INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Apoptosis, or programmed cell death, provides a natural and irreversible mechanism to remove damaged cells from tissue. Apoptosis is characterized by a series of cellular events which include cell shrinkage, membrane blebbing, chromatin condensation, and DNA fragmentation. The apoptotic process is divided into two phases: a long induction phase of quite variable length in which various types of signaling lead to activation of several "death proteins," and the more rapid execution phase in which degradation of certain cellular proteins occurs (10), which ultimately results in cell death. Over the last few years it has become clear that a family of unique cysteine proteases, collectively called caspases (1), are activated during the induction of apoptosis. These caspases carry out very specific cleavage events that most often will lead to loss of function of the target protein. A rapidly growing list of cellular caspase substrates is emerging which includes poly(ADP-ribose) polymerase (PARP), fodrin, lamins, the retinoblastoma tumor suppressor protein, and many others (10). Although activated caspases are thought to be the executioners of cell death, the biological relevance of many of these cleavage events is unclear, and it is still unknown how the execution phase can lead to such rapid cell death.

Little is known about the role of translational control during apoptosis. Since the components of the execution machinery of apoptosis, i.e., the caspases, are already present in the cell at induction, de novo protein synthesis is not required for induction of apoptosis in most systems (29, 33, 36, 38). On the other hand, it has been reported that in some cases, protein synthesis was necessary to induce apoptosis when nutrient deprivation or sterols were used to induce apoptosis (8, 9, 34). In these systems, expression of certain factors was necessary for induction of apoptosis, and apoptosis could be repressed by addition of translation inhibitors such as cycloheximide (8). However, even less is known about translation control events during the execution phase of apoptosis, other than the observation in several systems that protein synthesis is eventually inhibited in apoptotic cells (35).

Eukaryotic initiation factor 4F (eIF4F) is required to bind the vast majority of capped cellular mRNAs to ribosomes during the initiation step of translation. This factor contains three subunits: eIF4E, which specifically binds the 5' cap structure (m⁷GTP) present on cellular mRNAs; eIF4A, which is an ATP-dependent helicase; and eIF4G (formerly called p220), which functions as a molecular scaffold by simultaneously binding eIF4E, eIF4A, and eIF3, thus enabling mRNA to bind to ribosomes (22, 31). Several picornaviruses are known to cause shutoff of host cell translation via specific cleavage of eIF4G, thus disrupting the eIF4F complex and abolishing the ability of capped mRNA to bind ribosomes (2, 11, 13, 18, 27). Poliovirus (PV)-infected HeLa cells are the most thoroughly characterized of these viral systems. PV infection results in several morphological changes in the cell, and other proteins are cleaved in addition to eIF4G; however, it has long been thought that the rapid and complete shutoff of cap-dependent translation is one of the prime mechanisms which leads to rapid cell death and lysis. This is supported by the characterization of PV infections in a panel of K562 cell strains in which the ability of the virus to lyse cells instead of producing a persistent infection was linked to the ability to cause complete host translation shutoff (3, 4).

Since cleavage of eIF4G in viral infection results in rapid translation shutoff, and this is closely linked to cell death, we reasoned that eIF4G may be a prime target of caspases during apoptosis. Here, we show that eIF4G is targeted for cleavage during apoptosis, that eIF4G is a substrate for caspase 3, and that this cleavage event in vivo correlates with a drastic inhibition of cellular translation. This is the first report describing the cleavage of a translation initiation factor in apoptosis, and we propose that rapid and drastic inhibition of protein synthesis is a major mechanism of the execution phase of apoptosis, which leads to cell death.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Cell culture. HeLa S3 cells were grown at 37°C in S-MEM (minimal essential medium; Irvine Scientific) supplemented with 10% bovine calf serum, 0.5% fetal calf serum (Summit Biotech.), 100 U of penicillin, and 100 µg of streptomycin per ml (Sigma) in a humidified chamber containing 5% CO₂. For the induction of apoptosis, various concentrations of cisplatin (Aldrich) or etoposide (Sigma) stock solution were diluted with S-MEM and then incubated with cells at 37°C for the time indicated in each figure. For in vivo inhibition experiments, cells were first preincubated with the cell-permeable caspase inhibitor Z-VAD-fmk (Enzyme Systems Products) at the indicated concentrations for 1 h at 37°C before the cisplatin was added to the cell cultures.

Preparation of HeLa cell extracts. HeLa cells were washed with phosphate-buffered saline (PBS), resuspended in CHAPS {3[(3-cholamidopropyl)-dimethyl-ammonio]-1-propanesulfonate} lysis buffer (20 mM Tris [pH 7.2], 0.1 M NaCl, 1 mM EDTA, 10 mM dithiothreitol [DTT], 0.5% CHAPS, 10% sucrose) and incubated on ice for 30 min. Alternatively, for kinetics experiments (see Fig. 2), cells were resuspended in lysis buffer (10 mM KCl, 2.5 mM DTT, 1.2 mM MgAc₂, 20 mM HEPES [pH 7.4]), incubated on ice for 20 min, and lysed with 60 strokes in a Dounce homogenizer (Wheaton). In both cases, cell lysates were then centrifuged for 6 min at 10,000 × g at 4°C, and supernatants were collected and stored at 80°C.

