Tài liệu Báo cáo khoa học: Identification of different isoforms of eEF1A in the nuclear fraction of human T-lymphoblastic cancer cell line specifically binding to aptameric cytotoxic GT oligomers ppt

Identification of different isoforms of eEF1A in the nuclear fraction
of human T-lymphoblastic cancer cell line specifically binding
to aptameric cytotoxic GT oligomers
Barbara Dapas
1
, Gianluca Tell
2
, Andrea Scaloni
3
, Alex Pines
2
, Lino Ferrara
3
, Franco Quadrifoglio
1
and Bruna Scaggiante
1
1
Department of Biomedical Sciences and Technologies, University of Udine, Italy;
2
Department of Biochemistry, Biophysics
and Macromolecular Chemistry, University of Trieste, Italy;
3
Proteomics and Mass Spectrometry Laboratory, ISPAAM,
National Research Council, Naples, Italy
GT oligomers, showing a dose-dependent cytotoxic effect on
a variety of human cancer cell lines, but not on normal
human lymphocytes, recognize and form complexes with
nuclear proteins. By working with human T-lymphoblastic
CCRF-CEM cells and by using MS and SouthWestern
blotting, we identified eukaryotic elongation factor 1 alpha

(eEF1A) as the main nuclear protein that specifically
recognizes these oligonucleotides. Western blotting and
supershift assays confirmed the nature of this protein and
its involvement in forming a cytotoxicity-related complex
(CRC). On the contrary, normal human lymphocytes did
not show nuclear proteins able to produce CRC in a
SouthWestern blot. Comparative bidimensional PAGE and
Western-blotting analysis for eEF1A revealed the presence
of a specific cluster of spots, focusing at more basic pH, in
nuclear extracts of cancer cells but absent in those of normal
lymphocytes. Moreover, a bidimensional PAGE South-
Western blot demonstrated that cytotoxic GT oligomers
selectively recognized the more basic eEF1A isoform
expressed only in cancer cells. These results suggest the
involvement of eEF1A, associated with the nuclear-enriched
fraction, in the growth and maintenance of tumour cells,
possibly modulated by post-translational processing of the
polypeptide chain.
Keywords: aptameric oligonucleotides; eEF1A; proteomics;
CCRF-CEM cells; cytotoxicity.
Oligonucleotides, widely used as agents to specifically
inhibit gene expression by antisense [1] or antigene [2]
strategies, often display unexpected effects by interacting
with cellular proteins. In fact, they are able to bind to either
membrane or intracellular proteins, probably by their
polyanionic nature and/or by nonspecific or specific
sequence-related mechanisms [3]. In the last decade, oligo-
nucleotides have progressively gained aptameric function,
specifically recognizing proteins as natural or non-natural
ligands [4]. Constitutive proteins that bind to single-

stranded DNA oligomers are widely recognized to be
involved in important mechanisms associated with DNA
replication, repair and recombination [5–7]. Furthermore,
many reports evidenced that modulation of gene expression
[8,9], and stimulation or inhibition of cellular replication
[10,11], are influenced by single-stranded DNA sequences
specifically interacting with cellular proteins.
Oligonucleotides composed exclusively of G and T bases
have previously been shown to exert a specific, selective and
dose-dependent effect of cell growth inhibition on a variety
of human cancer cell lines [12]. The cytotoxic effect of these
GT oligomers was shown to be highly related to their ability
to form complexes with nuclear proteins, as measured by
UV cross-linking assays [12–15]. However, the nature of
these nuclear proteins behaving as single-stranded DNA-
binding proteins has not yet been identified [12–15]. A
protein isolated from fibroblasts with such an activity has
been already described [16], but it was able to tightly bind
either GA or GT oligomers. On the contrary, the nuclear
proteins binding to our GT oligomers did not specifically
recognize GA sequences [12]. More recently, it has been
shown that GT oligonucleotides, capable of forming
G-quartet structures, exerted a cytotoxic effect on human
cancer cell lines. By UV cross-linking assay, these oligomers
have been reported to interact with nucleolin, forming a
main complex of >100 kDa molecular mass [17]. This
complex was not formed when GT oligomers unable to
form a G-quartet structure were used [17,18]. The oligonu-
cleotide under our investigation (a 27-mer; see the Materials
and methods, below) did not present appreciable G-quartet

structures, as deduced by gel electrophoresis and circular
dichroism analysis [14]; on the contrary, it was able to form
a cytotoxic-related complex (CRC), with an apparent
molecular mass of 45 ± 7 kDa, with nuclear proteins of
different tumour cell lines [12–15]. Thus, the characteriza-
tion of these nuclear species seemed particularly interesting,
Correspondence to B. Scaggiante, Department of Biomedical Sciences
and Technologies, University of Udine, p.le Kolbe 4, 33100 Udine,
Italy. Fax: + 39 432 494301; Tel.: + 39 432 494311;
E-mail:
Abbreviations: CRC, cytotoxicity-related complex; CRS, control
rabbit total serum; eEF1A, eukaryotic elongation factor 1 alpha;
Egr1, early growth response protein 1; IPG, immobilized pH gradient;
PSD, postsource decay; TBP, TATA-binding protein.
(Received 7 March 2003, revised 27 May 2003, accepted 10 June 2003)
Eur. J. Biochem. 270, 3251–3262 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03713.x
either for using to elucidate new potential molecular targets
in tumour biology or to highlight the mechanism of action
of our cytotoxic GT oligonucleotides.
In this article, we report the identification of eukaryotic
elongation factor 1 alpha (eEF1A) as a nuclear component
of the CRC in T-lymphoblastic CCRF-CEM cancer cells.
In these cells, we found a striking relationship between the
growth-inhibition effect exerted by cytotoxic GT oligomers
and their selective binding to nuclear eEF1A. In fact, in
normal human lymphocytes no appreciable binding of GT
oligomer to nuclear eEF1A was shown and, accordingly,
these cells were not sensitive to its cytotoxic action. A
possible role for nuclear eEF1A in tumour cell growth or
maintenance is suggested by the selective identification of

more basic isoforms of eEF1A in cancer cells, but not in
normal lymphocytes.
Materials and methods
Oligonucleotides
Oligonucleotides were purchased from MWG Biotech
(Ebersberg, Germany) as HPLC pure species and their
purity was confirmed by electrophoresis on an 18%
polyacrylamide/7
M
urea gel. For cell cultures, oligonucleo-
tides were resuspended in water and sterilized by centri-
fugation on a spin-X tube provided with a 0.22-l
M
filter
(Costar, Cambridge, MA, USA). The GT oligomer
sequence was: 5¢-TGT TTG TTT GTT TGT TTG TTT
GTT TGT-3¢; and the control CT sequence was: 5¢-TCT
TTC TTT CTT TCT TTC TTT CTT TCT-3¢. The
oligomers were 5¢ end-labelled by [c-
32
P]ATP with T4
polynucleotide kinase (MBI, Fermentas, MGMBH, St
Leon-Rot, Germany).
Cell culture and cytotoxic assay
The T-lymphoblastic leukaemic cell line (CCRF-CEM) and
normal human lymphocytes, obtained from peripheral
blood by separation on Ficoll–Isopaque (Gibco BRL, Life
Technologies, Milan, Italy), were cultured in RPMI-1640
supplemented with 10% fetal calf serum (FCS), 2 m
M

