ORIGINAL ARTICLE
Journal of
Apoptosis Induction by eIF5A1
Involves Activation of the
Intrinsic Mitochondrial Pathway
Cellular
Physiology
ZHONG SUN, ZHENYU CHENG, CATHERINE A. TAYLOR,
BRENDAN J. MCCONKEY, AND JOHN E. THOMPSON*
Department of Biology, University of Waterloo, Waterloo, Ontario, Canada
The regulatory role of eukaryotic translation initiation factor 5A1 (eIF5A1) in apoptosis was examined using HT-29 and HeLa S3 cells.
eIF5A is the only known protein to contain the unusual amino acid, hypusine, and eIF5A1 is one of two human eIF5A family members. Two
observations indicated that eIF5A1 is involved in apoptosis. First, siRNA-mediated suppression of eIF5A1 resulted in inhibition of
apoptosis induced by various apoptotic stimuli, and second, adenovirus-mediated over-expression of eIF5A1 strongly induced apoptotic
cell death. A mutant of eIF5A1 incapable of being hypusinated also induced apoptosis when over-expressed indicating that unhypusinated
eIF5A1 is the pro-apoptotic form of the protein. Over-expression of eIF5A1 or of the mutant resulted in loss of mitochondrial
transmembrane potential, translocation of Bax to the mitochondria, release of cytochrome c, caspase activation, up-regulation of p53, and
up-regulation of Bim, a pro-apoptotic BH3-only Bcl-2 family protein. In addition, BimL and BimS, the pro-apoptotic alternative spliced
forms of Bim, were induced in response to over-expression of eIF5A1. Thus eIF5A1 appears to induce apoptosis by activating the
mitochondrial apoptotic pathway. Proteomic analyses indicated that, of 1,899 proteins detected, 131 showed significant changes in
expression ( P 0.05, expression ratio 1.5) within 72 h of eIF5A1 up-regulation. Among these are proteins involved in translation and
protein folding, transcription factors, proteins mediating proteolysis, and a variety of proteins known to be directly involved in apoptosis.
These observations collectively indicate that unhypusinated eIF5A1 plays a central role in the regulation of apoptosis.
J. Cell. Physiol. 223: 798–809, 2010. ß 2010 Wiley-Liss, Inc.
Eukaryotic translation initiation factor 5A (eIF5A) is a small
protein (18,000 Da) originally isolated from the ribosomal
fraction of rabbit reticulocyte lysates (Benne et al., 1978) and is
the only known cellular protein containing the unusual amino
acid, hypusine (Cooper et al., 1982). Formation of hypusine on
the eIF5A precursor is a polyamine-dependent reaction
involving two enzymes, deoxyhypusine synthase (DHS) and
deoxyhypusine hydroxylase (DOHH) (Park et al., 1982). In the
first step, the aminobutyl moiety of the polyamine, spermidine,
is transferred by DHS in an NADþ-dependent reaction to the
e-amino group of a specific lysine residue in eIF5A to form
deoxyhypusine (Park et al., 1982). This intermediate does not
accumulate and is immediately acted upon by DOHH, which
catalyzes the hydroxylation of the deoxyhypusine residue to
hypusine (Park et al., 1988). Hypusine-containing eIF5A is
thought to be ubiquitous in eukaryotes (Park et al., 1993a).
Although eIF5A was originally proposed to be a translation
initiation factor that stimulates the initiation phase of protein
synthesis by transient association with ribosomes (Benne et al.,
1978), this putative function has been questioned due to lack of
a clear correlation between eIF5A and general protein synthesis
(Kang and Hershey, 1994; Zuk and Jacobson, 1998). However,
there are two reports of hypusine- and translation-dependent
interaction of eIF5A with the structural components of the 80S
ribosome complex (Jao and Chen, 2006; Zanelli et al., 2006),
and a more recent study indicates that hypusinated eIF5A
promotes translation elongation (Saini et al., 2009). Numerous
other functions have also been attributed to eIF5A including
HIV Rev-mediated RNA export (Hofmann et al., 2001), mRNA
degradation (Zuk and Jacobson, 1998; Valentini et al., 2002), cell
proliferation (Schnier et al., 1991; Park et al., 1993b, 1998;
Hanauske-Abel et al., 1994), maintenance of yeast cell wall
integrity and actin cytoskeleton organization (Zanelli and
Valentini, 2005; Chatterjee et al., 2006), and regulation of
microtubule stability (Weir and Yaffe, 2004). In humans, eIF5A
is encoded by two genes, eIF5A1 (Smit-McBride et al., 1989) and
eIF5A2 (Guan et al., 2001; Jenkins et al., 2001). eIF5A1 mRNA
and protein appear to be present in all human tissues and cell
ß 2 0 1 0 W I L E Y - L I S S , I N C .
types. However, eIF5A2, which is 84% identical to eIF5A1 at the
amino acid level, exhibits restricted expression in normal
healthy tissue and is over-expressed in some cancers and
cancer cell lines (Jenkins et al., 2001; Clement et al., 2003, 2006),
prompting the suggestion that it may be an oncogene (Guan
et al., 2001, 2004; Jenkins et al., 2001; Clement et al., 2003).
Recent studies have also indicated a role for eIF5A in the
regulation of apoptosis in both a p53-dependent (Li et al., 2004)
and a p53-independent manner (Taylor et al., 2007). Apoptosis
is essential for organ development, tissue homeostasis, and the
elimination of defective cells in multi-cellular organisms
(Elmore, 2007). Two major apoptotic pathways have been
described for mammalian cells. The extrinsic death receptor
pathway is initiated by the association of death ligands with
membrane-bound receptors including TNFR1, Fas, and TRAIL
receptors 1 and 2, resulting in the formation of a death-inducing
signaling complex and subsequent activation of caspases
(Ashkenazi and Dixit, 1998; Peter and Krammer, 2003). The
intrinsic mitochondrial pathway is regulated by members of the
Bcl-2 protein family, which mediate mitochondrial outer
membrane permeabilization leading to the release of
apoptogenic mitochondrial proteins (Harris and Thompson,
2000; Zamzami and Kroemer, 2001) and caspase activation (Li
et al., 1997).
Additional Supporting Information may be found in the online
version of this article.
*Correspondence to: John E. Thompson, Department of Biology,
University of Waterloo, 200 University Ave. W., Waterloo,
Ontario, Canada N2L 3G1. E-mail: [email protected]
Received 4 August 2009; Accepted 11 January 2010
Published online in Wiley InterScience
(www.interscience.wiley.com.), 15 March 2010.
DOI: 10.1002/jcp.22100
798
eIF5A ACTIVATES INTRINSIC PATHWAY
Although over-expression of eIF5A1 has been shown to
induce apoptosis in cancer cell lines (Li et al., 2004; Taylor et al.,
2007), little is known about how eIF5A1-induced apoptosis
occurs. In this study, the pro-apoptotic function of
eIF5A1 was examined in HT-29 and HeLa S3 cells using
adenovirus-mediated delivery. Our data indicate that eIF5A1
induces apoptosis through activation of the intrinsic
mitochondrial pathway.
Materials and Methods
Cell culturing and transfection
HT-29, a human colon adenocarcinoma cell line, and HeLa S3, a
human cervical adenocarcinoma cell line, were kind gifts from Anita
Antes (University of Medicine and Dentistry of New Jersey).
HeLa S3 cells were cultured in Minimum Essential Medium Eagle
containing 0.1 mM non-essential amino acids, 1.0 mM sodium
pyruvate, and 10% fetal bovine serum (FBS). HT-29 cells were
maintained in RPMI-1640 supplemented with 1.0 mM sodium
pyruvate, 10 mM HEPES, and 10% FBS.
Cells were transfected with control siRNA, siRNA specifically
targeted to human eIF5A1 and siRNAs against human Bax and Bim
(Cell Signaling) using Lipfectamine 2000 (Taylor et al., 2004).
For adenovirus-mediated transfection, cells were seeded at
1 105 cells per well on a six-well tissue culture plate and infected
with adenovirus constructs [4,000 infectious unit (ifu)/cell for HT29 cells and 500 ifu/cell for HeLa S3 cells] 24 h later in medium
containing 2% FBS. Additional medium was added to the cells 4 h
after infection, and the concentration of FBS was brought to 10%.
Site-directed mutagenesis of eIF5A1 and construction of
adenoviral vectors
A site-specific mutation of eIF5A1 at lysine50 was obtained by PCR
using pHMeIF5A1 DNA (Taylor et al., 2007) as the template
following the protocol described previously (Ho et al., 1989). Using
the forward primer, 50 -GCCAAGCTTAATGGCAGATGATTTGG-30 , and the reverse primer, 50 -GTGCGCGCCAGTCTTCGAAGTAG-30 , a 158-bp fragment of the coding region of eIF5A1
containing a point-mutation-converting lysine50 to alanine as well as
an Sfu I restriction enzyme site was obtained. The remaining 347-bp
segment of the coding region was amplified with an Sfu I restriction
enzyme site using the forward primer, 50 -CTACTTCGAAGACTGGCGCGCAC-30 , and the reverse primer, 50 -CCTGAATTCCAGTTATTTTGCCATGG-30 . The two PCR products were
digested with Sfu I, ligated, and subcloned into the Hind III and EcoR I
sites of the pHM6 vector. The resultant construct containing
eIF5A1(K50A) was validated by sequencing.
Adenoviral (Ad) vectors (Adenovirus 5 serotype, E1 and
E3-deleted) expressing eIF5A1, eIF5A1(K50A), eIF5A2, or LacZ
were prepared using the AdMaxTM Hi-IQ system (Microbix
Biosystems Inc., Mississauga, ON) as described (Jin et al., 2008).
Induction and measurement of apoptosis
Apoptosis was induced in HT-29 cells or HeLa S3 cells by treatment
with Actinomycin D (0.5 mg/ml; Calbiochem, Gibbstown, NJ),
sodium nitroprusside (SNP; 3 mM; Sigma–Aldrich, Oakville, ON),
or the proteasome inhibitor MG-132 (5 mM; Sigma–Aldrich) or by
infection with Ad-eIF5A constructs. Cell death was measured using
the XTT cell viability assay (Roche Applied Science, Mississauga,
ON) or an ELISA kit (Roche Applied Science) according to the
manufacturer’s instructions. In some cases, cell death was also
measured using trypan blue. Trypsinized cells were stained for
5 min in phosphate-buffered saline (PBS) containing 0.1% (w/v)
trypan blue, and stained and unstained cells were counted using a
hemocytometer.