Metabolic labeling of proteins. After treatment of the cells with apoptosis inducers as indicated in the figures, the medium was replaced with methionine-depleted medium (1 ml), and cells were pulse-labeled with 27 µCi of Tran³⁵S (ICN) for 1 h at selected time points. Cytoplasmic extracts were prepared as described above with CHAPS lysis buffer and then analyzed (50 µg of protein) by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and autoradiography with Kodak Biomax MR film.

In vitro eIF4G cleavage assays. Lysates (50 µg of protein) from apoptotic cell cultures were incubated with 2 µl of ribosomal salt wash (U-RSW), which was prepared as described previously (7), for 18 h at 37°C (see Fig. 3A). eIF4G cleavage was determined by separation of proteins on SDS-PAGE (7% polyacrylamide) gels and subsequent immunoblot analysis with polyclonal N-terminal eIF4G-specific antiserum (27). For inhibitor studies (Table 1), lysates were first preincubated on ice for 10 min with various protease inhibitors, and then U-RSW was added and eIF4G cleavage was assessed as described above. The percent inhibition of cleavage was determined by quantitation of scanned immunoblots with NIH Image software. eIF4G cleavage by caspase 3 (see Fig. 6 to 8) was determined by incubation of either U-RSW (2 µl) or purified eIF4F (see below) with purified caspase 3 (see below) for 3 h at 37°C. The samples were then analyzed by immunoblotting with polyclonal antisera specific for either the N-terminal (27) or C-terminal (15) domains of eIF4G or by Coomassie staining.

Detection of PARP cleavage in apoptotic cells. HeLa cells treated with 100 µM cisplatin were washed with PBS and resuspended in ice-cold radioimmunoprecipitation assay (RIPA) buffer (PBS supplemented with 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS) with freshly added protease inhibitors (1 mM PMSF, 1 mM aprotinin). After incubation on ice for 30 min, cells were further disrupted by Dounce homogenization and centrifuged at 14,000 × g for 20 min at 4°C, and supernatants were collected (whole-cell extracts). Whole-cell extracts (50 µg protein) were analyzed for PARP cleavage on SDS-PAGE (10% polyacrylamide) gels and subsequent immunoblot analysis with a mouse monoclonal anti-PARP antibody (Zymed Labs). Blots were developed by Enhanced Chemiluminescence (ECL system) (Pierce).

Caspase assays. Cell lysates were analyzed for caspase activity with one of the colorimetric peptide substrates Ac-DEVD-pNA (acetyl-DEVD-para-nitroanilide), Ac-YVAD-pNA (Quality Controlled Biochemicals), and Ac-IETD-pNA (Biomol). Assay mixtures (0.1 ml) contained 20 µg of total protein from samples indicated in the figure legends and 0.2 mM pNA substrate (final concentration). Samples were incubated for 2 h at 37°C, and release of pNA was monitored at 405 nm with a Beckman DU-70 UV spectrophotometer.

Cell-free induction of apoptosis. HeLa S100 lysates were incubated with 400 nM cytochrome c (Sigma) and 1 mM dATP (Boehringer Mannheim) at 37°C for the time indicated in the figure legends. Incubated lysates were analyzed for eIF4G cleavage on SDS-PAGE (7% polyacrylamide) gels and subsequent immunoblot analysis as described above. For caspase 3 blots, samples were subjected to SDS-PAGE (13% polyacrylamide) gels, blotted onto nitrocellulose, and analyzed with anti-caspase 3 antibody (Santa Cruz Biotechnology) according to the manufacturer's protocol.

Expression and purification of caspase 3. The full-length cDNA encoding human caspase 3 (a kind gift from C. Vincenz) cloned into pET23b (Novagen) was expressed in Escherichia coli BL21(DE3)pLysS. The expressed protein was purified by affinity chromatography on TALON metal affinity resin (Clontech) according to the manufacturer's instructions.

