Apoptosis Rate Can Be Accelerated or Decelerated by Overexpression or Reduction of the Level of Elongation Factor-1α

EXPERIMENTAL CELL RESEARCH ARTICLE NO.

238, 168–176 (1998)

EX973819

Apoptosis Rate Can Be Accelerated or Decelerated by Overexpression or Reduction of the Level of Elongation Factor-1a Atanu Duttaroy, Denis Bourbeau, Xiao-Ling Wang, and Eugenia Wang1 The Bloomfield Center for Research in Aging, Lady Davis Institute for Medical Research, Sir Mortimer B. Davis Jewish General Hospital, and Department of Medicine, McGill University, Montre´al, Que´bec H3T 1E2, Canada

Peptide chain elongation factor-1a (EF-1a) is required for the binding of aminoacyl-tRNAs to acceptor sites of ribosomes during protein synthesis. More recently, EF-1a has been shown to be involved in cytoskeletal organization. The elongation factor functions in actin bundling and microtubule severing. Moreover, it can activate the phosphatidylinositol-4 kinase whose substrates are involved in regulation of actin polymerization. The expression level of EF-1a is regulated in many situations such as growth arrest, transformation, and aging. Because of this regulation of EF-1a in various states of cell life, and its key position in protein synthesis as well as cytoskeletal organization, we chose to investigate the effect of its expression levels on apoptosis. Apoptosis is a complex event regulated through numerous activators and inhibitors. In some situations, protein synthesis is required for apoptosis to be triggered. Investigation of the effect of altered levels of elongation factor-1a on apoptosis is of particular interest since it may affect both protein synthesis and cytoskeletal organization. For example, reduction of EF-1a leads to a reduced protein synthesis rate, which might reduce the presence of those ‘‘killer factors’’ triggering apoptosis. EF-1a involvement in cytoskeletal organization is another example, since cytoskeletal organization undergoes dramatic changes during apoptosis. Thus, this study has been planned to ascertain whether hypo- and hyperexpression of EF-1a protein, achieved by constructing expression vectors with the EF-1a cDNA in its antisense or sense orientation under the control of a cytomegalovirus promoter, can produce stable transfectants with either heightened or reduced responsiveness to apoptosis stimuli. Our results show the following: (1) induction of apoptosis by serum deprivation shows that antisense EF-1a provides cells significant protection from apoptotic cell death and (2) EF-1a overexpression causes a faster rate of cell death. These findings suggest that when EF-1a protein is abundant the cells are 1 To whom correspondence and reprint requests should be addressed at Lady Davis Institute for Medical Research, Sir Mortimer B. Davis Jewish General Hospital, 3755, Chemin de la Coˆte Ste.Catherine, Montre´al, Que´bec, Canada H3T 1E2. Fax: (514) 340-8295.

ECR 3819

/

6i2ah$$201

q 1998 Academic Press

INTRODUCTION

Programmed cell death (apoptosis) is a complex cellular event, operating in all multicellular organisms to selectively eliminate excess cells produced during development and also during maintenance of the metazoan body plan [review 1]. Apoptosis is a nonrandom event representing a highly predictable cellular mode, operated by the activation of a cellular suicidal plan made up of specific genetic signals. This scheduled elimination of extra cells is the organismal means of performing tissue organization, as well as fighting infection and DNA damage to get rid of unwanted or damaged cells. Many genes have been identified as factors directing the activation or suppression of apoptotic cell death. On the basis of their modes of action, these genes can be broadly categorized as apoptosis effector or protector genes. Noted examples of apoptosis effector genes include ced-3 and ced-4 in Caenorhabditis elegans, reaper gene in Drosophila, or interleukin-1b-converting enzyme (ICE) in mammals. Apoptosis protector genes like bcl-2, p35, and crmA, on the other hand, inhibit cell death with almost equal efficiency in cross-species cellular environments. Nonetheless, bcl2 is not ubiquitously active, being ineffective in Fas/APO-1-induced apoptosis in B-lymphocytes, thymocytes, or activated T-cells [2]. The inhibitory actions of p35 and crmA genes counteract the cell killing action of ICE-like cysteine proteases. In this study we chose to study the effect of altered levels of the peptide chain elongation factor-1a (EF-1a) on apoptosis. This elongation factor is an ubiquitously expressed protein that is highly conserved from Escherichia coli to man. As part of the translational elongation complex, EF-1a promotes GTP-dependent binding

168

0014-4827/98 $25.00 Copyright q 1998 by Academic Press All rights of reproduction in any form reserved.

AID

proapoptosis, and vice versa in low abundance the cells are in the mode of antiapoptosis. Therefore, changes in levels of EF-1a may be one of the global pivotal regulators modulating the rate of apoptosis.

