Báo cáo khoa học: Interactions of elongation factor EF-P with the Escherichia coli ribosome doc

Interactions of elongation factor EF-P with the Escherichia
coli ribosome
Hiroyuki Aoki
1
, John Xu
1
, Andrew Emili
2
, John G. Chosay
3
, Ashkan Golshani
4
and
M. Clelia Ganoza
1
1 University of Toronto, C.H. Best Institute, Canada
2 Terrance Donnelly Centre for Cellular and Biomolecular Research (Donnelly CCBR), Toronto, Canada
3 Pfizer Pharmaceuticals, Ann Arbor, MI, USA
4 Department of Biology, Carleton University, Ottawa, Canada
In all cells, ribosomes are inert in the absence of
several proteins that promote each facet of protein
synthesis. In eubacteria, initiation factors (IF-1, IF-2,
IF-3) bind to the 30S subunit, whereas elongation
factors (EF-Tu and EF-G) and termination factors
(RF-1, RF-2 and RF-3) bind to 70S ribosomes.
Reconstitution of synthesis using all of these homoge-
neous proteins revealed that they are insufficient for
synthesis directed by a native mRNA and that several
other proteins are required [1–6]. One of these pro-
teins, EF-P, stimulates peptide bond synthesis by
70S ribosomes [1,2,6–8]. The gene encoding EF-P,

efp, occurs at a unique site at 94.3 min on the Esc-
herichia coli chromosome [9]. Interruption of the efp
gene is lethal to the cell and results in an abrupt
cessation of protein synthesis, specifically in an
impairment of peptide-bond formation [10]. Cells
deleted for efp grow only in the presence of the efp
gene in trans. Thus, the efp gene is essential for cell
growth and viability [10,11].
Reconstitution experiments and the effect of various
antibiotics suggest that EF-P binds to a site on ribo-
somes that differs from the binding site of most other
translation factors. Thus, EF-P requires ribosomal pro-
tein L16 for its action, but not L7 ⁄ L12, L6 or L11
[12]. All of the G proteins, IF-2, EF-Tu, EF-G and
RF-3 act with L7 ⁄ L12 in the sarcin-ricin loop to trig-
ger GTP hydrolysis, and L6 is required for the action
of the release factors [6]. Several antibiotics perturb
the action of EF-P. Notably, streptomycin, which
impairs translational fidelity, is a strong inhibitor of
the EF-P reaction [12]. However, other antibiotics that
alter the fidelity of decoding such as neomycin and
kasugamycin, have no effect on the EF-P-mediated
Keywords
A-site; eIF5A; elongation factor EF-P;
ribosomes; translocation
Correspondence
H. Aoki, University of Toronto, C.H. Best
Institute, 112 College Street, Toronto,
Ontario M5G 1L6, Canada
Fax: +1 416 978 8528

Tel: +1 416 978 8918
E-mail:
(Received 1 August 2007, revised 3 Decem-
ber 2007, accepted 10 December 2007)
doi:10.1111/j.1742-4658.2007.06228.x
EF-P (eubacterial elongation factor P) is a highly conserved protein essen-
tial for protein synthesis. We report that EF-P protects 16S rRNA near the
G526 streptomycin and the S12 and mRNA binding sites (30S T-site).
EF-P also protects domain V of the 23S rRNA proximal to the A-site
(50S T-site) and more strongly the A-site of 70S ribosomes. We suggest
that EF-P: (a) may play a role in translational fidelity and (b) prevents
entry of fMet–tRNA into the A-site enabling it to bind to the 50S P-site.
We also report that EF-P promotes a ribosome-dependent accommodation
of fMet–tRNA into the 70S P-site.
Abbreviations
CMCT, 1-cyclohexyl-3-(2-morpholinoethyl)carbodi-imide metho-p-toluenesulfonate; DMS, dimethyl sulfate; EF-P, eubacterial elongation
factor P; IF, initiation factor; Ke, kethoxal; RF, termination factor.
FEBS Journal 275 (2008) 671–681 ª 2008 The Authors Journal compilation ª 2008 FEBS 671
synthesis of peptide bonds [12]. Also, translocation
inhibitors have little or no affect on the EF-P reaction
[12]. EF-P enhances the inhibition of peptide-bond for-
mation by chloramphenicol and lincomycin [12]. These
results imply that EF-P acts near the peptidyltransfer-
ase center of the ribosome.
EF-P is required for synthesis with native mRNA
templates and for synthesis of poly(Phe) primed by
N-acetyl-Phe-tRNA as the initiator [1,2,12]. The EF-P
protein is essential for the synthesis of certain fMet-ini-
tiated dipeptides [8,12]. Thus, the action of EF-P may
be similar to that of initiation factor eIF-5A in eukary-

otes, which preferentially promotes synthesis of the
first peptide bond in the protein sequence [13–15]. This
is further substantiated by the fact that EF-P harbors
84 and 64% sequence similarity with aIF5-A from
archaebacteria and eIF-5A from eukaryotes, respect-
ively [16–18].
Remarkably, the crystal structure of EF-P is an L,
or tRNA-like shape, characteristic of many proteins
that actuate translation [16–18].
The function of EF-P and eIF-5A are currently
unknown. It has been proposed that these proteins
stimulate the synthesis of certain proteins in the cell
[19]. In mammalian cells, including in humans, the
genes encoding eIF-5A and its isoform, eIF-5A2, are
oncogenes [19]. Despite these observations, the specific
mode of action of these proteins in translation remains
enigmatic.
Here, we examine interactions between EF-P and
70S ribosomes using several rRNA structure-specific
chemical probes [20,21]. The nature of the interactions
between specific rRNA domains and EF-P suggests a
plausible mode of action for this ubiquitous protein.
Results
Three approaches were undertaken to study the posi-
tion of the EF-P protein on the Escherichia coli
70S ribosome and its subunits.
Binding of labeled EF-P to ribosomes
To study the binding site of EF-P on ribosomes, we
first ascertained that ribosomes were free of EF-P. This
was accomplished by washing the ribosomes in 0.5 m

