are not naturally expressed in differentiated K562 cells. In contrast with the enhancement of IRES-mediated protein synthesis during differentiation, global protein synthesis is reduced, as judged by polysomal profiles and radiolabelled amino acid incorporation rate. The reduction in protein synthesis rate correlates with increased phosphorylation of the translation initiation factor eIF2a. Furthermore, IRES use is decreased by over-expression of the dominant-negative form of the eIF2a kinase, PKR, the vaccinia virus K3L gene, or the eIF2a-S51A variant which result in decreased eIF2a phos- phorylation. These data demonstrate a connection between eIF2a phosphorylation and activation of cellular IRES ele- ments. It suggests that phosphorylation of eIF2a, known to be important for cap-dependent transaltional control, serves to fine-tune the translation efficiency of different mRNA subsets during the course of differentiation and has the potential to regulate expression of IRES-containing mRNAs under a range of physiological circumstances. Keywords: differentiation; gene expression; initiation factor 2; IRES; translation initiation. Translation of eukaryotic gene expression is controlled both by global mechanisms that affect the overall rate of protein synthesis and by selective control mechanisms that affect the translation of subsets of mRNA molecules equipped with appropriate cis-regulatory elements. The global mechanisms are mostly based on controlling the availability of two rate- limiting components of the initiation step: eIF4E, the 5¢-cap binding protein, and eIF2, a GTP-binding protein that mediates the association of Met-tRNA i to the 40 S ribosomal subunit. Control of eIF4E activity is mediated by influencing its phosphorylation status and/or its interac- tion with the eIF4E binding proteins, as well as by affecting the integrity of eIF4G which serves as an adapter protein that bridges eIF4E, the RNA helicase eIF4A, poly(A) binding protein and eIF3 (reviewed in [1–3]). Control of eIF2 activity is mediated through reversible phosphoryla- tion of its a-subunit. When eIF2a is phosphorylated, the GDP-eIF2generatedattheendofeachinitiationstep becomes a competitive inhibitor of eIF2B, a rate-limiting guanine nucleotide exchange factor, resulting in a reduction of the exchange of eIF2-bound GDP for GTP. As GTP binding to eIF2 is a prerequisite to Met-tRNA i binding, phosphorylation of eIF2a effectively inhibits eIF2 recycling and consequently inhibits additional translation initiation steps (reviewed in [4]). The control mechanisms that govern the rate of global protein synthesis are responsive to a variety of conditions including nutrient deprivation, heat shock, apoptosis and viral infection. Under conditions that inhibit the initiation of global protein synthesis, subsets of mRNAs remain competent to be recruited by ribosomes. Depending on their specific cis-regulatory elements they may gain a translational advantage over other mRNA molecules. For instance, mRNAs encoding heat-shock proteins are translated efficiently under conditions of reduced eIF4E/4F activity due to their unstructured 5¢UTR (reviewed in [5]). Another example is the efficient translation of the yeast GCN4 mRNA under conditions of amino acid starvation due to leaky scanning of the upstream ORFs within its 5¢UTR (reviewed in [6]). While much data has been accumulated regarding the control of protein synthesis in response to various stress conditions, less is known about translational control mechanisms that are operative during cellular differenti- ation. Cells undergoing terminal differentiation exhibit extensive changes in the pattern of gene expression to acquire a specific biological function that is usually accom- panied by cessation of proliferation. In addition to the massive changes at the transcriptional level, mechanisms regulating overall inhibition of protein synthesis release most mRNAs from the polysomes and facilitate the translation of specific mRNAs that are important for the Correspondence to O. Elroy-Stein, Department of Cell Research & Immunology, George S. Wise Faculty of Life Sciences, Tel Aviv University, Tel Aviv 69978, Israel. Fax: +972 3 642 2046, Tel.: +972 3 640 9153, E-mail: ornaes@post.tau.ac.il Abbreviations: PDGF, platelet-derived growth factor; IRES, internal ribosomal entry site; VEGF, vascular endothelial growth factor; CMV, cytomegalovirus; TPA, 12-O-tetradecanoylphorbol-13- acetate; EMCV, encephalomyocarditis virus. (Received 18 January 2002, revised 3 April 2002, accepted 2 May 2002) Eur. J. Biochem. 269, 2810–2819 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.02974.x differentiation process [7]. However, there is little data regarding the molecular mechanisms that govern such global inhibition accompanied by activation of specific mRNA subpopulations. The proliferation-dependent association of polysomes with 5¢-terminal oligopyrymidine tract-containing mRNAs [8] and the 3¢UTR-mediated translational activation of 15-lipoxygenase mRNA during erythrocytic differentiation [9] are the best characterized examples. Whereas most mRNAs are translated efficiently, a subset of mRNAs is poorly translated due to the extraordinarily long, structured, upstream AUG-burdened 5¢UTRs that serve as barriers for ribosomal scanning. Interestingly, such translational inhibitors often belong to mRNAs encoding proteins involved in cell growth and differentiation such as growth factors, receptors, transcription factors, proto- oncogenes, and cytokines [10]. Using the 1022 nucleotide long 5¢UTR of platelet-derived growth factor (PDGF2) as a model, we were able to show previously that the cumber- some 5¢UTR is not a translational inhibitor, but rather a translational modulator that is sensitive to changes in the cellular milieu [11]. More specifically, the PDGF2 mRNA leader was shown to mediate efficient translation under conditions of megakaryocytic differentiation which provide a permissive milieu for activation of the PDGF2 internal ribosomal entry site (IRES) [12,13]. During the last decade, several cellular IRES elements have been discovered (reviewed in [14,15]), although the mechanisms of cellular IRES-mediated ribosome recruitment remain unclear. To further understand the mechanism of IRES activation, we wished to assess whether the mechanisms involved in regulation of global translation during differentiation have a role in the translational activation of the IRES-containing mRNA. Since internal translation is independent of the 5¢-cap structure of the mRNA, we did not focus our attention on the changes in the activity of the 5¢-cap binding protein, eIF4E, during differentiation. However, it remained of