01-14-98 09:41:48

eca

APOPTOSIS SUCCESS DEPENDS ON THE ABUNDANCE OF EF-1a

of aminoacyl-tRNA to ribosomes during peptide chain elongation [see review 3, 4]. The same protein is also involved in actin bundling, microtubule severing, and activation of the phosphoinositol-4 kinase [5], all activities related to cytoskeletal organization. The cytoskeleton undergoes dramatic, although poorly described, changes during apoptosis [6, 7]. We therefore suggest that due to its strategic position in protein translation and cytoskeletal organization, levels of EF-1a might affect the regulation of apoptosis. To investigate whether its level of expression affects responsiveness to apoptosis signaling, we have used transfection of sense- and antisense-EF-1a expression vectors to generate either hypo- or hyperexpression of this gene in Balb/C 3T3 mouse fibroblasts and evaluated the apoptotic potential of these cells. Measurement of apoptosis after serum deprivation showed a significantly reduced rate of cell death in EF-1a hypoexpressing cells. On the contrary, EF-1a hyperexpressing cells showed increased cell death under serum deprivation conditions. When the protein synthesis rate was measured, we found a reduced rate of protein synthesis in cells hypoexpressing EF-1a, while in hyperexpressing cells the increase in EF-1a does not lead to an increased protein synthesis rate. These results show that at least in the type of the apoptosis signaled by serum deprivation, increasing EF-1a level facilitates the execution of the apoptosis program. However, the change in apoptosis rate is not necessarily coupled with modulation of the protein synthesis rate, as evidenced by the discordance between EF-1a level and protein synthesis seen in the stable transfectant clone hyperexpressing EF-1a. This result leads to the suggestion that the other functions of EF-1a in regulating cytoskeletal organization may be as significant as its effect on protein synthesis during apoptosis. MATERIALS AND METHODS Cultures of mouse 3T3 fibroblasts and activation of apoptosis. A specific subclone line of Balb/C 3T3 strain of mouse fibroblasts, characterized to possess uniform response to contact inhibition-induced growth arrest, was developed; when this subclone reaches confluency, almost all cells in the monolayer assume a quiescent phenotype, showing no measureable evidence of DNA synthesis; no expression of immediate early genes such as c-fos, c-myc, or PCNA; and no phosphorylation of RB protein [8]. In addition, almost all cells in the contact-inhibited quiescent state show positive presence of statin, a nuclear protein found only in nongrowing cells [9]. Programmed cell death, or apoptosis, can be induced in the monolayer of these contact-inhibited quiescent cultures by total withdrawal of serum from the culture. Except in apoptosis-induction conditions, all cultures were incubated in Dulbecco’s minimal essential medium (DMEM) containing 10% fetal calf serum, at 377C in a humid atmosphere with 5% CO2 . Viability assay and statistical analysis. Approximately 1 1 105 cells were plated per well in six-well plates (three wells per time point for each clone) in 3 ml medium, and cells were grown to confluency in about 4 days. Once confluent, cells were washed in serum-free

AID

ECR 3819

/

6i2ah$$202

01-14-98 09:41:48

169

DMEM and maintained under serum-free conditions at 377C. Viability counting was done by harvesting the dead cells suspended in the medium, followed by detaching the remaining adherent cells by trypsinization. The two (suspended and adherent) cell fractions were combined and centrifuged at 3000 rpm for 5 min, and the cell pellets were resuspended in 1 ml DMEM. A 50-ml aliquot was withdrawn and mixed with trypan blue at 1:1 ratio. Cell counting was performed in a hemocytometer under an inverted microscope, where viable cells were seen as refractile, while dead cells appeared dark blue in color. Viability assays for experimental and control cells were performed at the same time. The percentage of viable cells in the total population was a pooled average of four independent counts in three separate cultures. To assess the significance of viability values, t tests assuming equal variance were performed on the primary data. Agarose gel analysis for fragmented oligonucleosomal profile. Monolayer cultures (attached and detached cells) were collected into 1.5 ml DMEM at different time points after removal of serum. The harvested cell materials were pelleted by centrifugation for 5 min and resuspended in 300 ml buffer containing 10 mM Tris– HCl (pH 8.0), 100 mM NaCl, 25 mM EDTA, 0.5% SDS, and 0.5 mg/ ml Proteinase K [10] and incubated at 377C overnight. DNA was precipitated with ethanol, resuspended in 200 ml Tris – EDTA buffer at pH 7.5 containing 50 mg/ml RNase, and incubated at 377C for an additional 2 h. Five micrograms of DNA samples was then electrophoresed on 2% agarose gel containing ethidium bromide. Assay for protein synthesis. To measure the activity of protein synthesis, cells were plated in triplicate and cultured to confluency. At confluency, cultures were washed twice in methionine-free DMEM (with 10% FBS) and incubated in the same medium for 30 min. Subsequently, 1 ml of the same medium carrying 100 mCi/ml [35S]methionine was added, and cells were incubated for 1 h at 377C. Any unincorporated radioactivity was removed by washing the cultures with phosphate-buffered saline (PBS) three times, and cells were harvested by scraping the monolayer cultures from the substratum. Labeled proteins were precipitated with 10% trichloroacetic acid (TCA) at 47C overnight. The TCA precipitates were washed three times with cold acetone and finally resuspended in PBS containing 0.1% SDS. Total protein content was measured by Bio-Rad protein assay (Bio-Rad Laboratories, Missisauga, Ontario, Canada), and radioactivity was determined by scintillation counting. The final cell incorporated count was standardized as counts per minute per microgram of protein. Cycloheximide-treated controls were used for some experiments; here cultures were treated with 50 mg/ml cycloheximide with preincubation for 1 h prior to [35S]methionine labeling. For the purpose of simultaneous measurement of protein synthesis and DNA content, cultures were labeled with [35S]methionine first (as above) and then washed in PBS, and the cell pellets were separated into two parts. One part was precipitated with 10% TCA, while the other was kept in 10% perchloric acid for DNA quantification. For this purpose, cells were incubated overnight in a glacial acetic acid solution containing 10% perchloric acid, 4% diphenylamine, and 40 mg/ml acetaldehyde. The optical density was read at 595 nm, and values obtained were converted to micrograms of DNA, using a standard DNA curve. Assay for cell growth activity. To assay for cell growth activity, we used the procedure described previously [11]. In brief, cultures were plated at subconfluent density at approximately 1000 cells per well of 96-well plates; at every 24 h for 6 days the medium was removed, and medium containing 2 mg/ml of 3-(4,5-dimethyl thiozol2-yl)-2,5 diphenyltetrazolium bromide (MTT) was added to the cultures for 3 h. Afterward, this medium was removed and 50 ml of dimethyl sulfoxide was added; cell materials were processed for optical density measurement by an ELISA reader for optical density at 490 nm. Protein extraction, gel electrophoresis (SDS–PAGE), and Western blotting. For protein sample preparation, a standard protocol from this laboratory was followed [8, 9]; this procedure is modified from

