ABSTRACT
To examine the function of the central pseudoknot in 16S rRNA, we have studied
Escherichia coli
30S subunits with the A
18
mutation in this structure element. Previously, this mutation, which changes
the central base pair of helix 2, C
18
-G
917
, to an A
18
*
G
917
mismatch, was shown to inhibit translation
in vivo
and a defect in initiation was suggested. Here, we find that the mutant 30S
particles are impaired in forming 70S tight couples and predominantly
accumulate as free 30S subunits. Formation of a 30S initiation complex, as
measured by toeprinting, was almost as efficient for mutant 30S subunits,
derived from the tight couple fraction, as for the wild-type control. However, the A
18
mutation has a profound effect on the overall stability of the subunit. The
mutant ribosomes were inactivated by affinity chromatography and high salt
treatment, due to easy loss of ribosomal proteins. Accordingly, the particles
could be reactivated by partial
in vitro
reconstitution with 30S ribosomal proteins. Mutant 30S subunits from the free
subunit fraction were already inactive upon isolation, but could also be
reactivated by reconstitution. Apparently, the inactivity in initiation of
these mutant 30S subunits is, at least in part, also due to the lack of
essential ribosomal proteins. We conclude that disruption of helix 2 of the
central pseudoknot by itself does not affect the formation of a 30S initiation
complex. We suggest that the
in vivo
translational defect of the mutant ribosomes is caused by their inability to
form 70S initiation complexes.
The central pseudoknot in 16S ribosomal RNA, first predicted by Pleij
et al.
(
1
) is a universally conserved structural element in small subunit RNAs (
2
-
5
). For
Escherichia coli
this pseudoknot is presented in Figure
1
. Models for the three-dimensional structure of 16S rRNA in the
E.coli
30S subunit predict three major domains. The 5' domain constitutes the body, the central domain is incorporated in the
platform and the 3' domain is present in the head of the subunit (
6
,
7
). The central pseudoknot structure, located in the center of the 16S rRNA
molecule, connects these three domains (
6
,
7
).
Escherichia coli
strain K5637 encodes the thermolabile [lambda] repressor (
c
I857), was constructed by Dr D. H. Miller and has been described (
14
). Cells were grown in LC medium (
15
). When appropriate, ampicillin (Sigma) was added to a final concentration of 100 mg/l. Plasmid pP
L
ASDX-Spc
R
-CATX, in this paper referred to as pASC, encodes the specialized ribosome
system. K5637 cells containing this plasmid were used as a source of
specialized 30S subunits. pASC was derived from plasmid pASDX-PSDX-hGH (
14
) and has been described (
12
). The
rrnB
operon on pASC encodes 16S rRNA with an altered ASD sequence and is under the
transcriptional control of the thermo-inducible P
L
promoter. In addition to the altered ASD sequence, the C residue at position
1192 of the 16S rRNA was changed into U, conferring resistance to spectinomycin
(
16
). Chromosomally encoded 30S subunits were derived from cells harboring pASC in
which the
Kpn
1-
Apa
1 896 bp fragment in the 16S rRNA gene was replaced by a 300 bp murine DNA
fragment. These cells do not produce specialized ribosomes.
30S subunits were isolated essentially as described before (
15
). Strain K5637 harboring pASC was used for isolating specialized 30S subunits
with a wild-type central pseudoknot in their 16S rRNA. A mutant derivative of this
plasmid was used for the isolation of specialized 30S subunits containing the A
18
mutation. 70S ribosomes and free 30S subunits were prepared by centrifuging the S30 extract through a 15-30% sucrose gradient in 10 mM Tris-acetate, pH 7.5, 4.2 mM Mg acetate, 60 mM NH
4
Cl, 0.1 mM EDTA. Fractions containing 70S tight couples or 30S free subunits
were collected and the magnesium acetate concentration was adjusted to 10 mM.
Fractions were pelleted by centrifugation for 5 h at 50 000 r.p.m. in a 50.2 Ti
rotor. To keep the isolation conditions comparable, we treated the free 30S
subunits in an identical way to the 70S tight couples. Pellets were dissolved
in 10 mM Tris-acetate, pH 7.5, 4.2 mM Mg acetate, 60 mM NH
4
Cl, 0.1 mM EDTA and dialyzed against the same buffer but with 1.2 mM Mg acetate to dissociate the couples (if present). 30S subunits were prepared by 10-30% sucrose gradient centrifugation in the same buffer. The fractions containing 30S subunits were
collected, the Mg
2+
concentration raised to 10 mM, and the 30S subunits pelleted by centrifugation
for 5 h at 50 000 rpm in a 50.2 Ti rotor. Pellets were resuspended overnight in 20 mM Tris-acetate, pH 7.5, 20 mM Mg acetate, 100 mM NH
4
Cl, 0,1 mM EDTA, 2 mM dithiothreitol (DTT). 30S subunits were reactivated by incubation for 15 min at 40oC and stored at -80oC.