L
-glutamine, 100 UÆmL
)1
penicillin and 100 lgÆmL
)1
strep-
tomycin (Euroclone, Celbio, Devon, UK).
CCRF-CEM cells in exponential growth phase, and
lymphocytes, were seeded at 10
4
cells in 100 lL of complete
medium containing 10% fetal clone serum (Euroclone,
Celbio) in a 96-well microtiter plate. The oligonucleotides
were added directly to the medium 4 h after seeding. After
24 h of incubation, 100 lL of fresh medium was added.
The cellular growth was evaluated 72 h after addition
of oligonucleotides by assessing the incorporation of
0.5 mgÆmL
)1
of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-
tetrazolium bromide (Sigma Chemical Co., St Louis, MO,
USA) into viable cells [19]. The percentage of viable cells in
the treated samples was estimated, taking, as 100% cellular
viability, that of the internal-control nontreated cells.
Preparation of total nuclear extracts
Total nuclear extracts were obtained from 2 · 10
7
cells by
using a minor modification of the Dignam’s method, as
previously described [12]. The protein content was deter-

mined by the Bradford method [19] using BSA as standard.
EMSA, UV crosslinking and supershift assays
IntheEMSA,1ngof[c-
32
P]ATP-labelled oligonucleotide
was incubated with 2 lg of total nuclear or cytoplasmic
extracts supplemented with protease inhibitors (2 lgÆmL
)1
apoprotinin, 1 lgÆmL
)1
pepstatin, 1 m
M
dithiothreitol)
(Sigma Chemical Co.) in 20 m
M
Hepes, 0.42
M
NaCl,
1.5 m
M
MgCl
2
,0.2m
M
EDTA, 25% glycerol, pH 7.0,
containing nonspecific competitors [1 lg of salmon-sperm
DNA or 1 lg of poly(dIdC)] (Pharmacia, Uppsala, Sweden)
and the indicated amounts of unlabelled specific CT
oligomer competitor. When indicated, the protein excised
from the Coomassie-stained gel was recovered in 50 m

M
Tris/HCl, pH 8.0, containing 0.1% SDS, 0.1 mgÆmL
)1
BSA, 0.2 m
M
EDTA, 2.5% glycerol. After two steps of
freeze/thawing, followed by precipitation with cold acetone,
the protein was rinsed with methanol, denatured with 8
M
urea and then renatured by overnight incubation in a fixed
volume of 50 m
M
Tris/HCl, pH 7.6, 100 m
M
KCl, 5 m
M
dithiothreitol, 0.1 m
M
phenylmethanesulfonyl fluoride. It
was not possible to quantify the amount of recovered
protein owing to the presence of a high molar excess of
BSA remaining in the buffer. Therefore, a fixed aliquot of
the protein was incubated with the indicated probes, as
previously described. After 25 min of incubation at room
temperature, the samples were loaded onto a native 7%
polyacrylamide gel in 20 m
M
Tris/borate/0.5 m
M
EDTA

buffer (TBE) and electrophoresed at 10 V cm
)1
,ata
temperature of 4 °C.
In the supershift gel-mobility assay, samples of total
nuclear extracts were diluted 1 : 5 (v/v) in water and 0.5 lg
of protein was incubated for 2.5 h at room temperature with
the indicated amounts of specific rabbit polyclonal anti-
eEF1A serum or with corresponding amounts of control
total serum obtained from unimmunized rabbits. Then,
2 ng of specific [c-
32
P]-labelled GT oligonucleotide was
added to 30 lLof20m
M
Tris/HCl buffer, pH 7.5,
containing 75 m
M
KCl, 5 m
M
dithiotreitol, 6 lgBSA,
0.1% Tween 20, 0.025 m
M
Escherichia coli DNA, and 15%
glycerol. After 30 min of incubation at room temperature,
the samples were loaded onto a 7% native polyacrylamide
gel in 20 m
M
TBE buffer and electrophoresed for 90 min
at 10 V cm

)1
,at15°C. The gel was then dried and auto-
radiographed.
In the UV cross-linking assay, the samples were incuba-
ted at room temperature, as described above for the EMSA
assay, and then irradiated at 302 nm for 10 min using a
transilluminator (Bio-Rad Laboratories). The samples were
denatured by adding Laemmli sample buffer and boiled
before electrophoresis through a 12% SDS polyacrylamide
gel, according to the procedure of Laemmli [20].
Electroblotting
Nuclear proteins separated by SDS/PAGE were electro-
phoretically transferred onto a 0.22-lm nitrocellulose
membrane (Schleicher & Schuell, Keene, NH, USA) in
50 m
M
Tris, 40 m
M
glycine, 0.4% SDS, 20% methanol
buffer using a transblot semidry apparatus system
3252 B. Dapas et al. (Eur. J. Biochem. 270) Ó FEBS 2003
(Amersham Pharmacia Biotech). The membrane was
stained with Ponceau S (Sigma Chemical Co.) and destained
with deionized water.
SouthWestern blotting analysis
Fifty micrograms of total nuclear protein was separated by
SDS/PAGE (29 : 1, acrylamide/bisacrylamide) (8% gel).
Proteins were transferred to nitrocellulose and then dena-
tured and slowly renatured by washing for 30 min at room
temperature consecutively with 6-, 4-, 3-, 2- and 1

M
guanidine hydrochloride solution in water. The membrane
was incubated overnight on a rocking shaker, at 4 °C, in
50 m
M
Hepes, pH 7.2, containing 0.1
M
KCl, 1 m
M
MgCl
2
,
5m
M
dithiothreitol, 1 m
M
EDTA and 10% glycerol.
Membranes derived from bidimensional PAGE blotting
were processed without the denaturation/renaturation pro-
cedure. Membranes were blocked by washing with the same
buffer, containing 5% nonfat dried milk and 5 m
M
dithiothreitol, for 1 h. Protein–DNA interaction was per-
formed overnight, at 4 °C, with 10 pmol of [c-
32
P]-labelled
oligonucleotide. Membranes were then washed between two
and four times for 10 min at room temperature, until the
background radioactivity started to decline, and were then
exposed to autoradiography.

Expression of recombinant eEF1A
Full-length eEF1A cDNA was cloned into pET11a (kindly
provided by Dr George M. C. Janssen, University of Leiden)
at the Nde1/BamH1 site for bacterial expression. Recom-
binant eEF1A protein was obtained from overexpression in
E. coli [21]. Briefly, E. coli BL21 cell culture (2 mL),
transformed with pET11a–eEF1A, was grown overnight
at 37 °C in LB (Luria–Bertani) medium supplemented with
50 lgÆmL
)1
ampicillin. Fresh and prewarmed (37 °C) LB
medium was inoculated with the overnight culture to an
absorbance (A) value of 0.05–0.1. The culture was grown
until the A reached a value of 0.5–0.7, then isopropyl thio-b-
D
-galactoside was added to a final concentration of 1 m
M
.
Expression of eEF1A protein was induced by culturing the
cells for additional 3 h at 37 °C. Cells were harvested by
centrifugation (10 000 g,10min,4°C), resuspended in
10 mL of lysis buffer (20 m
M
Tris/HCl, 5 m
M
dithiothreitol,
250 m
M
NaCl, 1 m
M

EDTA, 0.25% Tween-20, 0.3 lgÆlL
)1
lysozyme) per gram of bacterial pellet and disrupted by
sonication. The lysate was centrifuged (10 000 g,20min,
4 °C) and the recombinant eEF1A protein collected in the
supernatant as a soluble protein.
Western blotting analysis
The blotted membrane was blocked with 3% nonfat dried
milk in PBS (NaCl/P
i
) and incubated with eEF1A mono-
clonal antibody (mAb) (1 lg/mL) (Upstate Biotechnology,
Lake Placid, NY, USA) in NaCl/P
i
,overnight,at4°Cwith
constant rocking. Then, it was washed twice with deionized
water and incubated for 1.5 h with an anti-mouse IgG-
conjugated horseradish peroxidase secondary antibody
(Promega, Madison, WI, USA). After washing once with
NaCl/P
i
containing 0.05% Tween-20 and four times with
deionized water, the nitrocellulose blot was developed using
enhanced chemiluminescence detection (Pierce, Rockford,
IL, USA) according to the manufacturer’s protocols, and
thenexposedtoX-rayfilm.
The same filter was stripped by a 10-min incubation in
4
M
guanidine hydrochloride, rinsed with 10 volumes of