interest to determine eIF2a phosphorylation status during differentiation and to ascertain whether phosphorylation of eIF2a is involved in IRES activation. In this study we show that: (a) not only are differentiation conditions permissive for the recruitment of the PDGF2 IRES, but also for the vascular endothelial growth factor (VEGF), c-Myc and encephalomyocarditis virus (EMCV) IRES elements, although their mRNAs are not naturally expressed in differentiated K562 cells; (b) global protein synthesis rate is reduced during differentiation, correlating with increased eIF2a phosphorylation that is known to be important for cap-dependent translational control; and (c) inhibition of eIF2a phosphorylation during differentiation reduces the differentiation-induced IRES activation. MATERIALS AND METHODS Plasmids The pLL vector is composed of a fragment containing Renilla luciferase from pRL-null (Promega) as the first cistron, fused to a fragment encoding the cytosolic form of firefly luciferase from pGL3-basic (Promega) as the second cistron. A 22-base pair fragment containing StuIandNcoI sites separates the stop codon of the Renilla luciferase and the ATG initiator codon of the firefly luciferase. An NheI (filled)–SacI fragment containing both cistrons was ligated with a HindIII (filled)–SacIfragmentofpCL[12]to generate pLL, which contains the dual luciferase transcrip- tion unit downstream of the cytomegalovirus (CMV) promoter and upstream of the SV40 intron and polyade- nylation sites. The 5¢UTRs of VEGF and of PDGF2 were obtained as SpeI (filled)–NcoI fragments from pSKVLUC [16] and pCPL [12], respectively, and were ligated to the StuI–NcoI7.5-kbfragmentofpLLtogeneratepLVLand pLPL, respectively. The EMCV IRES fragment was obtained as Alw26I (filled)–NcoI fragment from pTM1 [17], and was ligated to the StuI–NcoI7.5-kbfragmentof pLL to generate pLEL. P2 c-Myc 5¢UTR (GeneBank acces- sion # J00120) was generated by RT-PCR using total RNA from K562 cells and the oligonucleotides 5¢-CCCCACTA GTAATTCCAGCGAGAGGCAGA-3¢ and 5¢-AATACC ATGGTCGCGGGAGGCTGCTG-3¢, and was ligated to the StuI–NcoI 7.5-kb fragment of pLL to generate pLML. pcK3L was generated by insertion of the NcoI (filled)– BamHI 0.3-kb fragment of pTM1-K3L [18] into the HindIII (filled)–BamHI sites of pcDNA3 (Invitrogen) under the control of the CMV promoter. pcPKRD6 ¼ p68D6-pcD- NAI/NEO [19] was used for expression of PKRD6, the dominant-negative variant of PKR under the control of the CMV promoter. pc2a-S51A (expresses eIF2a with Ser51 fi Ala mutation) was generated by PCR of the eIF2a cDNA from p51A and p51D [20], respectively, using the oligonucleotides 5¢-CTGGATATCATGCCGGGTCT AAGTTG-3¢ and 5¢-CTGCTCGAGTTAATCTTCAGCT TTGGC-3¢, followed by ligation into the EcoRV–XhoI sites of pcDNA3 (Invitrogen) under the control of the CMV promoter. PEGFP-N3 plasmid (Clontech) expressing GFP under the control of the CMV promoter was used as control plasmid for the cotransfection experiments. Cells and megakaryocytic differentiation The human chronic myelogenous leukemia cell line K562 was grown in RPMI 1640 medium (Biological Industries) supplemented with 50 U penicillinÆmL )1 ,50lgÆmL )1 strep- tomycin, 0.1 mgÆmL )1 kanamycin and 10% fetal bovine serum. Cells at a density of 5–7 · 10 5 cellsÆmL )1 or 1.2–1.5 · 10 6 cellsÆmL )1 were considered as logarithmically growing (log) or growth arrested (dense), respectively. Megakaryocytic differentiation was induced by dilution of cells at a density of 1.2 · 10 6 cellsÆmL )1 , to a final concen- tration of 5 · 10 5 cellsÆmL )1 , with medium containing 5 n M 12-O-tetradecanoylphorbol-13-acetate (TPA; Calbiochem) for 48 h. Plasmid transfections and luciferase assays Twenty or forty micrograms supercoiled DNA of each of the bi-cistronic vectors or the cotransfected plasmid, respectively, were used per 7.5 · 10 6 K562 cells for each electroporation sample. Electroporation was performed in 0.8 mL RPMI 1640 without serum by an electric pulse of 240 V and 1500 mF (Easy Ject1 Electroporator; Equibio). Immediately following the electric pulse, the cells were transferred to 10 mL RPMI 1640 medium supplemented with 20% fetal bovine serum for 24 h. The cells were diluted to a final concentration of 5 · 10 5 viable cellsÆmL )1 Ó FEBS 2002 eIF2a phosphorylation and IRES activity (Eur. J. Biochem. 269) 2811 (as determined by Trypan blue staining) in RPMI 1640 supplemented with 10% fetal bovine serum, with or without 5n M TPA, for 48 h. Transfection efficiency was 50–60%, as judged by the percentage of fluorescent GFP-expressing cells. The control and differentiated transfectants were harvested simultaneously, and assayed for Renilla and firefly luciferase activities using the Dual-luciferase reporter assay system (Promega) and TD-20e-Luminometer (Turner). RNA was isolated from cells transfected with the bicistronic constructs and analysed by Northern blotting using a LUC- specific probe to ensure that transcripts of the correct size were produced. Bi-cistronic transcript level from all con- structs was approximately five times higher in differentiated cells because of the increased activity of the CMV promoter [12]. 32 P i labelling, immunoprecipitation and Western analysis of eIF2a A total of 10 6 log, dense or megakaryocytic differentiated K562 cells were washed twice with Hepes/saline buffer (50 m M KOH/Hepes pH 7.0, 150 m M NaCl) and resus- pended in 0.5 mL Dulbecco’s modified Eagle medium lacking sodium phosphate (Sigma) supplemented with 10% dialysed fetal bovine serum. The cells were labelled for 2 h with 0.2 mCi 32 P i ÆmL )1 (Amersham,#PBS13),followedby twowasheswithcoldNaCl/P i containing 10 m M b-glycerophosphate and 50 m M NaF. Proteins were extrac- ted from the cell pellets by using 470 lL lysis buffer containing 25 m M KOH/Hepes pH 7.2, 0.5% Elugent TM (Calbiochem), 100 m M KCl, 0.05% SDS, 1 m M dithiothre- itol, 2 l M okadaic acid, 10 m M b-glycerophosphate, 50 m M NaF and protease inhibitor cocktail (Complete TM , Roche). For immunoprecipitation the sample was supplemented with 28 lL5 M NaCl and 0.5 lLanti-eIF2a mAb [21] and incubated for 1 h at 4 °C. Next, rabbit anti-(mouse IgG) Ig (Jackson Immuno Research) were added for further incubation of 1 h, followed by addition of 10 lLpacked protein A-Sepharose (Pharmacia Biotech) for an additional 1-h incubation. Following separation of the immunopre- cipitate by SDS/10% PAGE, the proteins were blotted onto a nitrocellulose membrane and quantified by phosphoi- mager. The membrane was then used for Western analysis using antibodies specific for Ser51-phosphorylated eIF2a (Research Genetics, Inc.), and following stripping mAb specific for total eIF2a were used. Polysome fractionation