Functional effects of deleting the coiled-coil motif
in
Escherichia coli
elongation factor Ts
Henrik Karring
1
, Asgeir Bjo¨ rnsson
2
, Søren Thirup
1
, Brian F. C. Clark
1
and Charlotte R. Knudsen
1
1
Department of Molecular Biology, Aarhus University, Denmark;
2
deCODE Genetics, Inc., Reykjavik, Iceland
Elongation factor Ts (EF-Ts) is the guanine nucleotide-
exchange factor for elongation factor Tu (EF-Tu) that is
responsible for promoting the binding of aminoacyl-
tRNA to the mRNA-programmed ribosome. The struc-
ture of the Escherichia coli EF-Tu–EF-Ts complex reveals
a protruding antiparallel coiled-coil motif in EF-Ts, which
is responsible for the dimerization of EF-Ts in the crystal.
In this study, the sequence encoding the coiled-coil motif
in EF-Ts was deleted from the genome in Escherichia coli
by gene replacement. The growth rate of the resulting
mutant strain was 70–95% of that of the wild-type strain,
depending on the growth conditions used. The mutant
strain sensed amino acid starvation and synthesized the

the aa-tRNA to the A-site of the mRNA-programmed
ribosome. Upon codon recognition by a cognate ternary
complex (EF-Tu–GTP–aa-tRNA), the ribosomal GTPase
centre stimulates the GTPase activity of EF-Tu and the
bound GTP is hydrolysed. The inactive EF-Tu–GDP is
released from the ribosome and recycled to the active EF-
Tu–GTP by the exchange of GDP with GTP [2]. Stimula-
tion of the guanine nucleotide release in EF-Tu by EF-Ts [3]
is required as the dissociation of GDP is otherwise very slow
(2 · 10
)3
s
)1
)[1,4].Invivo, the binding of GTP to the binary
EF-Tu–EF-Ts complex is favoured owing to the ninefold
higher concentration of GTP (0.9 m
M
) than GDP (0.1 m
M
)
[5]. The activation of EF-Tu is completed by the dissociation
of EF-Ts from EF-Tu–GTP. The equilibrium governing
EF-Tu is further driven to the GTP-bound state by the
formation of EF-Tu–GTP–aa-tRNA. Previous studies have
indicated the existence of a structural isomerization in the
EF-Tu–GDP–EF-Ts complex from a high- to a low-affinity
nucleotide binding conformation [1,6,7]. According to the
results published by Gromadski et al. [1], the structures
of the binary EF-Tu–EF-Ts complex and the nucleotide-
bound ternary complexes are different.

LB, Luria–Bertani; MCS, multiple cloning site; ppGpp, guanosine
5¢-diphosphate 3¢-diphosphate; pppGpp, guanosine 5¢-triphosphate
3¢-diphosphate; (p)ppGpp, ppGpp and pppGpp.
(Received 10 June 2003, revised 26 August 2003,
accepted 8 September 2003)
Eur. J. Biochem. 270, 4294–4305 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03822.x
complex used in the classical model of the elongation cycle
has remained 1 : 1 because only the heterodimer has been
detected in solution [10–12]. Thus, it is believed that
dimerization of EF-Ts, as observed in the crystal structure,
is not physiologically relevant.
The cellular response to amino acid starvation causes an
accumulation of the unusual guanine nucleotides guanosine
5¢-diphosphate 3¢-diphosphate (ppGpp) and guanosine
5¢-triphosphate 3¢-diphosphate (pppGpp) (stringent
response) [13,14]. These compounds are synthesized by the
relA gene product when activated by codon-specific A-site
binding of uncharged tRNA, reflecting the limitation of the
corresponding amino acid [15,16]. The guanosine 5¢-diphos-
phate 3¢-diphosphate and guanosine 5¢-triphosphate
3¢-diphosphate [(p)ppGpp] nucleotides mediate a reduction
in the elongation rate and influence the accuracy of
translation. Early results indicated that (p)ppGpp affect
translation by inhibition of elongation factors Tu, Ts and G
[17–19]. However, more recent studies suggest that
(p)ppGpp have no direct effects on either translation
elongation or error rates, but rather effect translation
indirectly by inhibition of mRNA synthesis during amino
acid starvation [20–23].
Kirromycin is one in a family of antibiotics that inhibits