Apoptosis was measured using the DeadEnd Fluorometric
TUNEL assay (Promega, Nepean, ON) according to the
JOURNAL OF CELLULAR PHYSIOLOGY
manufacturer’s instructions for adherent cells. The cells were
subsequently stained with 5 mg/ml Hoechst 33358 (Sigma–Aldrich)
for 10 min at room temperature and visualized by fluorescence
microscopy (Carl Zeiss, Toronto, ON) using a UV filter to observe
the Hoechst-stained nuclei and a fluorescein filter to observe
apoptotic cells stained by TUNEL. Apoptosis was also measured
using the Annexin V-FITC apoptosis detection kit I (BD Bioscience,
Mississauga, ON) according to the manufacturer’s instructions.
Cells were detached by trypsinization, stained with Annexin
V-FITC and propidium iodide (PI), and sorted using a BD
FACSVantage SE system (BD Bioscience) with an argon laser
source. Fifteen thousand cells were counted, and the data were
analyzed using WinMDI 2.8 software. Cells at early and late stages
of apoptosis were defined as cells that were Annexin V-positive,
PI-negative, and Annexin V-positive, PI-negative, respectively.
Preparation of cell fractions
Cell lysates were obtained by washing trypsinized cells twice with
PBS and resuspending them in boiling lysis buffer (2% SDS and
50 mM Tris–HCl, pH 7.4). The lysates were sonicated, and protein
concentration was determined using a bicinchoninic acid protein
assay kit (Sigma–Aldrich) following the manufacturer’s
instructions.
Cytosolic and mitochondrial fractions were obtained by
treatment with digitonin (Saikumar et al., 1998). The cells were
trypsinized, washed in PBS, and resuspended in sucrose buffer
(250 mM sucrose, 10 mM Hepes, 10 mM KCl, 1.5 mM MgCl2, 1 mM
EGTA, and 1 mM EDTA) containing 0.025% (w/v) digitonin (Sigma–
Aldrich) and 1 protease inhibitor cocktail (Sigma–Aldrich). The
suspension was incubated at room temperature for 1 min, and the
cell lysates were cleared by centrifugation at 14,000g for 10 min at
48C giving rise to a supernatant comprising the cytosolic fraction.
The pellet of permeabilized cells was resuspended in 150 ml of
complete lysis buffer [2 M Thiourea, 7 M Urea, 30 mM Tris, 4%
(w/v) CHAPS (Sigma–Aldrich), 0.5% (v/v) Triton X-100 (Bioshop,
Burlington, ON), and 1 protease inhibitor cocktail] and incubated
on ice for 10 min with gentle shaking. The sample was then
centrifuged at 14,000g for 10 min giving rise to a supernatant
comprising the mitochondrial fraction. Protein was assayed using a
Bradford assay (Bioshop) following the manufacturer’s
instructions.
SDS–PAGE and Western blotting
For Western blotting, 5 mg protein samples were fractionated by
SDS–PAGE and transferred to Hypond-P PVDF membrane
(Amersham, GE Healthcare, Baie d’Urfe, QC). The blots were
probed with primary antibodies for b-actin (Calbiochem), eIF5A1
(BD Bioscience), eIF5A2 (Novus Biologicals, Littleton, CO),
cytochrome c (BD Bioscience), as well as Bad, Bax, Bcl-2, Bid, Bim,
p53, and Puma (Cell Signaling). HRP-conjugated anti-mouse IgG
(Sigma–Aldrich), anti-mouse IgM (Calbiochem), or anti-Rabbit
IgG (Calbiochem) were used as secondary antibodies, and
immune-complexes were visualized by chemiluminescence (ECL
Plus Western blotting detection kit; Amersham). In some cases,
membranes were stripped by incubation in stripping buffer
[100 mM 2-mercaptoethanol, 2% (w/v) SDS, and 62.5 mM
Tris–HCl, pH 6.7] for 30 min at 508C and re-probed with a second
antibody.
Measurement of mitochondrial transmembrane potential
(DCm) and caspase activation
DiOC6(3), a dye that permeates the plasma membrane, was used
to detect the loss of DCm (Petit et al., 1995) following the
induction of apoptosis. Cells were trypsinized and diluted in
medium containing 50 nM DiOC6(3) (Sigma–Aldrich). The cell–dye
mixture was incubated in the dark at 378C for 20 min. After
incubation, the cells were washed twice with PBS, resuspended in
PBS, and analyzed on a BD FACSVantage SE system (BD
799
800
SUN ET AL.
Bioscience) with an argon laser source. For each sample,
15,000 cells were acquired in list mode and analyzed with
WinMDI 2.8 software (Joseph Trotter, Scripps Research Institute,
La Jolla, CA).
Caspase activation was measured using caspases 3, 8, and 9
detection kits (Calbiochem) according to the manufacturer’s
instructions. Cells were trypsinized and resuspended in complete
medium containing FITC–DEVD–FMK, FITC–IETD–FMK, or
FITC–LEHD. Cell suspensions were incubated at 378C for 1 h, and
then sorted using a Coulter Epics XL–MCL flow cytometry system
(Beckman Coulter, Mississauga, ON) with an argon laser source.
For each sample, 25,000 cells were acquired in list mode and
analyzed with WinMDI 2.8. To inhibit the activation of caspases,
cells were treated with the pan caspase inhibitor, Z–VAD–FMK
(20 mM; Calbiochem). Activation of caspase 3 was measured to
assess the inhibition.
Proteomic profiling
Analytical difference in-gel electrophoresis (DIGE) analyses were
performed according to the supplier’s instructions (GE Healthcare,
Baie d’Urfe, QC) as described (Cheng et al., 2009). Cell lysates
were harvested in cold lysis buffer (7 M urea, 2 M thiourea, 30 mM
Tris, 4% CHAPS, 1 protease inhibitor cocktail) 24, 48, and 72 h
after infection with Ad-eIF5A1 or Ad-LacZ, a negative control, and
cleared of cell debris by sonication and centrifugation. The samples
(internal standard, control sample, treated sample—50 mg protein
equivalent) were labeled with 400 pmol of different CyDye DIGE
fluors (GE Healthcare), mixed together and loaded on Immobiline
DryStrips (GE Healthcare) (pH 3-11NL, 24 cm). Isoelectric
focusing (IEF) performed using an Ettan IPGphor II (GE Healthcare)
was followed by SDS–PAGE fractionation on 12% gels. Four
analytical gels were run for each time point. Fluorescence was
measured using a Typhoon 9400 Imager system (GE Healthcare),
and matching and statistical analyses of the protein spots were
performed using DeCyder V 6.0 software (GE Healthcare).
Expression ratios were calculated as R ¼ (treated/control) for
up-regulated proteins, and as R ¼ (control/treated) for
down-regulated proteins. Statistical significance was calculated
using the DeCyder software and corrected for multiple
hypothesis testing. Proteins significantly up- or down-regulated
were deemed to be those with an expression ratio 1.5
( P 0.05).
Fractionation of proteins for sequencing was performed on
preparative 2D gels (Cheng et al., 2009). Cell lysates (0.5 mg of
protein) were loaded on Immobiline DryStrips (pH 3-11NL,
24 cm), fractionationed, and stained overnight with Sypro Ruby
(Bio-Rad Laboratories, Mississauga, ON). Spots of interests were
excised, washed with water, dehydrated with 100% acetonitrile
(ACN), and air-dried. The gel pieces were rehydrated for 10 min in
a 10 ng/ml trypsin solution (Promega) at a ratio of approximately
1:10 (w/w) of trypsin:protein. Fifty microliters of 50 mM
NH4HCO3 (pH 8.0) was added to each gel piece, and the proteins
were digested at 378C for 18 h. The peptides were extracted by
three cycles of vortexing and sonication (2 min of vortexing
followed by 5 min of sonication) and dried in a SpeedVac,
Instruments Inc., Hicksville, NY. The samples were resuspended in
5 ml of 50% ACN containing 0.1% formic acid and analyzed by
LC–MS/MS (mass spectrometry) using an Applied Biosystems QTRAP system (Applied Biosystems Inc., Foster City, CA). Peak lists
were generated and processed using Analyst software version 1.4.1
(Applied Biosystems). Proteins were identified using PEAKS
software 3.1 (Bioinformatics Solutions Inc., Waterloo, ON), which
combines auto de novo sequencing and database searching (NCBI
nr database was downloaded on February 28, 2007). The search
engines were programmed for parental and fragment mass errors
of 0.2 and 0.1 Da, respectively, with trypsin as the digestive enzyme
(one missed cleavage allowed) and carbamidomethylation and
methionine oxidation (post-translational modification) as the fixed
JOURNAL OF CELLULAR PHYSIOLOGY
and variable modifications, respectively. Protein identification was
further confirmed using MASCOT MS/MS ion search and
peptide-fingerprinting (Perkins et al., 1999). The database,
digestive enzyme, post-translational modification, missed cleavage
setting, and peptide error tolerance were the same as those
used for the PEAKS software analysis. The criterion for acceptance
of an identification score above the threshold in MASCOT
searches was P 0.05. When there was more than one
identification for a given protein, the identification selected was
based on the best match between calculated pI and molecular
mass values and those obtained empirically from the 2D gel. All
identified proteins contained at least one signature-specific,
non-redundant peptide.
Statistics
Independent Student’s t-test or one-way ANOVA followed by
post hoc Scheffe’s test were used for statistical analysis. A P-value
of 0.05 was considered to be statistically significant.