eca

170

DUTTAROY ET AL.

the protocol for processing protein samples as described [12]. Protein samples separated on a gel containing 12.5% polyacrylamide were transferred to nitrocellulose paper [13]; the nitrocellulose blots bearing the protein samples were incubated with a rabbit polyclonal antibody to EF-1a, designated HT7 (a generous gift from Dr. Reen Wu of the University of California, Davis). This antibody was produced from a peptide sequence 20 amino acids long from the extreme Cterminal end of the EF-1a protein [14]. Incubation with the HT7 antibody was carried out overnight at 47C, followed by incubation with goat anti-rabbit antibody (Cappel, Organon Teknika, NY) coupled to horseradish peroxidase. Positive bands were detected by further incubation with 4-chloronaphthol (Sigma-Aldrich, Oakville, Ontario, Canada) or by the ECL procedure (Amersham, Oakville, Ontario, Canada). Construction of antisense and sense EF-1a plasmids. A fulllength mouse EF-1a cDNA insert (1.7 kb) in the pBluescript plasmid was released by double digestion with EcoRI/ApaI and gel purified. Subsequently, the EF-1a cDNA insert was end filled and subcloned in the SmaI site of mammalian expression vector pBK-CMV which expressed neomycin resistance gene (Stratagene, PDI-Joldon, Aurora, Ontario, Canada). Five recovered subclones were partially sequenced from both ends using T3 and T7 primers, allowing us to confirm the antisense orientation of the EF-1a cDNA insert with respect to the cytomegalovirus promoter. For the purpose of EF-1a overexpression, the same EF-1a fulllength cDNA insert was subjected to PCR amplification, using primers carrying the BamHI and EcoRI sites in the 5* (forward) and 3* (reverse primer) ends, respectively. Subsequently, BamHI/EcoRI digestion of the PCR product facilitated its directional cloning into the pBK-CMV vector. Our ultimate goal was to overexpress EF-1a in mammalian cells, and therefore it was necessary to be certain that no aberrant mutation (premature stops) had been introduced into the EF-1a cDNA during PCR amplification. For this purpose, the amplified EF-1a cDNA was first subcloned into the pGEX-2T vector, where the induced expression of EF-1a peptide was confirmed

FIG. 2. (A) The reduced expression of EF-1a in the antisense EF-1a stable transfected cell clones, AS6 and AS7. Immunoblotting analyses were perform with the EF-1a-specific polyclonal antibody (HT7). One-hundred micrograms of total protein extracts were loaded onto each lane. (B) The data obtained after quantitative densitometric analysis of the positive EF-1a band intensity of the same blot (shown A). The EF-1a peptide levels in AS6 and AS7 cells were 45 and 50%, respectively, of those of wild-type 3T3 cells.