strain (named GRd.tsf) was characterized. In addition,
the activity of the coiled-coil deleted EF-Ts mutant in
promoting guanine nucleotide exchange in EF-Tu, and the
stability of the mutant EF-Tu–EF-Ts complex in the
presence of guanine nucleotides and kirromycin, were
examined in vitro.
The present study shows that the coiled-coil motif in
EF-Ts is not required for the guanine nucleotide-exchange
Fig. 1. Structure of the Escherichia coli elongation factor Tu (EF-Tu)–elongation factor Ts (EF-Ts) heterodimer. EF-Ts is drawn as ribbons,
indicating the secondary structure, whereas EF-Tu is drawn as a C-alpha trace. (A) The position of EF-Tu domains with respect to the N-terminal
and core domains of EF-Ts. The structure shows that EF-Ts interacts with domain 1 (G-domain) and domain 3 of EF-Tu. (B) The view in (A) is
rotated 90° around a vertical axis to show the positions of the protruding coiled-coil motif (residues 187–226) and the C-terminal module. The
terminal positions of the deletion of the coiled-coil motif and the insertion of the linker peptide, EPGGEA, are indicated by residues Asp184 and
Glu225, which are shown in Ôball & stickÕ representation. Coordinates were obtained from PDB entry [1EFU] and displayed using
MOLMOL
[64].
Ó FEBS 2003 Deletion of the coiled-coil motif of EF-Ts (Eur. J. Biochem. 270) 4295
reaction in EF-Tu. Instead, the motif plays a significant role
in the ability of EF-Ts to compete with guanine nucleotides
for the binding to EF-Tu. We propose that the coiled-coil
motif in bacterial EF-Ts is involved in a structural
isomerization step of the nucleotide-bound EF-Tu–EF-Ts
complex during the guanine nucleotide-exchange reaction.
Materials and methods
Construction of gene replacement vector
pMAK705-d.tsf
A tsf mutant, lacking the sequence encoding the coiled-coil
motif, was constructed by site-directed mutagenesis and
inserted into the gene replacement vector, pMAK705 [30],
for deletion of the motif in endogenous EF-Ts of E. coli. The

UPS (5¢-ATGCG
GGATCCAAGCTTGAGCTTACATC
AGTAAGTGACCGGGATGA-3¢) and reverse primer
Ts185 (5¢-TAGCAC
CCCGGGTTCGTCTTCCGGTTTG
ATGAATTCTGGCTTG-3¢). Primer tsf-UPS contains a
BamHI restriction site (underlined). Primer Ts185 contains
a unique AvaI restriction site (underlined) and part of a
sequence encoding the hexapeptide EPGGEA (bold text).
The downstream fragment was prepared using forward
primer Ts224 (5¢-TAGTTC
CCCGGGGGTGAAGCTGA
AGTTTCTCTGACCGGTCAGCCGTTC-3¢) and reverse
primer tsf-DOWNS (5¢-AGTCA
GGATCCGTCGACA
GAGCTTCGCCACTCAACTTAAGCAGAA-3¢). Like
primer Ts185, Ts224 contains a unique AvaI restriction site
(underlined) and part of a sequence encoding the hexapep-
tide EPGGEA (bold text). Primer tsf-DOWNS contains a
BamHI restriction site (underlined). Both PCR products
were digested with AvaI and ligated. The ligation was used
directly as a template in a final PCR using primers tsf-UPS
and tsf-DOWNS. Taq polymerase was used in all PCR
amplifications. The amplified fragment containing the
mutant tsf gene (d.tsf) was digested with BamHI and
inserted into pMAK705. The sequence of the insert in the
resulting pMAK705-d.tsf plasmid was verified by DNA
sequencing.
Deletion of the coiled-coil motif in endogenous
E. coli

L
-amino acids and other nutrients, except
Fig. 2. Schematic illustration of gene organizations in Escherichia coli
strains and the gene replacement plasmid pMAK705-d.tsf. (A) Elonga-
tion factor Ts (EF-Ts) wild-type strain UY211. (B) Gene replacement
plasmid pMAK705-d.tsf, used for the deletion of the coiled-coil motif
(cc) of endogenous EF-Ts. The plasmid contains an 1148-bp insert in
which the region in E. coli tsf (bp 556–675) encoding the coiled-coil
motif has been exchanged with a sequence encoding the hexapeptide
EPGGEA found at the corresponding position in mammalian EF-
Ts
mt
. (C) The coiled-coil deleted EF-Ts mutant strain GRd.tsf. Genes:
rpsB, gene encoding ribosomal protein S2; tsf, gene encoding wild-type
EF-Ts; cc, coiled-coil encoding region; smbA, gene encoding SmbA;
d.tsf, gene encoding coiled-coil deleted EF-Ts mutant. Restriction site
B: BamHI.
4296 H. Karring et al. (Eur. J. Biochem. 270) Ó FEBS 2003
for proline and thiamine [34]. LB medium was modified
by replacing bacto-tryptone with various amounts of
Casamino acid (1–3%). Overnight cultures were inoculated
into 40 mL of media to a final attenuance D
600
of  0.1 or
D
420
of  0.07. The D
600
culture was used to analyse growth
in media based on LB, while the D