Purification of eIF4F. eIF4F was purified from HeLa cells as described previously (23). In short, U-RSW (2 ml), as prepared above, was loaded on a 15 to 30% sucrose gradient in buffer F (20 mM HEPES [pH 7.6], 0.5 M KCl, 0.5 mM EDTA, 2 mM -mercaptoethanol) and centrifuged for 18 h at 3°C in a Beckman SW28 rotor. Fractions (1 ml) were collected and analyzed by immunoblotting for the presence of eIF4F. Fractions containing eIF4F were pooled, diluted fourfold with buffer A (20 mM MOPS [pH 7.6], 0.25 mM DTT, 0.1 mM EDTA, 50 mM NaF, 10% glycerol), and loaded onto a 7-methyl GTP-Sepharose 4B column (Pharmacia) equilibrated in buffer B110 (buffer A plus 110 mM KCl). eIF4F was eluted from the column with buffer B110 plus m⁷GTP (70 µM), and fractions were analyzed by Coomassie staining.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Cleavage of eIF4G occurs in a dose-dependent manner and correlates with the number of apoptotic cells. The antitumor agents cisplatin and etoposide are commonly used to induce apoptosis in a variety of cell types. We determined by flow cytometry with annexin V-Fluorescein isothiocyanate (Caltag) that in HeLa cells treated for 24 h at a concentration of 100 µM cisplatin or 50 µM etoposide, over 90% of the cells were apoptotic, as judged by positive annexin binding (data not shown). This was further confirmed by another apoptosis detection method, terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) (Boehringer Mannheim), showing that 95% of the treated cells were stained positive, indicative of DNA fragmentation, another hallmark described for apoptotic cells (data not shown). Thus, both concentrations of cisplatin and etoposide were effective in inducing apoptosis in greater than 90% of the HeLa cell cultures used in this study.

View larger version (55K): [in this window] [in a new window]   FIG. 1.   Dose-dependent cleavage of eIF4G during apoptosis correlates with translation shutoff. (A) HeLa cells were treated with various concentrations of cisplatin in S-MEM for 2 h and then incubated for 22 h with normal medium at 37°C. Cell lysates (50 µg of protein) were loaded on a 7% acrylamide gel and immunoblotted with polyclonal antisera specific for N-terminal eIF4G. Lane U, 2 µl of U-RSW from untreated cells; lane I, 4 µl of RSW from PV-infected cells (I-RSW). eIF4G and N-terminal eIF4G cleavage products (eIF4GcpN) are indicated on the right. (B) Twenty-four hours after treatment with cisplatin, HeLa cells were pulse-labeled with [35S]methionine for 1 h, lysed, and analyzed (50 µg of protein) on a 10% acrylamide gel. The concentrations of cisplatin used to treat the HeLa cells are indicated at the top. (C) HeLa cells were treated with increasing concentrations of etoposide as indicated at the top for 24 h at 37°C. Cell lysates (50 µg of protein) were analyzed as in panel A. The immunoblots in panels A and C and the autoradiograph in panel B were scanned with an Artec Viewstation A6000C, and the resulting images were labeled with Adobe Photoshop version 3.0.

eIF4G cleavage correlates with inhibition of cellular translation. Cleavage of eIF4G during PV infection has been clearly shown to be the major cause of shutoff of cellular translation (13, 17, 31). Therefore, we investigated whether the cleavage of eIF4G during apoptosis also correlated with inhibition of translation. Figure 1B shows that treatment with 50 µM cisplatin resulted in approximately 50% reduction of [³⁵S]methionine incorporation into protein, and higher doses of cisplatin resulted in both complete eIF4G cleavage and drastic inhibition of protein synthesis. The inhibition of translation observed correlates well with the degree of eIF4G cleavage observed, and it was similar to previous results in PV-infected HeLa cells, where greater than 80% cleavage was required to cause significant host translation shutoff (4, 13). Similar drastic inhibition of translation occurred when cells were treated with etoposide and again correlated with greater than 80% eIF4G cleavage (data not shown). In summary, the apoptosis inducers cisplatin and etoposide cause a dose-dependent cleavage of eIF4G which correlates well with a dose-dependent inhibition of cellular translation.

Kinetics of eIF4G cleavage during induction of apoptosis in HeLa cells treated with cisplatin. To determine the kinetics of eIF4G cleavage and cellular translation inhibition, HeLa cells were treated with 100 µM cisplatin, and aliquots were harvested over a 24-h period (Fig. 2A). Low levels of eIF4G cleavage products could be detected in control cell lysates, probably due to a small population of apoptotic cells in control cultures which were consistently detected with annexin staining (data not shown). However, as early as 4 h after administration of cisplatin (100 µM), significant new eIF4G cleavage products were detectable in lysates, and cleavage of eIF4G was complete by 16 h after drug treatment. Close examination also reveals that three sets of cleavage products appear, including a unique large-molecular-weight class (arrowhead), transient cleavage products which comigrate with those produced by PV infection (arrow), and the novel fast-migrating cleavage products (eIF4GcpN) shown in Fig. 1. In repeat experiments, the slower-migrating forms of eIF4GcpN consistently appeared transiently between 4 and 8 h, whereas the fastest-migrating forms eIF4GcpN were present at all times after induction of apoptosis and accumulated. Figure 2B shows the amounts of protein synthesis at various time points after addition of cisplatin and clearly indicates that the kinetics of translation inhibition correlates well with the decrease in levels of intact eIF4G. By 8 h after induction of apoptosis, protein synthesis significantly decreased, and it was almost completely abolished by 16 h.