NH
4
C1 buffer and isolating the 70S peak after two
successive sucrose-density gradient centrifugation steps
as described in Experimental procedures [22]. Western
blotting of the purified 70S preparations revealed that
the ribosomes were free of EF-P (data not shown). A
specific anti-EF-P mouse mAb was used to detect the
protein.
The purity of the EF-P was first assessed by isoelec-
tric focusing in the first dimension followed by SDS
electrophoresis as described previously [23] and by
subsequent immunoblotting with specific anti-(EF-P)
mAbs. A single band was observed on Coomasie Bril-
liant Blue-stained gels and on the corresponding im-
munoblots of the EF-P protein (Fig. 1A). Thus, the
EF-P protein appears to be homogeneous by these cri-
teria. This was verified by MS analysis of the tryptic
fragments of the protein (see Experimental proce-
dures). From the MS analysis of EF-P it was found
that the residue blocking Lys34 conforms to spermi-
dine within ± 1 Da (Table 1). No other known com-
pound was found to more closely fit this mass.
The N-terminus of the purified EF-P protein was
labeled by formylation with [
14
C]formate and the
protein was bound to partly dissociated 70S ribo-
somes containing 30S and 50S subunits, as described
in Experimental procedures. The ribosomeÆ[

14
C]EF-P
complex was sedimented on sucrose-density gradients
and the radioactivity of the different fractions was mea-
sured. EF-P was found associated with 70S ribosomes
and 30S and 50S subunits. Table 2 shows the amount
of EF-P that binds to each subunit and to 70S ribo-
somes. The labeled protein binds in an  1 : 1 molar
ratio to 70S ribosomes and to 30S and 50S subunits.
Identification of EF-P by immunoprecipitation
of ribosomes and their subunits
To examine whether EF-P occurs bound to native
polyribosomes, a mouse mAb specific to EF-P was
used to detect the protein. As shown in Fig. 1B, the
EF-P protein can be detected on 70S,30S and
50S particles and on polyribosomes. Fractions from
the sucrose-density gradients were collected and sub-
jected to isoelectric focusing in the first dimension
followed by 1D SDS electrophoresis and western
blotting. As shown in Fig. 1B, EF-P is bound to the
30S and 50S subunits, to 70S ribosomes and to poly-
ribosomes. The density of the immunoblots was
determined, as was the total area of each peak in the
ribosome profile. The percentage of bound EF-P
protein in each peak shows that EF-P is distributed
equally between 30S and 50S subunits and on
70S ribosomes. Approximately 17% of the protein
was found on 70S ribosomes and 7% occurred
bound to the 30S and 50S subunits, respectively.
The reminder of the protein was recovered in the

polyribosome fraction. The bound EF-P fraction
decreases as a function of the number of ribosomes
in the polyribosome fractions (Fig. 1C). This suggests
that EF-P acts in an initial stage of synthesis.
Ribosome interactions with elongation factor EF-P H. Aoki et al.
672 FEBS Journal 275 (2008) 671–681 ª 2008 The Authors Journal compilation ª 2008 FEBS
EF-P binding site on the ribosome
To further examine the interactions of EF-P, ribo-
somes were treated with various base-specific reagents
that react with unpaired bases at the ring nitrogen of
each exposed rRNA base. The sequence of the probes
used spanned the 16S and 23S rRNA regions that are
exposed to each reagent [20]. We first used dimethyl
A
BC
Fig. 1. (A) Immunoelectrophoretic analysis
of purified EF-P. Purified EF-P protein was
visualized with Coomasie Brilliant Blue and
detected on immunoblots of the electropho-
retic fractions using a specific mouse mAb.
The antibody was prepared as described in
Experimental procedures. Purified EF-P pro-
tein was subjected to isoelectric focusing
run in one dimension followed by SDS elec-
trophoresis in the second dimension using
conditions described previously [23]. (B) Im-
munoblots of polyribosomes treated with
anti-(EF-P) mAb. The optical density profiles
of fractions collected after sucrose-density
gradient centrifugation of polyribosomes are

shown. The position of the 70S,50S and
30S particles is indicated by arrows. An
immunoblot of EF-P, associated with differ-
ent fractions of the ribosome, is shown
below the ribosome profile. (C) Binding of
EF-P to different polyribosome fractions.
The density of the immunoblots was deter-
mined, as was the area of each peak. The
ratio of the density, which is related to
the amount of EF-P bound, to the area of
the peaks, which is related to the total
amount of the ribosome population in that
region of the polyribosome profile, is
plotted as a function of the percentage of
EF-P protein bound to each fraction.
Table 1. EF-P modification. MS analysis of tryptic fragments of EF-P protein was carried out as described in Experimental procedures.
Table 1 provides information regarding the quality of the search match, including the acquired MS ⁄ MS scan numbers, final candidate search
cross-correlation (X-corr) score, the normalized delta cross-correlation score (indicating the difference between the top-ranked and second
best match), the preliminary database (SIMS) search score, the matched protein identity, the matched peptide sequence with putative modi-
fication sites indicated with an @ symbol adjacent right hand to the target residue, and the predicted modification mass. The average modifi-
cation mass, and SD, is also shown. Modification mass, 143.77 ± 0.15.
Scan
X-corr
score DCn
SIMS
score Protein
Sequence
(@=modification)
Modification
mass