Results
siRNA suppression of eIF5A1 inhibits apoptosis
HT-29 cells were transfected with eIF5A1 siRNA or control
siRNA 48 h prior to treatment with SNP, a nitric oxide (NO)
donor that is known to induce cytotoxicity leading to apoptosis
(Wang et al., 2007). Twenty-four hours after SNP treatment,
cell viability was determined by the XTT assay. There was a 76%
loss in viability of control cells exposed to SNP, whereas cells
treated with eIF5A1 siRNA only exhibited a 25% loss in viability
(Fig. 1A). Moreover, the eIF5A1 siRNA-mediated protection
against SNP-induced cytotoxicity correlated with a strong
reduction in eIF5A1 protein levels (Fig. 1B).
siRNA suppression of eIF5A1 also provided protection
against MG-132-induced apoptosis. MG-132 is a potent
irreversible proteasome inhibitor that has been shown to
induce apoptosis in numerous cell lines (Drexler, 1997;
MacLaren et al., 2001). HT-29 cells were transfected with
eIF5A1 siRNA or control siRNA for 24 or 48 h prior to
treatment with MG-132 for 24 h. No protection against
MG-132-induced apoptosis was observed in cells transfected
with eIF5A1 siRNA for 24 h prior to MG-132 exposure
(Fig. 1C). However, for cells transfected with eIF5A1 siRNA for
48 h prior to MG-132 treatment, there was a 60% reduction
in the percentage of apoptotic cells compared with cells
transfected with control siRNA (Fig. 1C). Western blot analysis
indicated that the protein level of eIF5A1 was strongly reduced
in eIF5A1 siRNA-transfected cells, particularly 72 h after
transfection (Fig. 1D). Moreover, the protection against
MG-132-induced apoptosis conferred by eIF5A1 siRNA
correlated with the greater reduction of eIF5A1 protein
observed after 72 h of transfection compared to 48 h.
Suppression of eIF5A1 by siRNA also afforded protection
against Actinomycin D-induced apoptosis in HeLa S3 cells
(Supplementary Fig. 1). Immunoblot analysis of cell lysates using
eIF5A1-specific antibody confirmed that levels of eIF5A1
protein were dramatically reduced within 72 h in cells
transfected with eIF5A1 siRNA compared with those
transfected with control siRNA (Supplementary Fig. 1A).
Moreover, this correlated with strong protection against cell
death. Within 24 h of administering Actinomycin D, there
was 56% reduction in cell viability for cells transfected with
control siRNA, whereas cells treated with eIF5A1 siRNA only
exhibit 18% loss in viability (Supplementary Fig. 1B).
Over-expression of eIF5A1 induces apoptosis
eIF5A1 and eIF5A1(K50A), a mutant of eIF5A1 that cannot
be hypusinated, were over-expressed in HT-29 and HeLa S3
cells using adenovirus as a transfection agent. Adenovirus-
eIF5A ACTIVATES INTRINSIC PATHWAY
Fig. 1. Suppressing the expression of eIF5A1 using specific siRNA reduces SNP- and MG-132-induced cell death. A1, eIF5A1 siRNA; con, control
siRNA; un, untranfected; , without; R, with. A: HT-29 cells were transfected with either eIF5A1 siRNA or control siRNA. Forty-eight hours after
transfection, the cells were treated with 3 mM SNP for an additional 24 h. Cell viability was measured using the XTT assay. The data are normalized
to the value for untransfected cells, which was set at 100%. Values are means W SE (MP < 0.001; n U 3). B: Cell lysates from HT-29 cells treated as
described in (A) were collected and subjected to immunoblot analysis using an eIF5A1-specific antibody. Equal loading of the samples was evaluated
by probing with antiactin. The results shown are representative of those obtained in three independent experiments. C: HT-29 cells were
transfected with either eIF5A1 siRNA or control siRNA. After 24 or 48 h, the cells were treated with 5 mM MG-132 for another 24 h. Apoptosis was
measured using the Cell Death Detection ELISA kit. Data are expressed as a percentage of the value for cells transfected with control siRNA and
treated with 5 mM MG-132. Values are means W SE (MP < 0.001; n U 4). D: Immunoblot analysis of lysates of HT-29 cells treated as in (C) using specific
anti-eIF5A1 and antiactin. The results are representative of those obtained in three independent experiments.
expressing LacZ (Ad-LacZ) was used as a negative control.
b-galactosidase staining indicated that infection of HT-29 cells
and HeLa S3 cells with Ad-LacZ at 4,000 and 500 ifu/cell,
respectively, resulted in transfection rates of more than 90%
(data not shown). Adverse effects on cell morphology were not
observed at this level of infection. In HT-29 cells infected with
4,000 ifu/cell of Ad-eIF5A1 or Ad-eIF5A1(K50A), the protein
levels of eIF5A1 and eIF5A1(K50A) were strongly up-regulated
within 48 h of infection and remained high for at least 72 h after
infection (Fig. 2A,B). Similar levels of eIF5A1 up-regulation were
achieved in HeLa S3 cells following infection with adenovirus
constructs at 500 ifu/cell (Fig. 2B, Supplementary Fig. 2).
It has been demonstrated previously by 2D Western blot
analysis that in normally growing HT-29 cells eIF5A is present
almost exclusively in its hypusinated form (Taylor et al., 2007),
which is consistent with the fact that virtually all cellular eIF5A
protein is hypusinated immediately after synthesis (Park et al.,
1993a). In the same study, it was demonstrated that in
Ad-eIF5A1(K50A)-infected HT-29 cells, there is a large excess
of unhypusinated eIF5A1 and comparatively little hypusinated
eIF5A1 indicating that the rapidly synthesized trans-eIF5A1 is
not hypusinated (Taylor et al., 2007). This is to be expected
since the point mutation in eIF5A1(K50A) precludes
hypusination. However, there was also a large disproportionate
accumulation of unhypusinated relative to hypusinated eIF5A1
in HT-29 cells infected with Ad-eIF5A1 (Taylor et al., 2007),
which can be attributed to rate-limiting levels of DOHH and
DHS available to modify the large amounts of newly translated
eIF5A1 (Clement et al., 2006; Park et al., 2006). Thus, the
up-regulated pools of eIF5A1 in cells infected with Ad-eIF5A1
or Ad-eIF5A1(K50A) (Fig. 2A,B) contain very little hypusinated
eIF5A1 relative to unhypusinated eIF5A1.
JOURNAL OF CELLULAR PHYSIOLOGY
Adenovirus-mediated over-expression of eIF5A1 or
eIF5A1(K50A) induced apoptosis detected using the TUNEL
assay in HT-29 cells. Following infection with Ad-eIF5A1 or AdeIF5A1(K50A), approximately 23% of cells were undergoing
apoptosis 48 h later (Fig. 3A,B). In contrast, <1% of control cells
infected with Ad-lacZ were undergoing apoptosis within this
time frame (Fig. 3A,B). The cytotoxic effect of eIF5A1
adenoviral infection was confirmed by XTT assay 96 h after
infection. The viability of cells infected with Ad-eIF5A1 or
Ad-eIF5A1(K50A) was reduced by 77% in comparison with the
viability of cells infected with Ad-LacZ (Fig. 3C). HT-29 cells
infected with adenovirus constructs of eIF5A1 or
eIF5A1(K50A) were also stained with Annexin V/PI. Within
48 h of infection, 27% of the cells were undergoing apoptosis,
and by 72 h after infection, 62% of the cells were undergoing
apoptosis (Supplementary Fig. 3). Treatment with Actinomycin
D served as a positive control (Supplementary Fig. 3).
The induction of apoptosis in HeLa S3 cells by Ad-eIF5A1 and
Ad-eIF5A1(K50A) was also measured by flow cytometry after
staining with Annexin V-FITC/PI. The cells were infected with
500 ifu/cell adenovirus constructs, and apoptosis was measured
at intervals of 24, 48, and 72 h thereafter. Treatment with
Actinomycin D served as a positive control. There was minimal
induction of apoptosis within 24 h of infection (Fig. 3D,E), but by
48 h there was clear evidence of apoptosis for cells infected with
Ad-eIF5A1 and Ad-eIF5A1(K50A). Specifically, the proportion
of cells at early (Annexin Vþ/PI) and late stages (Annexin Vþ/
PIþ) of apoptosis was 21.3% for Ad-eIF5A1-infected cells and
32.3% for Ad-eIF5A1(K50A)-infected cells (Fig. 3E). At 72 h
after infection with Ad-eIF5A1 or Ad-eIF5A1(K50A), 60–65%
of the cells were undergoing apoptosis with a large percentage
at the late stage of apoptosis (Fig. 3E).
801
802
SUN ET AL.
Fig. 2. Infection with Ad-eIF5A1, Ad-eIF5A1(K50A), or Ad-eIF5A2
induces up-regulation of eIF5A expression. A1, Ad-eIF5A1; M,
Ad-eIF5A1(K50A); A2, Ad-eIF5A2; Z, Ad-lacZ. A: HT-29 cells were
infected with various adenovirus constructs at 4,000 ifu/cell. At 48 or
72 h after infection, cell lysates were collected and subjected to
immunoblot analysis using antibodies against eIF5A1, eIF5A2, and
actin. B: HeLa S3 cells and HT-29 cells were infected with 500 ifu/cell
and 4,000 ifu/cell of adenovirus constructs, respectively. Forty-eight
hours after infection, cell lysates were collected and subjected to
immunoblot analysis using antibodies against eIF5A1 and actin.
The results shown are representative of those obtained in three
independent experiments. Note that anti-eIF5A1 cross-reacted with
eIF5A2, and that anti-eIF5A2 cross-reacted with eIF5A1.
Over-expression of eIF5A1 induces activation of
caspases 3, 8, and 9
In order to determine whether Ad-eIF5A1- and AdeIF5A1(K50A)-induced apoptosis involves activation of
caspases, the effects of their over-expression on the activation
of executioner caspase 3 as well as initiator caspases 8 and 9
were examined in HeLa S3 cells over a period of 72 h. All three
caspases were activated in 25–45% of the cells, depending on
the caspase, within 48 h of infection with Ad-eIF5A1 or
Ad-eIF5A1(K50A), and in 55–75% of the cells, depending on
the caspase, within 72 h of infection (Fig. 4 and Supplementary
Fig. 4). Cells treated with Actinomycin D and cells infected with
Ad-lacZ served as positive and negative controls, respectively
(Fig. 4 and Supplementary Fig. 4). To determine whether
apoptosis induced by over-expression of eIF5A1 is caspasedependent, HeLa S3 cells infected with adenovirus eIF5A1 were
co-treated with the pan caspase inhibitor, Z–VAD–FMK. In
the presence of Z–VAD–FMK, the activation of caspase 3 and
the induction of apoptosis were inhibited by 90% in AdeIF5A1- and eIF5A1(K50A)-infected cells (Fig. 5), indicating that
eIF5A1-mediated apoptosis is caspase-dependent.