NaCl/P
i
, blocked with 5% nonfat dried milk and then
probed with b-actin antibody (VWR International Onco-
gene) for 1 h, at room temperature, followed by incubation
for 1 h with a goat anti-mouse IgM-conjugated horseradish
peroxidase secondary antibody (Sigma Chemical Co.). The
blot was developed by using the chemiluminescence detec-
tion kit. Band intensities were evaluated by scanning with
a Gel Doc2000 phosphoimager densitometer equipped
with a multianalyst PC software analysis system (Bio-Rad
Laboratories).
To test the nuclear enrichment, the presence of two
nuclear transcription factors – the early growth response
protein 1 (Egr1) and the TATA-binding protein (TBP) –
was confirmed by probing with specific rabbit antibodies
(Santa Cruz) on the cytoplasmic and nuclear extracts. The
membrane was incubated with the antibodies for 1 h at
room temperature. After three washes with NaCl/P
i
containing 0.1% Tween-20, the membrane was incubated
with anti-rabbit IgG–horseradish peroxidase conjugate
(Sigma Chemical Co.) for 60 min at room temperature.
The filter was then washed several times with NaCl/P
i
containing 0.1% Tween-20, and the blot was developed
using the enhanced chemiluminescence procedure (Amer-
sham Pharmacia Biotech).
MS analysis
Bands from SDS/PAGE were excised from the gel, tritu-

rated and washed with water. Proteins were in-gel reduced,
S-alkylated and digested with trypsin, as previously des-
cribed [22]. Digest aliquots were removed and used directly
or subjected to a desalting/concentration step on lZip-
TipC
18
(Millipore Corp., Bedford, MA, USA) before
analysis by MALDI-MS. Peptide mixtures were loaded
onto the MALDI target, using the dried droplet technique
and a-cyano-4-hydroxycinnamic as matrix, and analysed by
using a Voyager-DE PRO mass spectrometer (Applied
Biosystems, Framingham, MA, USA). Internal-mass calib-
ration was performed with peptides deriving from trypsin
autoproteolysis. The mass spectra were acquired in either
reflectron or linear mode with delayed extraction. Post-
source decay (PSD) fragment ion spectra were acquired for
intense signals after isolation of the appropriate precursor
by using timed ion selection. Fragment ions were refocused
onto the detector by stepping the voltage applied to the
reflectron in the following ratios: 1.000 (precursor ion
segment), 0.960, 0.750, 0.563, 0.422, 0.316, 0.237, 0.178,
0.133, 0.100, 0.075, 0.056 and 0.042 (fragment segments).
Individual segments were superimposed by using the
DATA
EXPLORER
4.0 software (Applied Biosystems). All precursor
ion segments were acquired at low laser power (variable
attenuator ¼ 1950), for less than 200 laser pulses, to avoid
saturating the detector. The laser power was increased to
200 units for all the remaining segments of the PSD

acquisitions. Typically, 300 laser pulses were acquired for
each fragment-ion segment. The PSD data were acquired
with an Acquiris digitizer at a digitization rate of 500 MHz.
Ó FEBS 2003 eEF1A binding to aptameric cytotoxic GT oligomers (Eur. J. Biochem. 270) 3253
PROTEINPROSPECTOR
and
PROWL
software packages were
used to identify spots unambiguously from independent
nonredundant sequence databases [23,24]. Candidates from
peptide-matching analysis were further evaluated by com-
parison with their calculated mass and pI using the
experimental values obtained from bidimensional PAGE.
Bidimensional gel analysis
Proteins of total nuclear extract (30 lg) were precipitated at
)20 °C with four volumes of acetone, washed with cold
methanol, and dried. Pellets were dissolved in 120 lLof
rehydration buffer (Amersham Pharmacia Biotech) con-
taining 8
M
urea, 2% CHAPS, 0.5% immobilized pH
gradient (IPG) buffer (pH 6–11), 65 m
M
dithiothreitol and
0.01% Bromophenol Blue, and used immediately in bidi-
mensional PAGE experiments. IEF was performed on 7-cm
IPG strips (range: pH 6–11) by using the IPGphor Isoelec-
tric Focusing System (Amersham Pharmacia Biotech). The
second dimension was performed on a 12% SDS/PAGE
system after equilibrating the strips for 10 min in SDS

Equilibration buffer containing 50 m
M
Tris/HCl (pH 8.8),
6
M
urea, 30% glycerol, 2% SDS, 2% dithiothreitol and
2.5% iodoacetamide. Gels were then used for Western or
SouthWestern blot analysis, as described above. As internal
normalizer, the presence of the nuclear protein, Ran-GTP,
was detected by Western blot on the same filters used to
analyse eEF1A protein, by using a specific mAb (BD
Pharmingen, CA, USA).
Results
Identification of eEF1A as a nuclear protein specifically
related to the cytotoxicity of GT oligomer in cancer cells
In order to identify the nuclear proteins that specifically
recognize cytotoxic GT oligomers, a 27-mer GT sequence
was used [12]. This oligomer forms a specific CRC with an
apparent molecular weight of 45 ± 7 kDa [12–15]. The
T-lymphoblastic CCRF-CEM cancer cell line was chosen
for this investigation as it was previously used to
demonstrate the specific cytotoxic action of the GT
oligomers [12–15]. Figure 1A shows that in SouthWestern
blots, the labelled GT oligomer bound (in a specific
manner) two main proteins, named P1 and P2, compared
with the binding of a labelled nontoxic CT oligomer, used
as a control. The latter showed a weak interaction with P1
and P2 proteins, whereas it preferentially bound to a
nuclear protein with a mass of 70 kDa (marked by an
asterisk), recognized to the same extent also by the GT

sequence. Binding of the GT oligomer was not a result of
DNA interaction with the more abundant components of
the nuclear extract, as revealed by Ponceau staining of the
immobilized proteins. In fact, many other bands, equally
or more intense than those recognized by the GT
oligomer, were also detected (data not shown). To test
for nuclear enrichment, nuclear and cytoplasmic extracts
were blotted and assayed for the nuclear proteins Egr1
and TBP. The results displayed in Fig. 1B clearly indicate
that the proteins Egr1 and TBP were detected only in the
nuclear fraction. This demonstrated that nuclear extracts
are effectively enriched in nuclear proteins. Moreover,
b-actin, used as a loading control, occurred at a higher
level in the cytoplasmic fraction, similar to the cellular
distribution of the protein.
P1 migrated with an apparent molecular mass similar to
that previously reported for the CRC (45 ± 7 kDa) [12].
In contrast, P2 showed a higher apparent molecular mass.
P1 and P2 were excised from a Coomassie-stained gel,
alkylated and digested. MALDI-MS analysis of the P1
digest yielded a series of peptide-mass values that were used
for nonredundant sequence database searching (Fig. 2A).
Fig. 1. SouthWestern blot analysis of GT oligomer binding to nuclear
proteins and immunoblot of the subcellular fractions. (A) SouthWestern
blot on the nuclear extract. Fifty micrograms of total nuclear extract
derived from CCRF-CEM cells was separated by SDS/PAGE (8%
gel) and then transferred, by semidry blotting, onto a nitrocellulose
filter. The proteins were denatured and renatured as described in the
Materials and methods. One half of the filter was tested for protein–
DNA interactions with a c

32
P-labelled GT probe (GT) and the other
half with a c
32
P-labelled CT oligomer (CT) as a control, at 4 °C(each
probe counted 650 000 c.p.m.). After incubation overnight, the filters
were rinsed and then exposed to Omat XAR Kodak film. (B) Immu-
noblotting of subcellular fractions. Twenty micrograms of cytoplasmic
(lane 1) or nuclear extract (lane 2) fractions of CCRF-CEM cells was
separated by SDS/PAGE (8% gel). After blotting onto nitrocellulose
membrane, the filter was probed with the nuclear-specific antibodies
anti-TBP or anti-Egr1, as described in the Materials and methods. As a
loading control, the presence of b-actin protein was also confirmed by
using specific antibody, as described in the Materials and methods.
3254 B. Dapas et al. (Eur. J. Biochem. 270) Ó FEBS 2003
The comparison of this peptide-mass fingerprint with the
theoretical ones, calculated by an in silico digestion of all
human sequences occurring in the databases, identified P1
as eEF1A. Furthermore, PSD experiments performed on
selected peptide precursor ions (i.e. m/z 1405.4 and 1780.9)
generated internal sequence tags that unambiguously con-
firmed the nature of this protein (data not shown).
Moreover, the spectrum reported in the figure showed the
occurrence of a series of signals that were not interpreted
simply on the basis of the eEF1A sequence, but according to
the post-translational modifications already described
for this protein [25]. It demonstrated the occurrence of
Ne-dimethyllysine (Lys55, Lys165) and Ne-trimethyllysine
(Lys79, Lys318) in the eEF1A sample purified from
T-lymphoblastic CCRF-CEM cancer cells. No data on

