ABSTRACT
Photolysis of [
3
H]tetracycline in the presence of
Escherichia coli
ribosomes results in an approximately 1:1 ratio of labelling ribosomal proteins and RNAs. In this work we
characterize crosslinks to both 16S and 23S RNAs. Previously, the main target
of photoincorporation of [
3
H]tetracycline into ribosomal proteins was shown to be S7, which is also part of
the one strong binding site of tetracycline on the 30S subunit. The crosslinks
on 23S RNA map exclusively to the central loop of domain V (G2505, G2576 and
G2608) which is part of the peptidyl transferase region. However, experiments
performed with chimeric ribosomal subunits demonstrate that peptidyltransferase
activity is not affected by tetracycline crosslinked solely to the 50S
subunits. Three different positions are labelled on the 16S RNA, G693, G1300
and G1338. The positions of these crosslinked nucleotides correlate well with
footprints on the 16S RNA produced either by tRNA or the protein S7. This
suggests that the nucleotides are labelled by tetracycline bound to the strong
binding site on the 30S subunit. In addition, our results demonstrate that the
well known inhibition of tRNA binding to the A-site is solely due to tetracycline crosslinked to 30S subunits and
furthermore suggest that interactions of the antibiotic with 16S RNA might be
involved in its mode of action.
The antibiotic tetracycline inhibits binding of tRNA to ribosomes (
1
). Specifically, it mainly influences binding to the A-site although some effects on the binding constant of Ac-Phe-tRNA to the P-site have also been observed (
2
,
3
). Tetracycline binds to a single strong binding site on the 30S ribosomal
subunit as well as to a number of weaker sites on both, the 30S and 50S
subunits (
2
,
4
-
6
). The precise mechanism of tetracycline inhibition is not known, but it is
generally assumed that inhibition is caused by binding of tetracycline to the
strong binding site on the 30S subunit (
2
,
5
,
6
). In a series of experiments where single proteins were omitted from the 30S
subunit it has been established that the high affinity site is dependent on the
presence of 16S RNA and the proteins S3, S7, S8, S14 and S19 (
7
). Of these proteins, S7 was the main target in experiments using [
3
H]tetracycline as a photoaffinity reagent (
5
).
In previous experiments, we have used a photoreactive benzophenone derivative of tRNA [3-(4'-benzoylphenyl)propionylphenylalanine transfer RNA (BP-Phe-tRNA)] to characterize the peptidyltransferase region on the 50S subunit (
8
,
9
). The photoreaction with the 23S RNA was completely inhibited by tetracycline, and tetracycline
itself crosslinked efficiently to the loop V region of 23S RNA (
9
,
10
). This was somewhat surprising since the data from several investigations suggested that tetracycline might incorporate mainly into ribosomal proteins (
5
,
11
,
12
). We therefore undertook a thorough analysis of the photoincorporation of tetracycline into ribosomal RNAs under conditions optimized to avoid non-specific binding and labelling due to tetracycline photoproducts. We show
that tetracycline can be crosslinked to 16S RNA as well as to 23S RNA but not
to 5S RNA. Activity data from the crosslinked subunits show that the inhibitory
effect results solely from the interaction of tetracycline with the small
subunit. This suggests that tetracycline crosslinks to 16S RNA from the strong
binding site and that it might act via interaction with the 16S RNA.
Tetracycline hydrocloride was purchased from Sigma, highly purified tetracycline
was a present from Dr George Ellestad (Wyeth-Ayerst, Pearl River, NY). [
3
H]Tetracycline was purchased from New England Nuclear (0.5 mCi/[mu]mol). All tetracycline solutions were stored frozen in the dark and replaced
frequently because the drug undergoes both thermal and photochemical
degradation. 70S Ribosomes were prepared from
Escherichia coli
MRE600 as described (
13
). 30S and 50S ribosomal subunits were isolated as in (
14
).
