Tài liệu hạn chế xem trước, để xem đầy đủ mời bạn chọn Tải xuống
1
/ 12 trang
THÔNG TIN TÀI LIỆU
Thông tin cơ bản
Định dạng
Số trang
12
Dung lượng
728,45 KB
Nội dung
Translationinitiationregiondependencyof translation
initiation inEscherichiacolibyIF1and kasugamycin
Serhiy Surkov
1
, Hanna Nilsson
1
, Louise C. V. Rasmussen
2
, Hans U. Sperling-Petersen
2
and Leif A. Isaksson
1
1 Department of Genetics, Microbiology and Toxicology, Stockholm University, Sweden
2 Department of Molecular Biology, Aarhus University, Denmark
Introduction
Translation initiation factor 1 (IF1), encoded by infA,
is a small protein consisting of 71 amino acids in
Escherichia coli. It is essential for cell viability [1], even
though the exact reason for this remains obscure [2,3].
IF1 is highly conserved, and homologous proteins are
present in all three domains of life (IF1 in bacteria,
aIF1A in archeaons, and eIF1A in eukaryotes) [4,5].
IF1 is the smallest of the three initiation factors in
E. coli. During initiationof translation, IF1 binds to
the 30S ribosomal subunit in the A-site region [6,7],
presumably through electrostatic interactions [8]. It
stimulates the action of the other two factors, espe-
cially translationinitiation factor 2 (IF2) [9,10]. A
crystal structure analysis ofIF1 bound to the 30S
ribosomal subunit shows that the factor is located in
a cleft formed between helix 44, the 530 loop of 16S
RNA, and ribosomal protein S12. Besides direct con-
formational changes in helix 44 and the neighboring
region, this binding induces small but significant
changes in overall 30S subunit conformation, tilting
the head of the subunit towards the A-site. It is pos-
sible that these conformational changes could be even
bigger in the absence of crystal lattice constraints
[11].
Keywords
bipA; cspA; infA(IF1); kasugamycin; yggJ
Correspondence
L. A. Isaksson, Department of Genetics,
Microbiology and Toxicology, Stockholm
University, S-10691 Stockholm, Sweden
Fax: +46 8 164315
Tel: +46 8 164197
E-mail: leif.isaksson@gmt.su.se
(Received 26 October 2009, revised 17
February 2010, accepted 17 March 2010)
doi:10.1111/j.1742-4658.2010.07657.x
Translation initiation factor 1 (IF1) is an essential protein in prokaryotes.
The nature ofIF1 interactions with the mRNA during translation initiation
on the ribosome remains unclear, even though the factor has several known
functions, one of them being RNA chaperone activity. In this study, we
analyzed translational gene expression in vivo in two cold-sensitive chromo-
somal mutant variants ofIF1 with amino acid substitutions, R40D and
R69L, using two different reporter gene systems. The strains with the
mutant IF1 gave higher reporter gene expression than the control strain.
The extent of this effect was dependent on the composition of the transla-
tion initiation region. The Shine–Dalgarno (SD) sequence, AU-rich ele-
ments upstream of the SD sequence and the region between the SD
sequence and the initiation codon are important for the magnitude of this
effect. The data suggest that the wild-type form ofIF1 has a translation
initiation region-dependent inhibitory effect on translation initiation. Kasu-
gamycin is an antibiotic that blocks translation initiation. Addition of
kasugamycin to growing wild-type cells increases reporter gene expression
in a very similar way to the altered IF1, suggesting that the infA mutations
and kasugamycin affect some related step intranslation initiation. Genetic
knockout of three proteins (YggJ, BipA, and CspA) that are known to
interact with RNA causes partial suppression of the IF1-dependent cold
sensitivity.
Abbreviations
IF1, translationinitiation factor 1; IF2, translationinitiation factor 2; SD, Shine–Dalgarno; TIR, translationinitiation region.
2428 FEBS Journal 277 (2010) 2428–2439 ª 2010 The Authors Journal compilation ª 2010 FEBS
Several functions are attributed to IF1. It facilitates
IF2-dependent fMet-tRNA binding to the P-site
[10,12,13], probably by stabilizing IF2 binding to the
30S subunit [14], and stimulates the GTPase activity of
IF2 [15]. It also increases binding of mRNA to the ini-
tiation complex in the presence of IF2 [16]. Together
with IF2, it stimulates drop-off of peptidyl-tRNAs
with short polypeptides from 70S ribosomes [17]. In a
recent study, it was shown that IF1 together with IF2
recognizes the formylmethionine moiety of initiator
aminoacyl-tRNA and discriminates against unformy-
lated and deacylated tRNA
f
Met
[3]. IF1 is necessary for
IF2 recycling after subunit joining and GTP hydrolysis
[13,18].
IF1 directly contacts domains III–V of IF2 during
initiation oftranslation [19,20]. Additionally, IF1 stim-
ulates both ribosomal dissociation and subunit associa-
tion without affecting the equilibrium point [21]. Even
though IF1 (together with fMet-tRNA) has many
effects on translation initiation, it is not crucial in the
in vitro translation system based on purified compo-
nents, in contrast to the other translation initiation
factors IF2 andtranslationinitiation factor 3 [22].
IF1 contains an oligomer-binding motif with high
homology to the RNA-binding domains of ribosomal
protein S1 and polynucleotide phosphorylase [23,24],
and the factor binds to different synthetic polynucleo-
tides in solution [25]. It has an RNA chaperone activ-
ity both in vivo andin vitro [26]. IF1 can act as a
transcriptional antiterminator in vivo, but this function
of the factor is not essential for cell growth [27]. Cold
shock stimulates expression ofIF1 at the levels of both
transcription andtranslation [28,29]. Other studies sug-
gest that heterologous expression of E. coliIF1 in
Bacillus subtilis can complement the double deletion of
the cold shock-inducible genes cspB and cspC [30].
