affinity purification of ribosomes with a lethal g2655c ... site zz domains fusion protein was...
TRANSCRIPT
Affinity purification of ribosomes with a lethal G2655C mutation in 23S rRNA
that affects the translocation
Andrei A. Leonov§, Petr V. Sergiev§#, Alexey A. Bogdanov, Richard
Brimacombe* and Olga A. Dontsova
Department of Chemistry, Moscow State University, Moscow, 119899, Russia *Max Planck
Institut fur Molekulare Genetik, Berlin, Germany
§Authors contributed equally to the work
#Corresponding author:
Tel: +7-095-9395418,
Fax: +7-095-9393181
e-mail: [email protected]
1
Copyright 2003 by The American Society for Biochemistry and Molecular Biology, Inc.
JBC Papers in Press. Published on May 1, 2003 as Manuscript M302873200 by guest on June 3, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Running Title: Affinity purified ribosomes with lethal mutation
Summary
A method for preparation of E. coli ribosomes carrying lethal mutations in 23S rRNA was
developed. The method is based on the site-directed incorporation of a streptavidin-binding tag
into functionally neutral sites of the 23S rRNA and subsequent affinity chromatography. It was
tested with ribosomes mutated at 23S rRNA position 2655 (the EF-G binding site). Ribosomes
carrying the lethal G2655C mutation were purified and studied in vitro. It was found in
particular that this mutation confers strong inhibition of the translocation process, but only
moderately affects GTPase activity and binding of EF-G.
2
by guest on June 3, 2018http://w
ww
.jbc.org/D
ownloaded from
Introduction
Now that atomic structures of ribosomal subunits have become available (1, 2, 3), the major
emphasis of ribosome research has shifted towards gaining an understanding of structure-
function relationships. Analysis of the role that specific rRNA nucleotides play in the translation
process is of primary interest. In this respect the application of site-directed mutagenesis of
rDNA has so far had one general limitation, namely that for the biochemical characterization of
mutant ribosomes the mutants need to be prepared in a pure form. An E. coli strain was created
in the laboratory of C. Squires, in which all the chromosomal rDNA operons were deleted (4).
With the help of such strain it is possible to produce mutant ribosomes, if these ribosomes are
able to support cell growth (5, 6). Obviously, most functionally interesting mutations are likely
to be lethal and could only be expressed if wild-type ribosomes are co-expressed to support
translation.
Two methods have been published describing attempts to create a pure population of ribosomes
carrying lethal mutations (7, 8, 9, 10). Both of them involve in vitro transcription of an rDNA or
rDNA fragment and subsequent reconstitution of the ribosomal subunit. However, the first
method (9) uses ribosomes that are not from E. coli, and the second (7) utilizes fragmented
rRNA, leading to very low activity (see 10, for discussion). The major problem here lies in the
absence of natural modifications as well as possible (at least partial) misfolding of the rRNA in
vitro, as compared with the in vivo assembly.
We aimed to create a system for affinity purification of ribosomes, assembled in vivo, via an
aptamer tag attached to the mutated rRNA. Whereas the purification of proteins carrying
genetically introduced affinity tags has become a routine, only a few such tags are known for
RNA (11, 12, 13), and none of them has so far been applied to a ribonucleoprotein particle with
3
by guest on June 3, 2018http://w
ww
.jbc.org/D
ownloaded from
the size of the ribosome.
One of the most challenging steps of the protein synthetic cycle is translocation, and the
coupling of GTP hydrolysis and translocation, stimulated by elongation factor G, has been
widely discussed in the literature. There are two known sites of interaction of EF-G with the
23S rRNA, namely the sarcin-ricin loop (SRL) and the GTPase-associated center (14). The
mutation G2655C of the SRL of 23S rRNA is known from the literature to be lethal (15), an
effect, which is related to the interaction of the elongation factor with the ribosome. In this
paper we decided to study this mutation in more detail, namely to find an exact stage in the
elongation factor G cycle affected by the G2655C mutation. For this purpose we applied a
method for the affinity purification of ribosomes harboring the G2655C lethal mutation, in
amount and purity sufficient for biochemical characterization.
Experimental Procedures
Plasmids
Introduction of the MS2 coat protein binding site into helices 9, 25, 45, and 98 of 23S rRNA has
been described by Matadeen et al. (16). The plasmid, coding for the MS2 coat protein dimer
TEV-cleavage site ZZ domains fusion protein was constructed from the pET-TEV-ZZ vector
(17) and the pCP12d1-12 plasmid (11), encoding the MS2 coat protein dimer. The streptomycin
binding aptamer (12) was introduced by sequential site-directed mutagenesis of M13 mp18
phage, carrying an SphI-BamHI fragment of the E. coli rrnB operon with the MS2 coat protein
binding already inserted site in helix 98. The resulting sequence of helix 98 of the 23 S rRNA is
2791-
GACCCUGGAUCGCAUUUGGACUUCUGCCCAGGGUGGCACCACGGUCGGAUCCA
GGGUC-2805. To introduce the streptavidin binding aptamer (13), an XbaI-SphI fragment of the
4
by guest on June 3, 2018http://w
ww
.jbc.org/D
ownloaded from
rrnB operon with the MS2 coat protein binding site already inserted at either helix 9, 25, or 45
was subcloned into an additional plasmid, where the KpnI and XmaIII sites of the MS2 insert
were unique. The plasmid, cut with KpnI and XmaIII, was ligated with annealed
oligonucleotides, coding for the aptamer, followed by subcloning of the XbaI-SphI fragment
back into the expression vector pLK1192U (18). The resulting sequence of helix 25 of the 23S
rRNA is 540-
CACGCUUGGGUACCGGCUGGGCCGACCAGAAUCAUGCAAGUGCGUAAG
AUAGUCGCGGGCCGGCCCGGCCGGGUACUCAGGC-550.
