Oncogene (1998) 17, 2921 ± 2931 ã 1998 Stockton Press All rights reserved 0950 ± 9232/98 $12.00 http://www.stockton-press.co.uk/onc Degradation of eukaryotic polypeptide chain (eIF) 4G in response to induction of apoptosis in human lymphoma cell lines

Michael J Clemens1, Martin Bushell2 and Simon J Morley2

1Department of Biochemistry, Cellular and Molecular Sciences Group, St George's Hospital Medical School, Cranmer Terrace, London SW17 0RE; 2Biochemistry Group, School of Biological Sciences, University of Sussex, Brighton BN1 9QG, UK

We have investigated the e€ect of inducing apoptosis in complex, which allows the mRNA molecule to BJAB and Jurkat cells on the cellular content of several associate with the 40S ribosomal subunit (reviewed in polypeptide chain initiation factors. Serum deprivation Merrick and Hershey, 1996; Morley et al., 1997). In results in inhibition of synthesis and induction of mammalian cells initiation factor eIF4G is a single apoptosis in BJAB cells; at early times, there is selective polypeptide chain of 154 kDa (Yan et al., 1992) which degradation of polypeptide initiation factor eIF4G but no exists in vivo at least partly in the form of a trimeric major losses of other key initiation factors. The complex with eIF4E and the ATP-dependent RNA disappearance of full length eIF4G is accompanied by eIF4A, constituting the large initiation factor the appearance of smaller forms of the protein, including eIF4F (Merrick, 1994). Recently a second form of a major product of approximately 76 kDa. Apoptosis human eIF4G has also been described (Gradi et al., induced by cycloheximide results in similar e€ects. Both 1998). eIF4G possesses domains for the binding of total cytoplasmic eIF4G and eIF4G associated with eIF4E and eIF3 in the N-terminal and central parts of eIF4E are degraded with a half-life of 2 ± 4 h under these its sequence respectively, and it also has binding sites conditions. Treatment of serum-starved or cyclohex- for eIF4A (Lamphear et al., 1995; Imataka and imide-treated cells with Z-VAD.FMK or Z- Sonenberg, 1997) and, at least in yeast, for the DEVD.FMK, which inhibit caspases required for poly(A) binding protein (PABP) (Tarun and Sachs, apoptosis, protects eIF4G from degradation and blocks 1996; Le et al., 1997; Tarun et al., 1997). Interaction of the appearance of the ca. 76 kDa product. Exposure of PABP with eIF4G or with the recently discovered BJAB cells to rapamycin rapidly inhibits protein eIF4G homologue PAIP (Craig et al., 1998) may allow synthesis but does not lead to acute degradation of functional association of the 3' end of an mRNA with eIF4G. In both BJAB and Jurkat cells induction of the 5' end, thus e€ectively causing `circularization' of apoptosis with anti-Fas antibody or etoposide also results the mRNA during protein synthesis (Tarun et al., in the selective loss of eIF4G, which is inhibitable by Z- 1997). VAD.FMK. These data suggest that eIF4G is selectively The association of eIF4G with eIF4E strongly targeted for cleavage as cells undergo apoptosis and is a enhances the binding of the latter to 5' mRNA cap substrate for proteases activated during this process. structures (Haghighat and Sonenberg, 1997). However this interaction is inhibited by a small family of 4E Keywords: apoptosis; eIF4G/initiation factors; initia- binding (4E-BPs) which can compete with tion of protein synthesis eIF4G for binding to eIF4E (Mader et al., 1995; Haghighat et al., 1995; Altmann et al., 1997; Matsuo et al., 1997). The extent of complex formation between the best studied of these proteins, 4E-BP1 (also known Introduction as PHAS-I) (Lawrence and Abraham, 1997), and eIF4E is regulated by the of 4E-BP1 The initiation of protein synthesis is regulated by (Graves et al., 1995; Von Manteu€el et al., 1996; complex interactions between a large number of Brunn et al., 1997). This phosphorylation is activated protein initiation factors and RNA molecules (Pain, by a pathway which can be inhibited by the macrolide 1996; Merrick and Hershey, 1996). Disruption of this immunosuppressant, rapamycin (Lin et al., 1995; Von process can lead to transformation of cells to an Manteu€el et al., 1996; Beretta et al., 1996a; Fadden et oncogenic phenotype, probably as a result of aberrant al., 1997; Brunn et al., 1997). As a result rapamycin expression of proteins involved in the control of cell inhibits cap-dependent initiation (Beretta et al., growth or cell death (Sonenberg, 1996). Many of the 1996a,b), as well as antagonizing increases in protein steps in the initiation pathway require speci®c protein ± synthesis brought about by mitogenic stimuli (Lin et protein interactions between individual initiation al., 1995). In some eukaryotic systems, notably the factors as well as between these factors and the yeast Saccharomyces cerevisiae, rapamycin also inhibits . It is now established that the large the basal rate of protein synthesis (Barbet et al., 1996), polypeptide chain initiation factor eIF4G plays a although it is not yet clear whether the same crucial role in these interactions by acting as a bridge mechanisms are responsible for the e€ects of rapamy- between other components such as the mRNA cap- cin in yeast and mammalian cells (Altmann et al., 1997; binding protein eIF4E and the multi-subunit eIF3 Thomas and Hall, 1997). Enhancement of protein synthesis at the translational level by mitogens is also associated with the increased phosphorylation of eIF4E at position Ser209 (Joshi et al., 1995; Flynn and Proud, Correspondence: MJ Clemens 1995), as well as the phosphorylation of eIF4G itself Received 10 April 1998; revised 18 June 1998; accepted 18 June 1998 (Morley and Pain, 1995a,b). Loss of eIF4G during apoptosis MJ Clemens et al 2922 In addition to regulation by alterations in the activity investigated whether down-regulation of the factor may of eIF4E and eIF4G via protein phosphorylation, also be seen under conditions where cells are subjected protein synthesis can be modulated by changes in the to growth inhibitory and death-inducing stimuli. To levels of these factors in the cell. Over-expression of study this we have chosen to use human B and T eIF4E or eIF4G in NIH-3T3 cells can result in the lymphocyte-derived tumour cell lines, viz. BJAB and transformation of these cells into a malignant Jurkat cells respectively, since such cell types are phenotype (Lazaris-Karatzas et al., 1990; Fukuchi- sensitive to a wide range of physiological signals that Shimogori et al., 1997) and it has been proposed that can inhibit proliferation and induce apoptosis (pro- this is due to the enhanced of normally grammed cell death). Examples of such conditions inecient mRNAs with complex secondary structures in include deprivation of serum growth factors (Kawa- their 5' untranslated regions (Koromilas et al., 1992). nishi, 1997), inhibition of protein synthesis with Conversely, down-regulation of eIF4E expression cycloheximide (Ishii et al., 1995, 1997), treatment with causes a strong inhibition of protein synthesis, antibodies to the CD95 Fas antigen (Oehm et al., 1992; decreases cell growth rates and impairs malignancy Juo et al., 1997), or exposure to etoposide, an inhibitor (De Benedetti et al., 1991; Gra€ et al., 1995). (Note, of DNA topoisomerase II (Sinha, 1995). Apoptosis is however, that the level of eIF4G also falls dramatically characterized by a series of cellular changes that lead when eIF4E is depleted (De Benedetti et al., 1991) so it ultimately to cell death (Hale et al., 1996; Rowan and is not possible to say which factor is the more important Fisher, 1997; Nagata, 1997; Salvesen and Dixit, 1997). in bringing about these phenotypic consequences). High An essential aspect of these events is the sequential levels of expression of eIF4E and, in some cases, of activation of a number of proteases with limited eIF4G have been found in a variety of malignant substrate speci®cities, the caspases (Fearnhead et al., human tumours (Miyagi et al., 1995; Li et al., 1997; 1995; Tewari et al., 1995; Enari et al., 1996; Salvesen Nathan et al., 1997a,b; Brass et al., 1997), but it is not and Dixit, 1997; Cohen, 1997; Longthorne and known whether this is due to enhanced synthesis or a Williams, 1997; Ohta et al., 1997). Intensive research slower rate of degradation of these initiation factors. in recent years has investigated the properties of these Elevated levels of mRNA for eIF4E are seen in cells enzymes, the nature of some of their substrates, and transformed with the oncogene c-myc (Rosenwald et al., their relative relationships in the protease cascades 1993), as well as in normal T lymphocytes after required for execution of the apoptotic programme. activation by mitogens (Mao et al., 1992). We report here that eIF4G is degraded as part of the Dramatic changes occur in the level and integrity of apoptotic response in BJAB cells and Jurkat cells. eIF4G in mammalian cells after infection with certain Cleavage of eIF4G to give smaller discrete products picornaviruses such as polio or foot-and-mouth-disease occurs contemporaneously with other events associated viruses. Within a few hours of infection eIF4G is with apoptosis and is blocked by caspase inhibitors, cleaved at speci®c internal sites by virally-encoded suggesting that the initiation factor may be a substrate proteases to yield de®ned N-terminal and C-terminal for apoptotic proteases in vivo. The signi®cance of the fragments (reviewed in Morley et al., 1997). Since the loss of intact eIF4G for changes in protein synthesis cleavages occur in a region between the eIF4E and during programmed cell death is discussed. eIF3 binding sites (Lamphear et al., 1993) they destroy the bridging function of eIF4G in protein synthesis and thus lead to the inhibition of cap-dependent initiation. Results However the remaining C-terminal fragment of eIF4G can still support the initiation of protein synthesis that E€ects of serum deprivation on protein synthesis, levels involves ribosome binding to internal sites on mRNAs of initiation factors and induction of apoptosis in BJAB (and is therefore cap-independent; Pestova et al., 1996). cells As a result picornavirus RNA-directed protein synth- esis, which occurs by internal initiation, is favoured in Protein synthesis in BJAB cells is very sensitive to cells in which eIF4G has been cleaved. It is not yet serum deprivation. As shown in Figure 1, the clear whether a similar mechanism is used to inhibit incorporation of [35S] into acid-insoluble cap-dependent protein synthesis in uninfected cells (in material declines by more than 60% within 24 h of which a small number of cellular mRNAs may also be placing the cells in 0.1% foetal calf serum. After 2 days translated by an internal initiation mechanism), of this treatment incorporation is reduced to 11% of although there are a few reports of the cleavage of the control value. We have investigated the extent to eIF4G under situations where protein synthesis is which this decrease in translational activity is down-regulated by physiological stimuli such as brain associated with the loss of key factors required for ischaemia and reperfusion (DeGracia et al., 1996) and polypeptide chain initiation. As shown in Figure 2a induced di€erentiation of erythroleukaemia cells (and quanti®ed in panel B), after 24 h of serum (Benton et al., 1996). The presence of several PEST deprivation there is a marked disappearance of sequences in eIF4G suggests that this factor might be initiation factor eIF4G to about 25 ± 30% of control subject to rapid degradation under some circumstances levels, with lesser changes in the levels of eIF4E, eIF4A (Morley et al., 1997). and 4E-BP-1. The level of eIF2a was also seen to Because potential increases in eIF4G activity (as a decrease over this time, albeit at a reproducibly slower result of enhanced expression or increased binding to rate. eIF4E) are associated with the responses of cells to In addition to inhibition of protein synthesis and cell mitogenic stimuli, and because eIF4G itself is growth, BJAB cells undergo apoptosis when placed in apparently able to induce cell growth when over- low serum. This is indicated by the appearance of a expressed (Fukuchi-Shimogori et al., 1997), we have speci®c 89 kDa cleavage product of the 112 kDa Loss of eIF4G during apoptosis MJ Clemens et al 2923