Kasugamycin is an aminoglycoside antibiotic that
selectively inhibits initiationoftranslationin prokary-
otes [31]. The antibiotic impairs binding of fMet-tRNA
to the P-site on the 30S subunit and on 70S ribosomes
[31]. Recent X-ray analyses have located a kasugamy-
cin-binding site on the 30S subunit or on the 70S ribo-
some in the regionof the mRNA-binding tunnel in the
E-site and P-site [32,33]. As no effect of kasugamycin
on mRNA binding to the 30S subunit was shown [34],
it was proposed that the antibiotic effectively distorts
the mRNA structure near the P-site codon, thus pre-
venting efficient fMet-tRNA binding [32,33,35]. Trans-
lation of different mRNAs is affected by kasugamycin
[36], depending on the nature of the nucleotides in
mRNA corresponding to the E-site [32]. Translation of
leaderless mRNA starting directly from AUG is insen-
sitive to kasugamycin action [37].
Expression of a reporter gene is increased in E. coli
strains that carry mutations in the chromosomal infA
(IF1) gene [38]. These mutant strains grow consider-
ably slower than the parental MG1655 strain, and
some of them are cold sensitive for growth. The repor-
ter gene is significantly overexpressed in two cold-sen-
sitive chromosomal IF1 mutant strains or through
addition ofkasugamycin to the corresponding wild-
type strain. In this study, we have made an extensive
in vivo analysis of the mRNA sequence composition
that causes such overexpression. We demonstrate simi-
lar effects on TIR-dependent gene expression of the
IF1 mutations and the antibiotic kasugamycin. The
IF1-dependent cold sensitivity is partly suppressed by
elimination by genetic knockout of some other pro-
teins involved in RNA recognition (YggJ, BipA, and
CspA).
Results
Increased gene expression by a mutant form of
IF1
IF1 is essential for cell viability, although the reason
for this is obscure. A set of mutants with mutations in
infA, giving an altered IF1, has been isolated [38].
Some of these mutants are cold sensitive for growth,
and have increased gene expression in an in vivo repor-
ter system. We wanted to characterize the determinants
of this effect, as they could reveal interactions between
the factor and the rest of the translation initiation
machinery in the growing cell. Comparison of several
mutants with different chromosomal infA mutations
motivated a closer study of two mutants, with an
R40D alteration (strain CVR40D) or an R69L alter-
ation (strain CVR69L), in IF1. Both of these mutants
are cold sensitive, with CVR40D being more sensitive
than CVR69L, and both can survive with the mutated
IF1 gene in a single chromosome. Furthermore, both
mutants have moderately increased expression of both
the b-galactosidase and the A¢ reporter gene systems,
indicating an effect of the altered IF1 [2].
The A¢ reporter gene system is based on a plasmid
with a test gene (3A¢) with a varied sequence and an
internal standard gene (2A¢) (see below). In this sys-
tem, the 3A¢⁄2A¢ ratio is dependent on sequence
changes introduced in 3A¢ as long as the control gene
2A¢ is not altered. As IF1 is involved in expression of
both the 3A¢ test gene and the 2A¢ internal control
gene, it was not clear whether an observed increase in
the 3A¢⁄2A¢ ratio in the infA mutant bacteria was the
result of an increase in 3A¢ expression or a decrease in
2A¢ expression, or both.
S. Surkov et al. TIR dependence oftranslationinitiationby IF1
FEBS Journal 277 (2010) 2428–2439 ª 2010 The Authors Journal compilation ª 2010 FEBS 2429
To address this question, we performed a radioactive
double-labeling experiment using the wild-type strain
and the two IF1 cold-sensitive mutants CVR40D and
CVR69L [2]. The pSS101 vector with the 3A¢ and 2A¢
genes was used to study protein A¢ expression in the
strains by gel scanning. Proteins in CVR40D and
CVR69 were labeled with [
3
H]lysine. The MG1655
parental strain was labeled during cultivation with
[
14
C]lysine. Cells were cultivated separately, but each
mutant was pooled with MG1655 when harvested. The
3A¢-encoded and 2A¢-encoded double-labeled proteins
in the mixture were purified and separated on gels.
The
3
H ⁄
14
C isotope ratios were determined for the 2A¢
and 3A¢ gel bands, and compared with the isotope
ratios of the total cellular protein, from the sample
taken before the purification step. As can be seen in
Fig. 1, values for 2A¢ reference gene expression in the
IF1 mutants were quite similar to the values for total
proteins. In contrast, 3A¢ expression was increased by
the IF1 mutations. The increased values for the
3A¢⁄2A¢ ratio agree with previous determinations
obtained by scanning of Comassie-stained gel bands.
The results suggest that the increased 3A¢⁄2A¢ expres-
sion ratio in the IF1 mutants studied here is mainly
the result of increased 3A¢ expression, whereas 2A¢ is
essentially not affected by the IF1 mutation. This
implication is supported by preliminary 2D gel analysis
of the CVR40D proteome showing that most cellular
proteins are not affected by the R40D mutation, even
though some are increased and a few are decreased in
expression (not shown). Taken together, the total pro-
tein mixture can be used as a reliable standard refer-
ence, suggesting that the two infA (IF1) cold-sensitive
mutant strains, CVR40D and CVR69L, both show sig-
nificant 3A¢ reporter gene overexpression, as compared
with the 2A¢ reference gene and total cellular proteins.
Plasmid copy numbers in MG1655 and CVR40D
were evaluated by spectrophotometric and electropho-
retic analysis. The results indicated that changes in
plasmid copy number do not play any role in the
observed increase in reporter gene expression in the
infA mutant strains. By use of a northern blotting
technique, it was found that the mRNA levels corre-
sponding to the 3A¢ and 2A¢ genes were not altered by
the infA mutations (not shown).
Expression ofIF1 can be increased under different
physiological conditions [28,29], even though infA is
not under auto-control [46]. By using IF1-specific
monoclonal antibodies [40], we have found that the
levels ofIF1in CVR40D and CVR69L, using EF-Tu
as a reference, are slightly higher (1.50- and 1.33-fold,
respectively) than in the wild-type strain. However, no
increase was seen in the 2A¢⁄total protein ratio as a
result of the IF1 mutations, whereas the 3A¢⁄total pro-
tein ratio was increased about two-fold for both of
them (Fig. 1). Preliminary data suggest that overpro-
duction of wild-type IF1 from a multicopy plasmid
does not cause any increase in 3A¢ expression. The
data suggest that the increased 3A¢⁄2A¢ ratio is mostly
the result of changed functionality of the mutant IF1
and not of an altered IF1 level. Because the 3A¢ and
the 2A¢ genes in the pSS101 plasmid are different in
their translationinitiationregion (TIR) composition,
the data suggest that there are altered functional inter-
actions between the mutated IF1 factor and some
sequence signals in the TIR region. We decided to ana-
lyze these signals.