by immunoblotting with HT7 antibody. Finally, this same EF-1a insert from pGEX-2T was digested with BamHI/EcoRI and subcloned into pBK-CMV for transfection purposes. Transfection procedure. Mouse 3T3 cultures were transfected with either pBK-CMV vector with no insert but bearing neomycin resistance gene or EF-1a sense or antisense plasmid also bearing neomycin resistance, using the Lipofectamine kit and following the instructions of the manufacturer (BRL). Approximately 8 1 104 cells were plated on six-well plates at 12 h prior to transfection. One microgram of plasmid DNA and 5.0 ml of lipofectamine reagent were mixed together; after incubation at room temperature for 30 min, the solution was further mixed with 400 ml DMEM and added onto the cultures. Transfection was performed for 6 hours at 377C, 5% CO2 , after which 1.0 ml of DMEM was added until the next day and then DMEM carrying the plasmid DNA was replaced with fresh DMEM and cultured at 377C for an additional 48 h. Transfected cell clones carrying stably integrated plasmid(s) were selected by keeping the cells in medium containing neomycin (Geneticin; BRL) at a concentration of 800 mg/ml for 3 weeks, after which the surviving cell clones were isolated and used for various analyses. FIG. 1. (A) The percentage of viable cells counted at 24-h intervals after activating apoptosis by removing serum totally from confluent cultures of Balb/C mouse 3T3 fibroblasts. Approximately 60% of cells undergo apoptosis within 24 h, and by 96 h 80% of the total cell population are dead. Triplicate cultures were used for viability measurement, determined by trypan blue exclusion assay. (B) The appearance of DNA fragmentation after inducing apoptosis by serum deprivation; in this case an added time point of 12 h was taken. The first resolvable oligonucleosomal ladder was found after 24 h of serum deprivation, and the ladder patterns showed gradual reduction in size with increasing time intervals of serum deprivation.

AID

ECR 3819

/

6i2ah$$202

01-14-98 09:41:48

RESULTS

Viability and DNA Fragmentation Assaying the Rate of Apoptosis Induced by Serum Deprivation in Mouse Balb/C 3T3 Fibroblasts Density growth-arrested cultures of mouse 3T3 fibroblasts show typical apoptotic cell death upon complete removal of serum. This effect of serum depriva-

eca

APOPTOSIS SUCCESS DEPENDS ON THE ABUNDANCE OF EF-1a

FIG. 3. Metabolic labeling with [35S]methionine was used to determine the rate of cellular protein synthesis activity. Protein synthesis in AS6 and AS7 is reduced, respectively, to 56.4 and 60.4% of the wild-type (WT) control. Comparison with cycloheximide-treated cells (CHX) showed that 50 mg/ml of cycloheximide reduces incorporation to 40% of that seen in the control cultures. (WT, wild type; V.A., transfected with vector alone; and AS6, AS7, two separate stable transfectant clones of EF-1a antisense plasmid).

tion-activated apoptosis can be assayed as percentage of viable cells in the total cell population (Fig. 1A). Death rate was estimated by the trypan blue exclusion viability test, at 24-h intervals starting from the time when serum was completely withdrawn from the culture. Sixty percent (60%) of the cells in serum-deprived cultures die within the first 24 h, and an additional 20% die within the next 3 days; the remaining 20% of cells remain viable till the end of the 96-h assaying period. Figure 1B shows the DNA fragmentation pattern of apoptotic cells after serum deprivation; kinetics of the DNA fragmentation pattern reveal a gradual reduction in the size of oligonucleosomes with increasing time after serum deprivation, representing a traditional hallmark of apoptotic cell death.

171

mouse fibroblast clones on the basis of their neomycin resistance phenotype. Subsequently, the expression level of EF-1a protein in all seven clones was checked by Western blot analysis with rabbit polyclonal antibody to EF-1a, the HT7 antibody (see Materials and Methods for details on HT7 antibody). Suppression of the endogenous EF-1a protein in antisense EF-1a clones AS6 and AS7 was verified in total protein extracts from density-arrested, confluent cultures of these two antisense transfectants and from two different control cultures composed of wild-type mouse 3T3 fibroblasts and neomycin-selected clones transfected with pBK-CMV vector carrying no insert. Immunoblotting results with the HT7 antibody confirmed that AS6 and AS7 cells have reduced quantities of EF-1a protein per microgram of total protein (Fig. 2A). Quantitative densitometry of the immunoblot showed that the EF1a protein level in AS6 and AS7 cells are reduced respectively to 45 and 50% of the wild-type control value (Fig. 2B). Determining the Rate of Protein Synthesis in AS6 and AS7 Cells Having established that cells of the AS6 and AS7 clones have reduced amounts of EF-1a protein abundance, we attempted to determine whether the rate of protein synthesis in these two clonal cultures was af-

Antisense Suppression of EF-1a by Stable Transfection Stable transfection of mouse 3T3 fibroblasts with the EF-1a antisense plasmid enabled us to isolate seven TABLE 1 Ratio of Protein Synthesis Activity Antisense clone 6/Balb/C 3T3 Antisense clone 7/Balb/C 3T3 CHXa-treated/untreated Balb/C 3T3 a

56.46% 60.41% 40.95%

50 mg/ml of cycloheximide (CHX).