A total of 3.5 · 10 7 log, dense, or megakaryocytic differ- entiated K562 cells were treated with 90 lgÆmL )1 cyclohex- imide for 10 min prior to harvest and used for fractionation of polysomes by sedimentation through 5–47% sucrose gradients [22]. Protein synthesis rate One million log, dense, or megakaryocytic differentiated K562 cells were re-suspended in 2 mL RPMI medium containing 10% fetal bovine serum. The cells were labelled for 20 min with 20 lCiÆmL )1 [ 35 S] L -methionine, [ 35 S] L - cysteine mix (NEN, #NEG072), followed by two washes with cold NaCl/P i . Proteins were extracted from the cell pellets using 50 lL lysis buffer containing 25 m M KOH/ Hepes pH 7.5, 1% Triton X-100, 100 m M KCl, 1 m M dithiothreitol, 2 l M okadaic acid, 10 m M b-glycerophos- phate, 50 m M NaF and protease inhibitor cocktail (Com- plete TM , Roche). Twenty micrograms total protein were applied onto 3 m M filter papers (Whatman) and washed three times for 1 min in boiling 5% (W/V) trichloroacetic acid containing traces of cold L -methionine and L -cysteine. Thefilterswerethenrinsedonceinethanol,driedand counted in a scintillation counter (Beckman). Cell cycle analysis and differentiation markers For cell cycle analysis 5 · 10 5 cells were harvested, washed with NaCl/P i and re-suspended in 0.5 mL NaCl/P i contain- ing 0.1% sodium azide. Following addition of 50 lLNaCl/ P i containing 1% Triton X-100 and 50 lL1mgÆmL )1 propidium iodide, the cell-cycle of the cells was analysed by Becton Dickinson FACSort, using the Cell Quest software. The Vav protein was used as a marker for differentiation. To detect Vav protein level the cells were lysed using a buffer containing 25 m M KOH/Hepes pH 7.5, 1% Triton X-100, 100 m M KCl, 1 m M dithiothreitol, 2 l M okadaic acid, 10 m M b-glycerophosphate, 50 m M NaF and protease inhibitors cocktail (Complete TM , Roche). 140 lgoftotal cell proteins were separated by SDS/10% PAGE, and blotted onto a nitrocellulose membrane which was then used for Western analysis using polyclonal antibodies specific for Vav (Santa Cruz) and polyclonal antibodies specific for CKIIa (a gift from D. Canaani, Tel Aviv University, Israel). RESULTS Favorable conditions for IRES-mediated translation are established during differentiation In previous studies we have demonstrated that in addition to transcriptional activation of PDGF2 during megakaryocytic differentiation, its IRES element undergoes functional activation during the differentiation process [12,13]. Interes- ted in elucidating the mechanism of IRES function in general, we have used the differentiation phase to learn more about the possible involvement of trans-acting factors. Viewing the differentiated state as a ÔpermissiveÕ environment for PDGF2-IRES mediated translation, we also wished to check the effect of differentiation conditions on the beha- viour of additional cellular and viral IRES elements. Although normally the mRNAs of VEGF and c-Myc are not present in differentiated megakaryocytes, it was still of interest to check the activity of their IRES elements in the differentiated K562 cells that are permissive for PDGF2 IRES use. The IRES elements of human VEGF, human c-Myc and EMCV were cloned into a CMV promoter-driven bicistronic vector, between the coding regions of Renilla and firefly luciferases, as illustrated in Fig. 1. K562 cells were transfected with each of the recombinant plasmids followed by incubation under normal or differentiation conditions for 48 h prior to measurements of Renilla and firefly luciferase enzymatic activities. As shown in Table 1, in differentiated cells we observed elevation in the activity of both luciferases, because of increased CMV promoter activity in this system which results in a fivefold increase in transcript levels (demonstrated in Fig. 3 [12]). However, upon differentiation, 2812 G. Gerlitz et al. (Eur. J. Biochem. 269) Ó FEBS 2002 the utilization of the first cistron increased only 2.2–2.7-fold whereas a 6–7.9-fold increase in utilization of the IRES- mediated second cistron was observed. This was in contrast with utilization of the second cistron from the IRES-less pLL vector. A 2.2- to 3.1-fold increase in the firefly per Renilla ratio was detected upon differentiation from the IRES- containing vectors, in contrast with the 0.8-fold increase observed for the IRES-less transcript from pLL. Reduction of global protein synthesis during differentiation is accompanied by eIF2a phosphorylation The terminal differentiation process is usually accompanied by arrest of cellular proliferation and by decreased global protein synthesis [7,8]. As megakaryocytic differentiated cells cease to proliferate [23], we wished to check the status of their global mRNA translation. The rate of radio- labelled amino acids incorporation in logarithmically growing cells was compared to that in density-arrested or differentiated cells. The incorporation rate was almost two times lower in both stationary (dense) and differentiated cells compared with logarithmically growing cells (Fig. 2A). The decrease in global protein synthesis was also evident from the differences in the polysomal profiles of the above cells. Fig. 2B demonstrates the reduced heavy polysomes levels upon growth arrest due to high density or differen- tiation. As the rate of protein synthesis in higher eukaryotes is commonly regulated at the level of eIF2a phosphorylation [4,24], we wished to check the status of eIF2a phosphory- lation in cells undergoing differentiation. For this purpose, the cells were metabolically labelled with 32 P i followed by immunoprecipitation using antibodies specific for eIF2a. Fig. 3A shows the radio-labelled phosphorylated eIF2a,the total amount of eIF2a as determined by Western analysis, and the ratio of phosphorylated eIF2a per total eIF2a.In dense cells, in which growth arrest was probably induced by depletion of essential nutrients/growth factors in the medium, a two-fold increase in eIF2a phosphorylation was observed compared to logarithmically growing cells. However, a more significant, 6.4-fold increase in eIF2a phosphorylation was detected upon growth arrest induced by the differentiation process. As the regulated phosphory- lation of mammalian eIF2a has been shown to occur only on Ser51 [25], we re-confirmed the 32 P i labelling results by using specific antibodies against phosphorylated Ser51. Fig. 3B shows the phosphorylated Ser51 and the total eIF2a levels as determined by Western analysis. In agreement with the labelling studies, the ratio of Ser51-phosphorylated eIF2a to total eIF2a revealed an increase of 3.3-fold and 7.9-fold in eIF2a phosphorylation level in dense and differentiated cells, respectively, compared