Effect of the downstream region composition
The influence of different sequences downstream of
AUG on gene expression at the translational level has
been well characterized [41]. Even though IF1 binds to
the A-site of the 30S subunit, different +2 codons do
not cause significantly changed levels of protein expres-
sion in different IF1 mutant strains [38], indicating a
0.0
0.5
1.0
1.5
2.0
2.5
3A
′
2A
′
Total protein
3
H/
14
C ratio
CVR40D/MG1655
CVR69/MG1655
MG1655 + ksg/MG1655
pSS101
3A′ TIR cuagcuaauaaauuaAGGAGGauuuaaauAUGaaaccucuagagucgacu
2A′ TIR cggauaacaauuucacacAGGAaacagaccAUGgaauugcaacacgauaag
Fig. 1. Protein expression from the pSS101 plasmid as measured
by
3
H ⁄
14
C double labeling. Proteins in the parental MG1655 strain
were labeled with [
14
C]lysine. Proteins in CVR40D and CVR69L or
in MG1655 in the presence of 175 lgÆmL
)1
kasugamycin were
labeled with [
3
H]lysine. Cultures were pooled together, and the iso-
tope ratios for the total cellular proteins, as well as the 3A¢ and 2A¢
protein bands in PAGE gels, were calculated. The isotope ratios for
the 3A¢ reporter gene and the 2A¢ reference gene are shown in
relation to the isotope ratio for the total cellular protein, which is
taken as unity. TIRs of the 3A¢ and 2A¢ genes in pSS101 are indi-
cated by the SDs in capital letters. The AUG initiation codon is in
underlined capital letters.
TIR dependence oftranslationinitiationbyIF1 S. Surkov et al.
2430 FEBS Journal 277 (2010) 2428–2439 ª 2010 The Authors Journal compilation ª 2010 FEBS
lack of codon specificity. Using CVR40D, we used the
3A¢⁄2A¢ test system to analyze other plasmids with dif-
ferent sequences downstream of the initiation codon in
3A¢ (downstream regions DR-A, DR-B, DR-C, and
DR-D) (Fig. 2) [42] but with a constant upstream
sequence. In CVR40D, the expression levels of these
3A¢ variants with the Shine–Dalgarno (SD)
+
sequence
were elevated, giving an approximately two-fold
increase (Fig. 2), which is similar to the observed value
for the original construct pSS101. This increased gene
expression in CVR40D relative to the parental strain
MG1655 is independent of the composition of the
downstream region.
Influence of the SD sequence and its upstream
sequence
IF1 inhibits the joining of 50S to the preinitiation
complex if the SD sequence in the mRNA is extended.
This effect is not seen for a four base SD [43]. For this
reason, we studied the influence on gene expression of
the length of the SD sequence in CVR40D and
CVR69L and their parental strain MG1655. A set of
constructs with different lengths of the SD sequence
(4–10 bases) was used. As shown in Fig. 3A, the gene
variants with a long SD sequence (6–10 bases) gave
increased 3A¢⁄2A¢ values in the mutants, whereas a
four base SD sequence gave very similar expression
values in the wild-type and mutant strains. This obser-
vation is in line with the fact that the 2A¢ reference
gene, with an SD
+
sequence that is four bases short
(Fig. 1), is expressed at an unchanged level in the infA
mutant strains (Figs 1 and 3A). However, the total
removal of the SD sequence from the test gene does
not abolish the elevated expression level in the infA
mutant strains, which remains 1.8-fold higher than in
MG1655 (Fig. 3B).
To further analyze the influence of the TIR sequence
on IF1-dependent gene expression, several reporter
gene variants were used. As can be seen in Fig. 3B for
CVR40D, all of the analyzed TIR sequences gave
higher 3A¢⁄2A¢ ratios than for the control strain
MG1655. This was particularly true for pSS201 with
its S1-binding site.
The sequence dependency on reporter gene expres-
sion was similar but less pronounced in CVR69L. The
exception was pSS201. However, at 30 °C (test
temperature for CVR69L), expression from this
plasmid is toxic for MG1655, which makes correct
evaluation of the expression levels difficult. The reason
for this toxic effect is not clear, and requires further
investigation.
Increased expression was also obtained by the addi-
tion ofkasugamycin to MG1655, as discussed below.
The effect of the length of the spacer between a canon-
ical SD
+
sequence and the initiation codon was also
analyzed. As shown in Fig. 3C, the longer spacers,
especially with a 12 base spacer, gave higher gene
expression in the infA mutants.
Comparison of two different reporter gene
systems
Relevant reporter gene sequences were also analyzed
by using the b-galactosidase assay system in MG1655
and CVR40D. For this purpose, the initiation region
of the b-galactosidase gene from the pCMS71 plasmid
[44] was replaced with some of the corresponding
DR-A cuagcuaauaaauuaAGGAGGauuuaaauAUGAAAGCAAUUUUCGUAc
DR-B cuagcuaauaaauuaAGGAGGauuuaaauAUGAGUGAAUCACAAGCCc
DR-C cuagcuaauaaauuaAGGAGGauuuaaauAUGAAAAAGGAGUCGACUc
DR-D cuagcuaauaaauuaAGGAGGauuuaaauAUGACCGAGGGUGUUUCCc
3A′
2A′
DR-A
3A′/2A′ 0.16
0.08
0.370.170.390.210.660.31
CVR40DMG1655CVR40DMG1655CVR40DMG1655CVR40DMG1655
DR-B
DR-C
DR-D
Fig. 2. PAGE analysis of reporter gene
expression in MG1655 and CVR40D. Protein
bands corresponding to the 3A¢ reporter
gene and 2A¢ reference gene products are
indicated. Sequences of TIRs of the reporter
gene are shown with the different down-
stream regions, DR-A, DR-B, DR-C, and
DR-D, in bold letters. The expression ratios
(3A¢⁄2A¢) for the two strains are given.
S. Surkov et al. TIR dependence oftranslationinitiationby IF1
FEBS Journal 277 (2010) 2428–2439 ª 2010 The Authors Journal compilation ª 2010 FEBS 2431
initiation regions of 3A¢ gene variants. The results
obtained by using the b-galactosidase or the 3A¢ repor-
ter gene systems are compared in Fig. 4. It can be seen
that the cold-sensitive CVR40D shows similarly
increased reporter gene expression, as compared with
the parental strain MG1655, for corresponding initia-
tion region sequences for both assay systems.