Over-expression of eIF5A1 induces loss of mitochondrial
transmembrane potential (DCm), release of
cytochrome c, and translocation of Bax
It is clear from the data in Figure 4B that caspase 9, the initiator
caspase for the mitochondrial apoptotic pathway, is activated in
HeLa S3 cells over-expressing eIF5A1. Accordingly, the
possibility that up-regulation of eIF5A1 results in the loss of
DCm, a well-established apoptotic trait (Petit et al., 1995),
was examined. DCm was measured in HeLa S3 cells infected
with Ad-eIF5A1 or Ad-eIF5A1(K50A) using the membrane
potential-sensitive dye, DiOC6(3), followed by fluorescenceJOURNAL OF CELLULAR PHYSIOLOGY
activated cell-sorting (FACS) analysis. Over the entire 72 h
experimental period, the dye was retained within mitochondria
in 97% of control cells infected with Ad-lacZ, reflecting
structural integrity of the mitochondrial membrane system
(Fig. 6). In contrast, a gradual decline in the number of cells with
normal DCm was observed following infection with the eIF5A1
or eIF5A1(K50A) adenovirus constructs with 50% of the cells
exhibiting loss of DCm within 72 h (Fig. 6). For cells treated with
Actinomycin D, which served as a positive control, there was a
similar decline in the proportion of cells exhibiting normal DCm
within 24 h (Fig. 6). Release of cytochrome c from mitochondria
into the cytosol is an early event in the mitochondrial pathway
of apoptosis. Once in the cytosol, cytochrome c binds to Apaf-1
to form the cytochrome c:Apaf-1 complex, which directly
induces the activation of caspase 9 (Li et al., 1997). To
determine whether the effects of up-regulated eIF5A1 on DCm
resulted in the induction of cytochrome c release, cytosolic and
mitochondrial fractions isolated from HeLa S3 cells infected
with Ad-eIF5A1 were subjected to immunoblot analysis using a
specific antibody against cytochrome c. Prior to infection with
Ad-eIF5A1, cytochrome c was detected only in the
mitochondrial fraction indicating that it was still entrapped in
the mitochondrial intermembrane space (Fig. 7A). Cytochrome
c remained in the mitochondrial fraction for up to 31 h after
infection with Ad-eIF5A1. However, by 48 h after infection,
translocation of cytochrome c from mitochondria to the
cytosolic fraction was evident, and after 55 h, there were almost
equal proportions of cytochrome c in the mitochondrial
fraction and in the cytosol (Fig. 7A). By 72 h after infection, the
bands corresponding to cytochrome c were weak (Fig. 7A and
Supplementary Fig. 2), presumably because the cells were dying.
These results corroborate the observation that loss of DCm,
which precedes cytochrome c release, was also not evident
until between 24 and 48 h after infection (Fig. 6B).
Localization of the pro-apoptotic Bcl-2 family protein, Bax,
which usually resides as a monomer in the cytosol but
oligomerizes and translocates to the mitochondrial outer
membrane in response to certain apoptotic stimuli (Schinzel
et al., 2004), was also examined following infection with AdeIF5A1. Prior to infection, most of the Bax protein was localized
in the cytosol, with only a small amount detectable in the
mitochondrial fraction (Fig. 7A). This is consistent with reports
that Bax can also loosely associate with the mitochondrial outer
membrane as a monomer in untreated HeLa cells (Antonsson
et al., 2001). However, within 24 h of infection with Ad-eIF5A1,
there was clear evidence for migration of Bax from the cytosol
to the mitochondrial fraction, and by 48 h, almost all of the Bax
protein was associated with the mitochondrial fraction
(Fig. 7A). Moreover, the translocation of Bax to the
mitochondrial fraction from Ad-eIF5A1-treated cells clearly
preceded the release of cytochrome c from the mitochondrial
fraction (Fig. 7A).
Effects of eIF5A1 over-expression on levels of Bcl-2
family proteins
The effects of eIF5A1 up-regulation on the expression levels of
the pro-apoptotic Bcl-2 family members, Bid, Puma, Bim, Bad,
and Bax, were examined in HeLa S3 cells. In spite of the ability of
eIF5A1 to induce apoptosis, the pro-apoptotic BH3-only Bcl-2
proteins, Bid and Puma, were down-regulated in
Ad-eIF5A1- and Ad-eIF5A1(K50A)-infected cells (Fig. 7B).
Moreover, cleaved Bid was not observed (Fig. 7B) indicating
that eIF5A-induced activation of caspase 8 does not directly
lead to activation of intrinsic mitochondrial pathway. Bim, a
pro-apoptotic BH3-only Bcl-2 family protein, was up-regulated
within 48 h in Ad-eIF5A1-infected cells, and there was also
induction of pro-apoptotic BimL and BimS, alternative splicing
forms of Bim, in both Ad-eIF5A1- and Ad-eIF5A1(K50A)-
eIF5A ACTIVATES INTRINSIC PATHWAY
Fig. 3. Over-expression of eIF5A induces apoptosis in HT-29 and HeLa S3 cells. A1, Ad-eIF5A1; M, Ad-eIF5A1(K50A); A2, Ad-eIF5A2; Z, Ad-lacZ;
UN, untreated; ActD, Actinomycin D. A: HT-29 cells were infected with eIF5A adenovirus constructs. After 48 h, the cells were fixed and stained
with TUNEL and Hoechst 33258. For each sample, the same field was observed by fluorescence microscopy using two different filters. A fluorescein
filter was used to visualize TUNEL-stained apoptotic cells and a UV filter was used to visualize Hoeschst-stained nuclei. The results shown are
representative of those obtained in three independent experiments. Magnification: 100T. B: The percentage of HT-29 cells undergoing apoptosis
was calculated by dividing the number of TUNEL-labeled apoptotic cells by the number of Hoechst-stained cells. The values are means W SE
(MP < 0.001; n U 3). C: HT-29 cells were infected with eIF5A adenovirus constructs. After 4 days, cell viability was measured using the XTT assay. The
values for Ad-eIF5A infected cells are normalized to that for Ad-lacZ infected cells, which was set at 100%. Values are means W SE (MP < 0.001; n U 3).
D: Flow cytometry analysis of Annexin V-FITC/PI-labeled HeLa S3 cells infected with eIF5A adenovirus constructs for 24, 48, or 72 h. Cells treated
with 0.5 mg/ml Actinomycin D served as a positive control. Results shown are representative of those obtained in 4 independent experiments.
E: Percentage of HeLa S3 cells treated as described in (D) undergoing apoptosis. Values shown are means W SE (n U 4).
infected cells (Fig. 7B). No significant changes in the levels of Bad
were apparent following infection (Fig. 7B). The expression
levels of Bax were also unchanged (Fig. 7B) notwithstanding the
fact that up-regulation of eIF5A1 induced migration of Bax from
the cytosol to mitochondria (Fig. 7A). However, it has been
JOURNAL OF CELLULAR PHYSIOLOGY
suggested previously that the localization of Bax on the
mitochondrial outer membrane is in itself sufficient to induce
mitochondrial outer membrane permeabilization
(Jürgensmeier et al., 1998; Shimizu et al., 1999). Levels of Bcl-2,
an anti-apoptotic member of the Bcl-2 family, were low in
803
804
SUN ET AL.
Fig. 4. Over-expression of eIF5A activates caspases 3, 8, and 9. HeLa
S3 cells were infected with eIF5A adenovirus constructs for 48 or 72 h,
or treated with 0.5 mg/ml Actinomycin D for 24 h. Cells were
subsequently collected, incubated with either FITC–DEVD–FMK
(detects activated caspase 3), FITC–IETD–FMK (detects activated
caspase 8), or FITC–LEHD–FMK (detects activated caspase 9) and
analyzed by flow cytometry. Data shown are percentages of cells with
activated caspases 3, 8, and 9. A1, Ad-eIF5A1; M, Ad-eIF5A1(K50A);
A2, Ad-eIF5A2; Z, AdlacZ; UN, untreated; ActD, Actinomycin D.
Values shown are means W SE (n U 4).
untreated HeLa S3 cells and remained low to undetectable
following infection with Ad-eIF5A1 or Ad-eIF5A1(K50A)
(Fig. 7C). This is in keeping with the finding that Bcl-2
transcription is curtailed in HeLa cells, the progenitor of HeLa
S3 cells (Cleary et al., 1986; Odin et al., 2001).
To determine whether eIF5A1-induced apoptosis is
dependent on Bax or Bim, their expression in HeLa S3 cells was
suppressed by treatment with specific siRNAs for 24 h before
infection with Ad-eIF5A1, Ad-eIF5A1(K50A), or Ad-LacZ.
Apoptosis was measured 48 h after infection. Apoptosis of cells
transfected with either Bax or Bim siRNA prior to infection
with Ad-eIF5A1 or Ad-eIF5A1(K50A) was reduced to levels
approximating the background levels obtained for untreated
cells and cells treated with control siRNA only, whereas cells
infected without prior siRNA treatment exhibited normal
levels of apoptosis (Fig. 8A). This correlated with strong
suppression of Bax and Bim expression by their respective
siRNAs (Fig. 8B,C). These observations suggest that both Bax
and Bim are involved in eIF5A1-induced apoptosis. Cells
pretreated with control siRNA and infected with Ad-LacZ, a
negative control, also exhibited background levels of apoptosis
(Fig. 8A).