The ratio of specialized to chromosomally encoded 30S subunits in the various
fractions was determined by primer extension on 16S rRNA (
17
,
18
). The procedure exploits the C
1192
-> U
1192
base substitution in specialized 16S rRNA (
12
).
CAT mRNA containing the specialized SD sequence 5' GUGUG was synthesized
in vitro
, using SP6 RNA polymerase (Pharmacia). Plasmid pGEMCAT-SDX (
15
), containing the specialized
cat
gene under control of the SP6 promoter, was linearized by restriction in the
Bam
HI site, situated about 750 bp downstream of the
cat
gene start.
In vitro
transcription was performed as described by Krieg and Mellon (
19
) but the DNase treatment was omitted. CAT mRNA was purified by phenol
extraction and column chromatography through Sephadex G50, precipitated with
ethanol and dissolved in H
2
O to a concentration of 0.5 [mu]g/[mu]l. Primer extension inhibition (toeprinting) was performed essentially as described by Hartz
et al
. (
20
). The primer, 5' GCAACTGACTGAAATGCCTC, complementary to residues 64-83 downstream of the
cat
gene startcodon, was end-labelled with [[gamma]-
32
P]ATP. Toeprinting reactions were performed in standard buffer (10 mM Tris-acetate, pH 7.5, 10 mM Mg acetate, 60 mM NH
4
Cl, 0.1 mM EDTA). Reaction mixtures contained 150 nM primer, 20 nM CAT mRNA, 30S
subunits as specified in Results, 1 [mu]M tRNA
f
Met
(Boehringer-Mannheim) and 0.5 U/[mu]l RNase inhibitor (RNAguard, Pharmacia). After incubation for 7 min
at 37oC, to allow the formation of initiation complexes, and addition of 0.05 U/[mu]l AMV reverse transcriptase (Promega), primer extension was performed
for 15 min at 37oC. Extension products were separated on an 8% polyacrylamide/8 M urea gel.
The relative toeprinting efficiency was determined by measuring the
radioactivity of the inhibited and uninhibited extension products in the gels,
using a Betascope 603 Blot Analyzer (Betagen), and defined as the ratio between
inhibited extension and total primer extension (
21
).
30S ribosomal proteins were extracted from chromosomally encoded 30S subunits
with LiCl and urea, as described by Leboy
et al.
(
22
). The extract was dialyzed against buffer A (30 mM Tris-acetate, pH 7.5, 20 mM Mg acetate, 500 mM NH
4
Cl, 2 mM DTT). The molar amount of ribosomal proteins in the extract was
according to the molar amount of 30S subunits, used for the extraction.
Reconstitution was performed by adding 0.1 vol buffer A, containing different
concentrations of ribosomal protein extract, to the complete toeprint reaction
mixture (see above), except for reverse transcriptase. The final concentrations
in the reconstitution-toeprint buffer were 13 mM Tris-acetate, pH 7.5, 12 mM Mg acetate, 110 mM NH
4
Cl, 0.1 mM EDTA, 0.2 mM DTT. The mixture was incubated for 10 min at 37oC to allow reconstitution and formation of the 30S initiation complex.
Then, reverse transcriptase was added and primer extension was performed as described above.
The oligodeoxynucleotides 5' TAAGGAGGT and 5' TAAGTGTGT, complementary to the 3' end of chromosomally encoded and specialized 16S rRNA,
respectively, were prepared on a Gene Assembler (Pharmacia) using
phosphoroamidite chemistry. The oligonucleotides were passed through Sephadex
G25 (medium) and 5'-end labelled with [[gamma]-
32
P]ATP. Excess [[gamma]-
32
P]ATP was removed by filtration through Sephadex G25 (medium). For hybridization
to 16S rRNA, 100 pmol of oligonucleotides were added to 25 pmol of 30S subunits
in standard buffer (see above) in a total volume of 50 [mu]l. After incubation for 20 min on ice, 10 min at 37oC and 10 min on ice, samples were filtered through presoaked
nitrocellulose (Schleicher & Schuell, BA 85, 0.45 [mu]m). Filters were washed twice with 1 ml ice-cold standard buffer, air dried, and counted in scintillation fluid
(Carboluma Lumac-LSC) using a Beckmann LS 5000TD scintillation counter.