with logarith- mically growing K562 cells. In summary, growth arrest is accompanied by elevation of eIF2a phosphorylation. However, megakaryocytic differentiation involves two- to threefold higher eIF2a phosphorylation than that resulting from growth arrest that is induced by increased cell density. Increased IRES-mediated translation during differentiation requires eIF2a phosphorylation The phenomenon of increased IRES-mediated translation under conditions of increased eIF2a phosphorylation, raised the notion that eIF2a phosphorylation confers a translational advantage on IRES-containing mRNAs. To test this hypothesis, we looked at the effect of expression of eIF2a phosphorylation inhibitors on IRES use. We used either the vaccinia virus K3L gene that encodes an eIF2a homologue and pseudo-substrate inhibitor of eIF2a protein Fig. 1. The bicistronic transcription units used. The bicistronic tran- scription unit expressing Renilla and Firefly luciferase reporter genes as the first and second cistrons, respectively, under the control of CMV promoter. The 5¢UTRs of the human PDGF2, VEGF, c-Myc or EMCV were placed in the intercistronic space as indicated to create plasmids pLPL, pLVL, pLML and pLEL, respectively. The bicistronic IRES-less vector pLL served as a control plasmid. Table 1. Effect of differentiation on IRES activity. Each of the bicistronic plasmids harbouring the IRES elements indicated in Fig. 1 was transfected into K562 cells followed by further incubation under nondifferentiation or differentiation conditions and subsequent analysis of Renilla (R) and Firefly (F) luciferase activity. Each value represents the mean ± SE of three independent experiments. The fold induction values represent the F/R ratio in differentiated cells relative to the F/R ratio in nondifferentiated cells. Non-differentiated cells Differentiated cells Fold F/R induction Renilla (U per 10 6 cells) Firefly (U per 10 6 cells) F/R Renilla (U per 10 6 cells) Firefly (U per 10 6 cells) F/R pLPL 59 ± 3.5 7.5 ± 0.6 0.13 ± 0.01 131 ± 18 53 ± 6 0.4 ± 0.05 3.1 ± 0.2 pLML 94 ± 10 6.4 ± 0.07 0.068 ± 0.01 253 ± 29 38 ± 5 0.15 ± 0.02 2.2 ± 0.3 pLVL 89 ± 10 42 ± 3.3 0.47 ± 0.06 240 ± 33 312 ± 30 1.3 ± 0.2 2.8 ± 0.1 pLEL 78 ± 9 3.4 ± 0.4 0.04 ± 0.006 205 ± 25 27 ± 3.5 0.13 ± 0.02 3.1 ± 0.4 pLL 52 ± 7 0.6 ± 0.08 0.01 ± 0.002 125 ± 15 1.0 ± 0.16 0.008 ± 0.001 0.8 ± 0.2 Ó FEBS 2002 eIF2a phosphorylation and IRES activity (Eur. J. Biochem. 269) 2813 kinases [18,26], or PKRD6, a dominant-negative variant of the dsRNA activated eIF2a-kinase, PKR [19]. To confirm the connection between eIF2a phosphorylation during differentiation and IRES activation, we also used a plasmid encoding a variant form of eIF2a in which Ser51 is replaced by an alanine residue (eIF2a-S51A). As this variant protein cannot undergo phosphorylation, it serves as a competitor that reduces the translational inhibitory effect of phosphor- ylated endogenous wild-type eIF2a [20]. Plasmids expres- sing K3L, PKRD6, eIF2a-S51A variant or GFP as control, were cotransfected along with the different IRES-containing bi-cistronic vectors into K562 cells followed by their incubation under normal or differentiation conditions for 48 h prior to measurements of Renilla and firefly luciferase enzymatic activities. Tables 2 and 3 show the effects of the transfected gene products on the absolute levels of the translation products of both cistrons. Overexpression of K3L, PKRD6oreIF2a-S51A led to enhanced translation of both cistrons in nondifferentiated cells, whereas in differen- tiated cells it led to decreased IRES-mediated translation of the second cistron. Fig. 4A summarizes the sensitivity of the differentiation-induced IRES activation to the various eIF2a phosphorylation inhibitors. Expression of eIF2a-S51A, K3L, or PKRD6 in differentiated cells reduced the level of eIF2a-P to 80%, 70% or 40% compared with GFP- transfected control, respectively (Fig. 4B). The effect of the various transfections on eIF2a phosphorylation is underes- timated as not all the cells were successfully transfected. The reduction in IRES use in differentiated cells by expression of K3L, eIF2a-S51A and PKRD6 was shown to be correlated with a reduction in the level of eIF2a-P. These data suggest that eIF2a phosphorylation is required for more efficient IRES use during the differentiation process. Fig. 3. Phosphorylation of eIF2a in logarithmically growing, dense and differentiated K562 cells. (A) A total of 10 6 logarithmically growing (Log), density-induced growth arrested (Dense) or differentiated (Dif f ) K562 cells were metabolically labelled with 32 Pi, followed by immunoprecipitation using an antibody specific for eIF2a.The immunoprecipitated phospholabeled proteins were separated by SDS/ 10% PAGE and blotted onto a nitrocellulose membrane. Phosphor- ylated eIF2a was observed following exposure of the membrane to an X-ray film and the intensities of the bands were determined using a phosphoimager. The same membrane was analysed for total eIF2a level by Western analysis and the intensities of the bands were deter- mined by densitometry. The [ 32 P]eIF2a/eIF2a ratio in logarithmically growing cells was set as 1. (B) Fifty lg total protein extract from Log, Dense or Diff K562 cells were separated by 10% SDS/PAGE and blotted onto a nitrocellulose membrane. Phosphorylated eIF2a was detected using antibodies specific for phosphorylated Ser51. The same membrane was stripped and used for Western analysis using antibodies specific for total eIF2a.TheeIF2a-P/eIF2a ratio in logarithmically growing cells was set as 1. Fig. 2. The effect of differentiation on the overall protein synthesis level. (A) Logarithmically growing (Log), density-induced growth arrested (Dense) or differentiated (Diff) K562 cells were metabolically labelled with [ 35 S] L -methionine and [ 35 S] L -cysteine followed by determination of their incorporation level by trichloro-acetic acid precipitation, as described in Materials and methods. The incorporation level (cpmÆlg )1 protein) in log cells was termed 100%. The values are means ± SE of three independent experiments. (B) A total of 3.5 · 10 7 log, dense or differentiated K562 cells were harvested and their cytoplasmic com- partments were subjected to fractionation on linear 5–47% sucrose gradients. Relative absorbance at 260 nm was monitored continuously as the gradient was collected. The vertical bars on the abscissa indicate the boundaries of the polysomal (P) and subpolysomal (SP) fractions. Peaks at the top of the gradient containing the 40 S, 60 S and 80 S ribosomal subunits are indicated. 2814 G. Gerlitz et al. (Eur. J. Biochem. 