EFP_115min.2009.2009.2.out 3.3168 0.4493 9 EC4040 Y.AVEASEFVK.P
EFP_115min.2147.2147.2.out 2.9467 0.3855 9 EC4040 Y.AVEASEFVK.P
EFP_115min.2172.2172.1.out 2.9151 0.3799 108 EC4040 Y.AVEASEFVK.P
EFP_115min.2109.2109.2.out 2.905 0.4137 9 EC4040 Y.AVEASEFVK.P
EFP_115min.2302.2302.2.out 2.8987 0.3769 9 EC4040 Y.AVEASEFVK.P
EFP_115min.2185.2185.2.out 2.7924 0.3979 9 EC4040 Y.AVEASEFVK.P
EFP_115min.2048.2048.2.out 2.703 0.4267 9 EC4040 Y.AVEASEFVK.P
EFP_115min.2224.2224.2.out 2.7001 0.4626 9 EC4040 Y.AVEASEFVK.P
EFP_115min.2523.2523.2.out 3.0763 0.2885 236 EC4040 143.9
EFP_115min.2870.2870.2.out 2.9992 0.2478 214 EC4040 143.6
EFP_115min.2483.2483.2.out 3.6326 0.2885 435 EC4040 Y.AVEASEFVKPGK@.G 143.8
H. Aoki et al. Ribosome interactions with elongation factor EF-P
FEBS Journal 275 (2008) 671–681 ª 2008 The Authors Journal compilation ª 2008 FEBS 673
sulfate (DMS), in the presence of borohydride and
aniline, to score for A, C and G modifications in the
presence or absence of EF-P, and probed the 16S and
23S rRNA. These results guided our further chemi-
cal probing with kethoxal (Ke), 1-cyclohexyl-3-(2-mor-
pholinoethyl)carbodi-imidemetho-p-toluenesulfonate
(CMCT) and diethyl pyrocarbonate (DEP). Recogni-
tion sites were identified by primer extension analysis
with reverse transcriptase and several synthetic deoxy-
oligonucleotide primers. The position of the modified
bases was detected by stops or pauses in the progress
of reverse transcription of the modified rRNA tem-
plate [20,21,24]. Artifact bands, presumably arising
from nicks in the template rRNA or from strong sec-
ondary structure features, were distinguished from
sites of chemical attack by their occurrence in tran-
scripts using unmodified control rRNA, which had

otherwise been subjected to identical treatment (data
not shown).
Footprinting experiments were performed with the
intact 70S ribosomes in the presence or absence
of EF-P protein using probes complementary to
16S rRNA. As shown in Fig. 2, EF-P markedly pro-
tects G527, U534 and G537 against CMCT treatment
of 70S ribosomes. Also, the reactivity of U531 is
enhanced by this treatment. CMCT reacts slowly with
G residues [21]; however, DMS protection of this
region confirmed these results (data not shown). This
region occurs near the G526 streptomycin binding site
adjacent to the A-site of the 30S subunit, close to the
S12 and mRNA binding sites [25].
Footprinting experiments were also performed on
the 23S rRNA modified in the presence or absence of
the ribosome-bound EF-P protein. As shown in Fig. 3,
EF-P protects A2564, G2524, C2507, G2505, G2502
Table 2. Stoichiometric binding of EF-P to 70S ribosomes and to
30S and 50S subunits. EF-P protein was labeled as described in
Experimental procedures. Approximately 240 pmol labeled EF-P
were added to 140 pmol 70S ribosomes and incubated for 10 min
at 37 °C prior to loading the proteinÆ ribosome complex on the gradi-
ents. The samples were centrifuged in a 0–40% sucrose gradient
in 10 m
M MgCl
2
,10mM Tris, HCl, pH 7.4 and 30 mM NH
4
Cl for

21 h at 4 °C. Samples were collected and the radioactivity and
A
260
values were determined.
Particle 70S 50S 30S
Particle pmol 18.9 22 17
[
14
C] EF-P bound pmol 19.7 20.0 14.7
Ratio
a
0.959 0.909 0.864
a
[
14
C] EF-P bound ⁄ 70S,50S or 30S.
Fig. 2. Primer extension analysis of CMCT-modified 16S rRNA in
the presence (+) or absence (-) of EF-P protein was conducted as
described in Experimental procedures. The positions of chemical
attack were determined using reverse transcriptase and a 22-nucle-
otide cDNA 5¢-AGATGCAGTTCCCAGGTTG-3¢ primer complemen-
tary to bases flanking the G526 streptomycin binding site. A, T, G,
C are dideoxy sequencing lanes. Experiments were performed in
triplicate. Ke and DMS treatment of the ribosomes verified the
location of the protected bases (data not shown). The data indicate
that EF-P protects the 16S rRNA near the A-site (see text).
Ribosome interactions with elongation factor EF-P H. Aoki et al.
674 FEBS Journal 275 (2008) 671–681 ª 2008 The Authors Journal compilation ª 2008 FEBS
and G2494 in the 23S rRNA against DMS treatment
of the ribosome. Significantly, the reactivity of A2572,

G2570, G2550 and C2498 is enhanced by the interac-
tion of EF-P with 70S ribosomes. It is clear from the
data that EF-P markedly protects domain V of the
50S subunit, which is associated with the peptidyl-
transferase functions of the ribosome.
Most important to the function of the EF-P protein
is the fact that most of the molecule’s interactions
occur at the aminoacyl-tRNAs (A-site). Indeed, one
set of bases protected by interactions of EF-P with the
70S ribosome occurs near the A-loop of the peptidyl-
transferase. In addition, the protected bases, A2564
and G2524, map adjacent to the A-site on the 50S sub-
unit near the ribosomal 530 loop that is involved in
decoding by EF-TuÆGTPÆaminoacyl-tRNA complex,
the T-site [26,27]. One of the weakly protected bases,
G537, also maps to the T-site of the 30S subunit. It is
possible that the T-site represents an alternate weak
binding site for EF-P (Fig. 5).
To further verify whether EF-P affects the ribosomal
A-site, the fMet–tRNA was bound under conditions
known to bind aminoacyl-tRNAs to either the ribo-
somal A-site or the P-site. The reaction was conducted
at 30 °C in the presence of GTP under conditions that
foster spontaneous translocation [28]. As shown in
Fig. 4A,C,D, GTP markedly stimulates the EF-P-
dependent synthesis of fMet–puromycin when the
fMet–tRNA is bound to the A-site. This effect is
greatly diminished when fMet–tRNA is bound under
conditions that bind it to the presumed P-site
(Fig. 4B). The oxazolidinone III antibiotic, which