Figure 1 Inhibition of protein synthesis by serum deprivation of BJAB cells. Exponentially growing cells were resuspended in RPMI 1640 medium containing either 10% or 0.1% (v/v) foetal calf serum and maintained in culture for the times indicated. Just Figure 2 Serum-deprivation promotes the degradation of eIF4G prior to harvest the cells were counted and triplicate 1 ml aliquots in BJAB cells. Cells were placed in 0.1% serum for various were incubated with [35S]methionine (2 mCi/ml) for 1 h. Protein periods of time, as described in Figure 1, and cytoplasmic extracts synthesis was measured as described in Materials and methods. were prepared (see Materials and methods). Equal quantities of Zero time values have been subtracted from the total acid- protein were subjected to SDS gel electrophoresis followed by insoluble radioactivity incorporated. Data from a typical immunoblotting for the proteins indicated. (a) Levels of eIF4G, experiment are shown and are the means+s.e.m. of three eIF4A, eIF2a, eIF4E, and 4E-BP1 in cell extracts at di€erent measurements at each time point times after serum deprivation. (b) Quanti®cation of relative levels of factors by scanning densitometry, expressed as percentage of level in control cells. The experiment was carried out three times and the error bars represent standard errors of the mean. (c) Parallel time-courses for cleavage of eIF4G and PARP at early protein poly(ADP-ribose) polymerase (PARP) (Kauf- times after serum deprivation. Full-length eIF4G and PARP and mann et al., 1993; Duriez and Shah, 1997), beginning their ca. 76 kDa and 89 kDa cleavage products are indicated within 1 h after initiation of serum deprivation (Figure 2c). We have observed that the loss of eIF4G under these circumstances is also accompanied by the appearance of a number of discrete fragments, the Cycloheximide treatment induces apoptosis and cleavage most prominent of which has an apparent size of about of eIF4G in BJAB cells 76 kDa (Figure 2c). These fragments are recognized by an anti-eIF4G serum directed against a C-terminal Incubation of BJAB cells with cycloheximide has region of the protein (amino acids 920 ± 1396) and must previously been shown to inhibit protein synthesis therefore contain at least part of this sequence. The and induce rapid apoptosis (Ishii et al., 1995, 1997). cleavages of eIF4G and PARP occur with essentially The e€ects of this apoptotic agent on the levels of identical time-courses. Because of these interesting several initiation factors are shown in Figure 4a. It can changes in the integrity of eIF4G our further studies be seen that eIF4G disappears with rapid kinetics after have focused largely on this factor. treatment of cells with cycloheximide. Densitometry of To determine whether the disappearance of eIF4G the immunoblots (Figure 4b) indicates that, after an is a consequence of the induction of apoptosis or initial lag, the level of the factor declines to about half whether it is a separate response to serum deprivation, its original value within 2 ± 4 h and eIF4G becomes we have investigated the e€ects of inhibition of virtually undetectable by 6 h. In contrast, eIF4E caspase activity in intact cells on the integrity of (quanti®ed in Figure 4b), eIF2a, 4E-BP1 and eIF4A eIF4G. BJAB cells were deprived of serum for 1 or 2 (Figure 4a) are quite stable after cycloheximide days, in the presence or absence of the broad treatment, except at long times of exposure to the speci®city caspase inhibitor Z-VAD.FMK. The in- drug. Similar e€ects of cycloheximide were obtained hibitor was able largely to prevent eIF4G degradation with cells that had been cultured in the presence of over this time period (Figure 3). The appearance of 0.1% serum, although in this case the initial level of the ca. 76 kDa product which occurred when BJAB eIF4G was already lower (data not shown). cells were incubated in 0.1% serum was also blocked We have also investigated whether eIF4G that is by Z-VAD.FMK (Figure 3), as was the cleavage of associated with eIF4E in the eIF4F initiation factor PARP (data not shown). In contrast, the levels of complex is degraded in the presence of cycloheximide. other minor cross-reacting bands which were present eIF4E and eIF4F were isolated by anity chromato- both before and after serum deprivation were not graphy on m7GTP-Sepharose and the levels of eIF4G a€ected by Z-VAD.FMK. present in eIF4F were determined by immunoblotting. Loss of eIF4G during apoptosis MJ Clemens et al 2924 As Figure 5 illustrates, eIF4G isolated by this procedure was also lost rapidly after cycloheximide treatment, with similar kinetics to the loss of total eIF4G (Figure 4). Quanti®cation of these data indicate that cleavage of eIF4G results in a decline in eIF4F levels, as shown by the marked decline in the ratio of eIF4G to eIF4E in the m7GTP-Sepharose-puri®ed material (Figure 5). As in the case of serum deprivation, the cycloheximide-induced disappearance of full-length eIF4G within 2 ± 4 h was accompanied by the appearance of a ca. 76 kDa fragment (Figure 6a, upper panel). The time-course of these changes exactly coincided with that of the cleavage of PARP, which yielded characteristic 89 kDa and 24 kDa cleavage products (lower panel). Preincuba- tion of BJAB cells with the caspase inhibitors Z- VAD.FMK or Z-DEVD.FMK protected eIF4G from cycloheximide-induced degradation and blocked the appearance of the ca. 76 kDa cleavage product (Figure 6b). However exposure of the cells to a

