Chemical and Enzymatic Probing of Antibiotic-Ribosome Complexes

Abstract

It has become clear over the past decade that many of the antibiotics that inhibit protein biosynthesis act at the level of rRNA. The earliest indications were the demonstrations that the host producers of some of the drugs protect themselves by chemically modifying a single nucleotide within one of then large rRNAs (1). In addition, there was early evidence, especially from genetic studies on mitochondria, that single nucleotide changes can produce drug resistance (reviewed in ref. 2).

23.1 Introduction

It has become clear over the past decade that many of the antibiotics that inhibit protein biosynthesis act at the level of rRNA. The earliest indications were the demonstrations that the host producers of some of the drugs protect themselves by chemically modifying a single nucleotide within one of then large rRNAs (1). In addition, there was early evidence, especially from genetic studies on mitochondria, that single nucleotide changes can produce drug resistance (reviewed in ref. 2).

This led to attempts to define rRNA binding sites of the many different types of ribosomal antibiotics (3-5), using chemical (or enzymatic) probing approaches. Several sites have been identified on the 16S rRNA and 23S rRNAs of Escherichia coli (6-13) using the latter approach, and some of the aminoglycoside antibiotics have been probed directly on fragments of 16S rRNA (14). A start has also been made in locating antibiotic sites on the 23S-like rRNAs of archaea (extreme halophiles) and eukaryotes (yeast) (13).

In principle, the experimental procedure is the same as that used for mapping protein binding sites on rRNA and ribosomal ligands on ribosomes where the reactivities of base-specific chemicals, or ribonucleases, are altered in the presence of a bound ligand (15). However, the antibiotic work is more technically demanding. Most drugs exhibit low binding affinities for ribosomes and often produce few, and weak, changes in chemical reactivity.

Examples of chemical probing data are illustrated for lincomycin and clindamycin bound to 23 s rRNA of E. coli ribosomes (from ref. 11) in Fig. 1 , and for the antibiotics anisomycin, bruceantin, griseoviridin and virginiamycin M₁ bound to the 23S rRNA of the archaeon Haloferax Mediterranei or to the 27s rRNA of Saccharomyces cerevisiae (from ref. 13) in Fig. 2 . Finally, data obtained by different chemical modification and enzymatic methods are summarized in Fig. 3 for a complex of thiostrepton and E. coli 23S rRNA (from ref. 8).

Fig. 1.

(A) Autoradiograms of a gel separation of primer extension reactions on 23S rRNA after lincomycin and clindamycin binding to 70s ribosomes followed by chemical modification with DMS. Extensions from a 2141–2157 primer on unmodified (lane K) and DMS-modified RNA templates (lanes 0–4) after binding of 0, 1, 10, 100, or 1000 µM antibiotic (lanes 0–4, respectively). Bands showing altered reactivities are indicated and numbered. They are displaced one nucleotide from the co-migrating sequence lanes. (B) Graphs showing how the DMS modification of A2058 and A2059 varies with the concentration of clindamycin and lincomycin. Band intensities in each lane were normalized relative to control lanes modified in the absence of drug. Each point on the curve represents an average of at least three experiments. Standard errors of the mean are shown as vertical bars. From ref. 11, with permission.

23.2 Materials

• 1. Polyacrylamide gel electrophoresis (PAGE) apparatus (15 × 15 × 0.1-cm plates) for primer purification.

• 2. Electrophoresis apparatus for sequencing (e.g., Bio-Rad 20 × 50-cm plates) and a 3000-V Power supply (e.g., Pharmacia, Uppsala, Sweden).

• 3. Bench thermostat for 30°, 37°, 50°, and 95°C.

• 4. Bench centrifuge and refrigerated bench centrifuge (e g., Eppendorf, Hamburg, Germany).

• 5. Speed-Vac centrifuge (e.g., Savant, Farmingdale, NY).

• 6. Minimonitor (g.m. meter, type 5.10 Mini-instruments, Burnham on Crouch, UK).

• 7. Laser microdensitometer (e.g., Hoeffer, San Francisco, CA)/

• 8. X-ray films (Fuji, Japan) and cassettes (Amersham, Little Chalfont, UK).

• 9. Enzymes polynucleotide kinase (e.g., BioLabs, Beverley, MA), AMV reverse transcriptase (e.g., Life Sciences, FL); ribonucleases T1 and T2 (Sankyo, Japan), ribonuclease V1 (Pharmacia, Uppsala, Sweden); lysozyme (Pharmacia).