Over-expression of eIF5A1 induces up-regulation of p53
To evaluate the effect of eIF5A1 up-regulation on the tumor
suppressor protein, p53 and its homologs, HeLa S3 cells were
infected with adenovirus constructs of eIF5A1 and
eIF5A1(K50A), and total protein was analyzed by Western
blotting using polyclonal antibody against p53 at intervals of 24,
48, and 72 h after infection. The level of p53 in untreated cells
was low (Fig. 7B). This reflects the fact that HeLa cells, the
progenitor of HeLa S3 cells, were transformed with human
papillomavirus (HPV) whose E6 gene product interacts with
p53, resulting in its rapid degradation through the
ubiquitin–proteasome pathway (Scheffner et al., 1990; Thomas
et al., 1996). However, a modest increase in p53 expression of
HeLa S3 cells was observed within 72 h of infection with
Ad-eIF5A1(K50A) (Fig. 7B). In addition, there was strong
up-regulation of an 80 kDa protein within 48 h in Ad-eIF5A1and Ad-eIF5A1(K50A)-infected cells (Fig. 7B). Since the p53
antibody used in this analysis is polyclonal, it is possible that this
up-regulated protein is p73, a homolog of p53. p73 is known to
be functional in HeLa cell lines because, unlike p53, it is not
inactivated by the products of viral oncogenes such as HPV E6
(Marin et al., 1998). There are two alternatively spliced p73
mRNAs giving rise to p73a, the full length protein (MW
80 kDa), and p73b, a smaller version of the protein (MW
70 kDa) (Kaghad et al., 1997; Marin et al., 1998). Only putative
p73b was detectable in untreated cells, but within 72 h of
infection with Ad-eIF5A1 or Ad-eIF5A1(K50A), both putative
p73a and p73b were up-regulated (Fig. 7B).
Proteomic analysis of eIF5A1-induced changes in
protein expression
Fig. 5. Ad-eIF5A-induced apoptosis is caspase-dependent. HeLa S3
cells infected with eIF5A adenovirus constructs were co-treated with
the pan caspase inhibitor, Z–VAD–FMK. Seventy-two hours later,
apoptosis (A) and caspase 3 activation (B) was assessed. Values shown
are means W SE (MP < 0.05; n U 4). A1, Ad-eIF5A1; M, AdeIF5A1(K50A); A2, Ad-eIF5A2; Z, Ad-lacZ; UN, untreated; ActD,
Actinomycin D.
JOURNAL OF CELLULAR PHYSIOLOGY
Proteomic tools were used to assess eIF5A1-induced changes
in protein expression in HeLa S3 cells. Within 24 h of
transfection with Ad-eIF5A1, 62 proteins out of a total number
of 1,846 that were detectable on analytical gels showed
significant changes in expression ( P 0.05, expression ratio
1.5). Of these, 32 were up-regulated, and 30 were
down-regulated. By 72 h after transfection, 131 proteins from a
total of 1899 detected showed significant changes in
expression, 90 displaying up-regulation and 41 downregulation. Of these, 49 were successfully identified by 2D
preparative gel electrophoresis and MS, and categorized into
eight groups (I–VIII) based on functional annotation
(Supplementary Fig. 5 and Supplementary Table 1).
eIF5A ACTIVATES INTRINSIC PATHWAY
Fig. 6. Effect of over-expression of eIF5A on mitochondrial transmembrane potential. A1, Ad-eIF5A1; M, Ad-eIF5A1(K50A); A2, Ad-eIF5A2; Z,
Ad-lacZ; UN, untreated; ActD, Actinomycin D. A: Histograms of labeled cells. HeLa S3 cells were infected with eIF5A adenovirus constructs. At
intervals of 24, 48, and 72 h after infection, cells were labeled with DiOC6(3) and analyzed by FACS. Cells treated with Actinomycin D or infected
with Ad-lacZ served as positive and negative controls, respectively. Results shown are representative of those obtained in four independent
experiments. B: Percentage of cells with normal mitochondrial transmembrane potential. Values shown are means W SE (n U 4).
Four eIF5A1 proteins were identified (proteins 1 through 4,
Supplementary Table 1 and Supplementary Fig. 5), all showing
up-regulation. MS sequencing indicated that protein 1 is
acetylated (K47) unhypusinated eIF5A1, protein 2 is
unacetylated unhypusinated eIF5A1, and proteins 3 and 4 are
truncated versions of proteins 1 and 2, respectively,
MS data illustrating this for proteins 1 and 2 are provided in
Supplementary Figure 6. Acetylation of eIF5A1 K47 has been
reported previously for HeLa cells (Klier et al., 1995). As well,
JOURNAL OF CELLULAR PHYSIOLOGY
the finding that all of the up-regulated eIF5A1 proteins are
unhypusinated is consistent with the fact that adenoviralmediated over-expression of eIF5A is known to inhibit
hypusination (Taylor et al., 2007). Proteins 3 and 4 were much
less abundant than proteins 1 and 2 (Supplementary Fig. 5),
which may reflect the fact that they appear to have been formed
from proteins 1 and 2. Proteins 5 through 8 of Group I are
heterogeneous nuclear ribonucleoproteins (hnRNP), including
hnRNP A2 and hnRNP L isoforms a and b, which were down-
805
806
SUN ET AL.
Fig. 7. Over-expression of eIF5A activates the mitochondrial apoptotic pathway. A: HeLa S3 cells were infected with Ad-eIF5A1. At the indicated
times after infection, cytosolic and mitochondrial fractions were isolated, and 5 mg of protein were fractionated by SDS–PAGE and analyzed by
Western blotting using antibodies against cytochrome c and Bax. C, cytosolic fraction; M, mitochondrial fraction. B: HeLa S3 cells were infected
with eIF5A adenovirus constructs. At the indicated times after infection, cell lysates were collected and subjected to Western blot analysis for p53,
Bid, Puma, Bim, Bax, Bad, or (C)Bcl-2. Each lane contained7 mg ofprotein. Results shown are representative ofthose obtained in three independent
experiments. A1, Ad-eIF5A1; M, Ad-eIF5A1(K50A); A2, Ad-eIF5A2; Z, Ad-lacZ.
regulated, and the apoptosis activator, hnRNP H1 (Garneau
et al., 2005), which was up-regulated (Supplementary Table 1).
Protein disulfide isomerase (protein 9) and ubiquilin (protein
11) of group II (protein folding and degradation), which both
showed increased expression, are known to be up-regulated in
the event of ischemia and thought to play roles in ER stressinduced apoptosis (Ko et al., 2002). ER-60 protease (protein 10)
and ubiquitin-conjugation enzyme E2D (protein 12) were also
up-regulated (Supplementary Table 1). Proteins involved in
transcription (Group III), mitochondrial proteins (Group IV),
stress proteins (Group V), and cell structure proteins (Group
VI) also showed changes in expression following overexpression of eIF5A1 (Supplementary Table 1). As well, galectin
and hemoglobin alpha, proteins 43 and 47, respectively, of
Group VIII, exhibited increased expression (Supplementary
Table 1) and are known to be pro-apoptotic (Perillo et al., 1995;
Blaser et al., 1998; Brachat et al., 2002). Albumin and fetuin
(Group VII) were detectable and showed changes in expression
following eIF5A1 over-expression (Supplementary Table 1), but
these are presumed to be components of the growth medium.
eIF5A2 mimics the pro-apoptotic effects of eIF5A1
Over-expression of eIF5A2, a second isoform of human eIF5A
that is rarely expressed, also induced apoptosis. Infection of
cells with Ad-eIF5A2 resulted in strong over-expression of
eIF5A2 (Fig. 2) and ensuing caspase-dependent induction of
apoptosis (Figs. 3 and 5). Moreover, as for eIF5A1-induced
apoptosis, this was accompanied by activation of caspases 3, 8,
and 9 (Fig. 4), loss of DCm (Fig. 6), and induction of proapoptotic BimL and BimS (Fig. 7B). eIF5A2-induced apoptosis
also proved to be dependent upon Bax and Bim (Fig. 8).
Discussion
A role for eIF5A1 in apoptosis is supported by the finding that
siRNA-mediated inhibition of its expression provides
JOURNAL OF CELLULAR PHYSIOLOGY
protection against apoptosis induced by exposure to TNF-a
(Taylor et al., 2004) and cytotoxic agents such as Actinomycin D
(Taylor et al., 2007). The efficacy of eIF5A1 siRNA as an
inhibitor of apoptosis was confirmed in this study using a NO
donor as well as a proteasome inhibitor as pro-apoptotic
stimuli. In both cases, suppression of eIF5A1 provided
protection against the induction of programmed cell death.
Conversely, over-expression of eIF5A1 has been shown to
induce apoptosis (Li et al., 2004; Taylor et al., 2007), further
supporting the contention that it is apoptogenic. This study
sheds some light on how eIF5A1 engages in apoptosis by
demonstrating that it facilitates activation of the intrinsic
mitochondrial apoptotic pathway. Moreover, these effects of
eIF5A1 are mirrored by eIF5A1(K50A), a mutant of the protein
that cannot be hypusinated, indicating that it is unhypusinated
eIF5A1 that is able to mobilize the intrinsic pathway of
apoptosis.
Unhypusinated eIF5A1 is apoptogenic
The notion that hypusinated and unhypusinated eIF5A may have
different functions stems in part from the fact that, whereas the
hypusination reaction has been correlated with cell
proliferation (Schnier et al., 1991; Byers et al., 1994; Park et al.,
1998), there is growing evidence that the accumulation of
unhypusinated eIF5A1 is correlated with apoptosis. For
example, inhibitors of DHS and DOHH, which reduce
hypusinated eIF5A1 and increase levels of unmodified eIF5A1,
have been shown to induce cell cycle arrest and, in some cell
lines, ensuing apoptosis (Park et al., 1993b; Hanauske-Abel
et al., 1994; Caraglia et al., 2003; Jin et al., 2003). Induction of
apoptosis in a leukemic cell line by treatment with a proteasome
inhibitor proved to be accompanied by enhanced levels of
unhypusinated eIF5A1 (Jin et al., 2003). Beyond such
correlations, the demonstration that mutants of eIF5A1
incapable of being hypusinated are apoptogenic provides more
eIF5A ACTIVATES INTRINSIC PATHWAY
Fig. 8. Bax and Bim are essential for eIF5A1-induced apoptosis. HeLa S3 cells transfected with Bax, Bim, or control siRNA for 24 h were infected
with different eIF5A adenovirus constructs. After an additional 48 h, cells were collected and analyzed. A1, Ad-eIF5A1; M, Ad-eIF5A1(K50A);
A2, Ad-eIF5A2; Z, Ad-lacZ; UN, untreated; ActD, Actinomycin D; con, control. A: Annexin V/PI assay of apoptosis. B: Western blot analysis
of Bax in cell lysates; each lane contained 8 mg of protein. C: Western blot analysis of Bim in cell lysates; each lane contained 8 mg of protein.