reported inhibitor of caspase-3, ZnCl2 (Perry et al., 1997), was without e€ect in these cells. The cleavage of PARP in the cycloheximide-treated cells was also inhibited by Z-VAD.FMK and Z-DEVD.FMK but

not by ZnCl2 (data not shown).

Is inhibition of protein synthesis sucient to destabilize Figure 3 Prevention of eIF4G cleavage in serum-deprived cells by inhibition of caspase activity. BJAB cells were incubated in eIF4G? 0.1% serum, in the presence or absence of Z-VAD.FMK (10 mM), The association of eIF4G with eIF4E in the eIF4F for the times indicated. Extracts were prepared as described in Materials and methods and the cleavage of eIF4G was monitored complex is inhibited by the eIF4E binding proteins, by immunoblotting (upper panel). Molecular weight markers are 4E-BP1 and 4E-BP2, which compete with eIF4G for indicated on the right. The arrow shows the ca. 76 kDa cleavage the cap-binding factor. The 4E-binding proteins are product. The lower panel shows the quanti®cation of the levels of reversibly phosphorylated in response to conditions full-length eIF4G obtained by scanning densitometry, expressed as percentages of the level in control cells. The experiment was which stimulate translation (Lin et al., 1994, 1995; Lin carried out three times and the error bars represent standard and Lawrence, 1996; Fadden et al., 1997; Lawrence errors of the mean and Abraham, 1997) and this process favours

Figure 4 Changes in levels of initiation factors in cycloheximide-treated BJAB cells. Exponentially growing cells were treated with cycloheximide (100 mg/ml) for the times indicated and cytoplasmic extracts prepared as described in Materials and methods. (a) Equal quantities of protein from each extract were subjected to SDS gel electrophoresis, followed by immunoblotting for eIF4G, eIF4A, eIF4E, eIF2a and 4E-BP1. (b) Kinetics of loss of eIF4G and eIF4E following treatment with cycloheximide, analysed by scanning densitometry and expressed as the level of protein remaining relative to the untreated cells. The experiment was carried out three times and the error bars represent standard errors of the mean Loss of eIF4G during apoptosis MJ Clemens et al 2925

+ + + + Figure 5 Cycloheximide induces loss of eIF4G recovered in ++ association with eIF4E. Aliquots of cell extracts prepared as in Figure 4 were subjected to anity chromatography on m7GTP- + Sepharose to purify eIF4E and associated proteins. Samples derived from equal amounts of total cytoplasmic protein were + + resolved by SDS gel electrophoresis followed by immunoblotting for eIF4G and eIF4E. The immunoblots were analysed by Figure 6 Appearance of eIF4G cleavage products in cyclohex- scanning densitometry to determine the relative levels of the imide-treated BJAB cells and inhibition by caspase inhibitors. two factors and the ratio of eIF4G:eIF4E was calculated (in Exponentially growing cells were treated with cycloheximide arbitrary units). The inset shows the blots from which these data (100 mg/ml) for the times indicated and cytoplasmic extracts were derived, which are representative of those obtained in three prepared as described in Materials and methods. (a) Equal separate experiments quantities of protein from each extract were subjected to SDS gel electrophoresis followed by immunoblotting for eIF4G and PARP. (b) BJAB cells were preincubated with or without Z- VAD.FMK (10 mM), Z-DEVD.FMK (25 mM) or ZnCl2 (100 mM) for 1 h and then incubated with or without cycloheximide dissociation of their complexes with eIF4E (Sonen- (100 mg/ml) for 3 h. Extracts were prepared as described in berg, 1996; Lawrence and Abraham, 1997). The Materials and methods and cleavage of eIF4G was monitored by immunoblotting. Molecular weight markers are indicated on the pathway leading to these events is inhibited by right of each panel and full-length eIF4G or PARP and their rapamycin, due to impairment of the phosphorylation cleavage products are indicated on the left. These data are of the 4E-BPs (Lin et al., 1995; Beretta et al., 1996a). representative of those obtained in three separate experiments In BJAB cells overall protein synthesis is particularly sensitive to rapamycin (Kay et al., 1996), suggesting that the e€ects of the drug are not limited to the translation of mRNAs with 5' terminal oligopyrimi- (Figure 7b). Rapamycin itself did not protect eIF4G dine tracts (Je€eries et al., 1994; Sonenberg, 1996) in from cleavage in cells simultaneously treated with this system. Thus it was of interest to determine cycloheximide (data not shown). These results indicate whether inhibition of protein synthesis by this agent both that inhibition of protein synthesis per se is not would also lead to increased degradation of eIF4G, as sucient to cause the loss of eIF4G and that the has recently been reported for Saccharomyces cerevi- dissociation of eIF4G from eIF4E under conditions siae (Berset et al., 1998). Figure 7a shows that in where apoptosis is induced by another agent does not BJAB cells rapamycin inhibits protein synthesis by prevent such loss. about 30% within 30 min, increasing to over 60% within 4 h. Consistent with the mode of action Induction of apoptosis by other agents correlates with described above, rapamycin caused rapid dephosphor- eIF4G cleavage ylation of 4E-BP1, together with increased association of 4E-BP1 and decreased association of eIF4G with We have used two other inducers of apoptosis in BJAB eIF4E. However there was no signi®cant loss of full- cells, viz. an agonistic antibody against the Fas antigen length eIF4G, or appearance of the ca. 76 kDa and the DNA topoisomerase inhibitor etoposide, to cleavage product, for up to 6 h after drug treatment determine whether eIF4G cleavage is a more general Loss of eIF4G during apoptosis MJ Clemens et al 2926 A