impedes binding of fMet–tRNA to the P-site and
inhibits translocation [29], interferes with the action of
EF-P by preventing the fMet–tRNA substrate from re-
binding to the P-site [29]. Oxazolidinone III has little
effect when fMet–tRNA is pre-bound to the P-site
(Fig. 4B), probably because fMet–tRNA out-competes
oxazolidinone III on the P-site. To learn whether
fMet–tRNA binding is affected by the action of GTP,
fMet–tRNA was bound in the presence of GTP and
EF-P for 20 min at 30 °C followed by incubation with
puromycin. As shown in Fig. 4C, the reaction proceeds
quantitatively. The controls show that the reaction
decreases by  70% when fMet–tRNA is bound to the
ribosome in the absence of GTP and EF-P, and these
reagents are added during the second reaction with
puromycin (Fig. 4D). Thus, EF-P may interact with
the A-site of the ribosome and may be required to
accommodate fMet–tRNA in the P-site of the
70S ribosome.
A computer-simulated structure of the 70S ribosome
of E. coli was used to fit the protected bases of the
EF-P protein relative to the A-, P-, and E-sites occu-
pied by their respective tRNAs as well as the mRNA-
binding sites (Fig. 5). As shown in Fig. 5, there appear
to be two possible binding sites for the protein. One
of these occupies the 50S and 30S A-site, whereas the
second, weaker binding site, appears to occur on the
adjacent T-site of the 70S ribosome.
Discussion
Biochemical studies have not revealed any effect of

EF-P on the initiation reaction. Thus, EF-P is not
Fig. 3. Primer extension analysis of DMS-modified 23S rRNA in
the presence (+) or absence (–) of EF-P protein. The positions of
chemical attack were determined using reverse transcriptase and a
22-nucleotide cDNA primer 5¢-TCTCCAGCGCCACGGCAGATAGG
GACC-3¢, complementary to domain V of the 23S rRNA as
described in Experimental procedures [20,21,24]. A, T, G, C are di-
deoxy sequencing lanes. Experiments were performed in triplicate.
The data indicate that EF-P protects domain V on the 23S rRNA,
which is close to the A-site of the peptidyltransferase center (see
text).
H. Aoki et al. Ribosome interactions with elongation factor EF-P
FEBS Journal 275 (2008) 671–681 ª 2008 The Authors Journal compilation ª 2008 FEBS 675
required for the binding of fMet–tRNA in the presence
or absence of the initiation factors, IF-1, IF-2 or IF-3
[1,7,12]. However, EF-P stimulates the initial rate of
translation programmed by poly(rU) when synthesis is
initiated by N -acetyl-Phe-tRNA [12] in the absence of
initiation factors. Under these conditions, EF-P does
not stimulate the synthesis of poly(Phe) from Phe-
tRNA dependent on EF-Tu and EF-G [12]. Thus, the
EF-P protein may not be required for initiation on
30S subunits or for elongation of protein chains. EF-P
is required for synthesis of several fMet-inititated di-
peptides [8,12]. Therefore, EF-P may function in the
formation of the initial peptides in the protein
sequence. Consistent with this idea is the fact that
EF-P is found bound to native polyribosomes as well
as to 70S,50S and 30S particles (Fig. 1B,C). The
EF-P protein occurs in 0.1 copies per ribosome and

may dissociate from ribosomes, perhaps after synthesis
of the first peptide bond [30].
Here, we present the results of chemical protection
footprinting analysis that provide clues concerning the
nature of the in vitro associations between rRNA
molecules and the EF-P protein. The entire length of
the exposed 16S and 23S rRNA molecules on the
70S ribosome was probed with DMS, Ke, DEP and
CMCT in the presence or absence of EF-P. A selected
number of bases is protected from chemical probes
by the complex. In several regions of the rRNA mole-
cules, the complex enhanced the reactivity of specific
bases, which is interpreted as being due to protein-
dependent conformational changes in the rRNA [20].
The evidence presented here indicates that the elon-
gation factor, EF-P, binds predominantly to the A-site
of 70S ribosomes. A possible second weak binding site
Fig. 4. Effect of binding fMet–tRNA to the
A- or to the P-site on the activity of EF-P.
f[
35
S]Met–tRNA (36 pmol) was incubated
with 70S ribosomes (20 pmol) for 20 min at
30 °C as described previously [22,28] using
4.8 m
M (A) or 7.2 mM Mg(Ac
2
) (B) for bind-
ing, presumably to either the A- or the P-site
of the ribosome, respectively. EF-P

(20 pmol) and puromycin (1 l
M) were added
and the reaction was continued at 30 °C for
5 min in the presence or absence of 0.1 m
M
GTP and, where indicated, 50 lM oxazolidi-
none III (oxo). (C) EF-P and GTP were added
during formation of the fMet–tRNAÆribo-
some complex (first reaction) and the reac-
tion was carried out for 20 min at 30 °C,
prior to the addition of puromycin (second
reaction). (D) The initiation complex was
formed without GTP or EF-P for 20 min at
30 °C, and GTP and EF-P were added with
puromycin and were incubated for 5 min
at 30 °C.
Ribosome interactions with elongation factor EF-P H. Aoki et al.
676 FEBS Journal 275 (2008) 671–681 ª 2008 The Authors Journal compilation ª 2008 FEBS
for the EF-P protein occurs proximal to the A-site on
the ribosomal T-site that is the recognition center for
the EF-TuÆGTPÆaminoacyl-tRNA on the 30S and
50S subunit. The footprints on the 16S rRNA reside
close to the mRNA binding site on the neck of the
30S subunit, adjacent to the G526 streptomycin and
the ribosomal protein S12 binding sites (Figs 2 and 5).
Interestingly, the structural paralog of EF-P, eIF5-A,
which shares strong sequence homology with
domains I and II of EF-P, binds to certain mRNAs in
a hypusine-dependent manner [31]. Our MS analysis of
EF-P indicates that Lys34 of the protein is modified