• 10. Modifying chemical reagents. dimethyl sulfate (DMS) (Aldrich-Gold brand Steinheim, Germany), kethoxal (ICN, Costa Mesa, CA), 1-cyclohexyl-3-[2-morpholinoethyl] carbodumide metho-p-toluene sulphonate (CMCT) (Sigma, St. Louis, MO).

• 11. Antibiotics (10-mM stocks).

• 12. Active ribosomes.

Fig. 2.

Autoradiograms showing some of the effects induced by anisomycin (ani), bruceantin (bru), griseoviridin (gri) and virginiamycin M₁ (vir) on the chemical reactivities of (A) kethoxal and (B) DMS of 23 S-like rRNAs of ribosomes isolated from the archaeon H. mediterranei and the eukaryote S. cerevisiae. Ru and Rm denote unmodified and modified ribosome samples, respectively, with no bound antibiotic and U, C, G, and A represent nucleotide sequencing tracks. Altered reactivities are indicated by arrows and sequence positions are given.

Fig. 3.

(A) Autoradiogram showing chemical (DMS and kethoxal [keth]), and enzymatic (RNase T₁ and V₁) probing of thiostrepton bound to the GTPase centre on 23S rRNA of E. coli ribosomes. (B) Data from the chemical (squares) and enzymatic (circles) probing (8), and hydroxyl radical (triangles) probing (22), of the thiostrepton-70S ribosome complex are superimposed on the secondary structure of the rRNA site (8).

23.3 Methods

23.3.1 Preparation of Active Ribosomes

• 1. Various procedures have been described for preparing ribosomes (16). It is important, when isolating vacant ribosomes, to prepare them in as active a form as possible. This generally involves performing high-salt washes to remove bound peptidyl tRNAs and the isolation of tight-couple ribosomes on sucrose density gradients (16).

• 2. It is important to work quickly during the ribosome preparation in order to minimize rRNA cutting. RNase cuts will appear on the final autoradiograms and may mask antibiotic-induced effects. Use an RNase-deficient strain, if available, for example MRE 600 for E coli.

• 3. Even tight-couple ribosomes contain low levels of bound peptidyl tRNA (10–30%). This residual tRNA can be removed by reversibly dissociating the ribosomal subunits. A procedure, described by Makhno et al (17) for E coli ribosomes (see also ref. 18), involves incubating tight-couple ribosomes in 20 mM Tris-HCl, pH 7.4–7.7, 3 mM MgCl₂ and 300 mM NH₄Cl for 15 min at 37°C and sedimenting them through a sucrose gradient (35–50% for 14 hrs at 100,000g) in the same buffer Equimolar levels of subunits are then recombined in 20 mM Tris-HCl, pH 7.4–7.7, 5 mM MgCl₂, and 50 mM NH₄Cl and centrifuged through a sucrose gradient in the same buffer to gave a 90–95% yield of 70S ribosomes.

• 4. Higher activities are obtained if the ribosomes are sedimented into a concentrated cushion of sucrose (1.1M), rather than letting them form a firm pellet.

• 5. Ribosomes should be stored in 25–50 µL aliquots at −80°C. Various storage buffers have been used and the optimal buffer will depend, to some extent, on the organism. For E. coli ribosomes, 10 mM Tris-HCl, pH 7.8, 10 mM magnesium acetate, 60 mM NH₄Cl and 6 mM 2-mercaptoethanol is satisfactory; for yeast 100 mM Tris-HCl, pH 7.4,12.5 mM MgCl₂ 80 mM KCl and 3 mM 2-mercaptoethanol has been used, and for archaeal extreme halophiles 20 mM Tris-HCl pH 7.4, 3M KCl, 60 mM magnesium acetate, and 6 mM 2-mercaptoethanol is appropriate (13) (see Note 1 ).

• 6. Some drugs, for example sparsomycin, do not bind to vacant ribosomes and require the presence of a bound peptidyl tRNA. Thus, a peptidyl tRNA must be prebound to the ribosomes, either enzymatically, or nonenzymatically, prior to adding the drug. A more straightforward experiment is to work with polysomes, where the peptidyl tRNA is bound.