Increased sensitivity ofIF1 mutants to
kasugamycin
The mRNA sequence-specific inhibition of translation
initiation by the antibiotic kasugamycin is well
documented [32,36]. It binds to the mRNA upstream
of the initiation codon, and X-ray crystallography has
AB
C
3A′/2A′ ratio
3A′/2A′ ratio
3A′/2A′ ratio
Fig. 3. (A) Influence of the SD sequence length on reporter gene expression. Strains and the numbers of bases in SD are indicated. (B) TIR-
dependent reporter gene expression in MG1655, CVR40D and CVR69L or in MG1655 in the presence of 175 lgÆmL
)1
kasugamycin (ksg).
Sequences with different TIRs are shown. The SD region is in bold capital letters, and the initiation codon is in underlined capital letters.
pSS301 and pSS101 represent SD
)
and SD
+
versions of the parental 3A¢ plasmid. pSS201 carries an extension by the ribosome-binding
sequence of ribosomal protein S1 (capital letters). The six bases upstream of SD
+
(capital letters) in pSS103 are from the 2A¢ gene in
pSS101. pSS144 carries an extension of four bases in the spacer downstream of the SD
+
, as compared with pSS103. pSS133 carries three
altered bases in the spacer as compared with pSS101. (C) Influence of the distance (Dis) between the SD sequence and the AUG initiation
codon on reporter gene expression in MG1655, CVR40D and CVR69L. The number of bases forming the distance are indicated. The SD
region is in bold capital letters, and the initiation codon is in underlined capital letters.
TIR dependence oftranslationinitiationbyIF1 S. Surkov et al.
2432 FEBS Journal 277 (2010) 2428–2439 ª 2010 The Authors Journal compilation ª 2010 FEBS
located the antibiotic to the ribosomal tunnel region.
As the effects of the R40D and the R69L mutations
are dependent on the sequence upstream of the initia-
tion codon, we wanted to analyze the effects of these
mutations on growth and reporter gene expression in
comparison with the action of kasugamycin.
Growth of CVR40D and that of its parental strain
MG1655 were compared in the presence of kasugamy-
cin (Fig. 5A). Addition ofkasugamycin reduced the
growth rate of MG1655. During the growth curve
acquisition for CVR40D cells in broth medium, it
became apparent that addition ofkasugamycin to a
concentration of 70 lgÆmL
)1
or higher had a
delayed bacteriostatic effect, stopping growth of
CVR40D at a D
590 nm
of 0.8–1.0. In comparison,
MG1655 did not show a bacteriostatic response unless
as much as 140 lgÆmL
)1
kasugamycin was used
(Fig. 5A). Minimal inhibitory concentrations were
determined for the three strains in minimal medium.
These values were 40, 50 and 75 lgÆmL
)1
for
CVR40D, CVR69L, and MG1655, respectively. In
summary, both CVR40D and CVR69L are more sensi-
tive to kasugamycin during growth than is MG1655.
Influence ofkasugamycin on gene expression
The effect ofkasugamycin on expression of the 3A¢
reporter gene in plasmid pSS101 in the parental strain
MG1655 was analyzed. The 3A¢⁄2A¢ ratio was
increased approximately two-fold at a kasugamycin
concentration of 175 lgÆmL
)1
, as determined by gel
scanning (Fig. 5B). The other bacteriostatic antibiotics
tested (chloramphenicol and tetracycline) markedly
decreased the growth rate of MG1655 cells but did not
influence the 3A¢⁄2A¢ ratio associated with pSS101
(Figs 1 and 5B). Thus, the increased expression caused
by kasugamycin was specific for this antibiotic and
was not caused by the other two antibiotics, which
also inhibit translation. Analysis of reporter gene
expression using plasmid pSS101 showed that the pres-
ence of 50 lgÆmL
)1
kasugamycin in LB medium had
almost no effect on the 3A¢⁄2A¢ ratio in MG1655, but
caused an increased 3A¢⁄2A¢ ratio in CVR40D and
CVR69L (Figs 1 and 5C). At this antibiotic concentra-
tion (50 lgÆmL
)1
), no growth rate reduction was visi-
ble for either MG1655 or the mutant strains. The
3A¢⁄2A¢ ratios were also measured for the different
strains during growth in the presence of 70 or
100 lgÆmL
)1
kasugamycin. At these concentrations,
cells were collected at a D
590
nm of 0.6, before the bac-
teriostatic effect of the antibiotic is seen. As can be
seen in Fig. 5C, the increased 3A¢⁄2A¢ ratios reveal
that reporter gene expression is more sensitive to kasu-
gamycin for both IF1 mutants than for MG1655.
In MG1655, kasugamycin at 175 lgÆmL
)1
caused an
increase in the 3A¢⁄2A¢ ratio in a TIR-dependent man-
ner (Fig. 3B). This was particularly true in the case of
pSS201, which contains the extended S1-binding site,
and pSS144, which has a 12 base spacer between the
SD sequence and the initiation codon. The same
sequences gave increased gene expression in CVR40D
and CVR69L in the absence of kasugamycin. Thus,
addition ofkasugamycin to MG1655 has very similar
effects on reporter gene expression as the R40D and
the R69L mutations (Figs 1, 3B and 5B).
Suppression of cold sensitivity of an IF1 mutant
by inactivation of other genes
We investigated whether the inactivation of any other
genes could suppress the cold-sensitive phenotype of
CVR40D. The less cold-sensitive mutant CVR69L was
not analyzed. The KEIO collection of E. coli gene
knockout strains was used to introduce nonessential
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
A
B
CR40D/MG1655 ratio
pSS201, S1 binding site
pSS101, SD
+
pSS301, SD
–
pSS103, mutated sequence upstream of SD
pSS133, mutated spacer
pSS144, spacer extended by 4 bases
Fig. 4. Comparison of gene expression measurements using two
different reporter gene systems. The ratios of gene expression val-
ues (3A¢⁄2A¢) in CVR40D and MG1655 as measured by the b-galac-
tosidase (A) and protein A¢ (B) assays are shown. The plasmids
used for the comparison are indicated, and their TIR sequences are
given in Fig. 3. Spacer refers to the sequence between the initia-
tion codon AUG and the SD region.