Mutations of nucleotide 2655 of the 23 S rRNA were made by site-directed mutagenesis of an
SphI-BamHI fragment of the E. coli rrnB operon, subcloned into M13 mp 18 phage.
Subsequently, the mutated fragment was subcloned back into the pLK1192U plasmid either with
or without the affinity tag.
The XL1 strain of E. coli was used for all genetic manipulations and phage growth. For
preparation of uracil-containing single-stranded DNA for mutagenesis, strain CJ236 was used
as a host for M13 phage. For translation fidelity measurements strain MC140 was used. Strain
AVS69009 (4) carrying no chromosome-encoded rRNA was transformed by pLK1192U
plasmids, carrying mutations. For displacement of originally present pHK-rrnC plasmid the
transformants were streaked on LB plates, containing spectinomycin (pLK1192U encodes a
mutation in 16S rRNA, causing resistance to spectinomycin). The loss of the pHK-rrnC plasmid
was monitored by loss of kanamycin resistance. The absence of wild-type rRNA in the cell was
confirmed by the primer-extension method, as in Sergiev et al. (5). The primer was
complementary to nucleotides 548-564 of the 23S rRNA. The extension was carried out in the
presence of dATP, dCTP, dTTP and ddGTP, thus proceeding up to nucleotide C544 in the wild
type case and until the last cytosine of the insert, 2 nucleotides from the primer binding site in
the case of ribosomes carrying the affinity tag.
5
by guest on June 3, 2018http://w
ww
.jbc.org/D
ownloaded from
Translational fidelity measurements and growth rate determination
Cells from strain MC140, transformed by both pLK1192U, carrying a mutation at nucleotide
2655 of the 23S rRNA and the pSG or CSH series of plasmids (19), were grown at 37°C until an
optical density at 600 nm of 0.3 was reached. β-Galactosidase activity was assayed essentially
according to O’Connor et al. (20).
Growth rates were measured at 37 °C in LB media, containing 50µg/ml ampicillin. Three
independent colonies were grown overnight after transformation of the XL1 strain or plasmid
substitution in the AVS69009 strain, and were used to inoculate a volume of medium 50 times
larger. Cell densities were monitored by absorbance at 600 nm.
Ribosome preparation and affinity purification
Ribosomes were prepared by the reassociation method (21). Affinity purification was carried out
in subunit dissociation buffer (20mM Hepes-K pH 7.5, 1 mM Mg(OAc)2, NH4C1 200mM,
4mM 2-mercaptothanol) at 4°C. Binding of 1000 pmol of ribosomes to 100 µl of streptavidin
Sepharose (Pharmacia) was performed overnight (16h) in a 2 ml tube, followed by 3 washes
with binding buffer. Elution was also carried out overnight, using binding buffer saturated with
biotin. After elution, the 50S ribosomal subunits were reassociated with 30S subunits from an
additional subunit preparation and the 70S ribosomes were separated by sucrose density gradient
centrifugation. The ratio of wild-type to mutant in the ribosome preparations was estimated by
primer extension, according to Sigmund et al., (22).
GTPase reaction
For the GTPase reaction 4 pmol of ribosomes were mixed with a preincubated (10 min, 37°C)
mixture of 4 pmol EF-G, 4 pmol GTP and 0.1 µCi γ-[32P]GTP for one round of hydrolysis and
6
by guest on June 3, 2018http://w
ww
.jbc.org/D
ownloaded from
400 pmol GTP for multiple rounds of hydrolysis. Buffer conditions for the reaction were 20 mM
Hepes-K pH 7.5, 10 mM MgCl2, 100 mM NH4Cl, 4mM 2-mercaptoethanol. Fusidic acid was
added if required to a final concentration of 1mM. 40% formic acid was used to stop the
reaction. The extent of GTP hydrolysis was assayed by thin-layer chromatography (23).
Recombinant EF-G was purified from the superproducing strain (24), provided by Dr. T. Ueda.
Footprinting
For footprinting experiments, the complexes of ribosomes and EF-G with or without
fusidic acid were prepared as for the GTPase reaction (above), but without radioactive GTP.
DMS modification was carried out by addition of 0.2 µl of 20% DMS solution in ethanol to 10
µl of the complex. Modification was allowed to proceed for 10 minutes at 37°C. The reaction was
stopped, rRNA extracted and primer extension carried out as described in Sergiev et al. (5). For
the analysis of the region around the SRL and GTPase associated center, primers
complementary to 23S rRNA nucleotides 1121-1140 and 2730-2749, respectively, were used.