by spermidine, which is a precursor of hypusine in the
eIF-5A protein [32]. The model in Fig. 5, in which the
orientation of the spermidine residue of EF-P is facing
the mRNA-binding pocket of the 30S subunit is con-
sistent with this result.
The EF-P footprints extend into the 50S subunit
and reside close to the L7 ⁄ L12 stalk. The 50 S T-site
harbors the ribosomal L12 protein (GAR), which
occurs in an open state when the A-site is empty or
filled with tRNA [33], but closes upon contact with the
EF-Tu ternary complex [33]. It is possible that EF-P
binds to the T-site in order to maintain a closed A-site.
This may insure that the fMet–tRNA is accommo-
dated on the P-site of the 70S ribosome prior to the
proof-reading that occurs with the EF-Tu ternary com-
plex and the ribosome.
The EF-P protein binds stoichiometrically to the
ribosome and stimulates accommodation of fMet–
tRNA, presumably to the 70S P-site. Thus, it is likely
that the protein binds to more than one site on the
ribosome depending on the course of synthesis. Most
bases protected upon binding of EF-P to the 70S ribo-
some appear to reside on the A-site of the 30S and
50S subunits (Figs 3 and 5).
It is noteworthy that EF-P protects bases that
occur at the A-site of the 50S subunit on 70S ribo-
somes (Figs 3 and 5). This is consistent with the EF-
P protein being a competitive inhibitor of puromycin
[34], an antibiotic known to act at the A-site of the
ribosome [35]. Reconstitution experiments, with core

particles of the 50S subunit lacking specific 50S pro-
teins, revealed that L16 is essential for the action of
the EF-P protein [36]. The crystal structure of the
ribosome, with bound aminoacyl-tRNA, indicates
that L16 occurs at the A-site of the 50S subunit [37].
Furthermore, the L16 protein binds aminoacyl-tRNA
[37] (and unpublished observations). The tRNA-like
shape of the EF-P protein is in keeping with its
interactions near the A-site of the 70S ribosome and
its subunits. The binding of EF-P at the A-site may
have important consequences for proper decoding of
the mRNA transcript. For example, by binding to
the A-site, EF-P may prevent the incorrect position-
ing of the fMet-RNA into the A-site of the 50S sub-
unit.
The binding site of EF-P to domain V is of special
importance because EF-P is predicted to enhance elon-
gation by influencing in some manner the peptidyl-
transferase center [12]. Several bases indicated in the
protection analysis (Fig. 3) demonstrate that EF-P
interacts close to the A-site of domain V of the
23S rRNA. The X-ray diffraction patterns of the
70S ribosomes of Thermus thermophilus, at 5.5 A
˚
reso-
lution, reveal that the binding site of the three tRNAs
is adjacent to the binding site of the elongation factors
[37]. Thus, the incoming aminoacyl-tRNAÆGTPÆEF-Tu
complex must be adjusted to the A-site in a position
where a peptide bond can be formed. By being at the

A-site, EF-P may prevent the spurious entrance of
aminoacyl-tRNAs, or of deacyl-tRNAs, which do not
occur in the ternary complex with EF-Tu, from prema-
turely entering the A-site prior to the proofreading
steps which precede accommodation of the incoming
aminoacyl-tRNA into the peptidyltransferase active
site.
Fig. 5. Computer-simulated 3D structure of the 70S ribosome of
E. coli showing the approximate location of the EF-P-binding site on
the 70S ribosome. Bases protected against chemical modification
by interactions with EF-P are shown in red; bases enhanced by the
interactions with EF-P are shown in yellow. The positions of the
exit (E) and peptidyltRNA (P) site on the 70S ribosome are shown.
EF-P (light green) clearly binds near the A-site of the 30S and 50S
subunits, as well as to the T-site of the ribosome. The probable
spermidine-binding site (on Lys34 of EF-P protein) is shown in
orange and points towards the 30S subunit (see text). The coordi-
nates for the E. coli ribosome were from Schuwirth et al. [43]. The
protections observed were also found to neighbor the A-site of the
70S ribosomes of T. thermophilus determined by X-ray diffraction
at the 5.5 A
˚
resolution [37].
H. Aoki et al. Ribosome interactions with elongation factor EF-P
FEBS Journal 275 (2008) 671–681 ª 2008 The Authors Journal compilation ª 2008 FEBS 677
We suggest that EF-P interacts with the A-site of
the peptide-bond-forming center of the large ribosomal
subunit and with the decoding center of the small sub-
unit. Thus, the binding site may span both subunits.
The oligonucleotide domain of the EF-P protein,

which harbors spermidine, may bind near the mRNA-
binding site (Fig. 5), while the hydrophobic residues
around the central loop of the molecule might be adja-
cent to the peptidyltransferase center. Most important
to the function of the EF-P protein is the fact that the
interactions of the molecule are in close proximity to
the peptidyltransferase center of domain V near the
positions of the A-site 3¢-CCA-termini of the bound
tRNAs (Figs 3 and 5).
The footprints of the EF-P protein on the 16S
rRNA are adjacent to the G526 streptomycin-binding
site [22,35] and the ribosomal protein S12 (Figs 2 and
5). Streptomycin is a potent inhibitor of EF-P-medi-
ated synthesis of peptide bonds [12]. Streptomycin
inhibits translation by increasing the error rate of syn-
thesis and interfering with the proofreading mecha-
nisms of the ribosome [25,35].
When a cognate tRNA binds to the A-site, the
30S subunit undergoes a conformational change from
an open to a closed form [25]. Some mutations that
affect accuracy either prevent or induce this conforma-
tional transition [25]. Streptomycin, for example, stabi-
lizes the closed form and induces errors in translation.
By contrast, certain mutations in S12 to streptomycin
resistance or dependence destabilize the closed form
[25].
Thus, by binding near the A-site, EF-P may prevent
the aminoacyl-tRNA in the EF-TuÆGTPÆaminoacyl-
tRNA complex from prematurely entering the A-site
prior to the proofreading functions carried out by the

ribosome [25–27]. This interaction would also help
poise the fMet–tRNA on the P-site of the 50S subunit
and may also result in increased accuracy of amino-
acyl-tRNA selection. Hydrolysis of GTP results in
ejection of EF-TuÆGDP from the ribosome and accom-
modation of the aminoacyl-tRNA into the A-site
[26,27]. A-site-bound aminoacyl-tRNA is predicted to
displace the EF-P protein from the ribosome. Thus,
one EF-P molecule would be used for each successfully
initiated round of translation.
The interactions of EF-P with the ribosome are
likely to result in rearrangements of the ribosomal
subunits that are essential for the transition between
the initiation and elongation stages of protein synthe-
sis. The conservation of this protein throughout spe-
cies and its obligatory requirement during synthesis
befit the essential nature of these functions in transla-
tion.
Experimental procedures
Materials
The materials used were as described in Xu et al. [5] and
[29].
RNA modification and primer extension analysis
The 70S ribosomes (8 pmol) were modified using DMS,
DEP, Ke or CMCT in the presence or absence of 20 pmol
EF-P protein. Adenines and cytosines were determined as
described previously [20,21]. The N-7 position of guanine
was detected with DMS using sodium borohydride and ani-
line to induce strand scission [21]. The DNA sequence was
performed by a modification of the double-strand dideoxy