Photolysis experiments were performed using a short arc mercury lamp (HBO 500 W/2 from OSRAM) having an output concentrated at 366 nm. Samples were irradiated in vertical tubes at a distance of ~200 mm from the lamp in the outer focal point (average luminance 3000 cd/cm
2
). Filters were chosen in such a way that any light below 300 nm was completely
eliminated. All photolyses were performed in standard TMK buffer (20 mM Tris-HCl, pH 7.4, 100 mM KCl, 6 mM MgCl
2
, 0.4 mM EDTA and 2 mM DTE) at 0oC. For the identification of the labelled nucleotides highly purified
tetracycline was used for crosslinking.
After photolysis, ribosome samples were separated into two equal parts. In one
of them ribosomal RNA was degraded by RNase T
1
. The ribosomal proteins were precipitated by addition of 0.1 vol 100 g/l BSA
and 1 vol 10% TCA, redissolved in 10 M urea and TCA precipitated again. The
precipitate was filtered through a GF/C (Millipore) filter and washed several
times with diethylether/ethanol (10:1) to remove unbound [
3
H]tetracycline. For determination of the amount of [
3
H]tetracycline photoincorporated into ribosomal RNA, the RNA was isolated by phenol/chloroform extraction and
precipitated. The pellet was dissolved in water and the radioactivity was measured. Virtually no background of [
3
H]tetracycline was detectable in non-irradiated control samples.
RNA isolated from the ribosomes as described above was used for reverse
transcriptase analysis according to (
15
). Primers used to investigate crosslinks on 16S and on 23S rRNA were the same
as used in (
16
).
tRNA was charged and acetylated as described (
17
). Ac-[
3
H]Phe-tRNA was purified by reversed phase high performance liquid chromatography
on nucleosil 300-5-C4-column (4 * 250 mm). Up to 30 nmol was typically applied to the
column. The eluting solvent had constant 400 mM NaCl, 10 mM Mg(CH
3
COO)
2
,
20 mM NH
4
-acetate, pH 5.0. The gradient steps had the following percentages of
methanol: 0%, 5 min; 0-9% in 5 min; 9-25% in 50 min; 25%, 5 min. The different tRNA species were
separated in the linear gradient from 9 to 25% of methanol. The fractions
containing Ac-[
3
H]Phe-tRNA were collected, desalted using an Econopac P6 desalting column
(BioRad) and dried in a speedvac.
Isolated 30S and 50S subunits were irradiated in the presence of tetracycline as
described above. To remove unbound tetracycline, the subunits were pelleted
twice (first 8 h at 31 000 r.p.m. for 30S or 5 h at 28 000 r.p.m. for 50S, then
18 h at 21 000 r.p.m. for 30S or 18 h at 15 000 r.p.m. for 50S) in a Beckman
ultracentrifuge using the SW50.1 rotor. The pellet was dissolved in T
20
M
20
N
400
buffer (20 mM Tris-HCl pH 7.5, 20 mM MgCl
2
, 400 mM NH
4
Cl, 4 mM [beta]-mercaptoethanol). After addition of an equimolar amount of the
complementary untreated ribosomal subunit the samples were incubated for 10 min
at 37oC. The samples were then centrifuged in a 10-30% sucrose gradient in T
20
M
10
N
100
buffer (20 mM Tris-HCl pH 7.5, 10 mM MgCl
2
, 100 mM NH
4
Cl, 6 mM [beta]-mercaptoethanol, 0.5 mM EDTA) for 18 h at 18 000 r.p.m. (Beckman SW28 rotor). The fractions containing
70S chimeric ribosomes were collected and centrifuged for 24 h at 24 000 r.p.m. (Beckman 45Ti rotor). The pellet was resuspended in T
20
M
10
N
100
(containing 50% glycerol) and stored at -70oC.