Translocation experiments
For translocation experiments, 2.5 pmol ribosomes were mixed with 1 µg polyU (Sigma) and a
two-fold excess of tRNAPhe in 10 µl volume of polyamine buffer (20mM Hepes-K pH 7.5, 6
mM Mg(OAc)2, 150mM NH4OAc, 2mM spermidine, 0.05 mM spermine, 4mM 2-
mercaptothanol) (21), and incubated at 37 °C 15 min. After incubation, a 1.2-fold excess of
AcPhe-tRNAPhe in the same buffer was added, followed by incubation for an additional 15
min. The same buffer with or without equimolar amounts of EF-G and GTP (2mM final
concentration) was added and incubation was continued for a further 10 min at 37 °C. In
7
by guest on June 3, 2018http://w
ww
.jbc.org/D
ownloaded from
parallel, the binding of AcPhe-tRNAPhe directly to the P-site was done the same way, but
without prior incubation of the ribosomes with the deacylated tRNA. Binding of AcPhe-
tRNAPhe to the ribosomal P or A sites was measured by filtration through nitrocellulose. The
puromycin reaction was performed at 37 °C for 10 min with a 5mM final concentration of
puromycin adjusted to pH 7.5. The radioactive aminoacid released was extracted with ethyl
acetate. The background of the puromycin reaction was measured in a parallel experiment
without puromycin.
Results
Choice of affinity tag
For the introduction of an affinity tag into E. coli 23 S rRNA we chose the internal 23S rRNA
helices 9, 25, 45 and 98, which are evolutionary variable and whose extension is known not to
affect protein biosynthesis (16). Several affinity tags were tested and the best results were
obtained for the streptavidin-binding aptamer. Other tags that we tested in preliminary
experiments were the MS2 coat protein binding site, and the aptamer to streptomycin.
Introduction of the MS2 coat protein binding site into rRNA has been previously described. A
recombinant fusion protein consisting of the MS2 coat protein dimer (11), a TEV-protease
cleavage site and a dimer of the Z domains of S. aureus protein A was created for the affinity
purification. Using this system, we were able to observe a quantitative binding of the fusion
protein to the affinity-tagged ribosomes and, independently, to the IgG Sepharose (data not
shown). However, the binding of ribosomes carrying the MS2 coat protein binding site to the
resin, mediated by the recombinant protein, was extremely low and relatively nonspecific (data
not shown). The reason was most probably in steric clash between ribosomes, protein and the
8
by guest on June 3, 2018http://w
ww
.jbc.org/D
ownloaded from
resin when all three are to bind each other simultaneously. The successfully applied MS2 coat
protein based systems for affinity separation of RNA and it’s complexes obviously involved
ribonucleoprotein particles of the much smaller size. We also cannot rule out that more flexible
and long linker between MS2 coat protein part and IgG Sepharose binding part of the fusion
protein could potentially enable the binding of tagged ribosomes to the resin.
The next affinity tag to be tested was the streptomycin-binding aptamer (12), which was
introduced into helix 98 of the 23 S rRNA. Purification of the ribosomes carrying this aptamer
on immobilized dihidrostreptomycin gave a very high level of co-purification of wild-type
ribosomes (data not shown). This might be explained by binding of the streptomycin derivative
to the 30S ribosomal subunit or by nonspecific electrostatic interactions between the
dihydrostreptomycin and rRNA. The latter is more probable, since the application of the 50S
subunits instead of the 70S ribosomes for the binding doesn’t increase the specificity of binding.
Finally, an aptamer to streptavidin (13) was introduced into the helices 9, 25 and 45 of the 23 S
rRNA. Strains (4) expressing only the tagged variant of 23 S rRNA had no growth defects in
rich medium, indicating that tagged ribosomes could support translation equally to the wild type
ones. Pure tagged ribosomes bound streptavidin Sepharose and could be eluted by biotin in a
similar manner as that described for yeast RNase P (13). The best binding was observed when
the aptamer was introduced into helix 25 (Fig. 1). In the equal conditions binding of other
variants of tagged ribosomes was 70% of the binding of the ribosomes carrying the aptamer
inserted into the helix 25. Under equal conditions, the ribosomes without the introduced aptamer
were not able to bind the resin to any detectable extent.
Affinity separation of ribosomes, carrying the lethal G2655C mutation
Plasmid pLK1192U, encoding 23S rRNA carrying both the streptavidin aptamer and the
G2655C mutation, was expressed in strain XL1 of E. coli. The plasmid used carries ampicillin
9
by guest on June 3, 2018http://w
ww
.jbc.org/D
ownloaded from
resistance gene and C1192U mutation in 16S rRNA, making ribosomes resistant to
spectinomycin. 70S ribosomes from this strain were prepared by the reassociation technique
(21). Affinity purification of tagged ribosomes by streptavidin Sepharose at 10mM magnesium
concentration lead to only 5 fold enrichment of initial mixture. Since binding of pure wild type
ribosomes (not as a mixture with tagged ones) to the resin was not observed, we concluded that
non-specific binding of non-tagged species arises from ribosome-ribosome rather then
ribosome-resin interactions. To avoid this effect we purified 50S subunits at a magnesium
concentration of 1mM, which led to a satisfactory separation; the initial mixture contained only
6% tagged ribosomes, which was enriched to a level of 90±3% in the final eluate from the resin
(Fig. 1). The affinity chromatography step was followed by reassociation of the subunits and
further purification of the 70S ribosomes by ultracentrifugation in a sucrose density gradient as
described by Blaha et al. (21).