the modification status of Lys36 (methylation), Glu301 and
Glu374 (glyceryl-phosphoryl-ethanolamine addition) were
inferred. MS analysis allowed a 50% coverage of the entire
eEF1A sequence. All signals occurring in the spectrum were
assigned to this protein, thus ruling out the possibility that
other polypeptide species comigrated in SDS/PAGE with
eEF1A.
Moreover, the possibility that P1 was the oncogenic
N-terminal truncated form of eEF1A protein, already
known as PTI-1 [26,27], seemed unlikely on the basis of the
signals occurring in the spectrum reported in Fig. 2A. In
fact, at least eight signals matched perfectly with those
expected for eEF1A (MH
+
at m/z 1492.9, 2501.8, 2516.9,
2997.6, 3023.6, 3151.9, 3980.3 and 4108.5) and demonstra-
ted the absence of seven of the eight amino acid substitu-
tions described for PTI-1 (Ala65Met, Glu66Gln, Arg67Ser,
Lys100Gln, Arg247Gly, Ala281Gly and Arg423Cys,
respectively). Moreover, clear MH
+
signals, corresponding
to the N-terminal region of eEF1A, were present; this region
is totally deleted in PTI-1. Similar considerations were taken
into account to exclude the possibility that P1 corresponded
to isoforms of eEF1A other than eEF1A1, already
described.
Similarly, peptide-mass fingerprint analysis by MALDI-
MS identified P2 as nucleolin (data not shown). Different
authors have already reported this protein as being able to

specifically generate a 100-kDa complex with GT oligomers
that form a G-quartet structure, thus exerting a cytotoxic
effect on human cancer cell lines [10,17]. PSD experiments
performed on selected precursor ions (i.e. m/z 2201.3 and
1649.7) allowed internal sequence tags to be obtained,
definitively demonstrating the nature of this species
(Fig. 2B).
The identity of P1 was also assayed by Western-blotting
experiments with a mAb for eEF1A. As illustrated in
Fig. 3A, the protein excised from the Coomassie-stained gel
was recognized by the specific eEF1A antibody (Fig. 3A,
lane 3). As controls, recombinant eEF1A protein (Fig. 3A,
lane 1) and a sample obtained from total nuclear extracts
(Fig.3A,lane2)weretested.
The EMSA with the protein eluted from the P1 band
excised from the Coomassie-stained gel of CCRF-CEM cell
nuclear extracts showed that this protein selectively recog-
nized the GT oligomer with respect to control CT sequence,
similarly to results obtained with the total nuclear extracts
[12–15]. It is noteworthy that all the EMSA and UV cross-
linking assays were performed using a buffer containing
25% glycerol to preserve the activity of eEF1A. As
illustrated in Fig. 3B, the eEF1A recovered from the
P1-excised band showed a stronger interaction when
incubated with the labelled GT oligomer (Fig. 3B, lane 1)
than when incubated with the labelled control CT oligomer
(Fig. 3B, lane 7). Moreover, the presence of a fivefold molar
excess of CT-unlabelled competitor (Fig. 3B, lane 5), did
not completely displace the GT oligonucleotide from the
protein interaction. On the contrary, only a fivefold molar

excess of GT-unlabelled oligonucleotide competitor
removed all the labelled CT control oligomer from the
complex (Fig. 3B, lane 8).
To explore the possibility that eEF1A was the protein
component present in the CRC [12], supershift assays were
performed under native conditions. The results shown in
Fig. 4 demonstrate that a rabbit polyclonal antibody
recognizing eEF1A elicited a specific supershift (marked
by an arrow; Fig. 4, lanes 4 and 5) from the complex. No
supershift resulted from incubation of the nuclear protein
extract with the same amounts of a total rabbit preimmune
serum, which displayed only nonspecific competition
(Fig. 4, lanes 6 and 7). The slight reduction in DNA-binding
Fig. 2. MALDI-MS analysis of P1 and P2 proteins. (A) MALDI-MS
analysis of component P1 following digestion with trypsin. The mass
values reported in the spectrum represent average values. Numbers in
parentheses indicate amino acid residues in the eEF1A sequence.
Possible methylation sites are shown and assigned based on the
observed mass values, eEF1A sequence and previously published
results [25]. Peptides originating from trypsin autoproteolysis are indi-
cated as open circles. (B) Postsource decay (PSD)-MALDI fragment
ion mass spectrum of the P2 tryptic peptide, GLSEDTTEETLK
ESFDGSVR, with MH
+
at m/z 2201.3 (average value). The mass
values reported in the spectrum are indicated as monoisotopic values.
Ó FEBS 2003 eEF1A binding to aptameric cytotoxic GT oligomers (Eur. J. Biochem. 270) 3255
activity shown in the presence of control rabbit total serum
(CRS) could be caused by nonspecific sequestration of
eEF1A by serum proteins.

More interestingly, as illustrated in Fig. 5A, with respect
to the protein of the nuclear extracts, the soluble eEF1A
recovered from the cytoplamic fraction did not bind to a GT
oligomer in SouthWestern blotting. Although comparable
quantities of the protein were loaded onto the gel, as
evidenced by Western blotting, in the cytoplasmic extract,
only the nucleolin band was evident. Similarly to this and to
the previous results [12], in the UV cross-linking assay the
specific CRC displayed by the nuclear extract was not
present in the cytoplasmic sample (Fig. 5B). On the
contrary, the cytoplasmic extract showed a band of about
28 kDa, previously demonstrated to bind to GT in a
nonspecific manner [12].
Characterization of eEF1A in normal and cancer cells
In order to compare the binding properties of the nuclear
eEF1A in normal cells compared with those in cancer
CCRF-CEM cells, human lymphocytes were isolated
from the peripheral blood of normal donors. These cells
were not sensitive to the cytotoxic effect of GT oligomers
Fig. 3. P1 Western blotting analysis and affinity measurements for GT
oligomer. (A) Western blotting analysis. Protein samples were separ-
ated by SDS/PAGE (12% gel) and then transferred onto a nitrocel-
lulose filter and incubated with mAb for eEF1A, as described in the
Materials and methods. Lane 1, bacterial recombinant eEF1A protein
(R eEF1A); lane 2, eEF1A protein from total nuclear extracts (NE
eEF1A); lane 3, P1 band excised from an SDS/PAGE gel (P1). (B) P1
affinity for GT oligomer. P1 protein, excised from an SDS/PAGE gel
loaded with 50 lg of total nuclear extract, was renatured as described
in the Materials and methods. Five microlitres of sample was then
incubated with 2 ng of [c-

32
P]-labelled GT probe (GT) in buffer
(200 m
M
Tris/HCl, pH 7.5, containing 750 m
M
KCl, 10 m
M
dithio-
threitol, 50 lgÆmL
)1
BSA) in the absence (lane 1) or in the presence of
10 ng (lane 2), 20 ng (lane 3), 50 ng (lane 4), 100 ng (lane 5) or 200 ng
(lane 6) of nonlabelled CT oligomer. An identical aliquot was incu-
batedwith2ngofc
32
P-labelled CT probe (CT) in the absence (lane 7)
or in the presence of 10 ng (lane 8), 20 ng (lane 9), 50 ng (lane 10),
100 ng (lane 11) or 200 ng (lane 12) of nonlabelled GT oligomer. The
two probes were added to the sample at the same specific activity
( 10 000 c.p.m.). Labelled GT oligomer incubated without P1 (lane
13) and labelled CT oligomer incubated without P1 (lane 14) were used
as controls. After 20 min of incubation at room temperature, the
samples were loaded onto an 8% polyacrylamide gel in 0.5 · Tris/
borate/EDTA (TBE) buffer and electrophoresed at 4 °C. The dried gel
was exposed to autoradiographic film. The arrow indicates the specific
complex.
Fig. 4. Supershift assay experiments. Proteins from CCRF-CEM cell
nuclear extracts (0.5 lg) were incubated with or without the indicated
amounts of specific polyclonal antibody (Ab eEF1A) or control rabbit