269) Ó FEBS 2002 Reduced phosphorylation of eIF2a has no significant effect on megakaryocytic differentiation The differentiation process is a cascade of events leading to major changes in gene expression. Reduced global protein synthesis is a consequence of upstream events, as inhibition of protein synthesis per se does not lead to differentiation. However, it seemed important to ascertain the effect of reduced eIF2a phosphorylation level on the differentiation process. As megakaryocytic differentiation involves growth arrest and polyploidy [27], DNA content evaluation by flow cytometry was chosen as a tool to detect the reduced number of cells in S-phase and enhanced number of cells harbouring two- to fourfold higher DNA content [23,28]. K562 cells were transfected with plasmids expressing K3L, PKRD6, eIF2a-S51A, or GFP under similar conditions to those used to assess the effect of inhibition of eIF2a phosphorylation on IRES activity. The transfection effi- ciency in these experiments was 50–60%, as judged by the percentage of fluorescent GFP-expressing cells. As shown in Fig. 5A, the differentiation process was not significantly affected by any of the transfected plasmids, as judged by the decreased number of cells in S-phase and increased number of polyploid cells. For additional confirmation we checked the level of the Vav proto-oncogene, which is known to increase early during megakaryocytic differentiation [27,29]. CKIIa protein level was used as a control. As shown in Fig. 5B, the level of Vav protein was increased due to differentiation, regardless of the transfected plasmid. These results suggest that interference with eIF2a phosphorylation does not prevent the early differentiation steps, e.g. the global changes in gene expression upstream of mRNA translation. Instead, it interferes with the ability to fine-tune the translation efficiency of specific mRNA groups. DISCUSSION Cells undergoing terminal differentiation exhibit extensive changes in the pattern of gene expression. Much data has been accumulated regarding transcriptional regulation, but less is known about the mechanisms that inhibit the translation of most transcripts while activating the transla- tion of specific mRNAs during the course of differentiation. During the early developmental stages of Xenopus, Caenor- habditis elegans and Drosophila, the translation of subclasses of mRNAs is regulated. However, during differentiation of mammalian cells, only a few individual mRNAs are known to be subjected to translational regulation due to their cis- regulatory elements (reviewed in [30]). Initial attempts to identify groups of translationally regulated genes during HL60 cell differentiation towards monocytes/macrophages has revealed that while most mRNAs are released from polysomes early in the differentiation process, a subset of transcripts is retained or even mobilized onto polysomes [7]. The data presented in this study suggest that mRNAs harbouring an IRES within long, structured, uORF- burdened 5¢UTRs, comprise a subgroup which is specific- ally translationally activated during differentiation (Fig. 1 Table 3. Effect of eIF2a-S51A expression on IRES activity. Absolute values of Renilla and Firefly activities from experiments described in Fig. 4A. Non-differentiated cells Differentiated cells Renilla (U per 10 6 cells) Firefly (U per 10 6 cells) F/R Renilla (U per 10 6 cells) Firefly (U per 10 6 cells) F/R Fold F/R induction pLPL + GFP 87 ± 6 7.8 ± 1.4 0.09 ± 0.02 184 ± 20 59 ± 5 0.32 ± 0.08 3.5 ± 0.3 + eIF2a-S51A 145 ± 11 14 ± 2 0.1 ± 0.02 200 ± 18 42 ± 3 0.2 ± 0.02 2.1 ± 0.2 pLVL + GFP 105 ± 11 42 ± 5 0.4 ± 0.06 273 ± 33 382 ± 37 1.4 ± 0.2 3.5 ± 0.2 + eIF2a-S51A 180 ± 20 72 ± 8 0.4 ± 0.1 304 ± 31 273 ± 28 0.9 ± 0.1 2.2 ± 0.2 Table 2. Effect of K3L and PKRD6 expression on IRES activity. Absolute values of Renilla and Firefly activities from experiments described in Fig. 4A. Non-differentiated cells Differentiated cells Fold F/R induction Renilla (U per 10 6 cells) Firefly (U per 10 6 cells) F/R Renilla (U per 10 6 cells) Firefly (U per 10 6 cells) F/R pLPL + GFP 53 ± 3 6.4 ± 0.6 0.12 ± 0.02 125 ± 17 50 ± 6 0.4 ± 0.05 3.2 ± 0.2 + K3L 184 ± 13 20 ± 2 0.1 ± 0 249 ± 27 48 ± 7 0.2 ± 0.03 2.0 ± 0.2 + PKRD6 132 ± 12 17 ± 2 0.13 ± 0.02 213 ± 16 37 ± 4 0.17 ± 0.03 1.2 ± 0.15 pLML + GFP 101 ± 12 7.3 ± 1.1 0.07 ± 0 235 ± 23 38 ± 14 0.16 ± 0.02 2.3 ± 0.3 + K3L 303 ± 27 20 ± 3.5 0.06 ± 0.01 318 ± 27 34 ± 3.6 0.1 ± 0.02 1.6 ± 0.2 + PKRD6 222 ± 25 22 ± 3 0.1 ± 0.03 286 ± 28 22 ± 3 0.08 ± 0.01 0.8 ± 0.1 pLVL + GFP 100 ± 11 52 ± 7 0.5 ± 0.1 253 ± 32 368 ± 40 1.43 ± 0.2 2.9 ± 0.2 + K3L 260 ± 21 130 ± 10 0.5 ± 0 390 ± 38 286 ± 20 0.73 ± 0.08 1.5 ± 0.15 + PKRD6 186 ± 20 120 ± 15 0.6 ± 0.1 300 ± 30 191 ± 5 0.6 ± 0.1 1.0 ± 0.1 pLEL + GFP 51 ± 6 2.0 ± 0.9 0.04 ± 0.01 226 ± 25 31 ± 5 0.13 ± 0.01 3.2 ± 0.4 + K3L 228 ± 24 8.5 ± 0.3 0.04 ± 0.01 330 ± 5 33 ± 4 0.11 ± 0.03 2.7 ± 0.25 + PKRD6 98 ± 10 5.7 ± 0.5 0.06 ± 0.01 192 ± 20 17 ± 3.3 0.09 ± 0.01 1.5 ± 0.2 Ó FEBS 2002 eIF2a phosphorylation and IRES activity (Eur. J. Biochem. 269) 2815 and Table 1) under conditions of eIF2a phosphorylation and substantial inhibition of protein synthesis (Figs. 2 and 3). Other recent studies demonstrate a correlation between differentiation with reduction of global protein synthesis and enhanced eIF2a phosphorylation [31–34]. The list of recently identified IRES elements within cumbersome 5¢UTRs of growth factors, cytokines, tran- scription factors and oncogenes is constantly growing (reviewed in [14,15]). Cellular IRES elements have been implied to confer a translational advantage under reduced levels of active 5¢-cap binding complex. However, the current study shows that under certain physiological conditions, for instance during differentiation, translation mediated by cellular IRES elements benefits from phos- phorylation of eIF2a. Supplementary mechanisms for inhibition of global protein synthesis, such as reduced availability of the 5¢-cap binding complex, may also take place during differentiation and contribute to the observed enhancement of 5¢-cap independent translation. However, such mechanisms were beyond the focus of this study. Phosphorylation of eIF2a which leads to decreased binding of initiator tRNA to the small ribosomal subunit, has been mostly studied in relation to growth inhibition induced in response to starvation for growth factors/nutrients, heat shock, and virus infection (reviewed in [35]). PKR, the interferon-induced (double-stranded) RNA-activated eIF2a kinase, has been implicated in cellular growth control, as well as in differentiation and apoptosis (reviewed in [36–39]). It seems likely that the regulatory function of eIF2 depends on the delicate balance of phosphorylated eIF2a with