AID

ECR 3819

/

6i2ah$$202

01-14-98 09:41:48

FIG. 4. Demonstration of slower growth rates in AS6 and AS7 stable transfectant cultures than in control cultures. Cultures taken every day were processed for growth rate measurements using 3(4,5-dimethyl thiozol-2-yl)-2,5 diphenyltetrazolium bromide assay. By day 6 the AS6 and AS7 cell density is half that of the wild-type (WT) control cultures or of cultures carrying stable transfectants of vector alone (V.A.).

eca

172

DUTTAROY ET AL.

FIG. 5. Comparison of apoptotic cell death rate after serum deprivation in stable transfectants AS6, AS7, and vector-transfected 3T3 fibroblast cultures as well as in the wild-type 3T3. Viability counting was determined by trypan blue exclusion assay, performed at 24-h intervals starting when serum was removed from confluent cultures of each cell type. As shown after 24 h of serum deprivation, an average 60% cell death was observed in wild-type (WT) and vector-transfected controls (V.A.), while only 14% of cells were dead in AS6 and AS7 cultures. Therefore, the cell death protection experienced by AS6 cells after 1, 2, 3, and 4 days of serum deprivation was 2.11, 2.11, 2.21, and 2.01 with respect to control 3T3 cells. Similar levels of apoptotic protection were also evident in AS7 cells, amounting to 2.11, 2.01, 1.91, and 1.61 the wild-type 3T3 culture levels. Significance was evaluated using t test assuming equal variance, as described in the text.

fected. Measurement of the protein synthesis rate was performed by [35S]methionine incorporation into total cellular proteins. Our results show that [35S]methionine incorporation was significantly lower in cultures of AS6 and AS7 stable transfectants than in cultures of their control counterparts (Fig. 3); the percentage ratios of protein synthesis activity between AS6 or AS7 and untransfected cultures were 56.4 and 60.41%, respectively (Table 1). Furthermore, this reduction of protein synthesis activity in AS6 and AS7 cells was comparable with the inhibitory action of 50 mg/ml of cycloheximide. Reduced Cell Proliferation Rate in AS6 and AS7 Cells Since the antisense EF-1a-transfected cell clones AS6 and AS7 have reduced rates of protein synthesis, we continued to characterize the growth rate of these cells to determine whether their proliferative potential is affected. Cell growth rate was determined by measuring the reduction of MTT, a substrate for mitochondrial dehydrogenases, which forms a blue product measured at 490 nm with an ELISA reader. The amount of MTT conversion is regarded as a measure of cell density. In Fig. 4, cell densities were plotted from the

AID

ECR 3819

/

6i2ah$$203

01-14-98 09:41:48

MTT conversion rate, showing that the AS6 and AS7 cells were growing slower than the control cultures. Finally on day 6, the end of our study period, respective cell densities in AS6 and AS7 cells reached 57 and 59% of the wild-type control value, suggesting slower growth in antisense EF-1a cells. The control transfectant with the vector alone shows some reduction in growth rate, reflecting possibly the effect of G418 used for stable transfectant selection. Stable Transfectants of Antisense EF-1a Protect from Serum Deprivation-Induced Apoptosis After establishing the reduced protein synthesis potentials in the antisense stable transfectants AS6 and AS7, we characterized further the susceptibility to apoptotic cell death by serum deprivation. Cell death rate was evaluated at 24-h intervals after collecting the dead (nonadherent) and live (adherent) cell fractions from serum-deprived cultures, which were mixed together and stained with trypan blue as described under Materials and Methods. Apoptosis activated by serum deprivation for 24 h causes on average 61.5% cell death in wild-type 3T3 cells, while 62% of cells were found dead in control stable transfectants of pBK-CMV vector

eca

APOPTOSIS SUCCESS DEPENDS ON THE ABUNDANCE OF EF-1a

173

tures varied between 77 and 82%, while AS6 and AS7 cultures showed 61% (P ° 0.002) and 63% (P ° 0.001) dead cells, respectively. All P values were determined using t test assuming equal variance. These data therefore suggest that in cultures of antisense EF-1a stable transfectants, there is a significant degree of protection from serum deprivation-induced apoptosis. Apoptotic Protection Is Reflected in the Reduced Level of DNA Fragmentation Apoptosis is generally associated with DNA fragmentation; the profile of such activity shows a typical ladder pattern of DNA on agarose gel with increasing intervals of serum deprivation. As shown in Fig. 6A, the first resolvable DNA ladder in the untransfected 3T3 fibroblast cultures appears after 24 h of serum deprivation, showing fragmented DNA of various size classes, ranging from 5.0 to 0.3 kb. After 48 h, the profile of fragmented DNA changes, with the majority of DNA fragments ranging from 1.0 kb to less than 0.3 kb. Finally, after 96 h of serum removal, oligonucleosome fragments are seen in the smallest size, ranging between 0.5 and 0.3 kb or less in size. We compared the DNA fragmentation profile of the AS7 stable transfectant after serum deprivation with wild-type 3T3 cells. Interestingly, in AS7 cells, the first resolvable DNA ladder did not appear until 48 h, as