We used K5637 cells harboring plasmid pASC as a source for specialized ribosomes
with a mutant or a wild-type central pseudoknot. For simplicity, specialized ribosomes harboring
the wild-type or the mutant pseudoknot will be referred to in this paper as wild-type or mutant ribosomes, respectively. The real wild-type ribosomes, encoded by the chromosome, will be called
chromosomally encoded ribosomes.
Cells were harvested 2 h after induction of specialized ribosome synthesis and the S30 extracts were prepared and layered on sucrose gradients
containing 4.2 mM Mg acetate. At this magnesium concentration, active 30S
subunits stay associated with the 50S subunit to form 70S tight couples. The
profiles obtained after centrifugation are shown in Figure
2
. Cells expressing ribosomes with the A
18
mutation showed a profile with very prominent free subunits peaks. Control
cells, expressing 30S subunits with the wild-type central pseudoknot showed a dominant peak of 70S tight couples, while
free 30S and 50S subunits were almost absent.
We measured the efficiency of 30S initiation complex formation on CAT mRNA,
containing a specialized SD sequence, using the toeprinting method (
20
). Figure
3
shows that formation of a ternary complex on the CAT mRNA is strictly tRNA
f
Met
dependent (lanes 1 and 2) and specific for specialized ribosomes (lanes 2 and
3). We calculated that specialized 30S subunits bind 20-30 times better to the CAT messenger than chromosomally encoded 30S
subunits (data not shown).
Accessibility of the 3' end of the 16S rRNA is obligatory for 30S initiation complex formation
on the mRNA. 30S subunits inactivated in ternary complex formation by low
magnesium treatment (
25
) or by the lack of ribosomal protein S21 (
26
) showed no oligonucleotide binding to their 16S rRNA 3' end (
27
).
Here, we tested the binding of a nonamer complementary to the 3' end of specialized 16S rRNA. The labelled oligonucleotide was added in
four-fold molar excess over 30S subunits that contain the wild-type or the mutant central pseudoknot. Bound nonamer was measured in
a nitrocellulose filter binding assay. As shown in Table
3
, mutant 30S subunits isolated from the free subunit or 70S tight couple
fraction, A
18
/30S and A
18
/70S, respectively, bound less oligonucleotide than wild-type 30S subunits, wt/70S. The oligonucleotide did not bind to
chromosomally encoded 30S subunits and the counts therefore did not need to be
corrected for the variable amounts of these 30S particles present in the
samples.
Table 3
As a control that the isolation procedure by itself was not harmful for the
quality of the ribosomes, we tested the binding of an oligonucleotide complementary to the 3' end of chromosomally encoded 16S rRNA. This nonamer bound stoichiometrically to the
chromosomally encoded 30S subunits in all of the tested fractions (data not
shown), indicating that the poor oligo binding to the mutant 30S subunits was
not due to the isolation procedure. We suspect loss of ribosomal proteins from
mutant ribosomes on the nitrocellulose filter and the consequent release of the
oligonucleotide.
We have investigated the activity of ribosomes in which helix 2 of the central
pseudoknot in 16S rRNA was disrupted by a mutation changing the central base
pair of this helix into a mismatch. This C
18
A mutation was shown by Brink
et al.
(
12
) to inhibit translation
in vivo
. By sucrose gradient analysis, these authors showed that the mutant 30S
subunits accumulate in the 30S fraction and do not form polysome complexes.
Therefore, a defect in translation initiation was suggested. Taking this into
account, we concentrated our research on the formation of the 30S initiation
complex.
We observed that mutant 30S subunits, when derived from the 70S tight couple
fraction, have an efficiency in initiation complex formation that is almost as
high as the control. Complementarity of the central basepair in helix 2 is
therefore, although obligatory for completing the initiation process
in vivo
(
12
), not essential for performing the first step in this process
in vitro
i.e. specific and efficient binding to the ribosome binding site on the mRNA
and facilitating codon-anticodon interaction in the P site of the 30S subunit (
28
).
The strong preference of the specialized CAT mRNA for specialized 30S compared
to chromosomally encoded 30S in the toeprint assay shows the importance of the
SD interaction for the formation of an initiation complex. The accessibility of
the ASD sequence in the 30S subunit can be tested by the binding of a
complementary oligonucleotide (
25
-
27
). We showed that mutant 30S subunits from the tight couple fraction were
inefficient in oligo-binding, implying an unavailable ASD sequence. On the other hand we had
measured efficient 30S initiation complex formation. This paradox needed an
explanation.