other cellular components, and on the physiological status of the cell. The data presented in this study support this idea. Inhibition of PKR activity by over-expression of its dominant-negative variant PKRD6, or reduction of eIF2a phosphorylation level by over-expression of its variant form, eIF2a-S51A, or the pseudosubstrate K3L, resulted in reduced IRES activity in differentiated cells. However, it did not have any inhibitory effect on IRES activity in nondif- ferentiated cells (Fig. 4). The fact that expression of PKRD6 had a greater impact on the levels of eIF2a phosphorylation and IRES-mediated translation compared with the efficient general eIF2a kinase inhibitor K3L suggests that PKR is the primary activated kinase during differentiation. Interest- ingly, the over-expression of the eIF2a phosphorylation Fig.4.EffectofeIF2a phosphorylation inhibitors on IRES activity. (A) Each of the bicistronic vectors pLPL, pLVL, pLML, pLEL (described in Fig. 1A) harbouring the IRES elements of PDGF2, VEGF, c-Myc or EMCV, respectively, was cotransfected into K562 cells together with a plasmid expressing the PKRD6, K3L, eIF2a Ser51 fi Ala mutant ( pc2a-S51A), or GFP coding region from the CMV promoter. The cells were further incubated under normal or differentiation con- ditions for 48 h and subsequently analysed for Renilla (R) and firefly (F) luciferase activity. The absolute values are presented in Tables 2 and 3. Each value represents the mean ± SE of three independent experiments. The fold induction values represent the F/R ratio in differentiated cells relative to the F/R ratio in nondifferentiated cells. The graph demonstrates the effect of K3L (stippled bars), PKRD6 (dark bars), or eIF2a-S51A (hatched bars) on the differentiation- induced IRES activation relative to the fold induction value with GFP that was set as 100% (light bars). (B) Fifty lg of total protein extract from differentiated cells transfected with plasmids expressing GFP, K3L, PKRD6oreIF2a-S51A were separated by 10% SDS/PAGE and blotted onto a nitrocellulose membrane. Phosphorylated eIF2a was detected using antibodies specific for phosphorylated Ser51. The same membrane was stripped and used for Western analysis using antibodies specific for total eIF2a.TheeIF2a-P/eIF2a ratio in GFP-transfected cellswassetas1. Fig. 5. The effect of eIF2a phosphorylation inhibitors on the differenti- ation process. (A) A total of 5 · 10 5 K562 cells were transfected by electroporation with a plasmid expressing GFP, eIF2a-S51A, K3L or PKRD6, as indicated. The transfected cells were incubated under normal (–TPA) or differentiation (+ TPA) conditions for 48 h, and subjected to DNA content analysis by flow cytometry. (B) One-hun- dred and forty lg of total protein extracted from the transfected cells as detailedin(A)wereseparatedbySDS/10%PAGE,blottedontoa nitrocellulose membrane, and subjected to Western analysis using antibodies specific for Vav and for CKIIa. 2816 G. Gerlitz et al. (Eur. J. Biochem. 269) Ó FEBS 2002 inhibitors did not interfere with the global process of megakaryocytic differentiation as judged by their morphol- ogy (not shown), cell-cycle, and enhanced Vav protein expression (Fig. 5). The latter result suggests that both eIF2a phosphorylation and IRES activation are late events during the differentiation process. The peak of PDGF2 IRES activation at 48 h after induction of differentiation [11,12] is in agreement with this notion. It is therefore conceivable that eIF2a phosphorylation serves to fine-tune the translation efficiency of specific mRNA groups. Increased translation of certain IRES-containing mRNAs has also been implicated in apoptosis [40–43], a cellular process that includes activation of eIF2a phos- phorylation [44–47]. Furthermore, the recently discovered cell cycle-dependent IRES elements are activated specifically at the G 2 /M boundary [48–50], when increased phosphory- lation of eIF2a is found in correlation with decreased overall rate of protein synthesis [51]. Moreover, the IRES elements of the amino acid transporter protein cat-1 and c-Myc mRNAs have recently been shown to function efficiently where there is an increase in eIF2a phosphory- lation, under conditions of amino acid starvation and genotoxic stress, respectively [52,53]. What is the mechanism underlying IRES-mediated translation under conditions of enhanced eIF2a phosphory- lation? In nondifferentiated cells, in which global translation is active, the IRES-containing mRNAs compete with the cap-dependent mRNAs for the translation machinery. The decrease in global protein synthesis and reduced competi- tion might be beneficial for IRES-mediated translation during differentiation. An interesting possibility may be the ability of IRES elements to direct efficient translation initiation in the absence of eIF2 and Met-tRNA i . Recently, internal initiation without Met-tRNA i has been demonstra- ted in two picorna-like insect viruses, Plautia stali intestinal virus and Cricket paralysis virus [54–56]. It is an open question whether cellular IRES elements known to contain conserved secondary and tertiary structural motifs can also direct internal translation from noncognate initiation codons in the absence of Met-tRNA i . Direct binding to the 40 S ribosomal subunit followed by joining of the 60 S subunit may provide a significant advantage to IRES elements that confer efficient translation under conditions of global translation inhibition mediated by eIF2a phosphory- lation. Another possibility could be that a mechanism exists that is similar to the translational regulation of GCN4 in yeast. In this case, the induction of GCN4 translation in response to eIF2a phosphorylation is modulated by four short uORFs in the 5¢UTR. Reduced rates of ternary complex formation leads to bypass of the uORFs and initiation at the downstream GCN4 major ORF [6]. A comparable mechanism of translational regulation was recently demonstrated for the stress-induced transcription factor ATF4, in mammalian cells under conditions of enhanced eIF2a phosphorylation due to stress [57]. Simi- larly, the short ORFs that furnish many of the cellular IRES elements may have a role in translational regulation, as seems to be the case for the activation of the cat-1 mRNA [52]. Internal ribosomal binding upstream of the translation initiator codon may be followed by subsequent scanning to the initiation codon. For instance, in the case of the PDGF2 5¢UTR, which contains three uORFs, the IRES has been mapped to the central part of the 5¢UTR at the vicinity of the first uORF [13]. The need to scan through the second and third uORFs towards the major coding region cannot be ruled out at this point. Another