The G2655C mutation affects cell growth and accuracy of translation
In agreement with previous reports (15), we found that G2655C mutation is recessive lethal. It
was not able to support cell growth if no wild type ribosomes were present in the cell, however
it could be tolerated if the translation is carried out by wild type ribosomes. In a strict sense the
mutation did not allow AVS69009 strain (4) to lose the plasmid encoding the wild type 23S
rRNA and remain viable. The study was extended to other substitutions of the 2655 nucleotide,
namely G2655A and G2655U. In contrast to G2655C these mutations were found to be viable
and were able to support growth, even when they represented 100% of the 23 S rRNA in
AVS69009 cells in which all the chromosomal rDNA operons are deleted. The doubling times
of the corresponding strains are listed in Table I. In agreement with previous reports (25), the
G2655A mutation showed little difference from the wild type, G2655C was the most defective,
and G2655U was intermediate.
10
by guest on June 3, 2018http://w
ww
.jbc.org/D
ownloaded from
The effect of the mutations at 23S rRNA position 2655 on the in vivo accuracy of translation
was tested with the help of a set of plasmids expressing the β-galactosidase gene (19). Each
plasmid of this set (except pSG25, used as a control) contain a mutation in lacZ gene which
would lead to non-functional protein if the translation of correspondent mRNA proceed without
mistakes. The active enzyme could only be synthesized as a result of translation error, such as
frameshifting, stop codon readthrough and misreading, compensating the mutation, encoded in
the DNA. Stop codon readthrough and frameshifting was increased in the strains expressing
mutant rDNA (Table II). The extent of the increase correlated well with the growth rates, being
minimal for G2655A and maximal for the lethal G2655C mutation. Curiously, the mutations
also increased the expression of the normal β-galactosidase gene, an effect that has been
described in the literature for several other mutations and explained by an increase of overall
wild type ribosome production in response to the expression of mutant rDNA genes (26).
The interesting observation is the decrease in Glu/Gln misreading in strains expressing the
G2655C and G2655U mutants (Table II). Altogether these effects could be explained if we
suppose that the G2655C mutant ribosomes move very slowly along the mRNAs. In this case,
the downstream ribosomes in the polysomes would be forced to move as slowly as the mutant
ones, thus gaining additional time for correct tRNA selection as well as for shifting the reading
frame.
Ribosomes, carrying the G2655C mutation stimulate GTPase of EF-G
The experiments were carried out with ribosomes, purified by affinity chromatography, carrying
the G2655C mutation. Ribosomes, purified by the same technique, with a wild-type 2655G
nucleotide were used as a control.
The multiple EF-G turnover reaction was measured at a ribosome:EF-G:GTP ratio 1:1:100. It
was shown that ribosomes carrying the G2655C mutation have a slightly slower GTPase
11
by guest on June 3, 2018http://w
ww
.jbc.org/D
ownloaded from
turnover rate (Fig. 2). However, in contrast to our expectations, it was found that the mutant
ribosomes do in fact stimulate the GTPase activity of elongation factor G. According to the
experiments of Munishkin & Wool (25), a sarcin-ricin loop oligonucleotide carrying the
G2655C mutation does not bind EF-G. We decided to check if the same is true in the context of
the ribosome. It is known that fusidic acid stabilizes the complex of EF-G with the ribosome,
where EF-G forms a contact with the SRL (14, 27). We checked the inhibition by fusidic acid
of GTPase turnover on the mutant ribosomes and found it to be the same as for the wild-type
ribosomes. Since fusidic acid prevents dissociation of EF-G and the ribosome this inhibition
proves that EF-G binds to the ribosome. This might indicate that an interaction of EF-G with
other parts of the ribosome, most probably with the GTPase-associated center, could overcome
the effect of the mutation and be sufficient for EF-G binding.
EF-G contacts GTPase-associated center of ribosomes carrying the G2655C mutation, while
the interaction with the sarcin-ricin loop is reduced
Stimulation of the GTPase activity already presumes that EF-G binds to the mutant
ribosomes. However, in order to demonstrate EF-G binding directly and to see if it contacts
both the SRL and the GTPase-associated center, we checked the protection by EF-G of 23S
rRNA residues from dimethylsulfate (DMS) modification, with and without fusidic acid.
We found that ribosomes carrying the G2655C mutation could bind EF-G (Fig. 3),
although the extent of binding is reduced in comparison with the binding to wild-type
ribosomes. We conclude, that EF-G contacts both the SRL and the GTPase associated center of
the mutant ribosomes, since both A1067 (in the GTPase associated center) and A2660 (in the
SRL) were protected by EF-G from DMS modification. Notably, protection of A1067 in both
the wild type and mutant ribosomes is almost the same, while the extent of A2660 protection
varied somewhat from one experiment to another, being on average 80% for the wild type and
12
by guest on June 3, 2018http://w
ww
.jbc.org/D
ownloaded from
31% for the mutant in the case of the complex with the fusidic acid. Without the antibiotic, EF-
G related protections on both the wild type and mutant ribosomes were reduced in agreement
with the literature data (14). It was not possible to compare the reactivity of nucleotide 2655 in
mutant vs. wild type ribosomes, since only mutant is reactive to DMS. However, it is
noteworthy that, whereas G2655 of wild-type ribosomes is protected by EF-G from
modification with kethoxal (14), C2655 of the mutant ribosomes is not protected from DMS
modification.