chain termination method of Sanger et al. with modified
T7 DNA polymerase and [
35
S]dATP[aS] [20,21,24].
Effect of EF-P binding on the chemical protection
footprints of 70S ribosomes
To locate the bases in 16S and 23S RNA on 70S ribosomes
that are in contact with EF-P, EF-P was first bound stoi-
chiometrically to the ribosome. Each of the synthetic oligo-
nucleotide primers was annealed to the rRNA in the
complex and was extended with reverse transcriptase in the
presence of [
35
S]dATP[aS] and the other three (unlabeled)
deoxynucleotide triphosphates. The labeled DNA tran-
scripts were resolved on a DNA sequencing gel along with
four dideoxy sequencing lanes on the same gel to help iden-
tify the modified bases. The entire length of the 16S or
23S rRNA on the 70S ribosomes was probed with DMS in
the presence or in the absence of the EF-P protein. Primers
complementary to sequences at  200-bp intervals were
then annealed to the ribosome [20]. Based on the results
with DMS, several different primers were utilized to further
study the specificity of the regions protected against chemi-
cal attack by Ke, CMCT and DEP. Primers to the 30S sub-
unit spanned the streptomycin binding sites at G526 and at
A915 [25]. Two additional primers were used to probe the
decoding region of the 16S rRNA. Two primers were used
to probe domain II of 23S rRNA; three primers were used
to probe domains IV and V and one primer was used to

probe domain VI of the 23S rRNA molecule.
Preparation of 70S ribosomes, 30S and 50S
subunits
Ribosomes and subunits were isolated from E. coli MRE-
600 mid-log cells at 4 °C as described previously [22].
Briefly, cells (100 g) were broken by grinding with 100 g
alumina (Alcoa Inc., Pittsburgh, PA, USA) and suspended
in buffer A (10 mm Tris ⁄ HCl, pH 7.4, 30 mm NH
4
Cl,
1mm dithiothreitol, 6 mm Mg(Ac)
2
), DNase (RNase-free;
Ribosome interactions with elongation factor EF-P H. Aoki et al.
678 FEBS Journal 275 (2008) 671–681 ª 2008 The Authors Journal compilation ª 2008 FEBS
1.0 lgÆmL
)1
) was added and the mixture incubated for
5 min at 4 °C. Unbroken cells and debris were removed by
two successive centrifugations for 20 min at 30 000 g. The
resulting supernatant was centrifuged at 78 000 g for 18 h
and the ribosomal pellets were suspended in buffer A con-
taining 0.5 m NH
4
Cl and the crude ribosomal suspension
was gently stirred for 2 h. The crude ribosomal suspension
was layered on sucrose density gradients (0–40% sucrose in
buffer A) and centrifuged at 78 000 g for 18 h. Fractions
containing the 70S ribosomes were combined and were cen-
trifuged for 24 h at 60 000 g. The ribosomal pellets were

suspended in 2–3 mL buffer A containing 0.5 m NH
4
Cl, the
suspension was stirred for 2 h at 4 °C and was subjected to
a second 0–40% sucrose density gradient centrifugation
step as above at 64 000 g. Absorbance at 260 nm was used
to identify the ribosomal fractions and the 70S peak was
pooled and centrifuged at 37 000 g for 18 h. Ribosomes
were collected and suspended in buffer A at a concentration
of 20 mgÆmL
)1
and stored at )80 °C.
Isolation of EF-P
The EF-P protein was assayed based on its ability to stimu-
late ribosomes to synthesize fMet–puromycin in the pres-
ence of 1.0 lm puromycin and the absence of organic
solvents [1] using f[
35
S]Met–tRNA bound to ribosomes as
described previously [22,38].
The EF-P protein was purified as described previously
[12]. Briefly, ribosomes were isolated as described above
and washed once with 0.5 m NH
4
Cl in buffer A containing
6mm Mg(Ac)
2
. The ribosomal wash, containing the EF-P
protein, was sequentially purified first through a column of
QAE–Sepharose, followed by hydroxylapatite, Sephacryl

S-300 and Mono-Q columns using FLPC. Fractions were
assayed after each step to locate the protein. The purified
protein was concentrated and was stored at )80 °C. Protein
concentration was determined using the method described
by Bradford [39].
Preparation of mouse mAbs to EF-P
mAbs to the purified EF-P protein were raised using nude
mice. This study received ethical approval by the University
of Toronto Animal Care Committee, which is in full com-
pliance with the Guidelines of the Canadian Council on
Animal Care and the Regulations of the Animals for
Research Act. Aliquots of the EF-P (1 mgÆmL
)1
) in buf-
fer A were injected into the mice intravenously. The immu-
nization schedule was 5 lg of the protein on day 1, 10 lg
on day 8, 20 lg on day 12 and 25 lg on day 16. After a
two-week rest, mice received a final 25 lg injection of the
EF-P protein. Three days after the final injection, blood
was obtained through the orbital vein. Three days after
this, the cells were harvested and fused with HAT-sensitive
mouse myeloma cells. The samples were tested first for anti-
body production using dot blots. Positive clones were ana-
lyzed on western blots.
Radioactive chemical labeling of the EF-P protein
The EF-P protein (700 pmol) was dissolved at 1 °Cin
0.1 m Na borate buffer (pH 9.0), 300 mm KC1 and 5 mm
b-mercaptoethanol. The [
14
C]formaldehyde (50 mCiÆmm