Amino acid starvation and measurement of (p)ppGpp
levels
Cells from overnight cultures were washed in glucose/Mops
and inoculated at a D
420
of  0.14 in 25 mL of glucose/Mops
that contained  15 lCiÆmL
)1
of [
32
P]phosphate and 40 l
M
of proline, which represents 0.1 · the recommended amino
acid concentration [34]. Cells were grown at 37 °Candthe
cell density was followed by measuring the D
420
. Cell samples
of 50 lL were withdrawn at appropriate time-points and
mixedwith50lLof2.0
M
formic acid. The labelled cell
samples were run on TLC sheets [poly(ethylenimine)] and
developed as described by Cashel et al. [36] with the only
exception that 10 lL of the cell-free supernatant was spotted
on the thin layer sheets. (p)ppGpp levels were determined by
phosphor-imaging. Every (p)ppGpp measurement was nor-
malized to the cell density by division with the D
420
of the
culture at the corresponding time-point. To correct for the

digestion of the PCR product with SapI, the resulting two
fragments were inserted into pAB146 in a three-fragment
ligation. The cloning was verified by DNA sequencing. The
resulting plasmid pAB146-tsf was used as a template in
PCR amplifications for the construction of the coiled-coil
deleted E. coli EF-Ts gene (d.tsf). The mutagenesis was
basically performed as described for the construction of
pMAK705-d.tsf, except that primers tsf-UPS and tsf-
DOWNS were replaced with the plasmid-specific forward
primer pT711 (5¢-TAATACGACTCACTATAGGGGA
ATTG-3¢) and reverse primer Int R (5¢-CCCATGACCT
TATTACCAACCTC-3¢), respectively. The final amplified
fragment was digested with XbaI and KpnI and inserted into
pAB146. The unique XbaI and KpnI restriction sites are
positioned immediately upstream and downstream of the
MCS of pAB146, respectively. The mutagenesis resulting in
the construct pAB146-d.tsf was verified by DNA sequen-
cing. Cloned Pfu polymerase (New England Biolabs) was
used in all PCR amplifications.
Expression and purification of
E. coli
EF-Tu and EF-Ts
Expression and purification of E. coli EF-Tu using plasmid
pGEXFXtufA was performed essentially as previously
described [39]. Strain B834(DE3) (Novagen) was trans-
formed with pAB146-tsf and pAB146-d.tsf for the expres-
sion and purification of wild-type and mutant EF-Ts,
respectively. LB medium containing 100 mgÆL
)1
of ampi-

cell extract was prepared by passing the cell slurry twice
through a French Press followed by centrifugation at
12 000 g for 20 min. The lysate was treated with DNase I
and loaded onto a chitin column (New England Biolabs),
equilibrated with Column Buffer, at 4 °C. The chitin
column was thoroughly washed with Column Buffer and
thereafter equilibrated with Precleavage Buffer (20 m
M
Tris/
HCl, pH 8.0, 50 m
M
NaCl, 0.1 m
M
EDTA). Then, the
column was incubated at 25 °C for 16–24 h after fast
equilibration with Cleavage buffer (Precleavage buffer
containing 60 m
M
dithiothreitol). Pure EF-Ts was eluted
with Elution Buffer (20 m
M
Tris/HCl, pH 8.0, 0.5
M
NaCl,
0.1 m
M
EDTA, 1.0 gÆL
)1
Triton-X-100) at 4 °Cand
dialysed against 20 m

and GDP. The EF-Tu–EF-Ts complex was
purified on a 5-mL HiTrap Q column (Pharmacia) using a
100-mL 100–375 m
M
KCl gradient in 20 m
M
Tris/HCl,
pH 7.6, 1 m
M
dithiothreitol. The design of fusion proteins
and the methods used for purification ensured that the
recombinant proteins have the same number of amino acids
as the native elongation factors.
Activity assays
The concentration of EF-Tu active in binding guanine
nucleotides, and the activity of EF-Ts in promoting the
exchange of guanine nucleotide with E. coli EF-Tu–GDP,
was determined essentially as described previously [41],
except that cellulose acetate filters (Gelman Sciences) were
used for filter-binding. EF-Tu was  50% active in guanine
nucleotide binding. Both wild-type and mutant EF-Ts were
estimated to be 100% active based on their ability to bind
EF-Tu, as judged by purification of the EF-Tu–EF-Ts
complex in excess of EF-Tu by anion-exchange chromato-
graphy, as described above. The ability of EF-Ts to
stimulate the exchange of EF-Tu-bound GDP with free
[
3
H]GDP was determined at 0 °C. The reaction mixtures
contained about 0.35 l