Ribosomes carrying the G2655C mutation are defective in EF-G-driven translocation
The most intriguing question to ask was whether EF-G could cause the translocation of tRNAs
on the mutant ribosomes. In order to test this we prepared a pre-translocation complex on
mutant polyU - programmed ribosomes, with the P-site filled by deacylated tRNAPhe and the
A-site by AcPhe-tRNAPhe. In parallel, AcPhe-tRNAPhe was directly mixed with polyU-
programmed ribosomes, so as to observe its maximal level of binding to the P-site, without
prior incubation with deacylated tRNAPhe. The proportion of P-site bound AcPhe-tRNAPhe
was measured by the puromycin reactivity of the AcPhe moiety related to the puromycin
reactivity of AcPhe-tRNAPhe bound directly to the P-site. The experiments were done
according to the modified Watanabe system (28, 21) in polyamine-containing buffer. As
expected, no detectable difference between G2655C and wild type ribosomes with regard to P
and A-site tRNA binding was observed. Binding of AcPhe-tRNAPhe to the P-site of polyU
programmed ribosomes was on average 98%, while the binding to the A-site of ribosomes,
carrying deacylated tRNAPhe in the P-site was 50%; both values being identical for wild-type
and G2655C ribosomes. Translocation, both spontaneous and EF-G related, was measured by
the puromycin reaction (Fig. 4) with and without addition of EF-G. No difference in 13
by guest on June 3, 2018http://w
ww
.jbc.org/D
ownloaded from
spontaneous translocation was detected, but a clear defect (circa 3-fold difference) in EF-G
driven translocation was observed for the G2655C ribosomes. Under the same incubation
conditions a 100-fold excess of GTP is completely hydrolysed by EF-G, stimulated by
ribosomes carrying the G2655C mutation. Thus, the principal result is that the G2655C mutation
moderately affects GTPase stimulation and binding of EF-G to the ribosome, but significantly
decreases translocation efficiency.
Discussion
The main goal of the affinity purification method described here was to make it possible to
obtain pure in vivo assembled ribosomes, carrying a lethal mutation. Three aptamer structures
were tested as affinity tags, introduced to the unconserved helices of the 23S rRNA. MS2 coat
protein binding site inserted into the 23S rRNA helices 9, 25, 45 and 98 was shown to be able to
direct specific binding of the MS2 coat protein to the tagged ribosomes. Unfortunately, the
attempt to bind such ribosomes to the resin via the fusion protein consisted with MS2 coat
protein and IgG binding domain failed. Among the possible explanations are steric clashes
between three components of the complex.
Streptomycin binding aptamer was introduced to the helix 98 of the 23S rRNA. It was shown
not to be suitable for the affinity preparation of mutant ribosomes due to the high background
binding of wild type ribosomes to the immobilized dihidrostreptomycin. Such a low specificity
might be explained by electrostatic binding of large ribosomal RNA to the positively charged
streptomycin.
An aptamer to streptavidin inserted to the helix 25 of the 23S rRNA performed the best as an
anchor to the mutant ribosomes. It was demonstrated that only limitation to the purity of the
mutant ribosomes, separated by this technique of affinity chromatography resides in ribosome-
ribosome interactions. Application of the low magnesium buffer, reducing unspecific
14
by guest on June 3, 2018http://w
ww
.jbc.org/D
ownloaded from
interactions between ribonucleoprotein particles allowed us to separate tagged 50S ribosomal
subunits sufficiently pure for the biochemical analysis in vitro.
The G2655C mutation in 23S rRNA was selected to test the method. Nucleotide G2655 is
protected by both EF-G and EF-Tu from chemical modification (14). This nucleotide is not
involved in extensive intraribosomal interactions (1), so that its mutation would not be expected
to cause any conformational changes outside the elongation factor binding site. The mutation
G2655C was first described for the sarcin-ricin loop (SRL) oligonucleotide (25) and was shown
to impede EF-G binding to this RNA fragment. Later, Macbeth & Wool (15) showed that
expression of 23S rRNA carrying this mutation, from highly efficient P1P2 rDNA promoters,
causes cell lethality. The mutant ribosomes were shown to be no more than 3% active in protein
synthesis (15). Mutations of other SRL nucleotides have defects in EF-G binding (29), EF-
Tu/AatRNA/GTP ternary complex binding (30), as well as effects on translation fidelity (31,
32). The mutations described here also affect the fidelity of translation, and they do so in
opposite directions; stop codon readthrough and frameshifting are increased, whereas
misreading is decreased. Both effects could be explained if one postulates a slower movement of
the mutant ribosomes along an mRNA molecule. An overall decrease in translation speed would
result in a wider time window for correct tRNA choice, as well as for frameshifting.