)1
)
was added in the above buffer to the EF-P protein, mixed
for 2 min. prior to the addition of 8 lg Na borohydride to
stop the reaction. The labeled EF-P was dialyzed against
10 mm HEPES buffer, pH 7.4, 10 mm b-mercaptoethanol
to eliminate the unreacted [
14
C] formaldehyde.
Tandem MS
Proteomic analysis of the samples was performed by
microcapillary electrospray LC-MS ⁄ MS. Briefly, an aliquot
of the EF-P protein was suspended in 100 mm
NH
4
HCO
3
⁄ 1mm CaCl
2
buffer, pH 8.5 and digested by
trypsin overnight at 37 °C using immobilized trypsin Porous
beads (PerSeptive Biosystems, Framingham, MA, USA).
The digested peptides were then fractionated on a 7.5 cm
(100 lm ID) reverse-phase C
18
capillary column attached
inline to a Thermo Finnigan LCQ-Deca quadrupole ion
trap mass spectrometer. The entire digested sample was
loaded as described previously [40] and the peptides eluted
by ramping a linear gradient from 2 to 60% solvent B over

90 min. Solvent A consisted of 5% acetonitrile, 0.5% acetic
acid and 0.02% heptofluorbutyric acid and solvent B con-
sisted of 80 : 20 acetonitrile ⁄ water (v ⁄ v) containing 0.5%
acetic acid and 0.02% heptofluorbutyric acid. The flow rate
at the tip of the needle was set to 300 nLÆmin
)1
by program-
ming the HPLC pump and use of a split line. The mass
spectrometer cycled through four scans as the gradient pro-
gressed. The first event was a full mass scan followed by
three tandem mass scans of the successive three most intense
precursor ions. Precursor ions (400–2000 m ⁄ z) were sub-
jected to data-dependent, collision-induced dissociation with
dynamic exclusion enabled. The spectra were searched
against UniProt protein sequences downloaded from the
European Bioinformatics Institute using the sequest com-
puter algorithm [41]. Precursor mass tolerance was set to
3 Da (with daughter mass ion tolerance set to the default of
0), enabling fully tryptic enzyme status, with single site
missed cleavages tolerated. The statistical probability of
each primary match was assessed using the statquest algo-
rithm [42], with candidate peptides filtered using a strict
95% confidence cut-off to minimize false positives.
Acknowledgments
We thank NSERC of Canada for financial support.
We thank Jennifer Yang and Ivona Kozieradsky for
H. Aoki et al. Ribosome interactions with elongation factor EF-P
FEBS Journal 275 (2008) 671–681 ª 2008 The Authors Journal compilation ª 2008 FEBS 679
excellent technical help during the initial stages of this
work. We are most grateful to Drs A. J. Becker and

K. Nierhaus for perceptive discussion of this study.
References
1 Glick BR & Ganoza MC (1975) Identification of a solu-
ble protein that stimulates peptide bond synthesis. Proc
Natl Acad Sci USA 72, 4257–4260.
2 Green RH, Glick BR & Ganoza MC (1985) Require-
ments for in vitro reconstruction of protein synthesis.
Biochem Biophys Res Commun 126, 792–798.
3 Ganoza MC, Cunningham C & Green RM (1985) Iso-
lation and point of action of a factor from Escherichia
coli required to reconstruct translation. Proc Natl Acad
Sci USA 82, 1648–1652.
4 Ganoza MD, Cunningham C & Green RM (1995) A
new factor from Escherichia coli affects translocation of
mRNA. J Biol Chem 270, 26377–26381.
5 Xu J, Kiel MC, Golshani A, Chosay JG, Aoki H &
Ganoza MC (2006) Molecular characterization of a
ribosome-dependent ATPase of Escherichia coli ribo-
somes. Nucleic Acids Res 34, 1158–1165.
6 Ganoza MC, Kiel MC & Aoki H (2002) Evolutionary
conservation of reactions in translation. Micobiol Mol
Biol Rev 66, 460–485.
7 Aoki H, Adams SL, Turner MA & Ganoza MC (1979)
Molecular characterization of the prokaryotic efp gene
product involved in a peptidyltransferase reaction.
Biochimie 71, 7–11.
8 Glick BR, Chladek S & Ganoza MC (1979) Peptide
bond formation stimulated by protein synthesis factor
EF-P depends on the aminoacyl moiety of the acceptor.
Eur J Biochem 97, 23–28.

9 Aoki H, Adams SL, Chung DG, Yaguchi M, Chuang
SE & Ganoza MC (1991) Cloning, sequencing and
overexpression of the gene for prokaryotic factor EF-P
involved in peptide bond synthesis. Nucleic Acids Res
19, 6215–6220.
10 Aoki H, Dekany K, Adams SL & Ganoza MC (1997)
The gene encoding the elongation factor P protein is
essential for viability and is required for protein synthe-
sis. J Biol Chem 272, 32254–32259.
11 Gerdes SY (2003) Experimental determination and sys-
tem level analysis of essential genes in Escherichia coli
MG 1655. J Bacteriol 185, 5673–5684.
12 Ganoza MC & Aoki H (2000) Peptide bond formation:
function of the efp gene product. Biol Chem 381, 553–
559.
13 Bartig D, Lemkemeier K, Frank J, Lottspeich F &
Klink F (1992) The archaebacterial hypusine-containing
protein. Structural features suggest a common ancestry
with eukaryotic translation initiation factor (eIF-5A).
Eur J Biochem 204, 751–758.
14 Smit-McBride Z, Deve TE, Merrick WC & Hershey
JWB (1989) Protein synthesis initiation factor (eIF-5A).
J Biol Chem 264, 1578–1583.
15 Zanelli CF & Valentini SR (2007) Is there a role for
eIF-5A in translation? Amino Acids 33, 351–358.
16 Kim KK, Yokota H, Kim R & Kim S-H (1997)
Cloning, expression and crystallization of a hyper-
thermophilic protein that is homologous to the
eukaryotic initiation factor, eIF-5A. Protein Sci 6,
2268–2270.