For P-site binding, 0.2 pmol/[mu]l 70S ribosomes were incubated in T
20
K
100
M
6
buffer (20 mM Tris-HCl pH 7.4, 100 mM KCl, 6 mM MgCl
2
, 0.4 mM EDTA and 2 mM DTE) for 10 min at room temperature in the presence of
0.2 pmol/[mu]l Ac-[
3
H]Phe-tRNA and 0.1 [mu]g/[mu]l poly(U). For A-site binding 0.2 pmol/[mu]l 70S ribosomes were pre-incubated in T
20
K
100
M
12
buffer (20 mM Tris-HCl pH 7.4, 100 mM KCl, 12 mM MgCl
2
, 0.4 mM EDTA and 2 mM DTE) for 3 min at 37oC in the presence of 0.2 pmol/[mu]l uncharged tRNA
Phe
and 0.1 [mu]g/[mu]l poly(U). Then 0.2 pmol/[mu]l Ac-[
3
H]Phe-tRNA was added and the sample was incubated for 10 min at room temperature. To remove unbound Ac-[
3
H]Phe-tRNA the samples were filtered through a nitrocellulose filter (NC-Filter, 45[mu]m, Milipore, Molsheim, France). The filter was washed several times with T
20
K
100
M
6
or T
20
K
100
M
12
, respectively. The radioactivity of the filter corresponded to the bound Ac-[
3
H]Phe-tRNA.
The puromycin reaction, the formation of Ac-[
3
H]Phe-puromycin from Ac-[
3
H]Phe-tRNA and puromycin, was used to measure peptidyltransferase activity. Ac-[
3
H]Phe-tRNA was bound to the P-site as described above and incubated with 1 mM puromycin for 10 min
at room temperature. After addition of 1 vol 0.3 M Na-acetate (pH 5.5) in saturated MgSO
4
the Ac-[
3
H]Phe- puromycin was extracted with ethyl acetate and the radioactivity measured
in a scintillation counter.
In the present work we reinvestigated the interaction of tetracycline with
ribosomal RNA. First, we determined the distribution of tetracycline
photocrosslinked to ribosomal RNA and proteins using [
3
H]tetracycline as photoaffinity label. 70S ribosomes were irradiated in the
presence of 100 [mu]M [
3
H]tetracycline for different periods of time. As shown in Figure
1
, the amount of tetracycline photocrosslinked to RNA and proteins increased with
increasing irradiation time. Tetracycline was photocrosslinked to both, RNA and
proteins, to about the same extent (Fig.
1
). This is in contrast to the results of Goldman and co-workers (
5
,
11
) who only found up to 10% of the tetracycline label incorporated into rRNA (see
Discussion). Using 25 s irradiation time, tetracycline was already incorporated
into 70S ribosomes at an approximate ratio of 1:1. Therefore, and to avoid
secondary reactions of tetracycline photoproducts we used an irradiation time
of just 30 s in all following experiments.
The analysis of the ribosomal proteins photolabelled by [
3
H]tetracycline confirmed the results of Goldman
et al
. (
5
), as we could also identify protein S7 as the major labelled protein (data not
shown). We therefore concentrated on the analysis of the sites of tetracycline-rRNA interactions using the primer extension method. The primers chosen
were spaced every ~200 nucleotides on the 16S, 23S and 5S RNA, so we were able to scan the
entire RNAs except the 3'-ends. The RNAs used for the templates were from 70S ribosomes
irradiated in the presence of different amounts of tetracycline (Fig.
2
). RNAs from non-irradiated 70S ribosomes and from ribosomes irradiated in the absence of
tetracycline were used as controls for random stops on the RNA template and
possible UV-induced internal RNA-RNA crosslinks. When a stop was observed, the crosslinked
nucleotide was taken to be the following nucleotide in the rRNA template (i.e.
the preceding one in the rRNA sequence). The numbers of photoaffinity labelled nucleotides increased with rising concentrations of tetracycline. The
half maximal inhibition of Ac-Phe-tRNA binding to the ribosome by tetracycline was reported to be 40 (
18
) or 4 [mu]M (
3
), respectively, whereas under our incubation conditions we observed a value of ~10 [mu]M. Therefore, only those nucleotides were considered to correlate well
with the inhibitory action of tetracycline which were labelled in the presence
of 40 [mu]M or lower concentrations of the antibiotic. Under these conditions, three
sites on the 16S RNA (Fig.