FIG. 6. (A) Protection from apoptotic cell death as reflected in the delayed appearance of DNA fragmentation in AS7 cells after removal of serum. Arrowheads show that after 24 h of serum deprivation, wild-type 3T3 cultures show clearly resolvable DNA ladders, whereas no DNA ladder appears in AS7 cultures in the same time interval. Furthermore, the fragmented DNA appeared smaller and smaller in subsequent hours in wild-type 3T3 cells, as opposed to AS7 cells which showed noticeable accumulation of DNA fragments in a higher molecular weight range. (B) The quantitative densitometric analysis of ethidium bromide-stained gels; the data are expressed in arbitrary units. After 24 h of serum deprivation, wild-type 3T3 cells have 3.671 more fragmented DNA than AS7 cells. This comparison is less dramatic in the later time points, since most of the DNA in control cultures was processed to too low molecular weight range to support meaningful comparison (as is clearly the case at the 96h time point). Therefore, the slower death pattern of AS7 cells is reflected in the DNA fragmentation pattern most dramatically in the first time point (24 h).

alone (Fig. 5). On the other hand, serum-deprived AS6 and AS7 clonal cultures showed little cell death, with only 14% dead cells after 24 h of serum deprivation (Fig. 5). After 48 h, respective cell death rates in wildtype 3T3 and vector-transfected control cells were 70 and 65%, while AS6 and AS7 cultures showed an average of 28 and 31% dead cells, respectively. Finally, at the end of our study period, 96 h following serum deprivation, the average cell death in the two control cul-

AID

ECR 3819

/

6i2ah$$203

01-14-98 09:41:48

FIG. 7. (A) Overexpression of EF-1a in stably transfected cell clone E10 compared with wild-type (WT), vector alone transfectant (V.A.), and antisense transfectants, clone AS6 and AS7. Immunoblotting was performed with HT7 antibody, after collecting cell extracts from confluent cultures. (B) Quantitative analysis of the same immunoblot shown in A shows that E10 cells express EF-1a at a level which is 1.5 times that of untransfected wild-type (WT) cells. Compared with antisense stably transfected AS6 and AS7 cells, the E10 value is 3.75 times and 6 times more than clones AS6 and AS7, respectively.

eca

174

DUTTAROY ET AL.

FIG. 8. Measurement of [35S]methionine incorporation in E10 as well as AS6 and AS7 clones, normalized per microgram of total DNA. When standardized with DNA, AS6 and AS7 clones showed a similar reduction of protein synthesis, as seen in Fig. 3. The hyperexpressing clone E10 showed no increase in protein synthesis compared to controls; the slight decrease is not significant according to t test. (WT, wildtype; V.A., transfected with vector alone; AS6, AS7, two separate stable transfectant clones of EF-1a antisense plasmid; and E10, stable transfectant clone of sense EF-1a.)

opposed to 24 h in the untransfected 3T3 cells, after serum deprivation (Fig. 6A). Furthermore, in comparison to the wild-type 3T3 cells, the degradation appeared to be much slower, as judged by the distribution of DNA size fragments. This was reflected after 96 h of serum deprivation, when the fragmented oligonucleosomes were still visible in the higher molecular weight range in AS7 cultures. We conclude that such a difference in the DNA fragmentation profile between control and AS7 stable transfectant is a true reflection of the slower rate of apoptosis in the latter. We have estimated the quantity of DNA fragmentation by densitometric scanning of the DNA gels stained with ethidium bromide (Fig. 6B); wild-type 3T3 cells carry 3.6 times more fragmented DNA than AS7 cells after 24 h of serum deprivation. Subsequently this ratio changed to 1.63, 1.52, and 0.8 after 48, 72, and 96 h of serum deprivation. The present data suggest that protection from apoptotic death seen in the AS7 stable transfectant is indeed reflected by the DNA fragmentation profile and is most noticeable in the 24 to 48 h time points after apoptosis is activated by serum deprivation. Overexpression of EF-1a May Produce Faster Cell Death The above results show that reduction in EF-1a expression by antisense transfection can slow down the

AID

ECR 3819

/

6i2ah$$203

01-14-98 09:41:48

rate of apoptotic cell death. We therefore asked what the effect would be on serum deprivation-induced apoptosis if EF-1a was overexpressed. Details of the construction of an EF-1a cDNA clone (E10) for overexpression purposes are described under Materials and Methods; this clone was one of the 10 stably transfected clones obtained after selection for neomycin resistance. Western analysis with HT7 antibody showed that clone 10 (referred to as E10) had the highest level of EF-1a expression (Fig. 7A). Quantitative densitometric analysis revealed that the EF-1a level in the E10 cell clone was 1.5 times higher than in the wild-type 3T3 cells and 1.3 times higher than in plasmid-transfected control cells, which is significantly higher than the antisense transfectant AS6 and AS7 cultures (Fig. 7B). However, this increased quantity of EF-1a was associated with no detectable increase in protein synthesis rate (Fig. 8). Induction of serum-deprived apoptosis in E10 cells showed an interesting cell death pattern, opposite to that in AS6 and AS7 antisense stable transfectants. For example, after 24 h of serum deprivation, 23.6% of E10 cells remained viable, compared to 38.5% viable cells in untransfected 3T3 fibroblast cultures (P ° 0.0006) and 37.6% in vector-transfected control cultures (P ° 0.0001) (Fig. 9). After 48 h, the viability