View larger version (50K): [in this window] [in a new window]   FIG. 2.   Kinetics of eIF4G cleavage during apoptosis correlate with kinetics of translation shutoff. (A) HeLa cells were treated with 100 µM cisplatin and incubated for the indicated time periods. Cells were lysed and proteins were immunoblotted as described for Fig. 1. eIF4G and dominant eIF4G cleavage products (eIF4Gcp) are indicated on the right. Lane U contains 2 µl of U-RSW; lane I contains 4 µl of I-RSW. The arrow and arrowhead denote alternate eIF4G cleavage products. (B) After treatment of the cells with 100 µM cisplatin for the indicated time periods, cells were labeled with [35S]methionine for 30 min at 37°C. The cells were then lysed and analyzed (50 µg of protein) by autoradiography. (C) HeLa cells were treated with 100 µM cisplatin and incubated for the indicated time periods. Cells were lysed with RIPA buffer, and proteins were immunoblotted with monoclonal anti-PARP antibody. The molecular masses of intact (116 kDa) and cleaved PARP (85 kDa) are indicated on the right. The immunoblots in panels A and C and the autoradiograph in panel B were scanned with an Artec Viewstation A6000C, and the resulting images were labeled with Adobe Photoshop version 3.0.

Temporal activation of eIF4G cleavage activity during apoptosis. To determine whether an activity that catalyzed eIF4G cleavage could be detected in an in vitro assay, apoptotic lysates (50 µg protein) from HeLa cells treated with cisplatin (100 µM) were incubated with 2 µl of U-RSW, a HeLa cell fraction enriched for translation initiation factors (including eIF4G), and incubated for 18 h at 37°C. Figure 3A, lanes U and C, shows the substrate (eIF4G present in U-RSW) before and after incubation at 37°C for 18 h, respectively. No cleavage of eIF4G was detected in vitro upon incubation. When incubated with lysates from apoptotic cells, an eIF4G cleavage activity was detected in lysates from cells treated for only 2 h, which generated the fastest-migrating form of eIF4GcpN observed previously. The activity was most prevalent in lysates derived from a 4- to 16-h treatment, causing complete cleavage of intact eIF4G during the assay. This is in agreement with Fig. 2, which shows that cleavage products of eIF4G can be detected as early as 4 h after induction of apoptosis. Interestingly, the eIF4G-specific cleavage activity diminishes by 16 to 24 h, perhaps due to the instability of the protease(s) involved during the assay. We also noticed that overall levels of cleavage activity were usually weaker in cell lysates than expected, which required longer than 8 h of incubation to cause complete eIF4G cleavage in vitro. The reason for this is unknown, but it may involve rapid turnover of proteases after cell lysis or binding of inhibitors. Generation of the slower-migrating eIF4G cleavage products shown in Fig. 2 was observed only inconsistently in in vitro assays, generally in lysates from earlier time points (2 h [Fig. 3A]).

View larger version (32K): [in this window] [in a new window]   FIG. 3.   Detection of eIF4G-specific-protease and caspase activities in lysates from apoptotic HeLa cells. (A) In vitro eIF4G cleavage assay. HeLa cell lysates (50 µg of protein) derived from cell cultures treated with cisplatin (100 µM) were incubated with 2 µl of U-RSW for 18 h at 37°C, and samples were then analyzed by immunoblotting. Lane U, U-RSW; lane C, incubated U-RSW; lane I, I-RSW. The numbers indicated above subsequent lanes correspond to lysates described for Fig. 2. (B) Colorimetric caspase assays. HeLa cell lysates (20 µg of protein) derived from those used for Fig. 2 were incubated in the presence of either 0.2 mM Ac-DEVD-pNA (solid bars), Ac-YVAD-pNA (hatched bars), or Ac-IETD-pNA (open bars) for 2 h at 37°C. Release of pNA was analyzed by optical density at 405 nm, and caspase activity is displayed as the number of nanomoles of pNA released per hour per total milligram of protein as calculated from a standard curve by using free pNA. The immunoblot in panel A was scanned with an Artec Viewstation A6000C, and the resulting image was labeled with Adobe Photoshop version 3.0.