total serum (CRS), for 2.5 h at room temperature. Then, 2 ng of c
32
P-
labelled GT oligomer was added to the samples, as reported in the
Materials and methods. After a further 30 min of incubation at room
temperature, the samples were loaded onto a 7% polyacrylamide gel in
0.5 · Tris/borate/EDTA (TBE) buffer and electrophoresed at 4 °C.
The gel was dried and exposed to Omat XAR Kodak film. c
32
P-
Labelled GT oligomer was incubated with buffer (lane 1), with 9 lgof
polyclonal anti-eEF1A (lane 2), with 9 lg of total CRS (lane 3),
with nuclear proteins and 9 lg of polyclonal anti-eEF1A (lane 4), with
nuclearproteinsand0.9lg of polyclonal anti-eEF1A (lane 5), with
nuclearproteinsand9lg of total CRS (lane 6), with nuclear proteins
and 0.9 lg of total CRS (lane 7) or with nuclear proteins only (lane 8).
The arrow indicates the supershift.
3256 B. Dapas et al. (Eur. J. Biochem. 270) Ó FEBS 2003
(Fig. 6A), and their nuclear proteins did not form the
CRC with the GT sequence (marked by arrow), as shown
by EMSA or UV cross-linking assays (Fig. 6B,C). To
investigate the binding properties of the lymphocyte
eEF1A protein, SouthWestern blots were performed. It
was found that the CCRF-CEM nuclear extracts con-
tained higher amounts of eEF1A than those of human
lymphocytes. In fact, Western blotting experiments dem-
onstrated that the relative amount of eEF1A recovered
from lymphocyte nuclear extracts was 2.7 ± 0.8-fold less
than that obtained from cancer CCRF-CEM T lympho-
blasts (mean of three independent experiments). For this

reason, the SouthWestern assay was performed after
Fig. 5. SouthWestern blot, Western blot and UV cross-linking analysis
of cytoplasmic extracts. (A) SouthWestern blot. Twenty-five micro-
grams of total protein from cytoplasm or nuclear extracts, previously
normalized by comparison on a Coomassie-stained gel, were separated
by SDS/PAGE (8% gel) and transferred onto a nitrocellulose filter, as
described in the Materials and methods. The proteins were denatured,
renatured and the filter hybridized with a c
32
P-labelled GT probe at
4 °C. After overnight incubation, the filter was rinsed and then
exposed to Omat XAR Kodak film, as described in the Materials and
methods. The same samples were used for Western blotting analysis
performed with the same amount of the cytoplasm and nuclear pro-
teins used in the SouthWestern blot. The specific protein was con-
firmed by using the monoclonal anti-eEF1A with the conditions
described in the Materials and methods. (B) UV crosslinking assay.
Two micrograms of total proteins derived from the cytoplasm or
nuclear extracts were incubated in buffer containing 25% glycerol,
with c
32
P-labelled GT probe in the presence of 1 lg poly(dIdC) and
1 lg of CT as competitors, as described in the Materials and methods.
After 25 min of incubation at room temperature, the samples were
cross-linked by UV exposure and then denatured and separated by
SDS/PAGE (12% gel). The dried gel was then exposed to Omat XAR
Kodak film. NE, nuclear extract; CE, cytoplasmic extract.
Fig. 6. Effect of GT oligomer on cellular growth and on nuclear protein
binding in human lymphocytes. (A) Effect of GT oligomer on cellular
growth or viability. A total of 10

4
CCRF-CEM cells or peripheral
normal human lymphocytes were seeded in 100 lL of complete
medium on 96-well microtiter plates. After 4 h of incubation, 7.5 l
M
of
GT oligomer or control CT sequence were added to the cells. The
percentage of viable cells was assayed after 72 h of incubation by
determining the incorporation of 3-(4,5-dimethylthiazol-2-yl)-2,5-di-
phenyl-tetrazolium bromide, as described in Materials and methods.
(B) EMSA assay. Two micrograms of total nuclear proteins derived
from CCRF-CEM cells or from human lymphocytes were incubated
with 1.5 lg of poly(dIdC), 1 lgofCTand1ngofc
32
P-labelled GT
oligomer in a buffer containing 25% glycerol, as described in the
Materials and methods. After incubation at room temperature for
30 min, the samples were loaded onto a 7% polyacrylamide gel in Tris/
borate/EDTA (TBE) buffer and run at 4 °C. (C) UV cross-linking
assay. Two micrograms of total nuclear proteins derived from CCRF-
CEM cells or from human lymphocytes were incubated with 1.5 lgof
poly(dIdC), 1 lgofCTand1ngofc
32
P-labelled GT oligomer, as
described in the Materials and methods. The samples were then
exposed for 10 min to a 302 nm UV light, added to SDS/PAGE
loading buffer and separated by SDS/PAGE (12% gel). The arrows
indicated the specific cytotoxicity-related complex.
Ó FEBS 2003 eEF1A binding to aptameric cytotoxic GT oligomers (Eur. J. Biochem. 270) 3257
normalizing the quantities of loaded proteins on a

Coomassie-stained gel by referring to the P1 band (known
tocorrespondtoeEF1A).AsreportedinFig.7A,inthe
CCRF-CEM sample the GT oligomer recognized nucleo-
lin (marked by a black arrow), eEF1A (marked by a
white arrow) as well as the nonspecific 70 kDa protein
(marked by an asterisk). On the contrary, in human
lymphocytes the GT oligomer did not bind to eEF1A, but
it significantly recognized nucleolin. To rule out that
protein-degradation artefacts might account for the lack
of specific recognition, the same amounts of protein used
in SouthWestern experiments were assayed by Western
blotting. As shown in Fig. 7B, the amount of eEF1A in
human lymphocytes was comparable to that in CCRF-
CEM cancer cells. On the contrary, b-actin showed
significant differences in the two samples, in agreement
with the higher quantity of total proteins loaded in the
lymphocyte sample, as mentioned above.
Previous data demonstrated that the eEF1A protein is
post-translationally modified [25,28,29]. In order to detect
whether this protein presented a different molecular nature
in normal and cancer cells, we performed a comparative
bidimensional PAGE analysis of nuclear extracts coupled
to Western blotting analysis with an eEF1A mAb. As
an internal normalizer of loading amounts and focusing
position, the nuclear protein, Ran-GTP, was used and
identified by a specific mAb. The data reported in Fig. 8A
clearly showed, in T-lymphoblastic CCRF-CEM cancer
Fig. 7. Comparative SouthWestern blot for CCRF-CEM cells and
normal lymphocytes. (A) SouthWestern blot. Twenty-five micrograms
of total nuclear protein from CCRF-CEM cells and 50 lgoftotal

nuclear protein from normal human lymphocytes were separated by
SDS/PAGE (8% gel) and transferred onto a nitrocellulose filter, as
described in the Materials and methods. The proteins were denatured
and renatured as described above, and the filter was hybridized with
c
32
P-labelled GT probe at 4 °C. After overnight incubation with the
probe, the filters were rinsed, as described in the Materials and
methods, and then exposed to Omat XAR Kodak film. (B) Western-
blotting analysis. Protein samples reported in (A) were analysed for
eEF1A and b-actin content.
Fig. 8. Bidimensional PAGE analysis of nuclear elongation factor 1
alpha (eEF1A). (A) Thirty micrograms of nuclear extracts from CCRF
CEM cells and normal human lymphocytes, calculated from evalua-
tion of the protein content in a Coomassie-stained gel, were analysed
by bidimensional PAGE, as described in the Materials and methods.
The presence of eEF1A was tested by Western blot analysis by using
the specific antibody anti-eEF1A. As an internal normalizer of loading
amount and focusing position, the presence of the constitutive nuclear
transporter, Ran-GTP, was also tested by using a specific monoclonal
antibody. (B) Bidimensional PAGE analysis of other samples, con-
firming the reproducibility of data obtained. Thirty micrograms of
nuclear extracts from normal human lymphocytes were similarly
analysed by bidimensional PAGE and compared with CCRF-CEM
nuclear extracts. The presence of eEF1A was confirmed by Western
blot analysis using the specific antibody, anti-eEF1A. Only the higher
magnification of IEF of the eEF1A region is reported.
3258 B. Dapas et al. (Eur. J. Biochem. 270) Ó FEBS 2003
cells, the presence of two different clusters of eEF1A
isoforms (cluster 1 and cluster 2). The apparent pI of cluster