eca

APOPTOSIS SUCCESS DEPENDS ON THE ABUNDANCE OF EF-1a

FIG. 9. Cell viability analysis of the stable EF-1a sense transfectant (E10), wild-type, and vector-transfected 3T3 cells after serum deprivation. After 24 h of serum deprivation, the viability in E10 cultures was 24%, while in wild-type (WT) and vector-transfected control (V.A.) viability was, respectively, 33 and 40%, indicating a faster cell death rate in E10 cells. At 48 h E10 viability was 15%, while in wild-type and vector-alone cultures it was, respectively, 30 and 27%. At 72 h of serum deprivation, only 7.8% viable cells were measured in E10 cultures, while the viability in wild-type and vectortransfected control was 24 and 19%, respectively. At 96 h of serum deprivation, close to 20% viability was observed in both control cultures, while few detectable viable cells were seen in the E10 cultures, reflecting an almost 0% viability. Significance was evaluated using t test assuming equal variance, as described in the text.

percentage for E10 cells was 15.3%, as opposed to 30.6% in wild-type (P Å 0.00005) and 26.5% in vectortransfected (P ° 0.0002) controls. Continued serum deprivation revealed that after 72 h, the E10 cultures showed only 7.8% viable cells, compared to 23.6% viable cells in wild-type 3T3 cells (P ° 0.01) and 18.8% viable cells in plasmid-transfected controls (P ° 0.002), respectively. By 96 h, there were barely any detectable viable cells in the E10 sense transfectant, while both control cultures showed close to 20% viability. All P values were determined using t test assuming equal variance. These data therefore suggest that EF-1a hyperexpression in 3T3 fibroblasts causes a cell death rate which is significantly faster than that of the controls. DISCUSSION

In this report, we have presented evidence that quantitative success in the specific type of apoptosis

AID

ECR 3819

/

6i2ah$$203

01-14-98 09:41:48

175

induced by serum deprivation is dependent upon the level of EF-1a, a translation factor involved in cytoskeletal organization and the elongation step of protein synthesis. This is based on the fact that when the protein level of endogenous EF-1a is repressed by antisense transfection, the viability of the transfected clones shows a significant increase, reflecting the retardation of the final death event, while the opposite is true with enhanced expression of EF-1a in sense transfectants, where the apoptotic rate is increased. These observations show that EF-1a levels affect the rate of the apoptotic process. Whether activation of apoptosis needs the presence of newly synthesized protein(s) has been one of the debatable areas in the literature, due largely to discordant results with the use of cycloheximide in different apoptotic systems. Although a significant number of reports show that cycloheximide’s inhibition of protein synthesis protects cells from apoptotic death, a number of other studies show that treatment with the same inhibitor can induce apoptosis [15–18]. We suggest that the contradictory results may stem from the fact that the same inhibitory effect of cycloheximide on protein synthesis may induce an opposite effect in different types of apoptosis. For those types requiring the upregulation of ‘‘killer’’ gene expressions, the inhibitory effect on protein synthesis is considered to be protective from apoptotic cell death. In contrast, for those types of apoptosis where the down-regulation of ‘‘survival’’ gene expressions is needed, the cycloheximide effect facilitates the removal of the survival factors and thus induces apoptotic death. We cannot preclude the possibility that the final death may be entirely dependent on the specific cell types, coupled with the specific mode of induction of apoptotic death. In our system, mouse 3T3 fibroblasts hypoexpressing EF-1a, the reduction of protein synthesis rate in the hypoexpressing transfectants correlates with an increased resistance to apoptosis induced by serum deprivation. Nonetheless, we cannot discriminate the effect of protein synthesis rate from potential effects of other activities of EF-1a, such as those involved with cytoskeletal organization. This rationale is based on the fact that in the hyperexpressing clone, protein synthesis is not increased by increasing levels of EF-1a. This was not a surprise, since EF-1a level is not a rate-limiting factor for protein synthesis in most circumstances. Nonetheless, its overexpression leads to an increased rate of apoptosis under serum deprivation. This result suggests that EF-1a may contribute to apoptosis regulation by its functions other than role in protein elongation. So far our results have shown that depriving serum from cultures of mouse 3T3 fibroblasts may be in essence removing the necessary growth factors for them to survive in the contact-inhibited, growth-arrested state. Depriving growth factors may indirectly activate

eca

176

DUTTAROY ET AL.