In vivo inhibition of eIF4G cleavage in HeLa cells treated with cisplatin or etoposide by Z-VAD-fmk. To further explore the relationship between caspases and eIF4G cleavage activity, we determined if eIF4G cleavage in vivo was reduced by treatment with the cell-permeable broad-spectrum caspase inhibitor benzyloxycarbonyl-VAD-fluoromethyl ketone (Z-VAD-fmk). Cell cultures were preincubated with 100 µM Z-VAD-fmk, followed by treatment with either cisplatin (100 µM) or etoposide (50 µM). Cell lysates analyzed by immunoblotting for eIF4G (Fig. 4A) show that cleavage of eIF4G was completely inhibited in cisplatin- or etoposide-treated HeLa cells when cells were preincubated with 100 µM Z-VAD-fmk, indicating that caspase activation is necessary for eIF4G cleavage to occur. In addition, when cells were pretreated with Ac-DEVD-CHO (100 µM), cleavage of eIF4G was also inhibited by approximately 50% (data not shown). Reduced inhibition with the latter reagent is most likely due to the fact that Ac-DEVD-CHO is less cell permeable than Z-VAD-fmk and is also a less potent inhibitor, since the CHO group binds reversibly, whereas the fluoromethylketone group provides irreversible binding. This is supported by experiments which show that Ac-DEVD-CHO was very effective in blocking eIF4G cleavage in in vitro assays (Table 1). However, pretreatment of HeLa cells with acetyl-YVAD-chloromethyl ketone (Ac-YVAD-cmk) (100 µM) before exposure to cisplatin had no effect on the cleavage of eIF4G (data not shown), suggesting involvement of caspase 3-like rather than caspase 1-like proteases. Furthermore, we analyzed the effect of treatment of HeLa cells with Z-VAD-fmk on the DEVD-pNA cleavage activity (caspase 3 related) present in cell lysates (Fig. 4B). The addition of Z-VAD-fmk to the cells reduced DEVD-pNA cleavage activity in cell lysates by approximately 75% for both cisplatin- and etoposide-treated HeLa cells. In conclusion, these results clearly support the involvement of caspases, possibly a caspase 3-like activity, either in the induction of eIF4G cleavage activity or in the catalysis of eIF4G cleavage. However, at this time it cannot be ruled out that multiple proteases, including some which are not caspases, may play a role in eIF4G cleavage.

View larger version (37K): [in this window] [in a new window]   FIG. 4.   Inhibition of caspase activity and eIF4G cleavage in cisplatin-treated HeLa cells by caspase inhibitors. (A) Inhibition of eIF4G cleavage in vivo. HeLa cell lysates (50 µg of protein) were analyzed by immunoblotting. Lane U, U-RSW; lane I, I-RSW; lane C, untreated cells; 0, and 100 indicate the amounts of Z-VAD-fmk used to preincubate (1 h) cells before addition of 100 µM cisplatin or 50 µM etoposide (16-h treatment). (B) In vivo inhibition of caspase activity. HeLa cells were preincubated with 100 µM of Z-VAD-fmk for 1 h and subsequently treated with 100 µM cisplatin for 16 h at 37°C. pNA assays were performed as described above. C, untreated cells. The immunoblot in panel A was scanned with an Artec Viewstation A6000C, and the resulting image was labeled with Adobe Photoshop version 3.0.

Characterization of the eIF4G cleavage activity present in apoptotic lysates. To examine the nature of the proteolytic activity responsible for the eIF4G cleavage, in vitro cleavage assays were performed in the presence of various inhibitors (Table 1). Overall, the data show that the eIF4G-specific cleavage activity measured here is inhibited by the same set of inhibitors previously shown to block several caspases. In particular, two caspase-specific inhibitors, Ac-DEVD-CHO and Ac-YVAD-CMK, inhibited the eIF4G cleavage activity at concentrations as low as 10 nM and 1 µM, respectively. This indicates that the eIF4G cleavage activity is more closely related to caspase 3-like proteases than to caspase 1-like proteases. In addition, serine protease inhibitors N-tosyl-L-phenylalanine chloromethyl ketone (TPCK), N-p-tosyl-L-lysine chloromethyl ketone (TLCK), and soybean trypsin inhibitor, which do not commonly inhibit cysteine proteases, are effective in blocking both eIF4G cleavage and caspases (5, 30, 32).

Induction of apoptosis in cell extracts results in cleavage of eIF4G. Recent reports have indicated that one pathway of apoptosis induction involves the release of cytochrome c from mitochondria into the cytosol (21, 26, 41), which in turn leads to the activation of caspase 3 (26). Therefore, we investigated whether induction of the apoptotic program in cell extracts by addition of cytochrome c would also lead to the cleavage of eIF4G. HeLa cell extracts (S100) were incubated in the presence of cytochrome c (400 nM) and dATP (1 mM) at 37°C for the indicated periods (Fig. 5). eIF4G cleavage was monitored by Western blot analysis, showing sequential cleavage of eIF4G starting at 60 min, which was complete by 18 h (1,028 min) (Fig. 5A). Interestingly, all of the eIF4G cleavage products observed here are identical to those observed in vivo (Fig. 2A), with the exception of the faint protein band indicated by the arrowhead. More so, these results clearly indicate that eIF4G is processed into eIF4GcpN by at least two cleavage events and could possibly indicate the involvement of more than one protease in the cleavage of eIF4G.