plotted as function of the added amount of EF-Ts. The
slope of the GDP exchange in the absence of EF-Ts was
subtracted from each slope value. The exchange of EF-Tu-
bound GDP with free [
3
H]GTP was performed as described
for [
3
H]GDP, except that pyruvate kinase and phos-
phoenolpyruvate were included in the [
3
H]GTP stock, so
that the final concentrations in the assay were 16.5 lgÆmL
)1
and 12 l
M
, respectively.
The ability of EF-Ts to stimulate EF-Tu in translation
in vitro was determined by poly(U)-directed polymerization
of phenylalanine in the polymix system, as described by
Ehrenberg et al. [42] and modified by Pedersen et al. [43].
Reaction mixtures were systematically titrated with E. coli
wild-type EF-Ts, yeast [
14
C]Phe-tRNA
Phe
, and ribosomes,
to ensure that the nucleotide exchange limited the transla-
tion. For the final experiments, the ribosome mix con-
tained 2.0 l

The stability of the E. coli EF-Tu–EF-Ts in the presence of
guanine nucleotides and kirromycin was studied using the
method of vertical zone-interference gel electrophoresis [45].
Agarose gels (1.5% and 2.0%), prepared in electrophoresis
buffer (20 m
M
Tris acetate, pH 7.6, 3.5 m
M
magnesium
acetate) were used in the experiments. Zone solutions
containing guanine nucleotide and kirromycin, at the
concentrations indicated, were prepared in electrophoresis
buffer containing 5% (v/v) glycerol. Sample solutions were
prepared in 10% (w/v) sucrose containing a trace of
bromphenol blue. Zone solutions of 80 lL, and sample
solutions of 5 lL, which contained 50 pmol of the EF-Tu–
EF-Ts complex, were pipetted into slots. The electropho-
resis was performed at 8 °C for 1 h at a constant voltage of
300 V, and the gel was stained as previously described [45].
Results
Deletion of the coiled-coil motif in endogenous EF-Ts
The sequence encoding the coiled-coil motif in tsf was
deleted by gene replacement in E. coli strain UY211, using
plasmid pMAK705-d.tsf (Fig. 2). The frequency of
co-integrates in the gene replacement procedure was in the
range of 10
)5
to 10
)4
. After regenerating free plasmid in the

in both groups of media.
Therefore, the growth rate of GRd.tsf ranges from  70–
95% of that of UY211 (Fig. 3). The lowest growth rate of
GRd.tsf (70%), relative to the growth rate of UY211, was
obtained in LB medium where both strains expressed their
maximal growth rate. The absence of the coiled-coil motif
does not affect growth at rates below 0.8 h
)1
. In contrast,
the mutant strain seems to be impaired in media supporting
higher growth rates. This disadvantage caused by deletion
4298 H. Karring et al. (Eur. J. Biochem. 270) Ó FEBS 2003
of the coiled-coil motif is dependent on growth rate, i.e. the
higher the potential growth rate (above 0.8 h
)1
), the larger
the impairment.
Phenotype during amino acid starvation
The phenotypes of UY211 and GRd.tsf during amino
acid starvation were determined by culturing cells at 37 °C
in glucose/Mops containing growth-limiting amounts of
proline (auxotrophic amino acid) (Fig. 4). Instantaneous
growth inhibition caused by proline exhaustion was observed
in both cases. However, strain GRd.tsf was found to be more
sensitive to the depletion of proline than the wild-type strain.
Proline exhaustion caused growth inhibition of GRd.tsf at a
significantly lower cell density compared with UY211
(Fig. 4A). Thus, the point of growth inhibition of GRd.tsf
(D
420

or during the full stringent response (Fig. 4B).
Growth rates in the presence of kirromycin
The effect of kirromycin on the growth of GRd.tsf was
investigated by measuring growth rates of cultures in LB
containing various concentrations of kirromycin. During
growth in microtitre trays, where aeration is suboptimal,
strains UY211 and GRd.tsf had similar growth rates
( 0.85 h
)1
) in the absence of kirromycin. The growth
rates of both strains decreased as the concentration of
kirromycin was increased from 0.5 to 5.0 l
M
. However, the
reduction in the growth rate caused by the presence of
kirromycin was larger for strain GRd.tsf than for the wild-
type strain (Fig. 5). Thus, at 50 l
M
kirromycin, the growth
rate of GRd.tsf, as a percentage of the corresponding
growth rate of the wild-type strain ( 0.32 h
)1
), was only
55%.
Fig. 3. Growth rates. Diagram showing the growth rate of strain
GRd.tsf as a function of the growth rate of the wild-type strain UY211.
Each point represents a specific medium either based on Luria–Bertani
(LB) medium (j) or supplemented minimal medium (h).
Fig. 4. Amino acid starvation and guanosine 5¢-diphosphate 3¢-diphosphate (ppGpp) and guanosine 5¢-triphosphate 3¢-diphosphate (pppGpp) [(p)ppGpp]
levels. (A) Growth curves of Escherichia coli UY211 (s)andGRd.tsf(j) in glucose/Mops initially containing 40 l