Values shown are means W SE (n U 4).
direct evidence of a role for unhypusinated eIF5A1 in apoptosis.
In this study, eIF5A1(K50A), a mutant of eIF5A1 in which
lysine50, the site of hypusination, has been replaced by alanine,
proved to be equally as pro-apoptotic as eIF5A1. Moreover,
eIF5A1 and eIF5A1(K50A) impacted common elements of the
apoptotic cascade. Key indices of apoptosis including
phosphatidylserine externalization, mitochondrial dysfunction,
DNA fragmentation, and plasma membrane permeabilization
were evident following infection with either Ad-eIF5A1 or AdeIF5A1(K50A). These indices are known to be directly or
indirectly attributable to the action of caspases (Hengartner,
2000). Consistent with this, Ad-eIF5A1 and Ad-eIF5A1(K50A)
both proved capable of activating initiator caspases 8 and 9 as
well as effector caspase 3, all of which underpin the execution of
apoptosis. These comparative data are thus consistent with the
contention that unhypusinated eIF5A1 is the form of the
protein that is apoptogenic.
Unhypusinated eIF5A1 facilitates activation of the
intrinsic mitochondrial apoptotic pathway
The finding that over-expression of eIF5A1 or eIF5A1(K50A)
induces activation of caspase 9 suggested that unhypusinated
eIF5A1 invokes apoptosis by facilitating activation of the
intrinsic mitochondrial pathway. This was confirmed by
demonstrating that infection with Ad-eIF5A1 or AdeIF5A1(K50A) resulted in translocation of Bax from the cytosol
to mitochondria, permeabilization of the outer mitochondrial
JOURNAL OF CELLULAR PHYSIOLOGY
membrane, and release of cytochrome c. Moreover, the
translocation of Bax preceded the release of cytochrome c,
which is consistent with reports that the redistribution of Bax
leads to mitochondrial outer membrane permeabilization
(Jürgensmeier et al., 1998; Shimizu et al., 1999) and is sufficient
for the activation of downstream signaling cascades inherent to
apoptosis (Antonsson et al., 2001). Of particular note is the fact
that siRNA-mediated suppression of Bax severely inhibited the
induction of apoptosis by over-expressed eIF5A1 or
eIF5A1(K50R), for this is further confirmation of the
contention that unhypusinated eIF5A1 facilitates activation of
the intrinsic mitochondrial pathway.
Enhanced expression of eIF5A1 or eIF5A1(K50A) also
induced up-regulation of Bim, a BH3-only pro-apoptotic
protein, including three isoforms, BimEL, BimL, and BimS
generated by alternative splicing. Bim has been shown to be
involved in apoptosis induced by a variety of stimuli including
DNA damage (Sunters et al., 2003) and death receptor ligation
(Han et al., 2006), and the isoforms of Bim formed by alternative
splicing are known to be translocated to mitochondria where
they facilitate the release of cytochrome c (O’Connor et al.,
1998; Gomez-Bougie et al., 2004). Moreover, as for Bax,
siRNA-mediated suppression of Bim strongly inhibited
eIF5A1-induced apoptosis consistent with the contention that
unhypusinated eIF5A1 facilitates activation of the intrinsic
mitochondrial apoptotic pathway.
The tumor suppressor, p53, is also involved in the
mitochondrial apoptotic pathway, and eIF5A1 has been shown
807
808
SUN ET AL.
to control p53 up-regulation in response to pro-apoptotic
stimuli (Li et al., 2004; Taylor et al., 2007). Infection of HeLa S3
cells with Ad-eIF5A1(K50A) in this study resulted in a modest
increase in p53 expression, although no increase in the p53
targets, Puma, Bid, and Bad, was observed. This likely reflects
the fact that p53 is inactivated in HeLa S3 cells due to the
presence of the HPV E6 oncogene (Scheffner et al., 1990, 1991;
Thomas et al., 1996). There was, however, a pronounced
accumulation of putative p73, which shares a high degree of
sequence and structural homology with p53 (Kaghad et al.,
1997) and likely cross-reacted with the p53 antibody used in the
present study. Unlike p53, p73 is not inactivated by HPV E6
(Marin et al., 1998) and is able to induce cell cycle arrest and
apoptosis irrespective of p53 status (Jost et al., 1997; Melino
et al., 2004). Moreover, the finding that eIF5A1 and
eIF5A1(K50A) induced apoptosis in both HeLa S3 cells, in which
p53 is inactivated (Scheffner et al., 1991), and HT-29 cells, which
express a non-functional p53 (Rodrigues et al., 1990), indicates
that apoptosis induced by eIF5A1 is likely not dependent on p53
activation.
The finding that over-expression of eIF5A2, a second
isoform of human eIF5A, simulated the apoptogenic effects of
eIF5A1 raises the possibility that the two proteins have some
degree of conserved functional potential. Indeed, in a previous
study using a syngeneic murine melanoma (B16-F0) model,
intratumoral injection of Ad-eIF5A1 or Ad-eIF5A2 significantly
inhibited subcutaneous melanoma tumor growth compared
with control mice, although the effect was more pronounced
with Ad-eIF5A1 (Jin et al., 2008). eIF5A2, like all eIF5A proteins,
contains the unique hypusination domain and is 84%
homologous with eIF5A1 at the amino acid level (Guan et al.,
2001; Jenkins et al., 2001). However, unlike eIF5A1, it is rarely
expressed in normal cells and tissues. To date, it has only been
detected in testis and parts of the brain (Jenkins et al., 2001).
Rather, eIF5A2 is over-expressed in certain colon and ovarian
cancers, has been mapped to 3q26, a chromosomal region that
is commonly amplified in human ovarian cancer, and has been
identified as an oncogene (Guan et al., 2001, 2004; Clement
et al., 2003; Marchet et al., 2007; Xie et al., 2008).
The proposed oncogenic function of eIF5A2 is not, however,
inconsistent with the observation in this study that it is overexpression of unmodified eIF5A2 that induces apoptosis. The
oncogenic activity of eIF5A2 is presumably attributable to the
hypusinated form of the protein. This is supported by the
findings that hypusine-modified eIF5A has been identified as a
marker of neoplastic growth (Cracchiolo et al., 2004), DHS,
which mediates the first step of hypusination, is up-regulated in
cancers (Clement et al., 2006), and DHS has been identified as a
marker for metastatic disease (Ramaswamy et al., 2003).
Moreover, as noted previously (Taylor et al., 2007) and
confirmed in the present study, adenoviral-mediated overexpression of eIF5A results in accumulation of unhypusinated
eIF5A. Indeed, accumulation of trans-hypusinated eIF5A in
mammalian cells can only be achieved by co-over-expression of
DHS and DOHH in the presence of eIF5A1 precursor (Park
et al., 2006). It can thus be presumed that the Ad-eIF5A2induced apoptosis observed in this study is attributable to an
accumulation of unhypusinated eIF5A2. It is also noteworthy
that the ability of over-expressed oncogenes to induce
apoptosis or increase sensitivity to apoptotic stimuli has been
well described (Harrington et al., 1994; Fearnhead et al., 1998).
eIF5A1 proteome
Proteomic analyses indicated that 131 proteins in HeLa S3 cells
underwent significant changes in expression (expression ratio
1.5, P 0.05) within 72 h of infection with Ad-eIF5A1. Of the
49 that were sequenced and identified, a number proved to be
pro-apoptotic factors including hnRNP H1, known to enhance
JOURNAL OF CELLULAR PHYSIOLOGY
the ratio of pro-apoptotic Bcl-xs to anti-apoptotic Bcl-xl by
modulating the alternative splicing of Bcl-x (Garneau et al.,
2005); hemoglobin alpha, known to enhance caspase activation
and DNA degradation (Brachat et al., 2002); galectin, a
b-galactoside-binding protein involved in cell–cell and
cell–matrix adhesion that inhibits cell growth and promotes
apoptosis (Perillo et al., 1995; Blaser et al., 1998); and cofilin, an
actin-binding protein that induces apoptosis (Chua et al., 2003).
Ubiquilin and protein-disulfide isomerase, which are both
induced in the event of ischemia and involved in ER
stress-induced apoptosis (Ko et al., 2002), were also
significantly up-regulated in Ad-eIF5A1-infected cells. Other
members of the eIF5A1 proteome including hnRNP L, hnRNP
A2, and basic transcription factor 3 showed significant downregulation (expression ratio 1.5, P 0.05) consistent with the
apoptogenic activity of eIF5A1. HnRNP A2, for example, is
required for cell proliferation (He et al., 2005) and downregulated during Fas-induced apoptosis (Thiede et al., 2001).
Indeed, suppression of hnRNP A2 together with its functional
homologue, hnRNP A1, has been shown to induce apoptosis in
a variety of cancer cell lines including HeLa S3 (Patry et al.,
2003). Basic transcription factor 3 is thought to function as an
anti-apoptotic modulator (Brockstedt et al., 1999; Thiede et al.,
2001), and its down-regulation would thus also facilitate the
execution of apoptosis.
Although global proteomic profiling coupled with targeted
Western blot analysis enabled identification of a number of
proteins that show altered expression in response to
adenoviral mediated up-regulation of eIF5A1, these proteins
likely correspond to members of the proteome for
unhypusinated eIF5A1 rather than hypusinated eIF5A1. In the
case of the Western blot analyses, this is evident from the fact
that the effects of eIF5A1 up-regulation were mirrored by upregulation of eIF5A1(K50A), which cannot be hypusinated. For
the proteomic data, this contention is supported by the fact that
an accumulation of unhypusinated, but not hypusinated, eIF5A1
was observed in Ad-eIF5A1-infected cells. Finally, many of the
proteins showing altered expression in response to Ad-eIF5A1
up-regulation are clearly capable of modulating apoptosis,
which is in keeping with the apoptogenic capability of
unhypusinated eIF5A1.