View larger version (51K): [in this window] [in a new window]   FIG. 5.   Cytochrome c- and dATP-dependent activation of the eIF4G cleavage activity in vitro. (A) Induction of eIF4G cleavage in HeLa S100 extracts. HeLa S100 extract (40 µl) was incubated with 400 nM cytochrome c and 1 mM dATP for the time periods indicated at 37°C. Reactions were stopped by addition of SDS-PAGE sample buffer, subjected to SDS-PAGE, and analyzed by immunoblotting. eIF4G and eIF4GcpN are indicated on the right. The arrow and arrowhead indicate intermediate cleavage products of eIF4G. (B) Activation of caspase 3 in HeLa S100 extracts. Reaction mixtures as described for panel A were loaded on 13% acrylamide gels and immunoblotted for caspase 3. The immunoblots in panels A and B were scanned with an Artec Viewstation A6000C, and the resulting images were labeled with Adobe Photoshop version 3.0.

In vitro cleavage of eIF4G can be induced by recombinant caspase 3. To further explore the possible involvement of caspase 3 in the cleavage of eIF4G, we generated recombinant caspase 3 proenzyme by overexpression in bacteria. Kinetics experiments have shown that as caspase 3 proenzyme expression increases in E. coli, a portion of the proenzyme becomes converted to the cleaved active form (data not shown [Fig. 8]). Figure 6A shows that incubation of control E. coli cell lysates with U-RSW did not induce eIF4G cleavage (lane 1). In contrast, lysates from bacteria expressing recombinant caspase 3 caused rapid and complete eIF4G cleavage (lane 2), generating the same sets of eIF4GcpN cleavage products previously noted in vivo and in vitro (Fig. 2 and 5). Figure 6B shows further analysis of this observed eIF4G cleavage by caspase 3 in kinetics experiments with highly purified recombinant caspase 3, which demonstrated sequential cleavage of eIF4G similar to that observed in Fig. 5A. These results imply that at least two observed sets of eIF4G cleavage products can be generated through direct cleavage of eIF4G by caspase 3. 

View larger version (21K): [in this window] [in a new window]   FIG. 6.   In vitro cleavage of eIF4G by caspase 3. (A) Cleavage of eIF4G by bacterially expressed caspase 3. U-RSW (4 µl) was incubated at 37°C for 3 h with control bacterial lysate [BL(DE3)pLysS; lane 1] or lysate from bacteria expressing caspase 3 [BL(DE3)pLysS/caspase 3; lane 2] and analyzed on a 9% acrylamide gel. Lane U, U-RSW; lane C, control U-RSW incubated for 3 h at 37°C. Asterisks indicate bacterial proteins which react with nonspecific antibodies present in the eIF4G serum. (B) Purified recombinant caspase 3 cleaves eIF4G. U-RSW (4 µl) was incubated with purified caspase 3 (25 µl), incubated for the indicated time points, and analyzed on a 7% acrylamide gel. Lane U, U-RSW; lane C, control U-RSW incubated for 3 h at 37°C; lane I, I-RSW. eIF4G and eIF4GcpN are indicated on the right. The arrow and arrowhead indicate intermediate cleavage products of eIF4G. The immunoblots in panels A and B were scanned with an Artec Viewstation A6000C, and the resulting images were labeled with Adobe Photoshop version 3.0.

Analysis of eIF4G C-terminal cleavage products in apoptotic lysates. In order to further elucidate the observed cleavage pattern of eIF4G, we also examined the cleavage by using an antibody specific for a peptide derived from amino acids 998 to 1023 of the C-terminal domain of eIF4G (Fig. 7A). The single eIF4G C-terminal cleavage product (eIF4GcpC) generated during a PV infection is shown both for comparison and as a marker (lane I). Incubation of U-RSW and purified recombinant caspase 3 for 3 h at 37°C resulted in the generation of two cleavage products containing the epitope which migrate near the 130- and 48-kDa eIF4G cleavage products, labeled eIF4GcpC1 and eIF4GcpC2, respectively (lane 1). Figure 7B shows that eIF4GcpC2 is also produced in vivo as well when HeLa cells are treated with 100 µM cisplatin for 16 h. The lack of detection of eIF4GcpC1 in vivo suggests that it does not accumulate and possibly is further processed into eIF4GcpC2 or it is totally degraded during apoptosis and therefore is not detectable in apoptotic lysates. Both cleavage fragments consistently stained weakly in immunoblots, possibly because the epitope recognized by the antisera lies very close to a potential caspase 3 site which may result in less antibody binding if cleavage occurs at that site.

View larger version (28K): [in this window] [in a new window]   FIG. 7.   In vitro and in vivo analysis of eIF4GcpC from apoptotic lysates. (A) In vitro analysis of eIF4GcpC. U-RSW (4 µl) was incubated with purified caspase 3 (25 µl) for 2 h at 37°C (lane 1), analyzed on a 7% acrylamide gel, and immunoblotted with an antiserum specific for the C-terminal portion of eIF4G. Lane I, I-RSW; lane C, control U-RSW incubated for 2 h at 37°C. eIF4G and C-terminal eIF4G cleavage products (eIF4GcpC1 and -2) are indicated on the right. (B) In vivo analysis of eIF4GcpC. HeLa cells were treated with 100 µM cisplatin and incubated for the indicated time periods. Cell lysates (50 µg of protein) were loaded on a 7% acrylamide gel and immunoblotted as in panel A. Lane I, I-RSW. The immunoblots in panels A and B were scanned with an Artec Viewstation A6000C, and the resulting images were labeled with Adobe Photoshop version 3.0.