(Fig. 6B). The activity in the translation assay was 6.2 pmol
[
14
C]Phe incorporated/(min · pmol EF-Ts) for both EF-Ts
species.
Dissociation of the EF-Tu–EF-Ts complex by guanine
nucleotides and kirromycin
The stability of the mutant EF-Tu–EF-Ts complex in the
presence of guanine nucleotides and kirromycin was
analysed using zone-interference gel electrophoresis
(Fig. 7). The principle of zone-interference gel electro-
phoresis, described by Abrahams et al. [45], was used for
the analysis to ensure that the concentrations of ligands
were maintained at a constant level during migration of
the elongation factors, thereby favouring equilibrium
conditions. The concentration of guanine nucleotides
required to dissociate the mutant EF-Tu–EF-Ts complex
was at least two orders of magnitude lower than that for
the wild-type complex. While the wild-type EF-Tu–EF-Ts
complex remained stable at 1 m
M
GDP (Fig. 7A, lane 8),
the mutant complex dissociated at 10 l
M
GDP (Fig. 7A,
lane 14). When the experiment was repeated with GTP
instead of GDP, the wild-type EF-Tu–EF-Ts complex
remained stable at 10 m
M
GTP (Fig. 7B, lane 7), while

M
GTP, the wild-
type complex still dissociated at 5 l
M
kirromycin (Fig. 7D,
lane 8), while the mutant complex appeared to dissociate at
Fig. 5. Growth in the presence of kirromycin. The growth rate of strain
GRd.tsf is presented as a percentage of the corresponding growth rate
of the wild-type strain, UY211, in the presence of kirromycin.
Fig. 6. Stimulation of the activities of elongation factor Tu (EF-Tu) by the coiled-coil deleted elongation factor Ts (EF-Ts) mutant. (A) Stimulation of
the exchange of EF-Tu-bound GDP with free [
3
H]GDP by wild-type EF-Ts (s) and the coiled-coil deleted EF-Ts mutant (j). (B) Stimulation of
the activity of EF-Tu in poly(U)-directed poly(Phe) synthesis by wild-type EF-Ts (s) and the coiled-coil deleted EF-Ts mutant (j). The nucleotide
exchange assays were repeated at least five times and the poly(Phe) assay was repeated six times. Each point in the figure represents a slope obtained
from a time curve containing six measurements. The data shown are representative of the results obtained on each occasion.
4300 H. Karring et al. (Eur. J. Biochem. 270) Ó FEBS 2003
50 l
M
kirromycin (Fig. 7D, lane 18). In this experiment, no
bandshift was observed.
Discussion
Phenotype of
E. coli
mutant GRd.tsf
The results presented in Fig. 3 show that the functional
activity of the coiled-coil motif of EF-Ts does not limit
growth at growth rates below 0.8 h
)1
. However, in media

accordance with the view of Glazier et al. [50], who
monitored the synthesis of (p)ppGpp in the temperature-
sensitive EF-Ts mutant strain, HAK88, of E. coli. There-
fore, the present results indicate that the deletion of the
coiled-coil motif in EF-Ts probably reduces the formation
of EF-Tu–GTP–aa-tRNA owing to a decrease in the
regeneration of EF-Tu–GTP.
A complete deletion of the tsf gene in bacteria has, to our
knowledge, never been reported. In comparison, deletion of
the single-copy gene, TEF5, in yeast, which encodes the
Fig. 7. Analysis of the elongation factor Tu (EF-Tu)–elongation factor Ts (EF-Ts) complex in the presence of guanine nucleotides and kirromycin by
zone-interference gel electrophoresis. (A) Wild-type and mutant EF-Tu–EF-Ts complexes in the presence of different concentrations of GDP
(0–1.0 m
M
). (B) Wild-type and mutant EF-Tu–EF-Ts complexes in the presence of different concentrations of GTP (0–10 m
M
). (C) Wild-type and
mutant EF-Tu–EF-Ts complexes in the presence of 10 l
M
GTP and different concentrations of kirromycin (0–50 l
M
). The bandshift described in
the text is indicated by a star. (D) Wild-type and mutant EF-Tu–EF-Ts complexes in the presence of 1 l
M
GDP and different concentrations of
kirromycin (0–50 l
M
). Sample solutions contained 50 pmol EF-Tu–EF-Ts complex.
Ó FEBS 2003 Deletion of the coiled-coil motif of EF-Ts (Eur. J. Biochem. 270) 4301
guanine nucleotide-exchange factor (eEF1Ba) correspond-