possibility may be that eIF2a phosphorylation induces the synthesis of a protein that interacts with the IRES. This can be achieved by direct translational regulation of its mRNA akin to GCN4/ATF4 mRNAs, or by regulating the translation of a GCN4/ ATF4-like transcription factor that activates the transcrip- tion of the potential IRES activator. Current experiments are designed to elucidate the mechanism(s) by which eIF2a phosphorylation serves to enhance IRES-mediated transla- tion. ACKNOWLEDGEMENTS This work was supported by the Israel Science Foundation adminis- tered by the Academy of Sciences and Humanities – the Charles H. Revson Foundation to O. E. S., by a grant from the Israeli Chief Scientist’s Office of the Ministry of Health to O. E. S. and NSF 9808401 to R. J. We thank B. White and G. Krause for the antibody against phosphorylated eIF2a, D. Canaani for antibody against CKIIa, N. Sonenberg for PKRD6 construct, R. J. Kaufman for eIF2a-S51A construct, and B Z. Levi for the pSKVLUC construct. We are grateful to T. Dever for comments on the manuscript. REFERENCES 1. Dever, T.E. (1999) Translation initiation: adept at adapting. Trends Biochem. Sci. 24, 398–403. 2. Mathews, M.B., Sonenberg, N. & Hershey, J.W.B. (2000) Origins and principles of translational control. In Translational Control of Gene Expression (Sonenberg, N., Hershey, J.W.B. & Mathews M.B., eds), pp. 1–32. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York. 3. Pain, V.M. (1996) Initiation of protein synthesis in eukaryotic cells. Eur. J. Biochem. 236, 747–771. 4. Hinnebusch, A.G. (2000) Mechanism and regulation of initiator methionyl-tRNA binding to ribosomes. In Translational Control of Gene Expression (Sonenberg, N., Hershey, J.W.B. & Mathews M.B., eds), 185–244. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York. 5. Schneider, R.J. (2000) Translational control during heat shock. In: Translational Control of Gene Expression (Sonenberg, N., Hershey, J.W.B. & Mathews M.B., eds), pp. 581–594. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York. 6. Hinnebusch, A.G. (1996) Translational control of GCN4: gene- specific regulation by phosphorylation of eIF2. In Translational Control (Hershey, J.W.B., Mathews, M.B. & Sonenberg, N., eds), pp. 199–244. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York. 7. Krichevsky, A.M., Metzer, E. & Rosen, H. (1999) Translational control of specific genes during differentiation of HL-60 cells. J. Biol. Chem. 274, 14295–14305. 8. Meyuhas, O. (2000) Synthesis of the translational apparatus is regulated at the translational level. Eur. J. Biochem. 267, 6321– 6330. 9. Thiele, B.J., Berger, M., Huth, A., Reimann, I., Schwarz, K. & Thiele, H. (1999) Tissue-specific translational regulation of alter- native rabbit 15-lipoxygenase mRNAs differing in their 3¢-untranslated regions. Nucleic Acids Res. 27, 1828–1836. 10. Gray, N.K. & Wickens, M. (1998) Control of translation initiation in animals. Annu. Rev. Cell. Dev. Biol. 14, 399–458. 11. Bernstein, J., Shefler, I. & Elroy-Stein, O. (1995) The translational repression mediated by the platelet-derived growth factor 2/c-sis mRNA leader is relieved during megakaryocytic differentiation. J. Biol. Chem. 270, 10559–10565. Ó FEBS 2002 eIF2a phosphorylation and IRES activity (Eur. J. Biochem. 269) 2817 12. Bernstein, J., Sella, O., Le, S.Y. & Elroy-Stein, O. (1997) PDGF2/ c-sis mRNA leader contains a differentiation-linked internal ribosomal entry site (D-IRES). J. Biol. Chem. 272, 9356–9362. 13. Sella, O., Gerlitz, G., Le, S.Y. & Elroy-Stein, O. (1999) Differ- entiation-induced internal translation of c-sis mRNA: analysis of the cis elements and their differentiation-linked binding to the hnRNP C protein. Mol. Cell. Biol. 19, 5429–5440. 14. Carter, M.S., Kuhn, K.M. & Sarnow, P. (2000) Cellular internal ribosomal entry site (IRES) elements and the use of cDNA microarrays in their investigation. In Translational Control of Gene Expression (Sonenberg, N., Hershey, J.W.B. & Mathews, M.B., eds), pp. 615–636. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York. 15. Hellen, C.U.T. & Sarnow, P. (2001) Internal ribosomal entry sites in eukaryotic mRNA molecules. Genes Dev. 15, 1593–1612. 16. Akiri, G., Nahari, D., Finkelstein, Y., Le, S.Y., Elroy-Stein, O. & Levi, B.Z. (1998) Regulation of vascular endothelial growth factor (VEGF) expression is mediated by internal initiation of translation and alternative initiation of transcription. Oncogene 17, 227–237. 17. Moss, B., Elroy-Stein, O., Mizukami, T., Alexander, W.A. & Fuerst, T.R. (1990) New mammalian expression vectors. Nature 348, 91–92. 18. Carroll, K., Elroy-Stein, O., Moss, B. & Jagus, R. (1993) Recombinant vaccinia virus K3L gene product prevents activation of dsRNA-dependent, eIF-2a-specific protein kinase. J. Biol. Chem. 268, 12837–12842. 19. Koromilas, A.E., Roy, S., Barber, G.N., Katze, M.G. & Sonen- berg, N. (1992) Malignant transformation by a mutant of the IFN-inducible dsRNA-dependent protein kinase. Science 257, 1685–1689. 20. Kaufman, R.J., Davies, M.V., Pathak, V.K. & Hershey, J.W.B. (1989) The phosphorylation state of eukaryotic initiation factor 2 alters translational efficiency of specific mRNAs. Mol. Cell. Biol. 9, 946–958. 21. Scorsone, K.A., Panniers, R., Rowlands, A.G. & Henshaw, E.C. (1987) Phosphorylation of eukaryotic initiation factor 2 during physiological stresses which affect protein synthesis. J. Biol. Chem. 262, 14538–14543. 22. Meyuhas, O., Thompson, E.A. & Perry, R.P. (1987) Glucocorti- coids selectively inhibit translation of ribosomal protein mRNA in P1798 lymphosarcoma cells. Mol. Cell. Biol. 7, 2691–2699. 23. Hoffman, R. (1989) Regulation of megakaryocytopoiesis. Blood 74, 1196–1212. 24. Hershey, J.W.B. & Merrick, W.C. (2000) Initiation of Protein Synthesis. In Translational Control of Gene Expression (Sonen- berg, N., Hershey, J.W.B. & Mathews, M.B., eds), pp. 33–88. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York. 25. Colthurst, D.R., Campbell, D.G. & Proud, C.G. (1987) Structure and regulation of eukaryotic initiation factor eIF-2. Sequence of the site in the alpha subunit phosphorylated by the haem-con- trolled repressor and by the double-stranded RNA-activated inhibitor. Eur. J. Biochem. 166, 357–363. 26. Davies, M.V., Elroy-Stein, O., Jagus, R., Moss, B. & Kaufman, R.J. (1992) The vaccinia virus K3L gene product potentiates translation by inhibiting double-stranded RNA-activated protein kinase and phosphorylation of the alpha subunit of eukaryotic initiation factor 2. J. Virol. 66, 1943–1950. 27. Nagata, Y., Nagahisa, H., Nagasawa, T. & Todokoro, K. (1997) Regulation of megakaryocytopoiesis by thrombopoietin and stromal cells. Leukemia 11, 435–438. 