Figure 8 Induction of eIF4G cleavage by apoptotic stimuli is prevented by inhibition of caspase activity. BJAB cells were incubated with anti-Fas antibody, etoposide or cycloheximide for 24 h in the presence or absence of Z-VAD.FMK (10 mM). Extracts were prepared and cleavage of eIF4G and PARP was monitored by immunoblotting. Lanes 1 ± 3, control cells; lanes 4 and 7, anti-Fas-treated cells; lanes 5 and 8, etoposide-treated cells; lanes 6 and 9, cycloheximide-treated cells. The extracts in lanes 3, 7, 8 and 9 were from cells pretreated with Z-VAD.FMK for 1 h before induction of apoptosis. These data are representative of those obtained in three separate experiments

eIF4G, the production of a ca. 76 kDa cleavage fragment (data not shown) and the cleavage of PARP (lower panel). In addition, anti-Fas antiserum Figure 7 Inhibition of protein synthesis in response to promoted the appearance of cells with less than the rapamycin treatment of BJAB cells does not lead to degradation G1 content of DNA, indicative of apoptosis (Figure of eIF4G. Exponentially growing BJAB cells were treated with 9b). Similar results were obtained using etoposide as rapamycin (10 nM) for the times indicated. Triplicate 1 ml aliquots were incubated with [35S] methionine (15 mCi/ml) for the apoptotic inducer (data not shown). Pretreament 1 h and protein synthesis was measured as described in Materials of cells with Z-VAD.FMK prevented the cleavage of and methods. Cytoplasmic extracts were prepared and subjected eIF4G and PARP (Figure 9a, lanes 4 ± 6) and the to SDS gel electrophoresis and immunoblotting for eIF4G. (a) appearance of cells with less than the G1 content of Protein synthesis in intact cells. The data are means+s.e.m. of three measurements. (b) Immunoblot of total cytoplasmic eIF4G. DNA (Figure 9b). However, cleavage of eIF4G in These data are representative of those obtained in three separate response to these stimuli was not prevented by use of experiments the proteosome inhibitor, MG132, or the calpain inhibitor, calpeptin (data not shown). On the other hand, rapamycin treatment did not cause any detectable loss of eIF4G or cleavage of PARP under phenomenon associated with apoptosis. Treatment of these conditions (compare lanes 1 and 7; see also BJAB cells with anti-Fas antiserum (Figure 8, upper Figure 9c, lanes 1 vs 4). As in the case of BJAB cells panel; compare lanes 1 and 4), etoposide (compare induction of apoptosis also depleted the amount of lanes 2 and 5) or cycloheximide (compare lanes 3 and eIF4G associated with eIF4E (Figure 9c, lanes 5 ± 8). 6) all resulted in the cleavage of eIF4G and generation In contrast there was no signi®cant reduction in the of the ca. 76 kDa cleavage product, albeit with varying levels of eIF4E, eIF4A or poly(A)-binding protein eciency. In addition, the cleavage of PARP observed (PABP; Figure 9c, lanes 1 ± 4), or of eIF4B or eIF2a following these treatments (lower panel) followed a (data not shown) in the same total cytoplasmic similar pattern. In each instance, incubation of cells in extracts. We did, however, observe a complete loss the presence of Z-VAD.FMK prevented the cleavage of the PABP that co-puri®ed with eIF4E after anti- of eIF4G (upper panel, lanes 7 ± 9) and PARP (lower Fas treatment (Figure 9c; compare lanes 5 and 6). The panel, lanes 7 ± 9). latter result may indicate that PABP associates with eIF4G in mammalian eIF4F complexes, as is the case in yeast (Tarun and Sachs, 1996; Le et al., 1997), or Induction of apoptosis in Jurkat cells results in loss of that interaction of PABP with the eIF4G homologue eIF4G PAIP (Craig et al., 1998) is also altered in cells To determine whether conditions which induce undergoing apoptosis. The e€ects of anti-Fas treat- apoptosis have an e€ect on eIF4G levels in another ment on eIF4G and the co-puri®cation of PABP with cell type we have studied the response of the human eIF4E were blocked by Z-VAD.FMK (Figure 9c, T-cell leukaemia-derived Jurkat cell line to various lanes 2 vs 3 and lanes 6 vs 7), under conditions where inducers. Figure 9a (upper panel) shows that induction this agent prevented the apoptotic response as of apoptosis in Jurkat cells with either anti-Fas measured by PARP cleavage (Figure 9a) or the antibody (compare lanes 1 and 2) or cycloheximide appearance of cells with less than the G1 content of (lane 3) resulted in the loss of total cytoplasmic DNA (Figure 9b). Loss of eIF4G during apoptosis MJ Clemens et al 2927 ylation of factors such as eIF2, eIF4E and the 4E-BPs controls protein synthesis (reviewed in Pain, 1996). However, rather little attention has been paid to the role of acute changes in the levels of individual initiation factors in translational control, in spite of the fact that the signi®cance of the relative amounts of key factors has been a subject for keen debate (Hershey, 1994; Rau et al., 1996). In this paper we have shown that exposure of human tumour cells of both B- and T-lymphoid origin to a variety of conditions that inhibit cell proliferation and induce apoptosis causes the rapid and selective disappearance of polypeptide chain initiation factor eIF4G. Under our conditions the e€ect appears relatively speci®c for eIF4G, with no major decreases in eIF4A or eIF4E but moderate e€ects on the level of eIF2a. In this study we have not investigated possible changes in the levels of the various subunits of eIF3, but results from other laboratories have shown that eIF3 can also be degraded rapidly under certain conditions (Naranda et al., 1996). In contrast to our results, Neumar et al. (Neumar et al., 1995) have reported that eIF4E is also susceptible to rapid decay, with a half-life of 15 ± 20 min, but this was assayed under conditions of brain ischaemia. Previous ®ndings have indicated that eIF4G is susceptible to rapid cleavage and disappearance of the full length protein, not only in picornavirus- infected cells (Lloyd et al., 1988; Devaney et al., 1988; Kirchweger et al., 1994; Ziegler et al., 1995a) but also in uninfected HeLa cells following antisense RNA- mediated down-regulation of eIF4E (De Benedetti et al., 1991), in brain following ischaemia and reperfusion Figure 9 Induction of apoptosis in Jurkat cells is associated with (DeGracia et al., 1996), in erythroleukaemia cells loss of eIF4G. Jurkat cells were preincubated with or without Z- induced to di€erentiate with haemin (Benton et al., VAD.FMK (10 mM) for 1 h and then treated with anti-Fas 1996) and in yeast subjected to nutritional deprivation antibody, cycloheximide, rapamycin or solvent alone for a further 20 h. Total cytoplasmic extracts and m7GTP-Sepharose-puri®ed (Berset et al., 1998). Thus eIF4G seems particularly samples were prepared and analysed by immunoblotting for susceptible to degradation, probably by a number of eIF4G, PARP, poly(A)-binding protein (PABP) and eIF4E as di€erent proteolytic pathways depending on the cell indicated. Aliquots of the cells were also examined by type and stimulus. The stability of eIF4G may be ¯uorescence-activated cell sorting for the appearance of