1( pH 9.0), was calculated theoretically by considering
the pH gradient linear, as indicated by the manufacturer. In
nuclear extracts from lymphocytes, a similar species focused
near the same pH. Interestingly, cluster 2, which was absent
in lymphocytes, focused at an apparent pH of 10.5. Thus,
it accounted for a more basic pI and for species different
from those present in lymphocytes. However, taking apart
the undoubtedly much more basic nature of the faster-
migrating isoform of eEF1A in cancer cells, the estimation
of its pI remains merely indicative. A magnified image of
this analysis, performed with a different nuclear extract, is
reported in Fig. 8B. It is clear that human lymphocytes
displayed only the main constitutive species that migrates
in correspondence to cluster 1 of CCRF-CEM cells (see
Fig. 8B). To elucidate the GT oligomer binding behaviour
of the isoforms found in the cancer cell sample, a
SouthWestern assay was performed after analysis by
bidimensional PAGE. The antibody recognition of the
proteins performed on the same filter was prevented by
the SouthWestern treatment. Therefore, identification of the
eEF1A–oligomer interaction was carried out by matching
the SouthWestern results with those of the Western blotting
of bidimensional PAGE performed on the same sample,
under identical experimental conditions. The perfect match
between the Western blot signals of bidimensional PAGE
(Fig. 9A) and the autoradiographic signals found in the
SouthWestern blot (Fig. 8B), unequivocally demonstrated
that the protein reacting in SouthWestern blots was eEF1A.
Furthermore, the SouthWestern blot, reported in Fig. 9B,
showed that the labelled GT oligomer mainly recognized the

more basic form of eEF1A, whereas a very weak interaction
was found for the isoform of eEF1A focusing at a pH of
9.0. The recognition seemed highly specific because, under
these experimental conditions, no other interactions were
detected on the filter. Moreover, no significant interaction in
bidimensional PAGE SouthWestern blots was found on
human lymphocytes at the position corresponding to
eEF1A (see Fig. 10), once more indicating that normal
eEF1A did not react with the GT oligomer.
Fig. 9. Comparative analysis of bidimensional PAGE analysis of
SouthWestern and Western blots for CCRF-CEM cells. Two samples of
50 lg of total nuclear protein from CCRF-CEM cells were separated,
in parallel, by bidimensional PAGE and blotted onto nitrocellulose
filters, as described in the Materials and methods. (A) Western blot-
ting. The filter was assayed using anti-eEF1A mAb, as described in the
Materials and methods. The position of the eEF1A protein was con-
firmed by revealing the presence of the Ran–GTP protein. (B)
SouthWestern blotting. The filter was assayed for SouthWestern
blotting, as described in the experimental section using, as probe, c
32
P-
labelled GT oligomer and then exposed to Omat XAR Kodak film.
Fig. 10. Comparative analysis of bidimensional PAGE, SouthWestern
and Western blotting for human lymphocytes. Two samples of 50 lgof
total nuclear protein of normal human lymphocytes, previously nor-
malized with respect to CCRF-CEM cell protein by a Coomassie-
stained gel, were separated, in parallel, by bidimensional PAGE and
blotted onto nitrocellulose filters, as described in the Materials and
methods. (A) Western blotting. The filter was assayed for Western blot
using anti-eEF1A mAb, as described in the Materials and methods.

The position of the eEF1A protein was focused by revealing the
presence of the Ran–GTP protein. (B) SouthWestern blotting. The
filter was assayed for SouthWestern blot as described in the Materials
and methods using, as probe, c
32
P-labelled GT oligomer and then
exposedtoOmatXARKodakfilm.
Ó FEBS 2003 eEF1A binding to aptameric cytotoxic GT oligomers (Eur. J. Biochem. 270) 3259
Discussion
Eukaryotic EF1A is a protein belonging to the GTP-
binding elongation factor family, which promotes the GTP-
dependent binding of aminoacyl-tRNA to the A-site of
ribosomes during protein biosynthesis and the capture of
deacylated tRNA at the exit site and its delivery to synthase
[30]. It has been suggested that eEF1A might serve also as a
downstream component of growth-signalling pathways,
possibly through its capability to interact with actin, thus
promoting cell transformation [31]. An overexpression of
eEF1A has been shown in many tumours [32], and eEF1A
has been associated with a highly metastatic potential in
cancer cells [33]. The regulation of eEF1A expression by
extracellular stimuli depends on human epidermal growth
factor (EGF) receptor family members that are widely
deregulated in human cancers [34].
eEF1A is the second most abundant protein in the cell;
whereas the b, c and d subunits of eEF1 are predominantly
located in the cytoplasm, a considerable fraction of the a
subunit can be found either in the cytoplasm or the nucleus
[35,36]. The involvement of eEF1A in the regulation of
nuclear processes includes the accumulation of a nuclear

complex with vigilin, for exporting tRNA [37], and with
ZPR1, for inducing cell proliferation upon mitogen stimu-
lation [38]. Thus, the elucidation of a possible role for the
nuclear fraction of eEF1A in modulating nuclear func-
tion and gene expression could gain new insights in
tumorigenesis.
In this manuscript, we demonstrate that eEF1A, isolated
from nuclear extracts of CCRF-CEM cancer cells, is
specifically recognized by a cytotoxic GT sequence. This
protein was found to be the polypeptide component of the
CRC,basedonMALDI-MSanalysis,Westernblotting
experiments and supershift assays. In contrast, the GT
oligomer did not bind to the eEF1A of normal human
lymphocytes and these cells were not sensitive to the
cytotoxic action of the GT. It should be noted that nucleolin
was also recognized by the GT oligomer [10,17]. However,
under native conditions, the more abundant CRC observed,
migrated with an apparent mass not associated with the
nucleolin–oligomer complex [12–15], probably because the
GT sequence used does not form, in appreciable quantity,
the G-quartet structure specifically recognized by this
protein, as revealed by gel electrophoresis and circular
dichroism studies [12,14]. Moreover, nucleolin, and not
eEF1A, was bound by the GT oligomer in lymphocyte
sample on one-dimensional SouthWestern assay; however,
lymphocyte viability was not affected by GT. In cytoplasmic
extracts, the nucleolin was found to bind to GT in a
SouthWestern assay, but not in EMSA or UV crosslinking
assays. Furthermore, we analysed a G-rich GT sequence
able to form the G-quartet structure and thus to bind to

nucleolin. We found that this oligomer did not elicit
cytotoxicity on CCRF-CEM cells, although it was effi-
ciently taken up by the cells. More interestingly, in UV
cross-linking competition experiments, this sequence did not
displace the labelled GT from the CRC but from the less
represented lower migrating complex (corresponding to
nucleolin) (data not shown). On this basis, we can hypo-
thesize a minor involvement of nucleolin in the mechanism
of cytotoxicity elicited by the GT oligomers [12]. It seemed
probable that the reactivity of the nucleolin was related
much more to the experimental conditions of the immobi-
lized protein on the SouthWestern assay, than to a native
binding affinity for the GT. Nevertheless, we cannot
completely exclude that other, less-abundant proteins can
contribute to this effect.
Binding assays by SouthWestern experiments demon-
strated that nuclear eEF1A affinity for GT oligomers was
significantly higher than that measured for the control CT
sequence. Moreover, in EMSA assays a significant quantity
of GT oligomer remained bound to the protein, derived
from the excised P1 band, in the presence of a 50-fold molar
excess of CT oligomer competitor, whereas only a fivefold
molar excess of GT oligomer was sufficient to release all
control CT oligomer from the complex. It is noteworthy
that the P1 band excised from the Coomassie-stained gel of
a normal lymphocyte sample failed to form complexes in
EMSA with GT oligomer (data not shown). Furthermore,
overloaded protein samples from normal lymphocytes did
not show significant interaction between the eEF1A protein
and the cytotoxic GT oligomer in SouthWestern assays.