Direct cleavage of eIF4G by caspase 3. To rule out the possibility that purified caspase 3 was activating other caspases present in a U-RSW, we tested highly purified caspase 3 for its ability to cleave highly purified eIF4G (as part of the translation initiation complex, eIF4F). Figure 8A shows the purified caspase 3 by Coomassie staining (lane 1) and immunoblotting (lane 2), respectively. The caspase 3 fraction consisted of both proenzyme and processed active forms (p17 and p10, respectively) of caspase 3 as well as some fast-migrating proteins which may be caspase degradation products. We have confirmed by immunoblotting (lane 2) the presence of the large subunit of caspase 3 (p17) in the fraction, indicating that activation of caspase 3 had occurred. We tested the ability of the purified caspase 3 to cleave purified eIF4G. The substrate, eIF4F, which contains the three components eIF4A, eIF4E, and eIF4G, is shown in lane U of Fig. 8B and contains faint high-molecular-weight bands which result from slight eIF4G degradation during purification of the eIF4F complex. Interestingly, incubation of caspase 3 with eIF4F resulted in the complete cleavage of the eIF4G component (lanes 1 and 2, two separate experiments) accompanied by the appearance of several putative cleavage products. Analysis of the sample in lane 2 by immunoblotting confirmed the generation of eIF4GcpN (lane 3) identical to those observed in vivo (compare lane 4 and Fig. 1 or 2) as indicated by the bar. In addition, no detectable cleavage of eIF4A and eIF4E was observed, suggesting that caspase 3 does not degrade all initiation factors. To address the possibility that eIF4G was cleaved by a contaminant protease present in the purified caspase 3 preparation, purified eIF4F and purified caspase 3 were incubated in the presence of specific caspase 3 inhibitor Ac-DEVD-CHO (Fig. 8C). Increasing concentrations of the inhibitor completely blocked the cleavage of eIF4G and coincided with the disappearance of the putative cleavage products. These results strongly suggest that caspase 3 is responsible for the observed cleavage of eIF4G.

View larger version (41K): [in this window] [in a new window]   FIG. 8.   Direct cleavage of eIF4G by caspase 3. (A) Purification of recombinant caspase 3. Caspase 3 was purified as described in Materials and Methods and analyzed by Coomassie staining (lane 1) and immunoblotting (lane 2). Pro-caspase 3 and active large subunit p17 are indicated on the right. (B) eIF4G cleavage by purified recombinant caspase 3. Purified eIF4F (100 µl) was incubated with purified caspase 3 (shown in panel A) for 3 h at 37°C and analyzed on a 9% acrylamide gel by Coomassie staining (lanes 1 and 2, two separate experiments) or immunoblotting (lane 3). Lane 4, cell lysate (50 µg) from HeLa cells treated with 100 µM cisplatin for 24 h; lane U, purified eIF4F; lane C, purified eIF4F incubated for 3 h at 37°C. eIF4A, eIF4E, and eIF4G are indicated on the right. Putative eIF4G cleavage products are indicated by dashes on the right. The bar denotes eIF4GcpN, as seen in Fig. 1 to 6. An asterisk denotes the putative 48-kDa cleavage product also observed in Fig. 7. (C) Inhibition of eIF4G cleavage by Ac-DEVD-CHO. Purified eIF4F (50 µl) was incubated with purified caspase 3 for 3 h at 37°C in the presence of increasing concentrations of Ac-DEVD-CHO, as indicated at the top. The samples were analyzed on a 9% acrylamide gel by Coomassie staining. Lane C, purified eIF4F incubated for 3 h at 37°C; lane D, purified eIF4F incubated for 3 h at 37°C in the presence of dimethyl sulfoxide (solvent control for inhibitor). The immunoblots in panels A and B and the Coomassie-stained gels in panels B and C were scanned with an Artec Viewstation A6000C, and the resulting images were labeled with Adobe Photoshop version 3.0.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Here we have shown that induction of apoptosis results in rapid and complete cleavage of eIF4G, and this cleavage coincides with a drastic inhibition of cellular translation. Furthermore, we have shown that caspase 3 could be responsible for the observed eIF4G cleavage, because it accounts for most of the changes observed. This is the first report describing the cleavage or inactivation of a major eukaryotic translation initiation factor during apoptosis. eIF4G is known to function in translation as a molecular bridge which is required to facilitate the binding of capped mRNA molecules to ribosomes. In PV-infected cells, cleavage of eIF4G is the hallmark of inhibition of host protein synthesis and is known to result in blockage of de novo mRNA binding by ribosomes. This translation inhibition is thought to be primarily mediated by eIF4G cleavage, since replenishment of in vitro translation systems prepared from infected cells with intact eIF4G partly restores cap-dependent translation (17). Furthermore, it is known that PV infection does not cause cleavage of other known initiation factors such as eIF4E, eIF4A, or eIF3 (12, 14, 25, 39). Thus, in apoptotic HeLa cells described here, it is likely that the observed eIF4G cleavage also constitutes a significant portion of the mechanism of translation inhibition which occurs. Since eIF4G is known to play such an important role in translation, the biological significance of eIF4G cleavage during apoptosis is apparent, and these data strongly suggest that this event and the resulting global disruption of protein homeostasis are likely to be crucial for the execution phase of apoptosis. In addition to cisplatin and etoposide, treatment of HeLa cells with tumor necrosis factor alpha, MG-132 (a proteasome inhibitor), and UV light (data not shown) all induced the same eIF4G cleavage products in HeLa cells during apoptosis, indicating that different inducers cause activation of the same eIF4G-specific protease(s). In addition, K562, HL-60, and Jurkat T cells also responded similar to cisplatin treatment, showing identical eIF4G cleavage products (data not shown), although more cell types must be examined to determine the universality of this response. This will help determine if eIF4G cleavage and translation inhibition are events required for the execution phase of apoptosis in certain cell types. Although we have not yet examined the fates of eIF4E and eIF4A in apoptotic cells, we have shown that similar to PV infection, caspase 3 does not cleave these factors under reaction conditions in which eIF4G was cleaved (Fig. 8B and C). Taken together, the results suggest that PV and apoptosis may inhibit translation by very similar mechanisms.