S. Surkov et al. TIR dependence oftranslationinitiationby IF1
FEBS Journal 277 (2010) 2428–2439 ª 2010 The Authors Journal compilation ª 2010 FEBS 2433
gene knockouts by P1 transduction into CVR40D and
kanamycin selection [45]. For screening, we chose 69
nonessential genes that are known to be associated
with ribosome function or maturation as well as cold
shock response. The double mutant strains were tested
for growth on LB plates at 18 °C. It was found that
inactivation of bipA, yggJ or cspA partly suppressed
the cold-sensitive phenotype of the IF1 mutant
CVR40D (Fig. 6).
Discussion
IF1 stimulates translationinitiationin an in vitro sys-
tem by promoting formation of the preinitiation com-
plex [9,22]. However, it was found by Croitoru et al.
[38] that cold-sensitive chromosomal infA mutants with
the mutations R40D and R69L had increased reporter
gene expression in vivo as compared with the wild-type
strain. This depends on the TIR in the mRNA [38].
The overexpression found in the mutant as compared
with the wild-type strain, using two different reporter
gene systems, depends on the composition of the TIR.
Such TIRs include the downstream region following
the initiation codon, the length of the SD
+
sequence,
provided that it is longer than four bases, and the dis-
tance between the initiation codon and the SD
+
sequence, provided that it is shorter than 15 bases but
longer than six bases.
IF1 is known to have antitermination activity during
transcription [27], and the intracellular level ofIF1 can
be increased by physiological treatments [28,29], but
infA itself is not subject to any feedback control [46].
We found that the reporter mRNA levels in the
mutants are not altered by the infA mutations. Analy-
sis with IF1-specific monoclonal antibodies suggests
that the level ofIF1 is slightly increased in CVR40L
and CVR40D. No corresponding increase is seen
for the 2A¢⁄total protein ratio as compared with the
Growth in the presence of
kasugamycin
0
1
2
3
4
5
6
A
0 200 400 600
Time (min)
D
590
MG1655
MG1655, ksg 70 µg·mL
–1
MG1655, ksg 140 µg·mL
–1
CVR40D
CVR40D, ksg 70 µg·mL
–1
CVR40D, ksg 140 µg·mL
–1
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
Tet 0.5 µg·mL
–1
Cam 2 µg·mL
–1
Cam 1 µg·mL
–1
Ksg 175 µg·mL
–1
Ksg 100 µg·mL
–1
0
3A′/2A′ ratio
Kasu
g
amycin (µ
g
·mL
–1
)
0
0.5
1
1.5
2
2.5
0 50 100
MG1655
CVR40D
CVR69L
3A′/2A′ ratio
B
C
Fig. 5. (A) Growth in LB medium in the presence of kasugamycin
(ksg). Filled symbols represent MG1655 and open symbols repre-
sent the mutant CVR40D. (B) Reporter gene expression in MG1655
in the presence ofkasugamycin (ksg) or other antibiotics. The plas-
mid was pSS101, as described in Fig. 1. Cam, chloramphenicol;
Tet, tetracycline. (C) Reporter gene expression (pSS101) in
MG1655 or in CVR40D and CVR69L in the presence of different
kasugamycin concentrations.
1
3
2
4
Fig. 6. Growth of CVR40D and derivatives. The indicated strains
were incubated on an LB plate at 18 °C for 4 days. 1, CVR40D; 2,
CVR40D DcspA; 3, CVR40D DbipA; 4, CVR40D DyggJ.
TIR dependence oftranslationinitiationbyIF1 S. Surkov et al.
2434 FEBS Journal 277 (2010) 2428–2439 ª 2010 The Authors Journal compilation ª 2010 FEBS
wild-type strain, whereas the 3A¢⁄total protein ratio is
increased about two-fold. Taken together, the data
suggest that most of the increase in the 3A¢⁄2A¢ ratio
in the mutants is the result of changed functionality of
IF1 and not overproduction of mutant IF1.
Normally, one would expect a mutationally altered
protein to show lowered, not increased, activity. Our
analysis of reporter gene expression in the two cold-
sensitive chromosomal IF1 mutants suggests that IF1
plays a role as a TIR-dependent repressor of transla-
tion initiation, and that this negative effect is less pro-
nounced in the mutants. It has been shown by Milon
et al. [43] that IF1 acts as an inhibitor of formation of
the initiation complex in vitro. In that study, inhibition
by IF1 was found for an mRNA with a six base SD
sequence, but not if this sequence was only four bases
long. Our in vivo results described here for the R40D
IF1 mutant are in line with their results, as the mutant
IF1 showed increased expression in the case of a six
base, but not a four base, SD sequence.
The results reported here were obtained using two
different reporter gene systems, two independent IF1
mutations and addition ofkasugamycin to growing
wild-type or mutant bacteria. The question can be
asked as to what extent the results are applicable to
normal genes. Preliminary proteomic analysis suggests
that most normal genes are unaffected by the IF1 alter-
ations, thus resembling the 2A¢ gene. Some genes, how-
ever, show increased expression, being similar to 3A¢,
and others show decreased expression. Kasugamycin
increases the 3A¢⁄2A¢ ratio in our reporter gene system.
The antibiotic is an inhibitor oftranslation initiation.
This suggests that the apparently increased expression
of 3A¢ relative to 2A¢ or to total protein caused by kas-
ugamycin reflects a lower sensitivity of 3A¢ than of 2A¢
to kasugamycin. Upon addition ofkasugamycin to
growing bacteria, expression of some natural genes is
either increased or decreased, also suggesting a differ-
ent sensitivity to the antibiotic (not shown).