that the reduced activity of the mutant EF-Ts in guanine
nucleotide exchange is caused by a reduction in the ability to
compete with guanine nucleotides for the binding to EF-Tu.
Previous mutational studies of E. coli EF-Ts, concerning
residues directly involved in the binding of EF-Tu, have
indicated a strong correlation between the abilities of the
mutated forms of EF-Ts to compete with GDP for binding
to EF-Tu and their activities in promoting guanine nucleo-
tide exchange [53]. In contrast, the present results indicate
that deletion of the coiled-coil motif in EF-Ts, which is not
in contact with EF-Tu in the EF-Tu–EF-Ts complex
(Fig. 1), strongly reduces the ability of EF-Ts to compete
with GDP as well as GTP for binding to EF-Tu. Moreover,
the activity in promoting guanine nucleotide exchange is
only slightly reduced.
Stability of the EF-Tu–EF-Ts complex in the presence
of kirromycin
The growth of GRd.tsf in the presence of kirromycin was
examined because the binding of kirromycin and EF-Ts to
EF-Tu has been shown to be mutually exclusive [27]. In the
presence of less than 0.5 l
M
kirromycininthemedium,
strain GRd.tsf was unaffected, while at concentrations
higher than 0.5 l
M
kirromycin, the growth rate of GRd.tsf
was reduced to a greater degree than that of the wild-type
strain. This suggests that the competition between kirro-
mycin and EF-Ts for the binding of EF-Tu is altered by

kirromycin to the mutant EF-Tu–EF-Ts complex does not
induce large conformational changes.
Stabilization of the mutant EF-Tu–EF-Ts complex by the
binding of kirromycin would reduce the pool of EF-Tu and
EF-Ts available for protein synthesis in the cell. This could
explain the additional reduction in the growth rate of
GRd.tsf compared with that of the wild-type strain in the
presence of the antibiotic. The formation of a quaternary
complex would probably require a conformational change
in the EF-Tu–EF-Ts complex, as the binding site of
kirromycin is not present in the EF-Tu–EF-Ts structure.
In addition, the structure of the wild-type EF-Tu–EF-Ts
complex [9], and that of the complex between EF-Tu–GDP
and aurodox (methyl kirromycin) [25], show that the
binding sites of EF-Ts and kirromycin on EF-Tu are not
overlapping. Thus, the suggested binding of kirromycin to
the mutant EF-Tu–EF-Ts complex either indicates that the
conformation of EF-Tu in the mutant complex is different
from that of the wild-type complex, or that the mutant
complex is more flexible than the wild-type complex, which
might facilitate the binding of the antibiotic.
Comparison of mutant EF-Ts and EF-Ts from bovine
mitochondria
Elongation factors equivalent to the prokaryotic EF-Tu and
EF-Ts are active during protein synthesis in mitochondria
and chloroplasts. The coiled-coil motif is preserved in
prokaryotes and chloroplasts, but absent in mammalian
mitochondria. Therefore, it was of interest to compare the
activities of mutant EF-Ts and EF-Ts
mt

logous EF-Tu–EF-Ts complex, the effect of making the
E. coli EF-Ts more mitochondrial-like by deleting the
coiled-coil region is a dramatic weakening of the complex.
Therefore, one would expect that an EF-Ts chimera
between the N-terminal half of EF-Ts
mt
and the
C-terminal half of E. coli EF-Tu should form an even
stronger complex with E. coli EF-Tu. This turns out to be
the case, as shown previously by Zhang et al. [58].
Previous studies have shown that mutations in EF-Ts
mt
,
which cause a weakening in the interaction with E. coli
EF-Tu, correlate with an increased ability to stimulate GDP
exchange as well as an increased activity in polymerization
[53,58]. The authors suggest that the strong interaction
between EF-Ts
mt
and E. coli EF-Tu makes it difficult for
guanine nucleotides to compete for interaction with EF-Tu,
thereby reducing the stimulatory effect of EF-Ts. Therefore,
the effect observed was probably caused by a shift in the
equilibrium governing the exchange reaction. We did not
observe a similar correlation in our studies, as both the
strength of the EF-Tu–EF-Ts complex, as well as the
stimulatory effect on guanine nucleotide exchange, was
decreased by deletion of the coiled-coil motif. This might
indicate that the coiled-coil motif is directly involved in
guanine nucleotide exchange.