Accordingly, GT oligomers did not elicit cytotoxic action on
these cells, and did not form the CRC with the nuclear
proteins when a fourfold increase in protein content was
loaded onto the gel (data not shown). These results
underline that eEF1A from CCRF-CEM cell nuclear
extracts displays specificity in recognizing GT oligomers,
and the selective cytotoxic action on CCRF-CEM cells
suggests a possible role for eEF1A in maintaining the
viability and proliferative activity of cancer cells. One
hypothesis may be that these oligomers exert their action by
blocking the binding of eEF1A to its ligand in cancer cells,
perhaps to zinc finger proteins involved in the modulation
of cell proliferation, as proposed by Gangwani et al. [38].
Bidimensional PAGE analysis of eEF1A combined with
a specific Western blotting assay showed the occurrence of
two distinct clusters of spots in T-lymphoblastic CCRF-
CEM cells, whereas normal lymphocytes presented only one
cluster. In particular, the newly occurring components in
cancer cells (cluster 2) focused at a more basic pH. This
result could hypothetically explain the higher affinity of the
protein towards oligonucleotides simply on the basis of a
charge increase at specific amino acids in its nucleotide-
binding site, but not its binding selectivity for the GT
sequences.
Different post-translational modifications have been
reported to occur in the eEF1A polypeptide chain, such as
phosphorylation, methylation and glyceryl-phosphoryl-
ethanolamine addition [25,27,29,39,40] but, to date, their
functional significance has not been totally solved. Differ-
ences in the level of phosphorylation of eEF1A have already

been reported to be associated with variation in binding
affinity towards viral genomic RNA, as well as to regulative
interconversion between active and inactive forms [41]. Our
findings should not sustain the hypothesis that eEF1A
propensity for recognition of GT oligomers in CCRF-CEM
cancer cells might be related merely to an increase of the
phosphorylation state. On the contrary, the modification of
eEF1A mobility on bidimensional PAGE by increasing its
pI value should be associated with the presence of other
post-translational modifications. Methylation of eEF1A has
been significantly associated with SV40 transformation in
3260 B. Dapas et al. (Eur. J. Biochem. 270) Ó FEBS 2003
mouse 3T3B cells [42]. It was suggested that this modifica-
tion should account for differences in growth properties for
the different cell types. Similarly, in Mucor racemous it was
observed that, during morphogenesis, a sixfold elevation in
eEF1A specific activity is accompanied by a dramatic
increase in protein methylation at as many as nine lysine
residues, without any change in the protein or mRNA levels
[43]. This modification, as well as other post-translational
modifications, not deduced by simple mass fingerprint
analysis on the resolved spots, cannot be excluded.
Furthermore, the eventual presence of specific amino acid
substitutions in the eEF1A gene, producing the more basic
isoform observed by bidimensional PAGE, has to be
considered. However, the possibility that this basic species
corresponded to the eEF1A2 isoform already known,
described in brain and skeletal muscle [44], or to PTI-1,
an oncogenic truncated form of eEF1A [26], was ruled out
on the basis of the peptide-fingerprint experiments reported

above.
By using SouthWestern assays on bidimensional PAGE
of CCRF-CEM cell nuclear extracts, we observed that the
more basic isoform of eEF1A binds to the GT oligomer,
whereas only a weak interaction was found for the eEF1A
migrating like the normal constitutive molecule. Accord-
ingly, eEF1A from normal human lymphocytes did not
recognize GT oligomer at all. On this basis, it could be
tempting to speculate that different isoforms of nuclear
eEF1A, in particular a more basic molecule, should
account for the different sensitivities to cytotoxic effect
exerted by GT oligomers. These differences might be
related to a variable degree of post-translational process-
ing, although the high shift in the apparent pI of the more
basic molecules cannot alone be simply explained by
modifications such as methylation of the lysine residues.
Even if theoretically possible, this phenomenon might
account for a low probable high number of methylated
lysine residues in a protein with a total low-density
charge. Thus, in these high basic proteins it might be that
either methylation of the lysine or specific substitutions in
the amino acid sequence, increasing the number of basic
residues of lysine, arginine and histidine, could occur.
However it seemed probable that the eEF1A isoforms
could act as a controlling event in maintaining the
viability and/or promoting the growth of T-lymphoblastic
tumour cells. Supporting evidence for the specific role of
this nuclear-associated eEF1A species come from the
observation that the soluble cytoplasmic eEF1A from
cancer cells did not bind cytotoxic GT oligomers in

EMSA, UV cross-linking or SouthWestern assays.
Further investigations on the structural characterization
of eEF1A in cancer cells could highlight differences among
its post-translational processing, as well as reveal gene
mutations between normal and transformed cells, to
elucidate the role of this protein in tumour growth
maintenance.
Acknowledgements
The authors thank Dr G. M. C. Janssen, for the gift of the polyclonal
eEF1A antibody and the eEF1A cDNA cloned into pET11a expression
vector, and Prof. G. Manzini, for the useful critical advice. This work
was supported by grants from National Research Council.
References
1. Baker, B.F. & Monia, B.P. (1999) Novel mechanisms for anti-
sense-mediated regulation of gene expression. Biochim. Biophys.
Acta 1489, 3–18.
2. Praseuth,D.,Guieysse,A.L.&He
´
le
`
ne, C. (1999) Identification of
a triplex DNA-binding protein from human cells. Biochim. Bio-
phys. Acta 1489, 181–206.
3. Agrawal, S. & Kandimalla, E.R. (2000) Antisense and/or immu-
nostimulatory oligonucleotide therapeutics. Mol. Med. Today 6,
72–81.
4. Roberts, R.W. & Ja, W.W. (1999) In vitro selection of nucleic acids
and proteins: What are we learning? Curr. Opin. Struct. Biol. 9,
521–529.
5. Seroussi, E. & Lavi, S. (1993) Replication protein A is the major

single-stranded DNA binding protein detected in mammalian cell
extracts by gel retardation assays and UV cross-linking of long
and short single-stranded DNA molecules. J. Biol. Chem. 268,
7147–7154.
6. Coverley, D., Kenny, M.K., Lane, D.P. & Wood, R.D. (1992) A
role for the human single-stranded DNA binding protein HSSB/
RPA in an early stage of nucleotide excision repair. Nucleic Acids
Res. 20, 3873–3880.
7. Yee, H.A., Wong, A.K.C., van de Sande, J.H. & Rattner, J.B.
(1991) Identification of novel single-stranded d[TC]n binding
proteins in several mammalian species. Nucleic Acids Res. 19,
949–953.
8. Michelotti,E.F.,Tomonaga,T.,Krutzsch,H.&Levens,D.(1995)
Cellular nucleic acid binding protein regulates the CT element of
the human c-myc protooncogene. J. Biol. Chem. 270, 9494–9499.
9. Yano-Yanagisawa, H., Li, Y., Wang, H. & Kowi, Y. (1995) Sin-
gle-stranded DNA binding proteins isolated from mouse brain
recognize specific trinucleotide repeat sequences in vitro. Nucleic
Acids Res. 23, 2654–2660.
10. Xu, X., Hamhouyia, F., Thomas, S.D., Burke, T.J., Girvan, A.C.,
McGregor,W.G.,Trent,J.O.,Miller,D.M.&Bates,P.J.(2001)
Inhibition of DNA replication and induction of S phase cell
cycle arrest by G-rich oligonucleotides. J. Biol. Chem. 276,
43221–43230.
11. Lang,R.,Hultner,L.,Lipford,G.B.,Wagner,H.&Heeg,K.
(1999) Guanosine-rich oligodeoxynucleotides induce proliferation
of macrophage progenitors in cultures of murine bone marrow
cells. Eur. J. Immunol. 29, 3496–3506.
12. Scaggiante,B.,Morassutti,C.,Dapas,B.,Tolazzi,G.,Ustulin,F.
& Quadrifoglio, F. (1998) Human cancer cell lines growth

inhibition by GTn oligodeoxyribonucleotides recognizing single-
stranded DNA-binding proteins. Eur. J. Biochem. 252, 207–215.
13. Morassutti,C.,Dapas,B.,Scaggiante,B.,Paroni,G.,Xodo,L.&
Quadrifoglio, F. (1999) Effect of oligomer length and base sub-
stitutions on the cytotoxic activity and specific nuclear protein
recognition of GTn oligonucleotides in the human leukemic
CCRF-CEM cell line. Nucleosides Nucleotides 18, 1711–1716.
14. Morassutti, C., Scaggiante, B., Dapas, B., Xodo, L., Tell, G. &
Quadrifoglio, F. (1999) Effect of phosphorothioate modifications
on the ability of GTn oligodeoxynucleotides to specifically
recognize single-stranded DNA-binding proteins and to affect
human cancer cellular growth. Biochimie 81, 1115–1122.
15. Scaggiante, B., Dapas, B. & Quadrifoglio, F. (2001) Alpha-
anomeric configuration of GT oligodeoxynucleotide leads to loss
of the specific aptameric and cytotoxic properties retained by
the beta-anomeric analog. Antisense Nucleic Acid Drug Dev. 11,
395–399.
16. Aharoni, A., Baran, N. & Manor, H. (1993) Characterization of a
multisubunit human protein which selectively binds single stran-
ded d[GA]n and d[GT]n sequence repeats in DNA. Nucleic Acids
Res. 21, 5221–5228.
Ó FEBS 2003 eEF1A binding to aptameric cytotoxic GT oligomers (Eur. J. Biochem. 270) 3261
17. Bates, P.J., Kahlon, J.B., Thomas, S.D., Trent, J.O. & Miller,
D.M. (1999) Antiproliferative activity of G-rich oligonucleotides
correlates with protein binding. J. Biol. Chem. 274, 26369–26377.
18. Monks, A., Scudiero, D., Skehan, P., Shoemaker, R., Paull, K.,
Vistica,D.,Hose,C.,Langley,J.,Cronise,P.,Vaigro-Wolff,A.,
Gray-Goodrich, M., Campbell, H., Mayo, J. & Boyd, M. (1991)
Feasibility of a high-flux anticancer drug screen using a diverse
panel of cultured human tumor cell lines. J. Natl. Cancer Inst. 83,