IF1 binds to different synthetic polynucleotides in
solution, and it contains an oligomer-binding motif
with high homology to the RNA-binding domains of
ribosomal protein S1 and polynucleotide phosphory-
lase [23–25]. The crystal structure ofIF1in complex
with the 30S ribosomal subunit suggests that IF1 could
directly contact mRNA nucleotides in the ribosomal
A-site [11]. IF1 affects the conformation of 16S rRNA,
causing a movement of helix 44 and a global confor-
mational change in the 30S subunit. This is visible as
the movement of the head of the subunit towards the
body and flipping of bases A1492 and A1493. This
flipping has been shown to constitute an important
part of the quality control signaling during tRNA or
RF factor recognition of the A-site [11]. The R40D
mutant described here is altered in its binding pocket
for the base A1493 [2]. This suggests a direct interac-
tion effect of the IF1 mutation. As IF1 influences the
splicing of a group I intron in vivo andin vitro, and
influences RNA annealing in vitro, the factor has an
RNA chaperone activity [26]. It is conceivable that the
mutant forms ofIF1 are less capable of setting the
intricate balance between favoring and disfavoring
higher RNA structures, in the rRNA, mRNA or both,
that are necessary for translational initiation. As a
result, the initiation machinery could be biased such
that the initiation conformation in the mutants is too
high, giving the observed decreased mutant growth
rate and cold sensitivity.
The very similar expression responses to a number
of different TIR sequences that were seen for CVR40D
and CVR69L as compared with the addition of kasu-
gamycin to the wild-type strain are compelling. The
antibiotic binds in the mRNA channel, just upstream
of AUG between bases G926 and A774 in 16S rRNA,
according to a current X-ray structure model. It dis-
torts the P-site, thereby disturbing the start codon
position and preventing fMet-tRNA binding during
the stage of 50S subunit joining [32,33]. Several aspects
of kasugamycin action are still unknown. However, it
is known that the resistance mutation (ksgA) affecting
modification of the 16S rRNA gives an effect at the
level of subunit joining [32]. In parallel with that, it
has been shown that the wild-type form ofIF1 reduces
subunit joining depending on the mRNA TIR in vitro
[43]. We have shown the apparent resemblance in the
TIR dependence ofIF1andkasugamycin action as
well as a synergistic effect on gene expression levels
between an infA mutation and addition of kasugamy-
cin. The results suggest that the actions of kasugamy-
cin andIF1 are dependent on a closely related target
or step during translation initiation. The correlation
suggests that those mechanisms underlying the actions
of the antibiotic andIF1 are similar, possibly at the
level of subunit joining. The effects of kasugamycin
and mutationally altered IF1 appear to be synergistic.
Using the Keio strain collection, we found that the
cold sensitivity associated with the R40D mutation is
partly suppressed by inactivation of yggJ, bipA,or
cspA. YggJ is a methylase that specifically modifies uri-
dine 1498 of the 16S rRNA. This base is located in the
mRNA channel upstream of AUG. The residue
directly contacts the kasugamycin molecule in the
X-ray structure [32]. It appears likely that it influences
the TIR selection or AUG adjustment specificity of
IF1. BipA is a protein that disrupts SD–antiSD
interactions in some mRNAs during the first steps of
S. Surkov et al. TIR dependence oftranslationinitiationby IF1
FEBS Journal 277 (2010) 2428–2439 ª 2010 The Authors Journal compilation ª 2010 FEBS 2435
translation [47]. Both of these two gene products are
connected to the TIR-specific response of the IF1
mutants, suggesting that the imbalance in the mRNA
translation is the primary reason for the cold sensitiv-
ity of the infA mutants studied here. It is conceivable
that elimination of yggJ or bipA could partially com-
pensate for the enhanced translationinitiation of
mRNAs that is observed in the IF1 mutant strains.
CspA facilitates translationinitiation at low temper-
atures by melting the mRNA secondary structure [39].
Cold shock stimulates expression ofIF1 at the levels
of both transcription andtranslation [28,29]. Thus,
cold shock constitutes a common denominator for
cspA and infA. The compensation of the cold sensitiv-
ity ofIF1 mutant strains by the inactivation of cspA is
another functional link to IF1.
In view of the finding that IF1 has RNA chaperone
activity [26], the effects of the elimination of BipA,
YggJ and CspA are conceivable, as all of them influ-
ence the structure of RNA. The results are all in line
with a model in which IF1, together with some other
proteins, recognizes and shapes the TIR region in
mRNA on the 30S ribosomal subunit.
Experimental procedures
Chemicals
All chemicals used were of the highest purity commercially
available. Enzymes were from New England Biolabs
(Ipswich, MA, USA), Invitrogen (Carlsbad, CA, USA),
Promega (Madison, WI, USA), and Fermentas Life
Sciences (Vilnius, Lithuania). DNA extraction kits were
from Qiagen (Hilden, Germany). Plasmids were prepared
using kits from Qiagen, Fermentas Life Sciences, and GE
Healthcare (Waukesha, WI, USA). Radioactive lysine was
from GE Healthcare.
Media and antibiotics
Strains were grown in LB broth with tryptone and yeast
extract or in M9 defined minimal medium [48] supplemented
with all amino acids except for lysine. Ampicillin was used at
a concentration of 100 lgÆmL
)1
. Other antibiotics were used
as indicated. Kasugamycin (Sigma, St Louis, MO, USA) was
dissolved in water at 12.5 mgÆmL
)1
, and the pH was adjusted
with NaOH.
Protein labeling
Overnight cultures of MG1655, CVR40D or CVR69L cells,
carrying the pSS101 plasmid with the 3A ¢ reporter system,
were grown in M9 minimal medium supplemented with
ampicillin and all amino acids except lysine. 50 lL of the
overnight cell culture were inoculated into 5 mL of the same
medium, and they were grown with intensive aeration to a
D
590 nm
of 0.2. At this point, 75 lL of a lysine solution
labeled with
14
C (50 lCiÆmL
)1
) was added to MG1655 cells,
and 35 lL of lysine solution labeled with
3
H (1 mCiÆmL
)1
)
was added to CVR40D and CVR69L cells. The cells were
grown to a D
590 nm
of 0.6. Then, cold lysine was added to the
final concentration of 0.25 mm, and the cultures were grown
for an additional 20 min at 37 °C for CVR40D and at 30 °C
for CVR69L. The difference in the growth temperature is
due to different cold sensitivities of the strains. CVR40D
shows a decreased growth rate, down to 50%, at 37 °C.
CVR69L is grown at 30 °C to obtain a similar decrease in
growth rate. The cultures were cooled, and MG1655 cells
were combined with CVR40D or CRV69L cells. The com-
bined cells were washed, harvested, and processed for pro-
tein A¢ analysis by gel electrophoresis [49]. The double-
labeled protein bands were excised and kept overnight for
extraction in 300 lL of a solution containing 30% H
2
O
2
and
1% NH
4
OH at 37 °C. The resulting solution was placed in a
scintillation vial containing 2 mL of Ultima Gold XR cock-
tail (Perkin Elmer, Norwalk, CT, USA), and shaken for 1 h
at room temperature. Radioactivity was counted in a 1219
Rackbeta scintillation counter (LKB, Bromma, Sweden).