cells with less than the normal G1 content of DNA as a measure of determined not only by the signalling pathways by apoptosis. (a) Total cytoplasmic extracts ± lanes 1 and 4, control which cells respond to stresses but also by the other cells; lanes 2 and 6, anti-Fas-treated cells; lanes 3 and 5, proteins with which the factor associates. Binding sites cycloheximide-treated cells; lane 7, rapamycin-treated cells. The for eIF4E, eIF4A, eIF3 and PABP have been extracts in lanes 4 ± 6 were from cells pretreated with Z- characterized on eIF4G from various eukaryotic VAD.FMK. In the lower blot the speci®c cleavage product of PARP (89 kDa) is indicated. (b) FACS analysis pro®les of cells sources (Tarun and Sachs, 1996; Morley et al., 1997; incubated with or without anti-Fas in the presence or absence of Le et al., 1997; Imataka and Sonenberg, 1997). The Z-VAD.FMK. (c) Total cytoplasmic extracts (lanes 1 ± 4) and in¯uence of these proteins on eIF4G stability in vivo m7GTP-Sepharose-puri®ed samples (lanes 5 ± 8) from cells remains unknown. However the eIF4G/eIF4E complex incubated with: lanes 1 and 5, no additions; lanes 2 and 6, anti- Fas antibody; lanes 3 and 7, anti-Fas plus Z-VAD.FMK; lanes 4 is a target for direct cleavage by rhinovirus 2A protease and 8, rapamycin. These data are representative of those obtained (Haghighat et al., 1996), and we have previously in three separate experiments suggested that eIF4E may be required to allow the cleavage of eIF4G by the foot-and-mouth disease virus L-protease (Ohlmann et al., 1997). In the present study we have shown that total cytoplasmic eIF4G Discussion disappears with similar kinetics to eIF4G that is associated with eIF4E. However this does not con¯ict The initiation of protein synthesis is controlled by a with the previous observations using L-protease since variety of mechanisms that respond to many physiolo- eIF4G probably exchanges rapidly in and out of gical signals. Dysregulation of this process (for example complexes with eIF4E in vivo. In addition, preliminary by over-expression of eIF4E or eIF4G (Lazaris- data suggest that the apoptosis-induced cleavage of Karatzas et al., 1990; De Benedetti and Rhoads, eIF4G occurs at a site(s) distinct to that of L-protease 1990; Fukuchi-Shimogori et al., 1997) or by mutation since the ca. 76 kDa cleavage fragment still binds to of the phosphorylation site in eIF2a (Donze et al., eIF4E and can itself be cleaved by the viral protease 1995) can lead to aberrant growth and oncogenic (M. Bushell, unpublished data). Further characteriza- transformation. Much e€ort has been devoted to tion of the apoptosis-speci®c cleavage of eIF4G will be understanding the means by which reversible phosphor- published elsewhere. Loss of eIF4G during apoptosis MJ Clemens et al 2928 The fact that several disparate inducers of studies with rapamycin (Figure 7) suggest that the apoptosis, which probably act by di€erent mechan- inhibition of protein synthesis per se is not sucient to isms (Boix et al., 1998), all lead to the loss of eIF4G cause any cleavage of eIF4G in BJAB cells. This is in indicates that the pathways activated by these agents contrast to the situation in yeast where the factor converge on this target protein. Low serum conditions disappears within 1 ± 2 h of rapamycin treatment activate apoptosis in immortalized or transformed cell (Berset et al., 1998). Another potential mechanism by lines (Kulkarni and McCulloch, 1994; Wang and which cycloheximide might destabilize eIF4G is by Pandey, 1995; Kawanishi, 1997), probably by depriv- promoting the phosphorylation of the 4E-BPs via ing the cells of growth factors necessary to promote activation of the mTOR/p70 S6 kinase signalling cell cycle progression (Ohta et al., 1997); cyclohex- pathway (Krieg et al., 1988; Sonenberg, 1996; Brunn imide is also an ecient inducer of apoptosis in a et al., 1997; Fadden et al., 1997; Lawrence and variety of cell types, perhaps because it blocks the Abraham, 1997). This would have the e€ect of synthesis of one or more proteins essential for the dissociating eIF4E from the 4E-BPs and favouring its prevention of apoptosis (Martin et al., 1990; Collins et association with eIF4G (Graves et al., 1995; Von al., 1991; Ledda-Columbano et al., 1992; Ishii et al., Manteu€el et al., 1996; Brunn et al., 1997), thus 1997); the CD95 (Fas) receptor acts by binding a potentially altering the conformation of the latter number of proteins to the `death domain' located on (Ohlmann et al., 1997). However, to date, there is no its intracellular portion; etoposide interacts with DNA evidence that the other inducers of apoptosis used in topoisomerase II and induces DNA strand breaks. our studies impinge upon the mTOR signalling The point of convergence of these pathways at the pathway and no reason to believe that the activation level of eIF4G is most likely the activation of one or of this pathway would be a€ected by caspase more `executioner' caspases (Longthorne and Wil- inhibitors. Moreover, as indicated earlier, although liams, 1997; Kamada et al., 1997; Cohen, 1997). This rapamycin blocks the activation of the mTOR/p70 S6 is consistent with the ability of the caspase inhibitors kinase signalling pathway it is unable to prevent the (Z-VAD.FMK and Z-DEVD.FMK; Dolle et al., 1994; cleavage of eIF4G that occurs after cycloheximide Fearnhead et al., 1995; Perry et al., 1997; Juo et al., treatment. 1997), but not calpain or proteosome inhibitors (data The physiological consequences of eIF4G loss in not shown), to block the disappearance of eIF4G apoptosing cells remain to be investigated. Depending regardless of the apoptosis-inducing stimulus. on the site(s) of cleavage and the stabilities of the The evidence that eIF4G is degraded by a caspase- products, fragments may be generated that in¯uence mediated mechanism in response to a variety of the spectrum of proteins synthesized by the cell (Joshi- apoptotic stimuli suggests that the initiation factor is Barve et al., 1992). This phenomenon is well a potential candidate to be added to the growing list of established in the case of cleavage by picornaviral proteins that are cleaved by caspases (Martin and proteases, where the C-terminal cleavage product can Green, 1995; Hale et al., 1996; Lavin et al., 1996; Chen still interact with eIF3 and eIF4A and can promote the et al., 1997; Widmann et al., 1998). However the e€ects translation of internally initiated and uncapped of the inhibitors Z-VAD.FMK and Z-DEVD.FMK are mRNAs at the expense of capped mRNAs (Lloyd et not sucient to indicate which caspase activity is al., 1987; Ohlmann et al., 1995, 1996; Borman et al., responsible, or indeed to establish that these enzymes 1997; Novoa et al., 1997). However, cleavage at other are the actual catalysts of eIF4G cleavage. Both Z- sites and/or further degradation of eIF4G to smaller VAD.FMK and Z-DEVD.FMK have relatively broad fragments may result in the factor rapidly becoming speci®cities against a number of caspases (Kidd, 1998). rate-limiting for overall protein synthesis. Thus it is