17 Hanawa-Suetsugu K, Sekine S, Sakai H, Takaho T,
Unzai S, Tame JRH, Shirouzu M & Yokoyama S (2004)
Crystal structure of the elongation factor P from Thermus
thermophilus. Proc Natl Acad Sci USA 101, 9595–9600.
18 Benson T, McCroskey M, Cialdella JI, Choi G & Pear-
son DP (2000) Structure of S. aureus elongation factor
P: another tRNA mimic in protein translation. Struc-
tural aspects of protein synthesis. Second Symposium,
Rensserlaeville, NY. Abstracts, p. 15.
19 Clement PMJ, Henderson CA, Jenkins ZA, Smit-
McBride Z, Wolff EC, Hershey JBW, Park MH & Joh-
nanson HE (2003) Identification and characterization of
eukaryotic initiation factor 5A-2. Eur J Biochem 270,
4254–4263.
20 Moazed D, Stern S & Noller HF (1986) Rapid chemical
probing of conformation in 16S ribosomal RNA and
30S ribosomal subunits using primer intention. J Mol
Biol 187, 399–417.
21 Christiansen J, Egebjerg J, Larsen N & Garrett R
(1990) Analysis of rRNA structure: experimental and
theoretical considerations. In Ribosomes and Protein
Synthesis. A Practical Approach (Spedding G, ed.),
pp. 229–252. IRL Press, Oxford.
22 Rheinberger HJ, Geigenmuller U, Wedde M & Nier-
haus KH (1988) Parameters for the preparation of Esc-
herichia coli ribosomes and ribosomal subunits active in
tRNA binding. Methods Enzymol 164, 658–670.
23 Kaltschmidt E & Wittman HG (1970) Two-dimensional
separation of ribosomal proteins. Proc Natl Acad Sci
USA 67, 1276–1282.

24 Sanger F, Nicklen S & Coulson AR (1977) DNA
sequencing with chain terminating inhibitors. Proc Natl
Acad Sci USA 74, 5463–5467.
25 Ogle JM, Murphy FV, Tarry MJ & Ramakrishnan V
(2002) Selection of tRNA by the ribosome requires a
transition from an open to a closed form. Cell 111,
721–732.
26 Powers T & Noller HF (1994) The 530 loop of 16S
rRNA: a signal to EF-Tu. Trends Genet 10, 27–31.
27 Blanchard SC, Gonzales RL Jr, Kim HD, Chu S &
Puglisi JD (2004) tRNA selection and kinetic proofread-
ing in translation. Nat Struct Mol Biol 11, 1008–1014.
28 Spirin A (1978) Energetics of the ribosome. Prog
Nucleic Acid Res Mol Biol 21, 39–62.
Ribosome interactions with elongation factor EF-P H. Aoki et al.
680 FEBS Journal 275 (2008) 671–681 ª 2008 The Authors Journal compilation ª 2008 FEBS
29 Aoki H, Ke L, Poppe SM, Poel TJ, Weaver EA, Gad-
wood RC, Thomas RC, Shinabarger DL & Ganoza
MC (2002) Oxazolidinoe antibiotics target the P site on
Escherichia coli ribosomes. Antimicrob Agents Chemo-
ther 46, 1080–1085.
30 An G, Click BR, Friesen JD & Ganoza MC (1980)
Identification and quantification of elongation factor
EF-P in Escherichia coli cell-free extracts. Can J
Biochem 58, 1312–1314.
31 Xu A, Jao DL & Chen KY (2004) Identification of
mRNA binds eukaryotic initiation factor 5A by affinity
purification and tandem display. Biochem J 384, 585–590.
32 Chen KY & Liu AY (1997) Biochemistry and function
of hypusine formation on eukaryotic initiation factor

5A. Biol Signals 6, 105–109.
33 Valle M, Zavialov A, Sengupta J, Rawat U, Ehren-
berg M & Frank J (2003) Locking and unlocking of
ribosomal motions. Cell 300, 123–134.
34 Swaney S, McCroskey M, Shinabarger D, Wang Z,
Turner BA & Parker CN (2006) Characterization of a
high-throughput screening assay for inhibitors of elon-
gation factor P and ribosomal peptidyl transferase
activity. J Biomol Screen 11, 736–742.
35 Spahn C & Prescott C (1997) Throwing a spanner in
the works: antibiotics and the translation apparatus.
J Mol Med 74, 423–439.
36 Baxter RM, Ganoza MC, Zahib N & Chung DG
(1987) Reconstruction of the peptidyltransferase activity
on 50S and 70S ribosomal particles by peptide frag-
ments of protein L16. Eur J Biochem 163, 473–479.
37 Yusupov MM, Yusapova GZ, Baucom A, Lieberman
K, Earnest TN, Cate JH & Noller HF (2001) Crystal
structure of the ribosome at the 5.5 A
˚
resolution.
Science 292, 883–896.
38 Chung D-G, Zahib ND, Baxter RM & Ganoza MC
(1990) Peptidyl transferase: the soluble protein EE-P
restores the efficiency of 70S ribosome-catalysed peptide
bond synthesis. In Ribosomes and Protein Synthesis: A
Practical Approach (Spedding G ed.), pp. 69–80. IRL
Press, Oxford.
39 Bradford MM (1976) Rapid and sensitive method for
quantification of microgram quantities of protein utiliz-

ing the principle of protein-dye binding. Anal Biochem
72, 248–254.
40 Gatlin CL, Kleemann GR, Hays LG, Link AJ & Yates
JR III (1998) Protein identification at the low femto-
mole level from silver-stained gels using a new fritless
electrospray interface for liquid chromatography-micro-
spray and nanospray mass spectrometry. Anal Biochem
263, 93–101.
41 Eng JK, McCormack AL & Yates JRI (1994) An
approach to correlate tandem mass-spectral data of
peptides with amino-acid-sequences in a protein data-
base. J Am Soc Mass Spectrom 11, 976–989.
42 Kislinger T, Rahman K, Radulovic D, Cox B, Rossant
J & Emili A (2003) PRISM, a generic large scale pro-
teomic investigation strategy for mammals. Mol Cell
Proteomics 2, 96–106.
43 Schuwirth BS, Borovinskaya MA, Hau CW, Zhang W,
Vila-Sanjurjo A, Holton JM & Cate JH (2005) Struc-
tures of the bacterial ribosome at 3.5 A
˚
resolution.
Science 310, 827–834.
H. Aoki et al. Ribosome interactions with elongation factor EF-P
FEBS Journal 275 (2008) 671–681 ª 2008 The Authors Journal compilation ª 2008 FEBS 681