The amounts of radioactivity corresponding to
3
H and
14
Cin
each sample were corrected according to the energy spectra
of the pure elements, giving the ratio of counts originally
derived from
3
H-labeled and
14
C-labeled lysine.
The radioactivity of total protein samples was measured
by placing 300 lL of the pooled
3
H-labeled and
14
C-labeled
cell cultures into 10% trichloroacetic acid with 1% casamino
acids. The precipitates were washed with 5% trichloroacetic
acid containing 0.1% casamino acids, whereafter the filters
were dried and radioactivity was measured in a scintillation
counter. Alternatively, the pooled
3
H-labeled and
14
C-labeled
cell lysate was loaded onto a polyacrylamide gel, and all of
the resulting protein bands in one lane were excised together
and counted as described above. Both methods gave similar
3
H ⁄
14
C ratios. Similar PAGE experiments were performed
with MG1655 cells expressing 3A¢ and 2A¢ proteins encoded
by the pSS101 plasmid in the presence or absence of
175 lgÆmL
)1
kasugamycin.
Protein A¢ assay
The protein A¢ reporter system has been extensively
described [49]. Briefly, a plasmid carries two genes under
the control of identical trc promoters. Both proteins are
composed of identical A¢ building blocks derived from the
IgG-binding domain (also known as the Z domain) of
Staphylococcus aureus protein A. One gene, encoding three
A¢ domains (3A¢, 21 kDa) is a reporter gene that can be
modified and used to study the influence of different
mRNA sequences on gene expression. The second gene in
the plasmid, encoding two A¢ domains (2A¢, 14 kDa) is an
TIR dependence oftranslationinitiationbyIF1 S. Surkov et al.
2436 FEBS Journal 277 (2010) 2428–2439 ª 2010 The Authors Journal compilation ª 2010 FEBS
internal control gene. The protein products of both genes
are purified by affinity chromatography using IgG Sepha-
rose, and the relative expression ratio 3A¢⁄2A¢ is estimated.
Protein A¢ is not toxic, and this system does not have any
transcriptional polarity effects.
Expression of the 3A¢ and 2A¢ genes was analyzed by
using 15% SDS ⁄ PAGE electrophoresis. Gels were stained
with Comassie Brilliant Blue R (Sigma) and scanned using
a LAS1000 plus (FujiFilm) camera. Bands corresponding to
3A¢ and 2A¢ proteins were quantified using image gauge
v. 4.0 (FujiFilm). All experiments were repeated at least
four times.
b-Galactosidase assay
Wild-type MG1655 and the IF1 mutants CVR40D and
CVR69L were grown overnight at 37 °C in M9 medium
supplemented with all amino acids at the recommended
concentrations and 100 lgÆmL
)1
ampicillin. These cultures
were used for inoculatation into the same medium at 37 °C.
Exponentially growing cells (D
590 nm
of 0.4–0.5) were har-
vested without isopropyl thio-b-d-galactoside induction, as
the trc promoter is leaky, giving significant expression even
in the absence of induction. b-Galactosidase acivity of the
lysed uninduced cells was then determined as described pre-
viously [50].
Plasmids and strains
All plasmids used are based on the pHN109 vector [44],
which carries the 2A¢ internal control gene and the 3A¢ test
gene. The different initiation regions in the 3A¢ test gene in
the plasmids used are shown in the corresponding figures.
P1 transduction was performed according to Miller [51].
The MG1655 strain (F
)
, ilvG, rfb-50, rph) was used as a
wild-type reference strain. Its derivatives CVR40D and
CVR69L have the same genotype except for the R40D
or R69L mutations, respectively, in the infA gene on the
chromosome [2].
Growth curves
Thirty microliter volumes of overnight cultures were inocu-
lated into 3 mL of fresh LB. Cells were grown with intensive
shaking to a D
590 nm
of 0.6, and then diluted in LB to obtain
a D
590 nm
of 0.05 in the presence or absence of kasugamycin;
this was followed by measurements of bacterial growth.
Cold sensitivity complementation test
The gene deletions tested were transferred from the KEIO
strain collection [45] by P1 transduction into CVR40D on
LB plates, with selection for kanamycin resistance. Plates
were incubated for 4 days at 18 °C.
Immunoblotting
Cells from overnight cultures were lysed by sonication, and
cell lysates were loaded into a 15% SDS ⁄ PAGE gel along
with IF1and EF-Tu standards. Proteins were transferred
to a nitrocellulose membrane by electroblotting. Buffer
[0.9% NaCl, 50 mm Tris ⁄ HCl (pH 7.5), 1% (v ⁄ v) Tween]
was used for the wash and incubation steps. Dry milk was
used for blocking, three types of mouse monoclonal anti-
bodies against IF1 (1BD3, 3AE12 and 2EF10 from [40])
and rabbit polyclonal antibodies against EF-Tu were used
as primary antibodies, and horseradish peroxidase-conju-
gated swine anti-(rabbit IgG) and horseradish peroxidase-
conjugated rabbit anti-(mouse IgG) were used as secondary
antibodies. The blot was developed using a GE Healthcare
ECL kit and exposed to X-ray film. The X-ray film was
scanned using a GS-800 calibrated densitometer (Bio-Rad,
Richmond, CA, USA), and analyzed with BioRad quantity
one software.
Acknowledgements
We thank V. Croitoru for the CVR strains, as well
as help and advice. This work was supported by
grants from the Swedish Science foundation (L. C. V.
Rasmussen) and from the Swedish Institute (Visby
Program). We thank Genkaku for strains from the
Keio collection.
References
1 Cummings HS & Hershey JW (1994) Translation initia-
tion factor IF1 is essential for cell viability in Escheri-
chia coli. J Bacteriol 176, 198–205.
2 Croitoru V, Bucheli-Witschel M, Hagg P, Abdulkarim
F & Isaksson LA (2004) Generation and characteriza-
tion of functional mutants in the translation initiation
factor IF1ofEscherichia coli. Eur J Biochem ⁄ FEBS
271, 534–544.