28. Cavalloni, G., Dane, A., Piacibello, W., Bruno, S., Lamas, E., Brechot, C. & Aglietta, M. (2000) The involvement of human-nuc gene in polyploidization of K562 cell line. Exp. Hematol. 28, 1432– 1440. 29. Bustelo, X.R., Rubin, S.D., Suen, K.L., Carrasco, D. & Barbacid, M. (1993) Developmental expression of the vav protooncogene. Cell Growth Diff. 4, 297–308. 30. Wickens,M.,Goodwin,E.,Kimble,J.,Stickland,S.&Hentze, M.W. (2000) Translational control of developmental decisions. In Translational Control of Gene Expression (Hershey, J.W.B., Mathews, M.B. & Sonenberg, N., eds), pp. 295–370. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York. 31. Aroor, A.R., Singh, L.P. & Wahba, A.J. (1995) Hexamethylene bisacetamide-induced differentiation of Friend virus-transformed murine erythroleukemia cells is associated with parallel changes is casein kinase II and guanine nucleotide exchange factor activities. Exp. Hematol. 23, 1204–1211. 32. Hensold, J.O., Barth-Baus, D. & Stratton, C.A. (1996) Inducers of erythroleukemic differentiation cause mRNAs that lack poly (A)- binding protein to accumulate in translationally inactive, salt- labile 80S ribosomal complexes. J. Biol. Chem. 271, 23246–23254. 33. Salzberg, S., Vilchik, S., Cohen, S., Heller, A. & Kronfeld-Kinar, Y. (2000) Expression of a PKR dominant-negative mutant in myogenic cells interferes with the myogenic process. Exp. Cell. Res. 254, 45–54. 34. Woldehawariat, G., Nekhai, S. & Petryshyn, R. (1999) Differential phosphorylation of PKR associates with deregulation of eIF- 2alpha phosphorylation and altered growth characteristics in 3T3- F442A fibroblasts. Mol. Cell. Biochem. 198, 7–17. 35. Rhoads, R.E. (1999) Signal transduction pathways that regulate eukaryotic protein synthesis. J. Biol. Chem. 274, 30337–30340. 36. Barber, G.N. (2000) The interferons and cell death: guardians of the cell or accomplices of apoptosis. Semin. Cancer Biol. 10,103– 111. 37. Jagus, R., Joshi, B. & Barber, G.N. (1999) PKR, apoptosis and cancer. Int. J. Biochem. Cell. Biol. 31, 123–138. 38. Kaufman, R.J. (2000) Double-stranded RNA-activated protein kinase, PKR. In Translational Control of Gene Expression (Sonenberg,N.,Hershey,J.W.B.&Mathews,M.B.,eds),pp.503– 527. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York. 39. Kronfeld-Kinar, Y., Vilchik, S., Hyman, T., Leibkowicz, F. & Salzberg, S. (1999) Involvement of PKR in the regulation of myogenesis. Cell. Growth Diff. 10, 201–212. 40. Coldwell, M., Mitchell, S., Stoneley, M., MacFarlane, M. & Willis, A.E. (2000) Initiation of Apaf-1 translation by internal ribosome entry. Oncogene 19, 899–905. 41. Henis-Korenblit, S., Levi-Strumpf, N., Goldstaub, D. & Kimchi, A. (2000) A novel form of DAP5 protein accumulates in apoptotic cells as a result of caspase cleavage and internal ribosome entry- site-mediated translation. Mol. Cell. Biol. 20, 496–506. 42. Holcik, M., Lefebvre, C., Yeh, C., Chow, T. & Korneluk, R.G. (1999) A new internal-ribosomal-entry-site motif potentiates XIAP-mediated cytoprotection. Nature Cell. Biol. 1, 190–192. 43. Stoneley, M., Chappell, S.A., Jopling, C.L., Dickens, M., Mac- Farlane, M. & Willis, A.E. (2000) c-Myc protein synthesis is initiated from the internal ribosome entry segment during apop- tosis. Mol. Cell. Biol. 20, 1162–1169. 44. Balachandran, S., Kim, C.N., Yeh, W.C., Mak, T.W., Bhalla, K. & Barber, G.N. (1998) Activation of the dsRNA-dependent pro- tein kinase, PKR, induces apoptosis through FADD-mediated death signaling. EMBO J. 17, 6888–6902. 45. Clemens, M.J., Bushell, M., Jeffrey, I.W., Pain, V.M. & Morley, S.J. (2000) Translation initiation factor modifications and the regulation of protein synthesis in apoptotic cells. Cell Death Diff. 7, 603–615. 46. Gil, J., Alcami, J. & Esteban, M. (1999) Induction of apoptosis by double-stranded-RNA-dependent protein kinase (PKR) involves the alpha subunit of eukaryotic translation initiation factor 2 and NF-kappaB. Mol. Cell. Biol. 19, 4653–4663. 2818 G. Gerlitz et al. (Eur. J. Biochem. 269) Ó FEBS 2002 47. Srivastava, S.P., Kumar, K.U. & Kaufman, R.J. (1998) Phos- phorylation of eukaryotic translation initiation factor 2 mediates apoptosis in response to activation of the double-stranded RNA- dependent protein kinase. J. Biol. Chem. 273, 2416–2423. 48. Cornelis, S., Bruynooghe, Y., Denecker, G., van Huffel, S., Tinton, S. & Beyaert, R. (2000) Identification and characterization of a novel cell cycle-regulated internal ribosome entry site. Mol. Cell 5, 597–605. 49. Honda, M., Kaneko, S., Matsushita, E., Kobayashi, K., Abell, G.A. & Lemon, S.M. (2000) Cell cycle regulation of hepatitis C virus internal ribosomal entry site-directed translation. Gastro- enterology 118, 152–162. 50. Pyronnet, S., Pradayrol, L. & Sonenberg, N. (2000) A cell-cycle- dependent internal ribosome entry site. Mol. Cell 5, 607–616. 51. Datta, B., Datta, R., Mukherjee, S. & Zhang, Z. (1999) Increased phosphorylation of eukaryotic initiation factor 2 alpha at the G2/M boundary in human osteosarcoma cells correlates with deglycosylation of p67 and a decreased rate of protein synthesis. Exp. Cell. Res. 250, 223–230. 52. Fernandez, J., Yaman, I., Merrick, W.C., Koromilas, A., Wek, R.C., Sood, R., Hensold, J. & M.Hatzoglou. (2001) Regulation of internal ribosomal entry site-mediated translation by eIF2a phosphorylation and translation of small uORF. J. Biol. Chem. 277, 2050–2058. 53. Subkhankulova, T., Mitchell, S.A. & Willis, A.E. (2001) Internal ribosomal entry segment–mediated initiation of c-Myc pro- tein synthesis following genotoxic stress. Biochem. J. 359, 183– 192. 54. Sasaki, J. & Nakashima, N. (2000) Methionine-independent initiation of translation in the capsid protein of an insect RNA virus. Proc. Natl Acad. Sci. USA 97, 1512–1515. 55. Wilson, J.E., Powell, M.J., Hoover, S.E. & Sarnow, P. (2000) Naturally occurring dicistronic cricket paralysis virus RNA is regulated by two internal ribosome entry sites. Mol. Cell. Biol. 20, 4990–4999. 56. Wilson, J.E., Pestova, T.V., Hellen, C.U.T. & Sarnow, P. (2000) Initiation of protein synthesis from the A site of the ribosome. Cell 102, 511–520. 57. Harding, H.P., Novoa, I., Zhang, Y., Zeng, H., Wek, R., Schapira, M. & Ron, D. (2000) Regulated translation initiation controls stress-induced gene expression in mammalian cells. Mol. Cell 6, 1099–1108. Ó FEBS 2002 eIF2a phosphorylation and IRES activity (Eur. J. Biochem. 269) 2819 . Phosphorylation of initiation factor-2a is required for activation of internal translation initiation during cell differentiation Gabi. pLL. Reduction of global protein synthesis during differentiation is accompanied by eIF2a phosphorylation The terminal differentiation process is usually accompanied by