Despite the large body of experimental data, the mystery of the function of the elongation
factors on the ribosome is still unsolved. One of the most intriguing questions is the coupling of
GTP hydrolysis with translocation. In the classical view (33, 34), translocation takes place
before GTP hydrolysis, the latter being only necessary to dissociate the factor. A more recent,
but often disputed view supports GTPase activation prior to translocation (35). Additional
support for this alternative hypothesis has come from biochemical (36) and cryo-EM (37)
studies, but has, however, been criticized by Cameron et al. (38).
Here, we present an unique analysis of lethal mutant (G2655C) ribosomes purified by affinity
15
by guest on June 3, 2018http://w
ww
.jbc.org/D
ownloaded from
chromatography, showing GTPase activity and EF-G binding accompanied by significantly
inhibited translocation. The binding of EF-G to the GTPase associated center was the same for
wild-type and the G2655C ribosomes, however the contact of EF-G with SRL of the mutant
ribosomes was weaker. This result could be explained in several ways. In the context of the
model of Rodnina et al. (35), our results indicate that EF-G is in contact with nucleotide 2655
after the GTPase reaction, but prior to translocation, as shown by the cryo-EM study (37). Thus,
translocation is not necessary for GTPase activation, simply because we observe GTPase
activity without translocation. However, the results obtained could also be explained in terms of
the "classical" model, which supposed a mechanism whereby GTP hydrolysis is activated only
after the translocation has taken place. Thus, it could be hypothesized that the mutation in
residue 2655 allows the necessity for translocation to activate the GTPase activity to be
"bypassed". In other words the mutation could make GTPase and translocation uncoupled, in the
fashion it happens with EF-G with deleted domain IV (39). It is also possible that the G2655C
mutation could impede an allosteric signal transition from the sarcin-ricin loop to the
peptidyltransferase center, thus making it impossible for the CCA ends of the tRNA molecules
to shift. Both explanations converge at the point that interaction between EF-G and SRL is not
primarily essential for the activation of GTP hydrolysis, however it is important for
translocation.
The method described for affinity chromatography of ribosomes will undoubtedly help to
explore more lethal mutations, giving insight into other critical steps in the ribosomal function.
Acknowledgements
Authors are very thankful to Dr. D.S. Peabody, Dr. T. Ueda and Dr. C. Squires for providing us
with their strains and plasmids. We would like to thank especially Albert Dahlberg, Michael
16
by guest on June 3, 2018http://w
ww
.jbc.org/D
ownloaded from
O’Connor and other members of their laboratory not only for their strains and plasmids, but also
for their inspiring discussions. Authors are thankful to Shura Mankin for his comments on the
manuscript and to Maria Rubtsova for her help.
The work was supported by grant #55000303 from Howard Hughes Medical Institute, RFBR
#02-04-48786 and RFBR-DFG #02-04-04001.
References
1. Ban, N., Nissen, P., Hansen, J., Moore, P.B., and Steitz, T.A. (2000) Science, 289, 905-920.
2. Schluenzen, F., Tocilj, A., Zarivach, R., Harms, J., Gluehmann, M., Janell, D., Bashan, A.,
Bartels, H., Agmon, I., Franceschi, F., and Yonath, A. (2000) Cell, 102, 615-623.
3. Wimberly, B.T., Brodersen, D.E., Clemons, W.M.Jr,, Morgan-Warren, R.J., Carter, A.P.,
Vonrhein, C., Hartsch, T., and Ramakrishnan, V. (2000) Nature, 407, 327-339.
4. Asai, T., Zaporojets, D., Squires, C., and Squires, C.L. (1999) Proc. Natl. Acad. Sci. USA, 96,
1971-1976.
5. Sergiev, P.V., Bogdanov, A.A., Dahlberg, A.E., and Dontsova, O. (2000) J. Mol.Biol, 299,
379-389.
6. Vila-Sanjurjo, A., Squires, C.L., and Dahlberg, A.E. (1999) J. Mol. Biol., 293, 1-8.
7. Samaha, R.R., Green, R., and Noller, H.F. (1995) Nature, 377, 309-314.
17
by guest on June 3, 2018http://w
ww
.jbc.org/D
ownloaded from
8. Green, R., and Noller, H.F. (1996) RNA, 2, 1011-1021
9. Khaitovich, P., Tenson, T., Kloss, P., and Mankin, A.S. (1999) Biochemistry, 38, 1780-1788.
10. Green, R., and Noller, H.F. (1999) Biochemistry, 38, 1772-1779.
11. Peabody, D.S., and Lim, F. (1996) Nucleic Acids Res., 24, 2352-2359.
12. Bachler, M., Schroeder, R., and von Ahsen, U. (1999) RNA, 5, 1509-1516.
13. Srisawat, C., and Engelke, D.R. (2001) RNA, 7, 632-641.
14. Moazed, D., Robertson, J.M., and Noller, H.F. (1988) Nature, 334, 362-364.
15. Macbeth, M.R., and Wool, I.G. (1999) J. Mol. Biol, 285, 965-975.
16. Matadeen, R., Sergiev, P., Leonov, A., Pape, T., van der Sluis, E., Mueller, F., Osswald, M.,
von Knoblauch, K., Brimacombe, R., Bogdanov, A., van Heel, M., and Dontsova, O. (2001) J.