Literature Cited
Antonsson B, Montessuit S, Sanchez B, Martinou JC. 2001. Bax is present as a high molecular
weight oligomer/complex in the mitochondrial membrane of apoptotic cells. J Biol Chem
276:11615–11623.
Ashkenazi A, Dixit VM. 1998. Death receptors: Signaling and modulation. Science 281:1305–
1308.
Benne R, Brown-Luedi ML, Hershey JW. 1978. Purification and characterization of protein
synthesis initiation factors eIF-1, eIF-4C, eIF-4D, and eIF-5 from rabbit reticulocytes. J Biol
Chem 253:3070–3077.
Blaser C, Kaufmann M, Müller C, Zimmermann C, Wells V, Mallucci L, Pircher H. 1998. Betagalactoside-binding protein secreted by activated T cells inhibits antigen-induced
proliferation of T cells. Eur J Immunol 28:2311–2319.
Brachat A, Pierrat B, Xynos A, Brecht K, Simonen M, Brüngger A, Heim J. 2002. A microarraybased, integrated approach to identify novel regulators of cancer drug response and
apoptosis. Oncogene 21:8361–8371.
Brockstedt E, Otto A, Rickers A, Bommert K, Wittmann-Liebold B. 1999. Preparative highresolution two-dimensional electrophoresis enables the identification of RNA polymerase
B transcription factor 3 as an apoptosis-associated protein in the human BL 60-2. Burkitt
lymphoma cell line. J Protein Chem 18:225–231.
Byers TL, Lakanen JR, Coward JK, Pegg AE. 1994. The role of hypusine depletion in cytostasis
induced by S-adenosyl-L-methionine decarboxylase inhibition: New evidence provided by
1-methylspermidine and 1,12-dimethylspermine. Biochem J 303:363–368.
Caraglia M, Marra M, Giuberti G, D’Alessandro AM, Baldi A, Tassone P, Venuta S, Tagliaferri
P, Abbruzzese A. 2003. The eukaryotic initiation factor 5A is involved in the regulation of
proliferation and apoptosis induced by interferon-alpha and EGF in human cancer cells.
J Biochem (Tokyo) 133:757–765.
Chatterjee I, Gross SR, Kinzy TG, Chen KY. 2006. Rapid depletion of mutant eukaryotic
initiation factor 5A at restrictive temperature reveals connections to actin cytoskeleton
and cell cycle progression. Mol Genet Genomics 275:264–276.
Cheng Z, Wei Y-YC, Sung WW, Glick BR, McConkey BJ. 2009. Proteomic analysis of the
response of the plant growth-promoting bacterium Pseudomonas putida UW4 to nickel
stress. Proteome Sci 7:18.
Chua BT, Volbracht C, Tan KO, Li R, Yu VC, Li P. 2003. Mitochondrial translocation of cofilin
is an early step in apoptosis induction. Nat Cell Biol 5:1083–1089.
Cleary ML, Smith SD, Sklar J. 1986. Cloning and structural analysis of cDNAs for bcl-2 and a
hybrid bcl-2/immunoglobulin transcript resulting from the t(14;18) translocation. Cell
47:19–28.
eIF5A ACTIVATES INTRINSIC PATHWAY
Clement PM, Henderson CA, Jenkins ZA, Smit-McBride Z, Wolff EC, Hershey JW, Park MH,
Johansson JE. 2003. Identification and characterization of eukaryotic initiation factor 5A-2.
Eur J Biochem 270:4254–4263.
Clement PM, Johansson HE, Wolff EC, Park MH. 2006. Differential expression of eIF5A-1 and
eIF5A-2 in human cancer cells. FEBS J 273:1102–1114.
Cooper HL, Park MH, Folk JE, Safer B, Braverman R. 1982. Posttranslational formation of
hypusine in a single major protein occurs generally in growing cells and is associated with
activation of lymphocyte growth. Cell 29:791–797.
Cracchiolo BM, Heller DS, Clement PMJ, Wolff EC, Park M-H, Hanauske-Abel HM. 2004.
Eukaryotic initiation factor 5A-1 (eIF5A-1) as a diagnostic marker for aberrant
proliferation in intraepithelial neoplasia of the vulva. Gynecol Oncol 94:217–222.
Drexler H. 1997. Activation of the cell death program by inhibition of proteasome function.
Proc Natl Acad Sci USA 94:855–860.
Elmore S. 2007. Apoptosis: A review of programmed cell death. Toxicol Pathol 35:495–516.
Fearnhead HO, Rodriguez J, Govek EE, Guo W, Kobayashi R, Hannon G, Lazebnik YA. 1998.
Oncogene-dependent apoptosis is mediated by caspase-9. Proc Natl Acad Sci USA
95:13664–13669.
Garneau D, Revil T, Fisette J-F, Chabot B. 2005. Heterogeneous nuclear ribonucleoprotein F/
H proteins modulate the alternative splicing of the apoptotic mediator Bcl-x. J Biol Chem
280:22641–22650.
Gomez-Bougie P, Bataille R, Amiot M. 2004. The imbalance between Bim and Mcl-1
expression controls the survival of human myeloma cells. Eur J Immunol 34:3156–3164.
Guan XY, Sham JS, Tang TC, Fang Y, Huo KK, Yang JM. 2001. Isolation of a novel candidate
oncogene within a frequently amplified region at 3q26 in ovarian cancer. Cancer Res
61:3806–3809.
Guan XY, Fung JM, Ma NF, Lau SH, Tai LS, Xie D, Zhang Y, Hu L, Wu QL, Fang Y, Sham JS.
2004. Oncogenic role of eIF-5A2 in the development of ovarian cancer. Cancer Res
64:4197–4200.
Han J, Goldstein LA, Gastman BR, Rabinowich H. 2006. Interrelated roles for Mcl-1 and BIM
in regulation of TRAIL-mediated mitochondrial apoptosis. J Biol Chem 281:10153–10163.
Hanauske-Abel HM, Park MH, Hanauske AR, Popowicz AM, Lalande M, Folk JE. 1994.
Inhibition of the G1-S transition of the cell cycle by inhibitors of deoxyhypusine
hydroxylation. Biochim Biophys Acta 1221:115–124.
Harrington EA, Fanidi A, Evan GI. 1994. Oncogenes and cell death. Curr Opin Genet Dev
4:120–129.
Harris MH, Thompson CB. 2000. The role of the Bcl-2 family in the regulation of outer
mitochondrial membrane permeability. Cell Death Differ 7:1182–1191.
He Y, Brown MA, Rothnagel JA, Saunders NA, Smith R. 2005. Roles of heterogeneous nuclear
ribonucleoproteins A and B in cell proliferation. J Cell Sci 118:3173–3183.
Hengartner MO. 2000. The biochemistry of apoptosis. Nature 407:770–776.
Ho SN, Hunt HD, Horton RM, Pullen JK, Pease LR. 1989. Site-directed mutagenesis by
overlap extension using the polymerase chain reaction. Gene 77:51–59.
Hofmann W, Reichart B, Ewald A, Muller E, Schmitt I, Stauber RH, Lottspeich F, Jockusch BM,
Scheer U, Hauber J, Dabauvalle MC. 2001. Cofactor requirements for nuclear export of
Rev response element (RRE)- and constitutive transport element (CTE)-containing
retroviral RNAs. An unexpected role for actin. J Cell Biol 152:895–910.
Jao DL, Chen KY. 2006. Tandem affinity purification revealed the hypusine-dependent binding
of eukaryotic initiation factor 5A to the translating 80S ribosomal complex. J Cell Biochem
97:583–598.
Jenkins ZA, Hååg PG, Johansson HE. 2001. Human EIF5A2 on chromosome 3q25–q27 is a
phylogenetically conserved vertebrate variant of eukaryotic translation initiation factor 5A
with tissue-specific expression. Genomics 71:101–109.
Jin BF, He K, Wang HX, Wang J, Zhou T, Lan Y, Hu MR, Wei KH, Yang SC, Shen BF, Zhang
XM. 2003. Proteomic analysis of ubiquitin-proteasome effects: Insight into the function of
eukaryotic initiation factor 5A. Oncogene. 22:4819–4830.
Jin S, Taylor CA, Liu Z, Sun Z, Ye B, Thompson JE. 2008. Suppression of primary and
disseminated murine tumor growth with eIF5A1 gene therapy. Gene Ther Mol Biol
12:207–218.
Jost CA, Marin MC, Kaelin WG, Jr. 1997. p73 is a simian [correction of human] p53-related
protein that can induce apoptosis. Nature 389:191–194.
Jürgensmeier JM, Xie Z, Deveraux Q, Ellerby L, Bredesen D, Reed JC. 1998. Bax directly
induces release of cytochrome c from isolated mitochondria. Proc Natl Acad Sci USA
95:4997–5002.
Kaghad M, Bonnet H, Yang A, Creancier L, Biscan JC, Valent A, Minty A, Chalon P, Lelias JM,
Dumont X, Ferrara P, McKeon F, Caput D. 1997. Monoallelically expressed gene related to
p53 at 1p36, a region frequently deleted in neuroblastoma and other human cancers. Cell
90:809–819.
Kang HA, Hershey JW. 1994. Effect of initiation factor eIF-5A depletion on protein synthesis
and proliferation of Saccharomyces cerevisiae. J Biol Chem 269:3934–3940.
Klier H, Csonga R, Joäo HC, Eckerskorn C, Auer M, Lottspeich F, Eder J. 1995. Isolation and
structural characterization of different isoforms of the hypusine-containing protein eIF-5A
from HeLa cells. Biochemistry 34:14693–14702.
Ko HS, Uehara T, Nomura Y. 2002. Role of ubiquilin associated with protein-disulfide
isomerase in the endoplasmic reticulum in stress-induced apoptotic cell death. J Biol Chem
277:35386–35392.
Li P, Nijhawan D, Budihardjo I, Srinivasula SM, Ahmad M, Alnemri ES, Wang X. 1997.