EF-Tu–EF-Ts complex exists as a heterodimer, rather
than as a heterotetramer in translation, and therefore
indicate that the crystallographic dimerization of E. coli
EF-Ts is only caused by the packing of the EF-Tu–EF-Ts
complex in the crystals. In contrast to E. coli EF-Ts, the
protruding antiparallel coiled-coil motif (helices 6 and 7)
of Thermus thermophilus EF-Ts is not involved in any
dimerization in the crystal structure of the EF-Tu–EF-Ts
complex [28]. However, the dimerization of the coiled-coil
motifs in the E. coli EF-Tu–EF-Ts crystals suggests an
affinity of the coiled-coil motif for an unidentified helical
structure. In this regard, the present results could indicate
that such a helical structure interacts with the coiled-coil
motif during a conformational change in one step of
the nucleotide exchange reaction and somehow stabilizes
the EF-Tu–EF-Ts complex in the presence of guanine
nucleotides.
Pressure relaxation studies of the nucleotide exchange
reaction have indicated that the ternary complex, EF-Tu–
thioGDP–EF-Ts, undergoes an isomerization step [7].
Likewise, Gromadski et al. [1] have proposed that the
binary complex, EF-Tu–EF-Ts, and the ternary complexes
containing EF-Tu, EF-Ts and GDP/GTP, are structurally
different. Inspection of the structure of E. coli EF-Tu–
EF-Ts suggests that the N-terminal domain of EF-Ts,
which consists of helical structures, might be a good
candidate as an interaction partner for the coiled-coil motif.
The N-terminal domain appears to be highly mobile and
disordered in free EF-Ts [28], and is separated from the core
domain by a highly accessible region [38]. Based on the

ed.), pp. 445–473. American Society for Microbiology, Washing-
ton, D.C.
6. Eccleston, J.F. (1984) A kinetic analysis of the interaction of
elongation factor Tu with guanosine nucleotides and elongation
factor Ts. J. Biol. Chem. 259, 12997–13003.
7. Eccleston, J.F., Kanagasabai, T.F. & Geeves, M.A. (1988) The
kinetic mechanism of the release of nucleotide from elongation
factor Tu promoted by elongation factor Ts determined by pres-
sure relaxation studies. J. Biol. Chem. 263, 4668–4672.
8.An,G.,Bendiak,D.S.,Mamelak,L.A.&Friesen,J.D.(1981)
Organization and nucleotide sequence of a new ribosomal operon
in Escherichia coli containing the genes for ribosomal protein S2
and elongation factor Ts. Nucleic Acids Res. 9, 4163–4172.
9. Kawashima,T.,Berthet-Colominas,C.,Wulff,M.,Cusack,S.&
Leberman, R. (1996) The structure of the Escherichia coli EF-Tu–
EF-Ts complex at 2.5 A resolution. Nature 379, 511–518.
10. Arai, K.I., Kawakita, M. & Kaziro, Y. (1972) Studies on poly-
peptide elongation factors from Escherichia coli. II. Purification of
factors Tu-guanosine diphosphate, Ts, and Tu-Ts, and crystal-
lization of Tu-guanosine diphosphate and Tu-Ts. J. Biol. Chem.
247, 7029–7037.
11. Nesper, M., Nock, S., Sedlak, E., Antalik, M., Podhradsky, D. &
Sprinzl, M. (1998) Dimers of Thermus thermophilus elongation
factor Ts are required for its function as a nucleotide exchange
factor of elongation factor Tu. Eur. J. Biochem. 255, 81–86.
12. Miller, D.L. & Weissbach, H. (1977) Factors involved in the
transfer of aminoacyl-tRNA to the ribosome. In Molecular
Mechanisms of Protein Biosynthesis (Weissbach, H. & Petska, S.,
eds), pp. 323–373. Academic Press, New York.
13. Cashel, M. & Gallant, J. (1969) Two compounds implicated in the

aminoacidstarvationinEscherichia coli. J. Mol. Biol. 236,
441–454.
23. Sørensen, M.A. (2001) Charging levels of four tRNA species in
Escherichia coli Rel(+) and Rel(–) strains during amino acid
starvation: a simple model for the effect of ppGpp on translational
accuracy. J. Mol. Biol. 307, 785–798.
24. Parmeggiani, A. & Swart, G.W. (1985) Mechanism of action of
kirromycin-like antibiotics. Annu. Rev. Microbiol. 39, 557–577.
25. Vogeley, L., Palm, G.J., Mesters, J.R. & Hilgenfeld, R. (2001)
Conformational change of elongation factor Tu (EF-Tu) induced
by antibiotic binding. Crystal structure of the complex between
EF-Tu, GDP and aurodox. J. Biol. Chem. 276, 17149–17155.
26. Fasano, O., Bruns, W., Crechet, J.B., Sander, G. & Parmeggiani,
A. (1978) Modification of elongation-factor-Tu–guanine-nucleo-
tide interaction by kirromycin. A comparison with the effect of
aminoacyl-tRNA and elongation factor Ts. Eur. J. Biochem. 89,
557–565.
27. Chinali, G., Wolf, H. & Parmeggiani, A. (1977) Effect of kirro-
mycin on elongation factor Tu. Location of the catalytic center for
ribosome-elongation-factor-Tu GTPase activity on the elongation
factor. Eur. J. Biochem. 75, 55–65.
28. Wang, Y., Jiang, Y., Meyering-Voss, M., Sprinzl, M. & Sigler,
P.B. (1997) Crystal structure of the EF-Tu–EF-Ts complex from
Thermus thermophilus. Nat. Struct. Biol. 4, 650–656.
29. Karring, H., Andersen, G.R., Thirup, S.S., Nyborg, J., Spremulli,
L.L. & Clark, B.F. (2002) Isolation, crystallisation, and pre-
liminary X-ray analysis of the bovine mitochondrial EF-Tu–GDP
and EF-Tu–EF-Ts complexes. Biochim. Biophys. Acta. 1601, 172–
177.
30. Hamilton, C.M., Aldea, M., Washburn, B.K., Babitzke, P. &