Mol. Biol, 307, 1341-1349.
17. Avdeeva, O.N., Myasnikov, A.G., Sergiev, P.V., Bogdanov, A.A., Brimacombe, R., and
Dontsova, O.A. (2002) FEBS Lett., 514, 70-73.
18. Powers, T., and Noller, H.F. (1990) Proc. Natl. Acad. Sci. USA, 87, 1042-1046.
18
by guest on June 3, 2018http://w
ww
.jbc.org/D
ownloaded from
19. O’Connor, M., and Dahlberg, A.E. (1993) Proc. Natl. Acad. Sci. USA, 90, 9214-9218.
20. O’Connor, M., Goringer, H.U., and Dahlberg, A.E. (1992) Nucleic Acids Res., 20, 4221-
4227.
21. Blaha, G., Stelzl, U., Spahn, C.M., Agrawal,R.K., Frank, J., and Nierhaus, K.H. (2000)
Methods Enzymol., 317, 292-309.
22. Sigmund, C.D., Ettayebi, M., Borden, A., and Morgan, E.A. (1988) Methods Enzymol., 164,
673-690.
23. Ehrenberg, M., Bilgin, N., and Kurland, C.G. (1990) In Spedding G. (ed.) Ribosomes am
Protein Synthesis: A Practical Approach. Oxford University Press, Oxford, UK, pp. 101-129.
24. Shimizu, Y., Inoue, A., Tomari, Y., Suzuki, T., Yokogawa, T., Nishikawa, K., and Ueda, T.
(2001) Nat. Biotechnol., 19, 751-755.
25. Munishkin, A., and Wool, I.G. (1997) Proc. Natl Acad. Sci. USA, 94, 12280-12284.
26. Moine, H., and Dahlberg, A.E. (1994) J. Mol. Biol., 243, 402-412.
27. Agrawal, R.K., Penczek, P., Grassucci, R.A., and Frank J. (1998) Proc. Natl. Acad. Sci.
USA, 95, 6134-6138.
28. Watanabe, S. (1972) J. Mol. Biol., 67, 443-457.
19
by guest on June 3, 2018http://w
ww
.jbc.org/D
ownloaded from
29. Macbeth, M.R., and Wool, I.G. (1999) J. Mol. Biol., 285, 567-580.
30. Tapprich, W.E., and Dahlberg, A.E. (1990) EMBOJ., 9, 2649-2655.
31. Bilgin, N., and Ehrenberg, M. (1994) J. Mol Biol., 235, 813-824.
32. O’Connor, M., and Dahlberg, A.E. (1996) Nucleic Acids Res., 24, 2701-2705.
33. Glukhova, M.A., Belitsina, N.V., and Spirin, A.S. (1977) Mol Biol (Mosk). 11, 1190-1200.
34. Kaziro, Y. (1978) Biochim. Biophys. Acta, 505, 95-127.
35. Rodnina, M.V., Savelsbergh, A., Katunin, V.I., and Wintermeyer, W. (1997) Nature, 385,
37-41.
36. Rodnina, M.V., Savelsbergh, A., Matassova, N.B., Katunin, V.I., Semenkov, Y.P., and
Wintermeyer, W. (1999) Proc. Natl. Acad. Sci. USA, 96, 9586-9590.
37. Stark, H., Rodnina, M.V., Wieden, H.J., van Heel, M., and Wintermeyer, W. (2000) Cell,
100, 301-309.
38. Cameron, D.M., Thompson, J., March, P.E., and Dahlberg, AE. (2002) J. Mol. Biol, 319,
27-35.
39. Martemyanov, K.A., and Gudkov, A.T. (1999) FEBS Lett. 452, 155-159.
20
by guest on June 3, 2018http://w
ww
.jbc.org/D
ownloaded from
40. Harms, J., Schluenzen, F., Zarivach, R., Bashan, A., Gat, S., Agmon, I., Bartels, H.,
Franceschi, F., and Yonath, A. (2001) Cell, 107, 679-688.
Figure legends
Fig. 1. Affinity purification of ribosomes. A. The secondary structure of the 23S rRNA region,
carrying streptavidin binding aptamer inserted at the helix 25. B. The place of the streptavidin
binding tag in the structure of the large ribosomal subunit. Shown is the structure of the large
ribosomal subunit of D. radiodurans (40) as viewed from the solvent side. A secondary structure
element, correspondent to the E. coli 23S rRNA helix 25 is marked black. C. The principle of
the primer extension analysis of the ratio between the affinity tagged and wild type ribosomes.
D. Autoradiograph of primer extension products, separated by 16% PAGE. Primer extension
proceeded up to the first cytosine after the primer binding site on the 23 S rRNA. The "Mutant"
stop corresponds to the tagged ribosomes carrying the streptavidin-binding aptamer at 23S
rRNA helix 25, while the "WT" stop corresponds to the wild type ribosomes. The "Initial" lane
contains primer extension products of the mixture of chromosomally encoded ribosomes with
affinity tagged ribosomes. This mixture was purified from the cells and applied to streptavidin
sepharose. The "Flow through" lane contains primer extension products of those ribosomes
which did not bind to the resin. The "Wash" contains primer extension products of the
ribosomes that were washed from the resin by excess loading buffer. The "Elution" lane contains
primer extension products of the ribosomes that were eluted from the affinity resin by biotin.