ZnCl2 is an inhibitor of caspase-3 but it is not clear possible that the loss of eIF4G contributes to the whether it is also active against other caspases (Perry et marked shut-o€ of translation that occurs as apoptosis al., 1997). Although it is likely that several di€erent proceeds (Deckwerth and Johnson, 1993; Zakeri et al., caspases become activated under the conditions of our 1996; S. Morley, unpublished data).

experiments the lack of e€ect of ZnCl2 on eIF4G Finally, it is not known whether high level cleavage might suggest that caspase-3 activity is not expression of eIF4G might protect cells against required. Studies in vitro with puri®ed or recombinant apoptosis, as has been shown in the case of eIF4E enzymes will be necessary to establish whether eIF4G (Polunovsky et al., 1996). Over-expression of the factor is a direct substrate for caspases. has been reported to cause malignant transformation In addition to speci®c degradation it is possible that (Fukuchi-Shimogori et al., 1997). Intriguingly, a the synthesis of eIF4G may be severely impaired in fragment of a gene that encodes a mammalian apoptotic cells since conditions such as serum homologue of the C-terminal part of eIF4G, variously deprivation, cycloheximide, anti-Fas or etoposide called p97, DAP-5 or NAT1 (Imataka et al., 1997; treatment all inhibit overall translation. However, the Levy-Strumpf et al., 1997; Yamanaka et al., 1997), has prevention of the loss of eIF4G by the caspase been reported to protect HeLa cells against interferon- inhibitors in cycloheximide-treated cells cannot be g-induced apoptosis. The functional relationship, if attributed to any increase in synthesis of the factor any, of p97/DAP-5/NAT1 to the ca. 76 kDa eIF4G since cycloheximide completely inhibits translation cleavage fragment observed in apoptotic cells remains when present at the concentration used here (100 mg/ to be established, but preliminary data suggest that the ml). It is also unlikely that the decrease in the level of properties of the two proteins are distinct. The 76 kDa eIF4G under conditions where protein synthesis is fragment does not appear to protect cells against inhibited is merely a re¯ection of normal rapid apoptosis since it is observed only under conditions in turnover of the protein since this would not be which this process is taking place. If intact eIF4G (or a blocked by inhibiting caspase activities. Furthermore, related protein) can indeed inhibit apoptosis (perhaps Loss of eIF4G during apoptosis MJ Clemens et al 2929 by promoting the synthesis of crucial anti-apoptotic precipitated with 10% trichloroacetic acid in the presence proteins), but is also itself a target for apoptotic of 0.5 mg bovine serum albumin carrier protein. Precipi- cleavage, giving rise to the ca. 76 kDa form, the tates were washed with 5% trichloroacetic acid and progressive loss of the full-length factor might be industrial methylated spirit and the radioactivity deter- expected to lead to irreversible cellular catastrophe and mined by scintillation counting. the acceleration of cell death. Preparation of cell extracts Cytoplasmic extracts of cells were prepared for immuno- Materials and methods blotting by lysing the cells in a bu€er containing a cocktail of protease and protein phosphatase inhibitors. The Reagents composition of the lysis bu€er was: 50 mM MOPS (3-[N- Morpholino]propanesulphonic acid), pH 7.4, 50 mM NaCl, Tissue culture materials were purchased from Life 2mM EDTA, 2 mM EGTA, 50 mM b-glycerophosphate, 35 Technologies. [ S]methionine was purchased from ICN. 1 mM microcystin, 2 mM benzamidine, 2 mM Na vanadate, Cycloheximide and etoposide were obtained from Sigma- 5mM p-nitrophenylphosphate, 1 mM PMSF, 0.1 mM GTP Aldrich. Anti-Fas antibody (clone CH-11) was provided by and 7 mM 2-mercaptoethanol. The cells were washed by Upstate Biotechnology. The caspase inhibitors Z- centrifugation and resuspension twice in phosphate- VAD.FMK and Z-DEVD.FMK were purchased from bu€ered saline containing 50 mM b-glycerophosphate, and Alexis Biochemicals or Calbiochem. Rapamycin was a once in lysis bu€er without detergent. After a ®nal gift from Dr J Kay (University of Sussex, UK). Antibodies centrifugation for 1 min in a refrigerated microfuge the to the 4E-BPs and poly(A)-binding protein were kind gifts cells were drained and resuspended in lysis bu€er (5 mlper from Dr AA Thomas (University of Utrecht) and Dr D 106 cells) and Nonidet P-40 or Igepal CA-630 (Sigma) Schoenberg (Ohio State University, USA), respectively. added to a ®nal concentration of 0.2% (v/v). The Antibody to poly(ADP-ribose) polymerase (PARP) was suspensions were vortexed vigorously and centrifuged for obtained from Boehringer. A rabbit antiserum was raised 5 min in a microfuge at 48C to remove nuclei. The against the C-terminal portion of eIF4G (amino acids supernatants were rapidly frozen in small aliquots at 920 ± 1396; Pestova et al., 1996). The antibodies against 7808C. eIF4E, eIF4A and eIF2a were as described previously (Scorsone et al., 1987; Ohlmann et al., 1997). Isolation of eIF4E and associated proteins by m7GTP-Sepharose chromatography Cell culture Initiation factor eIF4E and proteins bound to it were BJAB cells are an EBV-negative cell line derived from a isolated from cell extracts by anity chromatography on human Burkitt's lymphoma. Jurkat J6 cells are a T-cell line m7GTP-Sepharose as previously described (Morley and derived from a human lymphoblastic leukaemia. Both cell Pain, 1995a,b). Brie¯y, cytoplasmic extracts were diluted lines were grown in stationary suspension culture in RPMI with lysis bu€er and mixed with m7GTP-Sepharose at 48C 1640 medium supplemented with glutamine (300 mg/l) or for 10 min. After washing twice with 250 mloflysisbu€er GlutaMAXTM (446 mg/l) (Life Technologies), 10% (v/v) the proteins were eluted from the resin with SDS sample foetal calf serum, penicillin (100 units/ml) and streptomy- bu€er. cin (100 mg/ml), in a 5% CO2 atmosphere at 378C. Cells were maintained at densities between 2.5 and 86105 cells per ml. Immunoblotting and densitometry Cell extracts and m7GTP-Sepharose-puri®ed fractions were Induction and assay of apoptosis subjected to electrophoresis on 7.5%, 10% or 15% polyacrylamide SDS gels (according to the size of the Cell growth arrest and apoptosis were induced in BJAB proteins to be analysed) and transferred to PVDF cells by replacing the growth medium with an otherwise membranes (Millipore) using a semi-dry blotting appara- identical medium containing only 0.1% (v/v) foetal calf tus (Hoefer). In each experiment equal amounts of protein serum for up to 2 days, or by adding anti-Fas antibody were loaded in each lane of the gel (within the range 3 ± (150 ng/ml), etoposide (100 mg/ml) or cycloheximide 10 mg protein per sample in the case of total cytoplasmic (100 mg/ml), as described in the ®gure legends. Apoptosis extracts, depending on the sensitivity of the antibody to be in Jurkat cells was induced by similar treatment with used). Fractions puri®ed from equal amounts of cell extract cycloheximide or anti-Fas antibody. Where indicated cells (ca. 40 mg protein) by anity chromatography on m7GTP- were preincubated with agents previously shown to inhibit Sepharose were analysed similarly. Blots were blocked, caspase activity, viz. Z-VAD.FMK, Z-DEVD.FMK or incubated with the appropriate primary antibodies (see ZnCl2, before exposure to the apoptosis-inducing condi- Figure legends) and developed using alkaline phosphatase- tions. The progress of apoptosis was monitored in intact linked secondary antibodies with NBT as the substrate, cells by the cleavage of the caspase substrate PARP under standard conditions (Morley and Pain, 1995a). The (Kaufmann et al., 1993) and/or by the appearance of cells blots were analysed by densitometry using a scanning laser with less than the G1 content of DNA (analysed by Personal Densitometer SI (Molecular Dynamics) employ- ¯uorescence-activated cell sorting; Milner et al., 1995). ing ImageQuant software or a Sharp JX-325 scanner with Image Master software (Pharmacia Biotech). In all cases, Measurement of protein synthesis detection was in the linear range of the antiserum employed for decoration of resolved proteins. The overall rate of protein synthesis in intact cells was measured by the incorporation of [35S]methionine into trichloroacetic acid-insoluble material. After pulse-labelling for 1 h with up to 15 mCi/ml of the radioactive Acknowledgements (in the presence of the normal level of methionine in the This work was initiated during a period of sabbatical leave cell culture medium), cells were brie¯y centrifuged, washed granted to MJC by St George's Hospital Medical School. once in cold phosphate-bu€ered saline containing 100 mM We are grateful to all members of the VM Pain and SJ unlabelled methionine, dissolved in 0.3 M NaOH and Morley research groups for valuable discussions. The Loss of eIF4G during apoptosis MJ Clemens et al 2930 research was supported by grants from The Royal Society, Industrial CASE Studentship in collaboration with Roche the Wellcome Trust, the Cancer Prevention Research Trust Discovery (Welwyn Garden City) and SJM is a Senior and the Leukaemia Research Fund. MB is funded by an Research Fellow of the Wellcome Trust.