2
A; G693, G1300 and G1338) and three sites on the 23S RNA (Fig.
2
B; G2505, G2576 and G2608) were photoaffinity labelled by tetracycline. No
incorporation of tetracycline into 5S RNA could be detected (data not shown).
Several additional labelled nucleotides on 16S and 23S RNA could be identified
when 80 or 120 [mu]M tetracycline were used and some of them are discussed later.
Next we investigated the effect of tetracycline photocrosslinks on ribosomal
function using chimeric ribosomes. To distinguish the effects of tetracycline
crosslinked to the 30S subunit from those of tetracycline crosslinked to the
50S subunit we performed the following experiments. Isolated 30S or 50S
ribosomal subunits were irradiated in the presence of tetracycline. Then we immediately removed the unbound tetracycline by centrifugation. The ribosomal subunits were reconstituted to 70S particles with the
complementary untreated subunit. The chimeric ribosomes were then isolated by
density gradient centrifugation and the effects of the photomodified subunits on binding of peptidyl-tRNA to the ribosome and on peptidyltranferase activity were investigated. As shown in Figure
4
, the photocrosslinked 50S ribosomal subunit had no effect on binding of Ac-[
3
H]Phe-tRNA either to the ribosomal A- or to the P-site when compared with untreated 70S ribosomes.
The Ac-[
3
H]Phe-puromycin formation, which measures peptidyltranferase activity, was also unaffected by tetracycline photocrosslinked
to the 50S subunit. In contrast, chimeric ribosomes with tetracycline
photocrosslinked to the 30S subunit had diminished ability of binding Ac-[
3
H]Phe-tRNA to the ribosomal A-site compared with untreated 70S ribosomes, whereas the P-site binding and the Ac-[
3
H]Phe-puromycin formation remained unaffected. These results correlate well with the published data on the
inhibition of A-site binding of tRNA by tetracycline (
2
,
3
).
The experiments described in this paper show that tetracycline can be
photocrosslinked not only to ribosomal proteins, but also to rRNA. We found an
approximately 1:1 incorporation of radioactivity in ribosomal proteins and RNA,
respectively. [
3
H]Tetracycline has been previously used in extensive studies to characterize
ribosomal binding sites for this antibiotic (
5
,
11
,
12
). In these experiments up to 90% of the radioactivity was found to be
incorporated in ribosomal proteins with S7 being the main protein labelled (
5
). The difference in the distribution of the label may result from the different
irradiation conditions used. We used ~500 W for only 30 s (15 kJ) whereas Cooperman's group used 30 W and irradiation times between 60 and 90 min (108-162 kJ) (
11
). These workers also showed that upon irradiation tetracycline photoproducts were generated which could further react with the ribosome (
5
). Therefore, we took care to avoid long irradiation times. Furthermore, as our
results did not change upon addition of [beta]-mercaptoethanol which has been used to avoid light independent
incorporation of tetracycline photoproducts and as our protein labelling
pattern conforms to the one published previously (
5
), we are confident that the crosslinks observed derive from genuine
tetracycline.
The experiments were performed with increasing concentrations of tetracycline; however, only those crosslinks have been described which
appear <= 40 [mu]M tetracycline. The number of crosslink sites increased with higher
concentrations (e.g. on 23S RNA three more at 80 [mu]M, and additional seven at 120 [mu]M tetracycline), in accordance with a large number of low affinity
binding sites for tetracycline observed on both the 30S and 50S subunits (
4
-
6
).
We would like to thank I. Halama and S. Dorner for help during preparation of
the manuscript, B. Weiser for the XRNA program, R. Schroeder, K. Nierhaus and Z. Rattler for invaluable discussions and infinite
patience. This work was supported by a grant (P09454-MIB) from the `Österreichischer Fonds zur Förderung der wissenschaftlichen Forschung' and from the `Anton Dreher Gedächtnisschenkung für Medizinische Forschung' to A.B.
REFERENCES
Return