Surprisingly, there is very little other published data concerning alterations in translation control during apoptosis. There has been a debate about whether ongoing translation is required for apoptosis induction to occur, with evidence from some experimental systems suggesting that apoptosis does not require de novo protein synthesis (29, 33, 36, 38). Conversely, other reports have shown that de novo protein synthesis is required for the cells to become apoptotic in response to nutrient deprivation or steroids (8, 9, 34). Even though this conflict likely reflects differences in the multiple pathways which can trigger apoptosis, our results do not contribute to this debate, since translation was not inhibited for more than 8 h after induction of apoptosis began.

eIF4G cleavage in PV-infected cells is known to separate the eIF4E- and eIF3-binding domains on eIF4G (22). Interestingly, the predominant N-terminal cleavage products observed in apoptotic cells are smaller than those generated by PV infection, thus suggesting that apoptosis results in cleavage of eIF4G at a site closer to the N terminus than where PV causes cleavage. It is unknown at this time whether this new cleavage site lies within the eIF4E-binding domain (28). In addition, it is clear that eIF4G is cleaved sequentially in apoptotic cells, as shown by the transient eIF4G cleavage products resulting from cleavage at one (or more) other site(s) (Fig. 2), which could also be observed in a cell-free system by using cytochrome c and dATP (Fig. 5). Close examination of the eIF4GI amino acid sequence reveals several potential cleavage sites (DXXD) for caspase 3 at positions 333 to 336, 659 to 662, 786 to 789, 961 to 964, 974 to 977, and 977 to 980. In addition, sequence analysis of eIF4GII indicates potential cleavage sites for caspase 3 at positions 557 to 560, 848 to 851, 975 to 978, and 1159 to 1162. Only two of these sites (positions 659 to 662 versus 848 to 851 and 786 to 789 versus 975 to 978 in eIF4GI and eIF4GII, respectively) are conserved between the two forms of eIF4G. Cleavage at one or more of the potential sites would most likely lead to loss of function of eIF4G and subsequently result in inhibition of translation initiation in apoptotic cells, although this remains to be demonstrated experimentally.

Initial in vitro protease inhibition studies indicated that the eIF4G cleavage activity may be closely related to caspase 3 because of its high sensitivity to Ac-DEVD-CHO. Further examination showed indeed that eIF4G was a substrate for caspase 3. In addition, it was shown that caspase 3 processed eIF4G sequentially, producing two sets of cleavage products and providing evidence for multiple cleavage sites for caspase 3 on eIF4G, which was supported by close examination of the eIF4G sequence, which contains several DXXD motifs (see above). Two of the sets of eIF4GcpN cleavage products produced by caspase 3 were also produced in vivo (Fig. 2); however, a third set of in vivo transient cleavage products which comigrated with PV-induced cleavage products were not produced by caspase 3. This might suggest that another unidentified cellular protease exists which cleaves eIF4G near amino acid 486 and contributes to the processing of eIF4G in apoptotic cells in vivo. Identification of other proteases and of the cleavage sites utilized by caspase 3 is currently under investigation.