References

Altmann M, Schmitz N, Berset C and Trachsel H. (1997). Haghighat A, Mader S, Pause A and Sonenberg N. (1995). EMBO J., 16, 1114 ± 1121. EMBO J., 14, 5701 ± 5709. Barbet NC, Schneider U, Helliwell SB, Stans®eld I, Tuite HaghighatA,SvitkinY,NovoaI,KuechlerE,SkernTand MF and Hall MN. (1996). Mol. Biol. Cell, 7, 25 ± 42. Sonenberg N. (1996). J. Virol., 70, 8444 ± 8450. Benton PA, Barrett DJ, Matts RL and Lloyd RE. (1996). J. Haghighat A and Sonenberg N. (1997). J. Biol. Chem., 272, Virol., 70, 5525 ± 5532. 21677 ± 21680. Beretta L, Gingras AC, Svitkin YV, Hall MN and Sonenberg Hale AJ, Smith CA, Sutherland LC, Stoneman VEA, N. (1996a). EMBO J., 15, 658 ± 664. Longthorne VL, Culhane AC and Williams GT. (1996). Beretta L, Svitkin YV and Sonenberg N. (1996b). J. Virol., Eur. J. Biochem., 236, 1±26. 70, 8993 ± 8996. Hershey JWB. (1994). Biochimie, 76, 847 ± 852. Berset C, Trachsel H and Altmann M. (1998). Proc. Natl. Imataka H, Olsen HS and Sonenberg N. (1997). EMBO J., Acad. Sci. USA, 95, 4264 ± 4269. 16, 817 ± 825. Boix J, Fibla J, Yuste VJ, Piulats JM, Llecha N and Comella Imataka H and Sonenberg N. (1997). Mol. Cell. Biol., 17, JX. (1998). Exp. Cell Res., 238, 422 ± 429. 6940 ± 6947. Borman AM, Kirchweger R, Ziegler E, Rhoads RE, Skern T Ishii H, Tamauchi H and Gob GC. (1997). Int. J. Exp. and Kean KM. (1997). RNA, 3, 186 ± 196. Pathol., 78, 123 ± 131. Brass N, Heckel D, Sahin U, Pfreundschuh M, Sybrecht GW Ishii HH, Etheridge MR and Gob GC. (1995). Immunol. Cell and Meese E. (1997). Hum. Mol. Genet., 6, 33 ± 39. Biol., 73, 463 ± 468. Brunn GJ, Hudson CC, Sekulic A, Williams JM, Hosoi H, Je€eries HBJ, Reinhard C, Kozma SC and Thomas G. Houghton PJ, Lawrence Jr JC and Abraham RT. (1997). (1994). Proc. Natl. Acad. Sci. USA, 91, 4441 ± 4445. Science, 277, 99 ± 101. Joshi B, Cai A-L, Keiper BD, Minich WB, Mendez R, Beach Chen LH, Marechal V, Moreau J, Levine AJ and Chen JD. CM, Stepinski J, Stolarski R, Darzynkiewicz E and (1997). J. Biol. Chem., 272, 22966 ± 22973. Rhoads RE. (1995). J. Biol. Chem., 270, 14597 ± 14603. Cohen GM. (1997). Biochem. J., 326, 1 ± 16. Joshi-Barve S, De Benedetti A and Rhoads RE. (1992). J. Collins RJ, Harmon BV, Souvlis T, Pope JH and Kerr JFR. Biol. Chem., 267, 21038 ± 21043. (1991). Br. J. Cancer, 64, 518 ± 522. Juo P, Kuo CJ, Reynolds SE, Konz RF, Raingeaud J, Davis Craig AWB, Haghighat A, Yu ATK and Sonenberg N. RJ, Biemann H-P and Blenis J. (1997). Mol. Cell. Biol., 17, (1998). Nature, 392, 520 ± 523. 24 ± 35. De Benedetti A, Joshi-Barve S, Rinker-Schae€er C and Kamada S, Washida M, Hasegawa J, Kusano H, Funahashi Rhoads RE. (1991). Mol. Cell. Biol., 11, 5435 ± 5445. Y and Tsujimoto Y. (1997). Oncogene, 15, 285-290. De Benedetti A and Rhoads RE. (1990). Proc. Natl. Acad. Kaufmann SH, Desnoyers S, Ottaviano Y, Davidson NE and Sci. USA, 87, 8212 ± 8216. Poirier GG. (1993). Cancer Res., 53, 3976 ± 3985. Deckwerth TL and Johnson EM. (1993). J. Cell Biol., 123, Kawanishi M. (1997). Virology, 228, 244 ± 250. 1207 ± 1222. Kay JE, Smith MC, Frost V and Morgan GY. (1996). DeGracia DJ, Neumar RW, White BC and Krause GS. Immunol., 87, 390 ± 395. (1996). J. Neurochem., 67, 2005 ± 2012. Kidd VJ (1998). Ann. Rev. Physiol., 60, 533 ± 573. Devaney MA, Vakharia VN, Lloyd RE, Ehrenfeld E and Kirchweger R, Ziegler E, Lamphear BJ, Waters D, Liebig H- Grubman MJ. (1988). J. Virol., 62, 4407 ± 4409. D, Sommergruber W, Sobrino F, Hohenadl C, Blaas D, Dolle RE, Hoyer D, Prasad CVC, Schmidt SJ, Helaszek CT, Rhoads RE and Skern T. (1994). J. Virol., 68, 5677 ± 5684. Miller RE and Ator MA. (1994). J. Med. Chem., 37, 563 ± Koromilas AE, Lazaris-Karatzas A and Sonenberg N. 564. (1992). EMBO J., 11, 4153 ± 4158. Donz O, Jagus R, Koromilas AE, Hershey JWB and Krieg J, Hofsteenge J and Thomas G. (1988). J. Biol. Chem., Sonenberg N. (1995). EMBO J., 14, 3828-3834. 263, 11473 ± 11477. Duriez PJ and Shah GM. (1997). Biochem. Cell Biol., 75, Kulkarni GV and McCulloch CA. (1994). J. Cell Sci., 107, 337 ± 349. 1169 ± 1179. Enari M, Talanian RV, Wong WW and Nagata S. (1996). Lamphear BJ, Yan R, Yang F, Waters D, Liebig H-D, Nature, 380, 723 ± 726. Klump H, Kuechler E, Skern T and Rhoads RE. (1993). J. Fadden P, Haystead TAJ and Lawrence Jr JC. (1997). J. Biol. Chem., 268, 19200 ± 19203. Biol. Chem., 272, 10240 ± 10247. Lamphear BJ, Kirchweger R, Skern T and Rhoads RE. Fearnhead HO, Dinsdale D and Cohen GM. (1995). FEBS (1995). J. Biol. Chem., 270, 21975 ± 21983. Lett., 375, 283 ± 288. Lavin MF, Watters D and Song Q. (1996). Experientia, 52, Flynn A and Proud CG. (1995). J. Biol. Chem., 270, 21684 ± 979 ± 994. 21688. Lawrence JC and Abraham RT. (1997). Trends Biochem. Fukuchi-Shimogori T, Ishii I, Kashiwagi K, Mashiba H, Sci., 22, 345 ± 349. Ekimoto H and Igarashi K. (1997). Cancer Res., 57, Lazaris-Karatzas A, Montine KS and Sonenberg N. (1990). 