3 Antoun A, Pavlov MY, Lovmar M & Ehrenberg M
(2006a) How initiation factors maximize the accuracy of
tRNA selection ininitiationof bacterial protein synthe-
sis. Mol Cell 23, 183–193.
4 Kyrpides NC & Woese CR (1998) Universally con-
served translationinitiation factors. Proc Natl Acad Sci
USA 95, 224–228.
5 Sorensen HP, Hedegaard J, Sperling-Petersen HU &
Mortensen KK (2001) Remarkable conservation of
translation initiation factors: IF1 ⁄ eIF1A and
IF2 ⁄ eIF5B are universally distributed phylogenetic
markers. IUBMB Life 51, 321–327.
6 Moazed D, Samaha RR, Gualerzi C & Noller HF
(1995) Specific protection of 16 S rRNA by
S. Surkov et al. TIR dependence oftranslationinitiationby IF1
FEBS Journal 277 (2010) 2428–2439 ª 2010 The Authors Journal compilation ª 2010 FEBS 2437
[...]... usage of infA in response to cold shock inEscherichiacoli FEBS Lett 580, 539–544 Giangrossi M, Brandi A, Giuliodori AM, Gualerzi CO & Pon CL (2007) Cold-shock-induced de novo transcription andtranslationof infA and role ofIF1 during cold adaptation Mol Microbiol 64, 807–821 Weber MH, Beckering CL & Marahiel MA (2001) Complementation of cold shock proteins bytranslationinitiation factor IF1in vivo... Binding of formylmethionyl-tRNA to 30S ribosomal sub-units Nature 217, 1013–1016 Kay AC & Grunberg-Manago M (1972) The mechanism of action ofinitiation factor IF1: non-analogy with elongation factor EFTs Biochim Biophys Acta 277, 225–230 Stringer EA, Sarkar P & Maitra U (1977) Function ofinitiation factor 1 in the binding and release ofinitiation factor 2 from ribosomal initiation complexes in Escherichia. ..TIR dependence oftranslationinitiationbyIF1 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 S Surkov et al translational initiation factors J Mol Biol 248, 207– 210 Dahlquist KD & Puglisi JD (2000) Interaction oftranslationinitiation factor IF1 with the E coli ribosomal A site J Mol Biol 299, 1–15 Celano B, Pawlik RT & Gualerzi CO (1988) Interaction ofEscherichiacoli translation- initiation factor... structure of the translational initiation factor IF1 from E coli contains an oligomer-binding motif EMBO J 16, 1436–1443 Gribskov M (1992) Translational initiation factors IF-1 and eIF-2 alpha share an RNA-binding motif with prokaryotic ribosomal protein S1 and polynucleotide phosphorylase Gene 119, 107–111 Schleich T, Verwolf GL & Twombly K (1980) A circular dichroism study ofEscherichiacoli initiation. .. Differential inhibition of coliphage MS2 protein synthesis by ribosome-directed antibiotics J Mol Biol 70, 41–55 37 Moll I & Blasi U (2002) Differential inhibition of 30S and 70S translationinitiation complexes on leaderless mRNA bykasugamycin Biochem Biophys Res Commun 297, 1021–1026 38 Croitoru VV, Bucheli-Witschel M & Isaksson LA (2005) In vivo involvement of mutated initiation factor IF1in gene expression... Cooperative effects by the initiation codon and its flanking regions on translationinitiation Gene 273, 259–265 42 Jin H, Zhao Q, Gonzalez de Valdivia EI, Ardell DH, Stenstrom M & Isaksson LA (2006) In uences on gene expression in vivo by a Shine–Dalgarno sequence Mol Microbiol 60, 480–492 43 Milon P, Konevega AL, Gualerzi CO & Rodnina MV (2008) Kinetic checkpoint at a late step intranslationinitiation Mol... CM, Jin H, Major LL, Tate WP & Isaksson LA (2001) Codon bias at the 3¢-side of the initiation TIR dependence oftranslationinitiationbyIF1 45 46 47 48 49 50 51 codon is correlated with translationinitiation efficiency inEscherichiacoli Gene 263, 273–284 Baba T, Ara T, Hasegawa M, Takai Y, Okumura Y, Baba M, Datsenko KA, Tomita M, Wanner BL & Mori H (2006) Construction ofEscherichiacoli K-12 inframe,... IF1and IF2 synergistically remove peptidyl-tRNAs with short polypeptides from the P-site of translating Escherichiacoli ribosomes J Mol Biol 281, 241–252 Chae YB, Mazumder R & Ochoa S (1969) Polypeptide chain initiationin E coli: isolation of homogeneous initiation factor E2 and its relation to ribosomal proteins Proc Natl Acad Sci USA 62, 1181–1188 Brock S, Szkaradkiewicz K & Sprinzl M (1998) Initiation. .. & SperlingPetersen HU (2005) Initiationof protein synthesis in bacteria Microbiol Mol Biol Rev 69, 101–123 Antoun A, Pavlov MY, Lovmar M & Ehrenberg M (2006b) How initiation factors tune the rate ofinitiationof protein synthesis in bacteria EMBO J 25, 2539– 2550 Carter AP, Clemons WM Jr, Brodersen DE, MorganWarren RJ, Hartsch T, Wimberly BT & Ramakrishnan V (2001) Crystal structure of an initiation. .. Initiation factors of protein biosynthesis in bacteria and their structural relationship to elongation and termination factors Mol Microbiol 29, 409–417 Allen GS, Zavialov A, Gursky R, Ehrenberg M & Frank J (2005) The cryo-EM structure of a translationinitiation complex from Escherichiacoli Cell 121, 703–712 van der Hofstad GA, Buitenhek A, van den Elsen PJ, Voorma HO & Bosch L (1978) Binding of labeled . Translation initiation region dependency of translation initiation in Escherichia coli by IF1 and kasugamycin Serhiy Surkov 1 , Hanna Nilsson 1 , Louise C. V. Rasmussen 2 , Hans U. Sperling-Petersen 2 and. TIRs of the 3A¢ and 2A¢ genes in pSS101 are indi- cated by the SDs in capital letters. The AUG initiation codon is in underlined capital letters. TIR dependence of translation initiation by IF1. of bases forming the distance are indicated. The SD region is in bold capital letters, and the initiation codon is in underlined capital letters. TIR dependence of translation initiation by IF1