A possible clue came with the inactivation of mutant particles by the affinity
chromatography treatment. This method can potentially be used to prepare a
homogeneous fraction of specialized 30S subunits without loss of activity in
initiation complex formation (
15
). We showed here that the inactivation of the mutant 30S subunits by the column
was probably due to the loss of ribosomal proteins since we could reactivate
the particles by adding a total 30S ribosomal protein extract. In the oligo-binding assay, we also diluted the 30S fractions during the washing step of the
nitrocellulose filtration procedure. We suspect that this dilution step,
possibly in combination with the adsorption to the filter surface, is detrimental to the mutant particles. Understandably, in the toeprint assay, where such a step is absent, we found no decrease in
activity.
The observation that the loss of ribosomal proteins inactivates a (mutant)
ribosome, implies that these proteins are important for initiation complex
formation and oligo binding to the 16S rRNA 3' end. For the mutant 30S from the free subunit fraction, we analyzed the
ribosomal protein content and found a decreased presence of S1, S2, S18 and S21
(
29
). S1 and S21 are essential for initiation complex formation (
26
,
30
) and S21 is also important for an accessible 3' end (
27
). The reduced affinity in the mutant 30S subunits for S1 and S21 may therefore
account for their conditional inactivity.
The mutant 30S active in forming an initiation complex were derived from 70S
tight couples and represented 20% of the total mutant population. This number
is significantly higher than the almost negligible amount of ribosomes found in
the polysomal fractions by Brink
et al.
(
12
). A similar phenomenon was observed in the study of another pseudoknot in
E.coli
16S rRNA, formed by the interaction between nucleotides G
570
U
571
and C
866
A
865
(
11
). In the translation incompetent mutant A
571
, formation of polysomes was severely inhibited while tight couple formation was
not disturbed. Unlike polysomes, 70S tight couples are not supposed to contain
mRNA or tRNA. The similar behavior of these two translationally inactive
mutants suggests more stringent demands on 30S subunits for incorporation into
a programmed 70S complex than into a tight couple.
In the group of Brakier-Gingras, the implications for translational activity of substitutions U
13
A, A
914
U, and the double mutation were examined
in vivo
and
in vitro
(
31
-
33
). These pseudoknot mutations (for their positions, see Fig.
1
) impaired growth when cells depended on the mutant ribosomes (
33
). In toeprint experiments, using a 50/50 mix of mutant and wild-type 30S subunits, a 25% decrease in 30S initiation complex formation was
observed. Also, the 915-921 region in the mutant 16S rRNA showed a modest increase in
accessibility for oligonucleotide binding and for modification of G
917
by kethoxal. Since helix 2 is formed by basepairing to a part of this region,
the authors suggest that an undisrupted central pseudoknot is necessary for
ternary complex formation (
33
). However, the incorporation of mutant ribosomes into polysomes was still 60% of the incorporation observed in the wild-type control (
32
). If the major defect
in vivo
in these mutant 30S subunits is in initiation, one would expect their presence
to be less prominent in the polysomal complexes. Here, we studied 30S subunits
with a more seriously disrupted helix 2. Translational activity is absent and
the mutant ribosomes do not form polysomes (
12
). Nevertheless, the mutant 30S are almost fully capable of forming a 30S
initiation complex, indicating that disruption of helix 2 causes a defect other
than ternary complex formation.
A
18
mutant 30S subunits do not form programmed 70S complexes and were therefore
suggested to have a defect in translation-initiation (
12
). Given our finding that 30S initiation complex formation is not affected, the
most obvious translational defect would be the association of this complex with
the 50S subunit.
Baudin
et al.
(
13
) showed that in 16S rRNA the 770-930 region together with nucleotides 19 and 20 were protected from
chemical modification upon association with the 50S subunit. This is in
agreement with structure models for the 30S subunit (
6
,
7
) in which the central pseudoknot region is located at the interface between the
30S and 50S subunit. Disruption of the central basepair of helix 2 might
therefore interfere with the interaction between the subunits. The instability
of the mutant particles, resulting in an easy loss of several ribosomal
proteins predicts also an important role of this pseudoknot in the overall
architecture of the 30S subunit. The proposed position of the central
pseudoknot at the junction of the three major domains in the 16S rRNA (
6
,
7
) would agree with such a structural function.
Sequence/ fraction
a
% 30S active in oligo binding
b
chrom/70S
c
0
wt/70S
69 +- 4
A
18
/70S
16 +- 4
A
18
/30S
17 +- 2
REFERENCES
Return