5041 ± 5044. Nature, 345, 544 ± 547. GradiA,ImatakaH,SvitkinYV,RomE,RaughtB,Morino Le H, Tanguay RL, Balasta ML, Wei CC, Browning KS, S and Sonenberg N. (1998). Mol. Cell. Biol., 18, 33 Ð 342. Metz AM, Goss DJ and Gallie DR. (1997). J. Biol. Chem., Gra€ JR, Boghaert ER, De Benedetti A, Tudor DL, Zimmer 272, 16247 ± 16255. CC, Chan SK and Zimmer SG. (1995). Int. J. Cancer, 60, Ledda-Columbano GM, Coni P, Faa G, Manenti G and 255 ± 263. Columbano A. (1992). Am.J.Path.,140, 545 ± 549. Graves LM, Bornfeldt KE, Argast GM, Krebs EG, Kong X, Levy-Strumpf N, Deiss LP, Berissi H and Kimchi A. (1997). Lin TA and Lawrence Jr JC. (1995). Proc. Natl. Acad. Sci. Mol. Cell. Biol., 17, 1615 ± 1625. USA, 92, 7222 ± 7226. Loss of eIF4G during apoptosis MJ Clemens et al 2931 Li BDL, Liu L, Dawson M and De Benedetti A. (1997). Ohlmann T, Rau M, Morley SJ and Pain VM. (1995). Nucleic Cancer, 79, 2385 ± 2390. Acids Res., 23, 334 ± 340. Lin T-A, Kong X, Haystead TAJ, Pause A, Belsham G, Ohlmann T, Rau M, Pain VM and Morley SJ. (1996). EMBO Sonenberg N and Lawrence Jr JC. (1994). Science, 266, J., 15, 1371 ± 1382. 653 ± 656. Ohlmann T, Pain VM, Wood W, Rau M and Morley SJ. Lin T-A, Kong X, Saltiel AR, Blackshear PJ and Lawrence (1997). EMBO J., 16, 844 ± 855. Jr JC. (1995). J. Biol. Chem., 270, 18531 ± 18538. Ohta T, Kinoshita T, Naito M, Nozaki T, Masutani M, Lin T-A and Lawrence JC,Jr. (1996). J. Biol. Chem., 271, Tsuruo T and Miyajima A. (1997). J. Biol. Chem., 272, 30199 ± 30204. 23111 ± 23116. Lloyd RE, Jense HG and Ehrenfeld E. (1987). J. Virol., 61, Pain VM. (1996). Eur. J. Biochem., 236, 747 ± 771. 2480 ± 2488. Perry DK, Smyth MJ, Stennicke HR, Salvesen GS, Duriez P, Lloyd RE, Grubman M and Ehrenfeld E. (1988). J. Virol., Poirier GG and Hannun YA. (1997). J. Biol. Chem., 272, 62, 421 Ð 4223. 18530 ± 18533. Longthorne VL and Williams GT. (1997). EMBO J., 16, Pestova TV, Shatsky IN and Hellen CUT. (1996). Mol. Cell. 3805 ± 3812. Biol., 16, 6870 ± 6878. Mader S, Lee H, Pause A and Sonenberg N. (1995). Mol. Polunovsky VA, Rosenwald IB, Tan AT, White J, Chiang L, Cell. Biol., 15, 4990 ± 4997. Sonenberg N and Bitterman PB. (1996). Mol. Cell. Biol., Mao X, Green JM, Safer B, Lindsten T, Frederickson RM, 16, 6573 ± 6581. Miyamoto S, Sonenberg N and Thompson CB. (1992). J. Rau M, Ohlmann T, Morley SJ and Pain VM. (1996). J. Biol. Biol. Chem., 267, 20444 ± 20450. Chem., 271, 8983 ± 8990. Martin SJ, Lennon SV, Bonham AM and Cotter TG. (1990). Rosenwald IB, Rhoads DB, Callanan LD, Isselbacher KJ J. Immunol., 145, 1859 ± 1867. and Schmidt EV. (1993). Proc. Natl. Acad. Sci. USA, 90, Martin SJ and Green DR. (1995). Cell, 82, 349 ± 352. 6175 ± 6178. Matsuo H, Li HJ, McGuire AM, Fletcher CM, Gingras AC, Rowan S and Fisher DE. (1997). Leukemia, 11, 457 ± 465. Sonenberg N and Wagner G. (1997). Nature Struct. Biol., Salvesen GS and Dixit VM. (1997). Cell, 91, 443 ± 446. 4, 717-724. Scorsone KA, Panniers R, Rowlands AG and Henshaw EC. Merrick WC. (1994). Biochimie, 76, 822 ± 830. (1987). J. Biol. Chem., 262, 14538 ± 14543. Merrick, W.C. and Hershey, J.W.B. Translational Control. Sinha BK. (1995). Drugs, 49, 11 ± 19. Hershey JWB, Mathews MB and Sonenberg N. (eds). Sonenberg N. Translational Control. Hershey JWB, Math- Cold Spring Harbor, New York: Cold Spring Harbor ews MB and Sonenberg N. (eds). Cold Spring Harbor, Laboratory Press, 1996, pp. 31 ± 70. NY: Cold Spring Harbor Laboratory, pp. 245 ± 269. Milner, A.E., Wang, H. and Gregory, C.D. Applications of Tarun Jr SZ, Wells SE, Deardor€ JA and Sachs AB. (1997). ¯ow cytometry in biotechnology. Al-Rubai M (ed). New Proc. Natl. Acad. Sci. USA, 94, 9046 ± 9051. York: Marcel Dekker, 1995, pp. 56 ± 76. Tarun Jr SZ and Sachs AB. (1996). EMBO J., 15, 7168 ± MiyagiY,SugiyamaA,AsaiA,OkazakiT,KuchinoYand 7177. Kerr SJ. (1995). Cancer Lett., 91, 247 ± 252. Tewari M, Quan LT, O'Rourke K, Desnoyers S, Zeng Z, Morley SJ, Curtis PS and Pain VM. (1997). RNA 3, 1085 ± Beidler DR, Poirier GG, Salvesen GS and Dixit VM. 1104. (1995). Cell, 81, 801 ± 809. Morley SJ and Pain VM. (1995a). J. Cell Sci., 108, 1751 ± Thomas G and Hall MN. (1997). Curr. Opin. Cell Biol., 9, 1760. 782 ± 787. Morley SJ and Pain VM. (1995b). Biochem. J., 312, 627 ± Von Manteu€el SR, Gingras AC, Ming XF, Sonenberg N 635. and Thomas G. (1996). Proc. Natl. Acad. Sci. USA, 93, Nagata S. (1997). Cell, 88, 355 ± 365. 4076 ± 4080. Naranda T, Kainuma M, MacMillan SE and Hershey JWB. Wang E and Pandey S. (1995). J. Cell. Physiol., 163, 155 ± (1997). Mol. Cell. Biol., 17, 145 ± 153. 163. Nathan CA, Carter P, Liu L, Li BD, Abreo F, Tudor A, Widmann C, Gibson S and Johnson GL. (1998). J. Biol. Zimmer SG and De Benedetti A. (1997a). Oncogene, 15, Chem., 273, 7141 ± 7147. 1087 ± 1094. Yamanaka S, Poksay KS, Arnold KS and Innerarity TL. Nathan CAO, Liu L, Li BD, Abreo FW, Nandy I and De (1997). Genes and Dev., 11, 321 ± 333. Benedetti A. (1997b). Oncogene, 15, 579 ± 584. Yan R, Rychlik W, Etchison D and Rhoads RE. (1992). J. Neumar RW, DeGracia DJ, White BC, McDermott PJ, Biol. Chem., 267, 23226 ± 23231. Evans DR and Krause GS. (1995). J. Neurochem., 65, Zakeri Z, Quaglino D, Latham T, Woo K and Lockshin RA. 1391 ± 1394. (1996). Microsc. Res. Tech., 34, 192 ± 201. Novoa I, Martnez-Abarca F, Fortes P, Ortn J and Carrasco Ziegler E, Borman AM, Deliat FG, Liebig HD, Jugovic D, L. (1997). Biochemistry, 36, 7802 ± 7809. Kean KM, Skern T and Kuechler E. (1995a). Virology, Oehm A, Behrmann I, Falk M, Pawlita M, Maier G, Klas C, 213, 549 ± 557. Li-Weber M, Richards S, Dhein J, Trauth BC, Ponsting H Ziegler E, Borman AM, Kirchweger R, Skern T and Kean and Krammer PH. (1992). J. Biol. Chem., 267, 10709. KM. (1995b). J. Virol., 69, 3465 ± 3474.