757–766.
19. Bradford, M.M. (1976) A rapid and sensitive method for the
quantitation of microgram quantities of protein utilizing the
principle of protein-dye binding. Anal. Biochem. 72, 248–254.
20. Laemmli, U.K. (1970) Cleavage of structural proteins during the
assembly of the head of bacteriophage T4. Nature 227, 680–685.
21. Tell, G., Pines, A., Paron, I., D’Elia, A., Bisca, A., Kelley, M.R.,
Manzini, G. & Damante, G. (2002) Mitochondrial localization of
APE/Ref-1 in thyroid cells. J. Biol. Chem. 277, 14564–14574.
22. Allegrini, S., Scaloni, A., Ferrara, L., Pesi, R., Pinna, P., Sgarrella,
F., Camici, M., Eriksson, F. & Tozzi, M.G. (2001) Bovine cyto-
solic 5¢-nucleotidase acts through the formation of an aspartate
52-phosphoenzyme intermediate. J. Biol. Chem. 276, 33526–33532.
23. Clauser, K.R., Baker, P.R. & Burlingame, A.L. (1999) Role of
accurate mass measurement [+/- 10 ppm] in protein identification
strategies employing MS or MS/MS and database searching. Anal.
Chem. 71, 2871–2882.
24. Zhang, W. & Chait, B.T. (2000) ProFound: an expert system for
protein identification using mass spectrometric peptide mapping
information. Anal. Chem. 72, 2482–2489.
25.Dever,T.E.,Costello,C.E.,Owens,C.L.,Rosenberry,T.L.&
Merrick, W.C. (1989) Location of seven post-translational
modifications in rabbit elongation factor 1 alpha including
dimethyllysine, trimethyllysine, and glycerylphosphorylethanol-
amine. J. Biol. Chem. 264, 20518–20525.
26. Gopalkrishnan, R.V., Su, Z.Z., Goldstein, N.I. & Fisher, P.B.
(1999) Translational infidelity and human cancer: role of the PTI-1
oncogene. Int. J. Biochem. Cell. Biol. 31, 151–162.
27. Sun, Y., Lin, J., Katz, A.E. & Fisher, P.B. (1997) Human prostatic
carcinoma oncogene PTI-1 is expressed in human tumor cell lines

and prostate carcinoma patient blood samples. Cancer Res. 57,
18–23.
28. Sherman, M. & Sypherd, P.S. (1989) Role of lysine methylation in
the activities of elongation factor 1 alpha. Arch. Biochem. Biophys.
275, 371–378.
29. Venema, R.C., Peters, H.I. & Traugh, J.A. (1991) Phosphoryla-
tion of elongation factor 1 (EF-1) and valyl-tRNA synthetase by
protein kinase C and stimulation of EF-1 activity. J. Biol. Chem.
266, 12574–12580.
30. Petrushenko, Z.M., Negrutskii, B.S., Ladokhin, A.S., Budkevich,
T.V., Shalak, V.F. & El’skaya, A.V. (1997) Evidence for the for-
mation of an unusual ternary complex of rabbit liver EF-1alpha
with GDP and deacylated tRNA. FEBS Lett. 407, 13–17.
31. Tatsuka,M.,Mitsui,H.,Wada,A.,Nagata,H.,Nojima,H.&
Okayama, H. (1992) Elongation factor-1 alpha gene determines
susceptibility to transformation. Nature 359, 333–336.
32. Grant, A.G., Flomen, R.M., Tizard, M.L. & Grant, D.A. (1992)
Differential screening of a human pancreatic adenocarcinoma
lambda gt11 expression library has identified increased transcrip-
tion of elongation factor EF-1 alpha in tumour cells. Int. J. Cancer
50, 740–745.
33. Duttaroy, A., Bourbeau, D., Wang, X.L. & Wang, E. (1998)
Apoptosis rate can be accelerated or decelerated by overexpression
or reduction of the level of elongation factor-1 alpha. Exp. Cell
Res. 238, 168–176.
34. Talukder, A.H., Jorgensen, H.F., Mandal, M., Mishra, S.K.,
Vadlamudi, R.K., Clark, B.F., Mendelsohn, J. & Kumar, R.
(2001) Regulation of elongation factor-1alpha expression by
growth factors and anti-receptor blocking antibodies. J. Biol.
Chem. 276, 5636–5642.

35. Sanders, J., Brandsma, M., Janssen, G.M., Dijk, J. & Moller, W.
(1996) Immunofluorescence studies of human fibroblasts demon-
strate the presence of the complex of elongation factor-1
beta gamma delta in the endoplasmic reticulum. J. Cell Sci. 109,
1113–1117.
36. Janssen, G.M., van Damme, H.T., Kriek, J., Amons, R. & Moller,
W. (1994) The subunit structure of elongation factor 1 from
Artemia. Why two alpha-chains in this complex? J. Biol. Chem.
269, 31410–31417.
37. Kruse, C., Grunweller, A., Willkomm, D.K., Pfeiffer, T., Hart-
mann, R.K. & Muller, P.K. (1998) tRNA is entrapped in similar,
but distinct, nuclear and cytoplasmic ribonucleoprotein com-
plexes, both of which contain vigilin and elongation factor 1 alpha.
Biochem. J. 329, 615–621.
38. Gangwani, L., Mikrut, M., Galcheva-Gargova, Z. & Davis, R.J.
(1998) Interaction of ZPR1 with translation elongation factor-
1alpha in proliferating cells. J. Cell Biol. 143, 1471–1484.
39. Das, T., Mathur, M., Gupta, A.K., Janssen, G.M. & Banerjie,
A.Y. (1998) RNA polymerase of vesicular stomatitis virus spe-
cifically associates with translation elongation factor-1 alpha-
betagamma for its activity. Proc. Natl Acad. Sci. USA 95, 1449–
1454.
40. Ransom, W.D., Lao, P.C., Gage, D.A. & Boss, W.F. (1998)
Phosphoglycerylethanolamine posttranslational modification of
plant eukaryotic elongation factor 1alpha. Plant Physiol. 117,
949–960.
41. Blackwell, J.L. & Brinton, M.A. (1997) Translation elongation
factor-1 alpha interacts with the 3¢ stem–loop region of West Nile
virus genomic RNA. J. Virol. 71, 6433–6444.
42. Coppard, N.J., Clark, B.F. & Cramer, F. (1983) Methylation of

elongation factor 1 alpha in mouse 3T3B and 3T3B/SV40 cells.
FEBS Lett. 164, 330–334.
43. Fonzi, W.A., Katayama, C., Leathers, T. & Sypherd, P.S. (1985)
Regulation of protein synthesis factor EF-1 alpha in Mucor
racemosus. Mol. Cell. Biol. 5, 1100–1103.
44. Chambers, D.M., Peters, J. & Abbott, C.M. (1998) The lethal
mutation of the mouse wasted [wst] is a deletion that abolishes
expression of a tissue-specific isoform of translation elongation
factor 1alpha, encoded by the Eef1a2 gene. Proc. Natl Acad. Sci.
USA 95, 4463–4468.
3262 B. Dapas et al. (Eur. J. Biochem. 270) Ó FEBS 2003