38. Bøgestrand, S., Wiborg, O., Thirup, S. & Nyborg, J. (1995)
Analysis and crystallization of a 25 kDa C-terminal fragment of
cloned elongation factor Ts from Escherichia coli. FEBS Lett. 368,
49–54.
4304 H. Karring et al. (Eur. J. Biochem. 270) Ó FEBS 2003
39. Knudsen,C.R.,Clark,B.F.,Degn,B.&Wiborg,O.(1992)One-
step purification of E. coli elongation factor Tu. Biochem. Int. 28,
353–362.
40. Sottrup-Jensen, L. (1993) Determination of halfcystine in proteins
as cysteine from reducing hydrolyzates. Biochem. Mol. Biol. Int.
30, 789–794.
41. Miller, D.L. & Weissbach, H. (1974) Elongation factor Tu and
the aminoacyl-tRNA–GTP complex. Methods Enzymol. 30,219–
232.
42. Ehrenberg, M., Bilgin, N. & Kurland, C.G. (1988) Design and use
of a fast and accurate in vitro translation system. In Ribosomes and
Protein Synthesis, a Practical Approach (Spedding, G., ed.),
pp. 101–128. Oxford University Press, Oxford.
43. Pedersen, G.N., Rattenborg, T., Knudsen, C.R. & Clark, B.F.
(1998) The role of Glu259 in Escherichia coli elongation factor Tu
in ternary complex formation. Protein Eng. 11, 101–108.
44. Spedding, G. (1990) Isolation and analysis of ribosomes from
prokaryotes, eukaryotes, and organelles. In Ribosomes and Protein
Synthesis, a Practical Approach (Spedding, G., ed.), pp. 1–29.
Oxford University Press, Oxford.
45. Abrahams,J.P.,Kraal,B.&Bosch,L.(1988)Zone-interference
gel electrophoresis: a new method for studying weak protein–
nucleic acid complexes under native equilibrium conditions.
Nucleic Acids Res. 16, 10099–10108.
46. Gordon, J. & Weissbach, H. (1970) Immunochemical distinction

drial protein synthesis elongation factors. Identification and initial
characterization of an elongation factor Tu-elongation factor Ts
complex. J. Biol. Chem. 264, 19125–19131.
57. Xin, H., Leanza, K. & Spremulli, L.L. (1997) Expression of bovine
mitochondrial elongation factor Ts in Escherichia coli and char-
acterization of the heterologous complex formed with prokaryotic
elongation factor Tu. Biochim. Biophys. Acta 1352, 102–112.
58. Zhang, Y., Sun, V. & Spremulli, L.L. (1997) Role of domains in
Escherichia coli and mammalian mitochondrial elongation factor
Ts in the interaction with elongation factor Tu. J. Biol. Chem. 272,
21956–21963.
59. Zhang, Y. & Spremulli, L.L. (1998) Roles of residues in mam-
malian mitochondrial elongation factor Ts in the interaction with
mitochondrial and bacterial elongation factor Tu. J. Biol. Chem.
273, 28142–28148.
60. Gray, M., Burger, G. & Lang, B.F. (1999) Mitochondrial evolu-
tion. Science 258, 1476–1481.
61. Uhlin, U., Cox, G.B. & Guss, J.M. (1997) Crystal structure of the
epsilon subunit of the proton-translocating ATP synthase from
Escherichia coli. Structure 5, 1219–1230.
62. Maesaki,R.,Ihara,K.,Shimizu,T.,Kuroda,S.,Kaibuchi,K.&
Hakoshima, T. (1999) The structural basis of Rho effector
recognition revealed by the crystal structure of human RhoA
complexed with the effector domain of PKN/PRK1. Mol. Cell. 4,
793–803.
63. Biou, V., Yaremchuk, A., Tukalo, M. & Cusack, S. (1994) The 2.9
A crystal structure of T. thermophilus seryl-tRNA synthetase
complexed with tRNA (Ser). Science 263, 1404–1410.
64. Koradi, R., Billeter, M. & Wuthrich, K. (1996) MOLMOL: a
program for display and analysis of macromolecular structures.