Fig. 2. EF-G catalyzed GTPase reaction, stimulated by affinity-purified ribosomes. A. The
figure shows an autoradiograph of thin layer chromatography of the GTP hydrolysis products.
21
by guest on June 3, 2018http://w
ww
.jbc.org/D
ownloaded from
Positions of inorganic phosphate and unhydrolysed GTP are indicated. The hydrolysis time was
30 seconds. "-Rs". GTPase reaction, catalyzed by EF-G without ribosomes. "WT" GTPase
reaction, catalyzed by EF-G and stimulated by affinity purified ribosomes, without any
additional mutations; GTP:EF-G:ribosome ratio 100:1:1. "M" GTPase reaction, catalyzed by
EF-G and stimulated by affinity purified ribosomes carrying the G2655C mutation; GTP:EF-
G:ribosome ratio 100:1:1. "WT+ Fus" Same as WT, but in the presence of 100 µM fusidic acid.
"M+ Fus" Same as M, but in the presence of 100 µM fusidic acid. B. A diagram shows
inorganic phosphate accumulation upon GTP hydrolysis by EF-G, stimulated by wild type and
mutant ribosomes. The conditions are equal to that of "A". Counts per minute are depicted
along Y axis. X value corresponds to the time of reaction in minutes. Black squares correspond
to the wild type ribosomes, black triangles to the G2655C mutant ribosomes. Grey squares and
triangles indicate the extent of hydrolysis by wild type and mutant ribosomes in the presence of
fusidic acid. Circles correspond to the hydrolysis without the ribosomes.
Fig. 3. EF-G footprinting. Autoradiograph of 23S rRNA sequencing and primer extension
products, separated by 10% PAGE. A. Probing of the GTPase associated center. B. Probing of
the sarcin-ricin loop. Lanes, marked with A, C, G and U correspond to the 23S rRNA sequence.
The "Unmod" lane contains primer extension products of unmodified ribosomes. The "-EF-G"
lane corresponds to ribosomes without EF-G, modified with DMS. The "EF-G" and "EF-G +
Fus" lanes contain primer extension products of 23S rRNA extracted from ribosomal complexes
with EF-G in the absence or presence of fusidic acid, respectively, treated with DMS. The
"WT" group of lanes represents footprinting of affinity purified ribosomes, carrying nonmutated
G2655. The "G2655C" group corresponds to the affinity purified mutant ribosomes.
Fig. 4. Puromycin reactivity of AcPhe-tRNAPhe in functional complexes of affinity-purified 22
by guest on June 3, 2018http://w
ww
.jbc.org/D
ownloaded from
ribosomes. The Figure shows the ratio between puromycin-reactive and total bound AcPhe-
tRNAPhe. Black boxes correspond to affinity-purified ribosomes carrying the G2655C
mutation. White boxes correspond to affinity-purified ribosomes carrying the affinity tag only.
Complex 1: ribosomes, polyU, AcPhe-tRNAPhe. Complex 2: ribosomes, polyU, tRNAPhe in
the P-site, AcPhe-tRNAPhe in the A-site. Complex 3: same as 2, after addition of EF-G and
GTP.
23
by guest on June 3, 2018http://w
ww
.jbc.org/D
ownloaded from
Tables
Table I. Doubling times of strains AVS69009 and XL1, harboring pLK1192U plasmids
expressing 23S rDNA with (or without) mutations of nucleotide 2655.
AVS69009 XL1
wt 85 min 77 min
G2655A 86 min 88 min
G2655U 117 min 130 min
G2655C lethal 177 min
Table II. Effects of mutations at position 2655 of 23 S rRNA on the readthrough of stop-codons,
frameshifting and missense aminoacid incorporation.
Stop codon read through wt G2655A G2655U G2655C
pSG 853 UAA 1 1.85±0.13 3.1±0.28 3.04±0.31
pSG12-6UAG 1 0.91±0.04 1.13±0.12 1.62±0.23
pSG 3/4 UGA 1 1±0.01 1.39±0.21 1.86±0.29
Frameshifting
pSGlac7(+l) 1 1.23±0.15 1.33±0.07 2.22±0.29
pSG12DP(-1) 1 1.1±0.1 1.54±0.17 1.68±0.14
Misreading
CSH103 (Glu/Gln) 1 1.03±0.02 0.78±0.01 0.62±0.07
Total translation
pSG 25 (wt) 1 1.08±0.03 1.46±0.18 1.53±0.03
24
by guest on June 3, 2018http://w
ww
.jbc.org/D
ownloaded from
A. DontsovaAndrei A. Leonov, Petr V. Sergiev, Alexey A. Bogdanov, Richard Brimacombe and Olga
affects the translocationAffinity purification of ribosomes with a lethal G2655C mutation in 23S rRNA that
published online May 1, 2003J. Biol. Chem.
10.1074/jbc.M302873200Access the most updated version of this article at doi:
Alerts:
When a correction for this article is posted•
When this article is cited•
to choose from all of JBC's e-mail alertsClick here
by guest on June 3, 2018http://w
ww
.jbc.org/D
ownloaded from