Cytochrome c and dATP-dependent formation of Apaf-1/caspase-9 complex initiates an
apoptotic protease cascade. Cell 91:479–489.
Li AL, Li HY, Jin BF, Ye QN, Zhou T, Yu XD, Pan X, Man JH, He K, Yu M, Hu MR, Wang J, Yang
SC, Shen BF, Zhang XM. 2004. A novel eIF5A complex functions as a regulator of p53 and
p53-dependent apoptosis. J Biol Chem 279:49251–49258.
MacLaren AP, Chapman RS, Wyllie AH, Watson CJ. 2001. p53-dependent apoptosis induced
by proteasome inhibition in mammary epithelial cells. Cell Death Differ 8:210–218.
Marchet A, Mocellin S, Belluco C, Ambrosi A, De Marchi F, Mammano E, Digito M, Leon A,
D’Arrigo A, Lise M, Nitti D. 2007. Gene expression profile of primary gastric cancer:
Towards the prediction of lymph node status. Ann Surg Oncol 14:1058–1064.
Marin MC, Jost CA, Irwin MS, DeCaprio JA, Caput D, Kaelin WG, Jr. 1998. Viral oncoproteins
discriminate between p53 and the p53 homolog p73. Mol Cell Biol 18:6316–6324.
Melino G, Bernassola F, Ranalli M, Yee K, Zong WX, Corazzari M, Knight RA, Green DR,
Thompson C, Vousden KH. 2004. p73 Induces apoptosis via PUMA transactivation and Bax
mitochondrial translocation. J Biol Chem 279:8076–8083.
O’Connor L, Strasser A, O’Reilly LA, Hausmann G, Adams JM, Cory S, Huang DC. 1998. Bim:
A novel member of the Bcl-2 family that promotes apoptosis. EMBO J
17:384–395.
JOURNAL OF CELLULAR PHYSIOLOGY
Odin JA, Huebert RC, Casciola-Rosen L, LaRusso NF, Rosen A. 2001. Bcl-2-dependent
oxidation of pyruvate dehydrogenase-E2, a primary biliary cirrhosis autoantigen, during
apoptosis. J Clin Invest 108:223–232.
Park MH, Cooper HL, Folk JE. 1982. The biosynthesis of protein-bound hypusine (N epsilon(4-amino-2-hydroxybutyl)lysine). Lysine as the amino acid precursor and the intermediate
role of deoxyhypusine (N epsilon-(4-aminobutyl)lysine). J Biol Chem
257:7217–7222.
Park MH, Wolff EC, Abbruzzese A, Folk JE. 1988. Post-translational formation of hypusine:
Biogenesis of translation initiation factor eIF-4D. Adv Exp Med Biol 231:633–640.
Park MH, Wolff EC, Folk JE. 1993a. Hypusine: Its post-translational formation in eukaryotic
initiation factor 5A and its potential role in cellular regulation. Biofactors 4:95–104.
Park MH, Wolff EC, Lee YB, Folk JE. 1993b. Antiproliferative effects of inhibitors of
deoxyhypusine synthase. Inhibition of growth of Chinese hamster ovary cells by guanyl
diamines. J Biol Chem 269:27827–27832.
Park MH, Joe YA, Kang KR. 1998. Deoxyhypusine synthase activity is essential for cell viability
in the yeast Saccharomyces cerevisiae. J Biol Chem 273:1677–1683.
Park J-H, Aravind L, Wolff EC, Kaevel J, Kim YS, Park MH. 2006. Molecular cloning,
expression, and structural prediction of deoxyhypusine hydroxylase: A HEAT-repeatcontaining metalloenzyme. Proc Natl Acad Sci USA 103:51–56.
Patry C, Bouchard L, Labrecque P, Gendron D, Lemieux B, Toutant J, Lapointe E, Wellinger R,
Chabot B. 2003. Small interfering RNA-mediated reduction in heterogeneous nuclear
ribonucleoparticule A1/A2 proteins induces apoptosis in human cancer cells but not in
normal mortal cell lines. Cancer Res 63:7679–7688.
Perillo NL, Pace KE, Seilhamer JJ, Baum LG. 1995. Apoptosis of T cells mediated by galectin-1.
Nature 378:736–739.
Perkins DN, Pappin DJC, Creasy DM, Cottrell JS. 1999. Probability-based protein
identification by searching sequence databases using mass spectrometry data.
Electrophoresis 20:3551–3567.
Peter ME, Krammer PH. 2003. The CD95(APO-1/Fas) DISC and beyond. Cell Death Differ
10:26–35.
Petit PX, Lecoeur H, Zorn E, Dauguet C, Mignotte B, Gougeon ML. 1995. Alterations in
mitochondrial structure and function are early events of dexamethasone-induced
thymocyte apoptosis. J Cell Biol 130:157–167.
Ramaswamy S, Ross KN, Lander ES, Golub TR. 2003. A molecular signature of metastasis in
primary solid tumors. Nat Genet 33:49–54.
Rodrigues NR, Rowan A, Smith ME, Kerr IB, Bodmer WF, Gannon JV, Lane DP. 1990. p53
mutations in colorectal cancer. Proc Natl Acad Sci USA 87:7555–75559.
Saikumar P, Dong Z, Patel Y, Hall K, Hopfer U, Weinberg JM, Venkatachalam MA. 1998. Role
of hypoxia-induced Bax translocation and cytochrome c release in reoxygenation injury.
Oncogene 17:3401–3415.
Saini P, Eyler DE, Green R, Dever TE. 2009. Hypusine-containing protein eIF5A promotes
translation elongation. Nature 459:118–121.
Scheffner M, Werness BA, Huibregtse JM, Levine AJ, Howley PM. 1990. The E6 oncoprotein
encoded by human papillomavirus types 16 and 18 promotes the degradation of p53. Cell
63:1129–1136.
Scheffner M, Munger K, Byrne JC, Howley PM. 1991. The state of the p53 and retinoblastoma
genes in human cervical carcinoma cell lines. Proc Natl Acad Sci USA 88:5523–5527.
Schinzel A, Kaufmann T, Schuler M, Martinalbo J, Grubb D, Borner C. 2004. Conformational
control of Bax localization and apoptotic activity by Pro168. J Cell Biol 164:1021–
1032.
Schnier J, Schwelberger HG, Smit-McBride Z, Kang HA, Hershey JW. 1991. Translation
initiation factor 5A and its hypusine modification are essential for cell viability in the yeast
Saccharomyces cerevisiae. Mol Cell Biol 11:3105–3114.
Shimizu S, Narita M, Tsujimoto Y. 1999. Bcl-2 family proteins regulate the release of
apoptogenic cytochrome c by the mitochondrial channel VDAC. Nature 399:483–487.
Smit-McBride Z, Dever TE, Hershey JW, Merrick WC. 1989. Sequence determination and
cDNA cloning of eukaryotic initiation factor 4D, the hypusine-containing protein. J Biol
Chem 264:1578–1583.
Sunters A, Fernández de Mattos S, Stahl M, Brosens JJ, Zoumpoulidou G, Saunders CA, Coffer
PJ, Medema RH, Coombes RC, Lam EW. 2003. FoxO3a transcriptional regulation of Bim
controls apoptosis in paclitaxel-treated breast cancer cell lines. J Biol Chem 278:49795–
49805.
Taylor CA, Senchyna M, Flanagan J, Joyce EM, Cliche DO, Boone AN, Culp-Stewart S,
Thompson JE. 2004. Role of eIF5A in TNF-alpha-mediated apoptosis of lamina cribrosa
cells. Invest Ophthalmol Vis Sci 45:3568–3576.
Taylor CA, Sun Z, Cliche DO, Ming H, Eshaque B, Jin S, Hopkins MT, Thai B, Thompson JE.
2007. Eukaryotic translation initiation factor 5A induces apoptosis in colon cancer cells and
associates with the nucleus in response to tumour necrosis factor alpha signalling. Exp Cell
Res 313:437–449.
Thiede B, Dimmler C, Siejak F, Rudel T. 2001. Predominant identification of RNA-binding
proteins in Fas-induced apoptosis by proteome analysis. J Biol Chem 276:26044–26050.
Thomas M, Matlashewski G, Pim D, Banks L. 1996. Induction of apoptosis by p53 is
independent of its oligomeric state and can be abolished by HPV-18 E6 through ubiquitin
mediated degradation. Oncogene 2:265–273.
Valentini SR, Casolari JM, Oliveira CC, Silver PA, McBride AE. 2002. Genetic interactions of
yeast eukaryotic translation initiation factor 5A (eIF5A) reveal connections to poly(A)binding protein and protein kinase C signaling. Genetics 160:393–405.
Wang X, Sun H, Li C. 2007. Nitric oxide induces promyelocytic cell growth arrest and
apoptosis through deactivation of Akt pathway. Leuk Res 31:653–660.
Weir BA, Yaffe MP. 2004. Mmd1p, a novel, conserved protein essential for normal
mitochondrial morphology and distribution in the fission yeast Schizosaccharomyces
pombe. Mol Biol Cell 15:1656–1665.
Xie D, Ma NF, Pan ZZ, Wu HX, Liu YD, Wu GQ, Kung HF, Guan XY. 2008. Overexpression
of EIF-5A2 is associated with metastasis of human colorectal carcinoma. Hum Pathol
39:80–86.
Zamzami N, Kroemer G. 2001. The mitochondrion in apoptosis: How Pandora’s box opens.
Nat Rev Mol Cell Biol 2:67–71.
Zanelli CF, Valentini SR. 2005. Pkc1 acts through Zds1 and Gic1 to suppress growth and cell
polarity defects of a yeast eIF5A mutant. Genetics 171:1571–1581.
Zanelli CF, Maragno AL, Gregio AP, Komili S, Pandolfi JR, Mestriner CA, Lustri WR, Valentini
SR. 2006. eIF5A binds to translational machinery components and affects translation in
yeast. Biochem Biophys Res Commun 348:1358–1366.
Zuk D, Jacobson A. 1998. A single amino acid substitution in yeast eIF-5A results in mRNA
stabilization. EMBO J 17:2914–2925.
809