the hypothetical killer gene expressions needed for apoptotic cell death to occur. Our previous work shows that this is indeed the case, since apoptosis-dependent proteolysis of the cytoplasmic protein terminin indeed occurs during serum deprivation-induced apoptosis [19]. We provide evidence that DNA fragmentation activity can be slowed down significantly by the repression of EF-1a protein level, indirectly affecting either the synthesis or the activation of these putative nucleases and therefore delaying final death. According to the reduction of protein synthesis rate in the hypoexpressing cells, we are tempted to suggest that in serum deprivation-induced apoptosis, death requires the synthesis of newly synthesized protein, such as those proteases needed to process the proteolytic action for commitment to apoptotic death [1]. Lower rates of synthesis of new protein would extend the time needed for a certain activator of apoptosis to reach its threshold for activity. However, in the overexpression clone of EF-1a, the increased level of EF-1a increases the susceptibility to apoptosis without increasing the protein synthesis rate. Therefore, EF-1a may exert an apoptotic activity independent of protein synthesis. This activity may be related to EF-1a interactions with actin, tubulin, or phosphatidylinositol-4 kinase [20–26]. At this point, it is not known which property of EF-1a is responsible for this action. Moreover, we cannot disregard the reduction of this activity in the hypoexpressing cells, which may account in part or completely for the delay of apoptosis in those cells. Future experiments on the hyper- and hypoexpression clones of EF-1a stable transfectants in the context of changes of cytoskeletal organization and signal induction will provide leads to dissecting the bifunctional role of EF-1a in regulation of apoptosis susceptibility. REFERENCES 1. White, E. (1996) Genes Dev. 10, 1–15. 2. Strasser, A., Harris, A. W., Huang, D. C. S., Krammer, P. H., and Cory, S. (1995) EMBO J. 14, 6136–6147. 3. Merrick, W. C. (1992) Microbiol. Rev. 56, 291–315.

4. Lee, S., Duttaroy, A., and Wang, E. (1996) In Cellular Aging and Cell Death (Holbrook, Martin, and Lochsin, Eds.), pp. 139– 151, Wiley–Liss, New York. 5. Condeelis, J. (1995) Trends Biochem. Sci. 20, 169–170. 6. Brancolini, C., Benedetti, M., and Schneider, C. (1995) EMBO J. 14, 5179–5190. 7. Malorni, W., Rivabene, R., Straface, E., Rainaldi, G., Monti, D., Salvioli, S., Cossarizza, A., and Franceschi, C. (1995) Biochem. Biophys. Res. Commun. 207(2), 715–724. 8. Wang, E., and Pandey, S. (1995) J. Cell. Physiol. 163, 155–163. 9. Pandey, S., and Wang, E. (1995) J. Cell. Biochem. 58, 135–150. 10. Caellus, C., Helmberg, A., and Karin, M. (1994) Nature 370, 220–223. 11. Carmichael, J., Degraft, W. G., Gazder, A. F., Minna, J. D., and Mitchell, J. B. (1986) Cancer Res. 47, 936–942. 12. Laemmli, U. K. (1970) Nature 227, 680–685. 13. Towbin, H., Staehelin, T., and Gordon, J. (1979) Proc. Natl. Acad. Sci. USA 76, 4350–4354. 14. Liu, C. H. L., Liu, S., and Wang, E. (1993) Biochem. Biophys. Res. Commun. 195(3), 1371–1378. 15. Baxter, G. D., Collins, R. J., Harmon, B. V., Kumar, S., Pretentice, R. L., Smith, P. J., and Lavin, M. F. (1989) J. Pathol. 158, 123–129. 16. Agarwal, M. L., Clay, M. E., Harvey, E. J., Evans, H. H., Antunez, A. R., and Oleinick, N. L. (1991) Cancer Res. 51, 5993– 5996. 17. Takano, Y. S., Harmon, B. V., and Kerr, J. F. R. (1991) J. Pathol. 163, 329–336. 18. Miura, M., Friedlander, R. M., and Yuan, J. (1995) Proc. Natl. Acad. Sci. USA 92, 8318–8322. 19. He´bert, L., Pandey, S., and Wang, E. (1994) Exp. Cell Res. 210, 10–18. 20. Ohta, K., Toriyama, M., Miyazaki, M., Murofushi, H., Hosoda, S., Endo, S., and Sakai, H. (1990) J. Biol. Chem. 265, 3240– 3247. 21. Yang, W., Burkhart, W., Cavallius, J., Merrick, W. C., and Boss, W. F. (1993) J. Biol. Chem. 268, 392–398. 22. Yang, W., and Boss, W. F. (1994) J. Biol. Chem. 269, 3852– 3857. 23. Herrera, F., Correia, H., Triana, L., and Fraile, G. (1991) Eur. J. Biochem. 200, 321–327. 24. Hayashi, Y., Urade, R., Utsumi, S., and Kito, M. (1989) J. Biochem. (Tokyo) 106, 560–563. 25. Edmonds, B. T., Murray, J., and Condeelis, J. (1995) J. Biol. Chem. 270, 15222–15230. 26. Yang, F., Demma, M., Warren, V., Dharmawardhane, S., and Condeelis, J. (1990) Nature 347, 494–496.

Received February 19, 1997 Revised version received August 13, 1997

AID

ECR 3819

/

6i2ah$$204

01-14-98 09:41:48

eca