• 7. Polysomes can be prepared in high yield from E. coli by the procedure of Ron et al (19). The crucial step in this procedure is the cell lysis. It involves mixing 30–100 mL midlog phase culture with an equal volume of ice. Centrifuge 5 min at 10,000g. Resuspend cells in cold 0.01M Tris-Cl, pH 7.8, 15 mM Mg acetate containing 0.5 mg RNase-free lysozyme. Freeze in an ethanol-dry me bath, and thaw carefully in cool water. Perform two cycles and then incubate with 15 mL 10% Na deoxycholate for 3 min at 0°C to complete the lysis.

• 8. Activity tests should be performed on the ribosomes to ensure a high level of activity. Poly (U)-dependent polyphenylalanine assays can be performed for bacteria and eukaryotes (16) and archaea (20) (see Note 2 ).

23.3.2 Preparation of Antibiotic-Ribosome Complexes

Many antibiotics have now been shown to bind to the rRNA components of ribosomes of the three primary kingdoms, and it is likely that most of the ribosomal inhibitors are rRNA-binding ligands. Several cytotoxins have also been isolated that inactivate ribosomes. At least some of these affect a highly conserved terminal loop in a stem-loop structure of the 23S-like rRNA either by introducing a site-specific cut (e.g., α-sarcin) or by causing a depurination (e.g., ricin). Their sites of action, and their binding sites on the rRNA, can also be examined (21) by the procedures outlined here.

• 1. In a typical experiment, 20-µg ribosomes are diluted into 100 µL of modification buffer HEPES buffer is generally used (its pK of 7.5 is convenient for modification experiments, see Note 5 ). For ribosomes from E. coli and S. cerevisiae, and most other organisms, 50 mM HEPES-KOH, pH 7.8, 10mM MgCl₂, 15 mM KCl, 15mM NH₄Cl, 1 mM dithiothreitol (DTT), and 0.1 mM EDTA can be used ( Figs. 1 , 2B , and 3 ). For extreme halophile archaea 70 mM HEPES-KOH, pH 7.8, 60 mM magnesium acetate, 3M KCl, 6 mM 2-mercaptoethanol is appropriate ( Fig. 2A ).

• 2. Ribosomes are incubated for 20 min at 30°C. The antibiotic is then added from a 10 mM stock solution to a final concentration of 0.1 mM, or higher (see Notes 3 and 4 ). A control sample without antibiotic is included Antibiotic-ribosome complexes and the control are incubated for 30 min at 30°C and allowed to cool slowly on me so that the local structure around the bound drug can renature, this should produce higher yields of complex.

• 3. After modification, the ribosomes can be run over a sucrose gradient, and a new selection made for ribosomes, or subunits, to ensure that the level of dissociated ribosomes, or degraded material, is minimal in the samples to be analyzed further. This will lead to higher quality data on the final autoradiograms.

23.3.3 Chemical and Enzymatic Modification

The procedures for chemical and enzymatic modification of rRNA and protein-rRNA complexes, and primer extension, have been described in detail (15) and these procedures are appropriate for antibiotic-ribosome complexes. The hydroxyl radical procedure for ribosome modification (22) has been used for monitoring the C1’ and C4’ atoms in ribose residues of antibiotic-ribosome complexes. Whereas the base-specific chemicals and the ribonucleases (with the exception of RNase V₁— Fig. 3 ) monitor alterations in the reactivities of unpaired bases, the hydroxyl radical procedure will monitor most of the RNA backbone (see Fig. 3B ). The main details of the procedures and those that have been altered from ref. 15, are outlined below.

• 1. Base specific chemical reagents. DMS - G(N7) > A(N1) > C(N3)—1 µL (1∶1 in ethanol), 10 min at 30°C (see Note 6 ); Kethoxal—G (N1, N2)—5 µL (35 mg/mL in 20% ethanol), 10 min. at 3°C, CMCT—U(N3) > G(N1) - 100 µL (42 mg/mL in ribosome modification buffer) are added to 100 µL antibiotic-ribosome complex and a 100 µL control sample (without antibiotic) and incubated for 20 min at 30°C (see Note 7 ).

• 2. Ribonuclease probes are added to 20 µg antibiotic-ribosome complex in 100 µL modification buffer. 0.02 RNases T₁ (G-specific), 0.01 U RNase T₂ (low specificity with a preference for As) and 0.3 U RNase V₁ (specific for double helical regions).

• 3. An unmodified sample should be included in the analyses, as well as a modified sample, that are incubated without antibiotic, and treated in the same way as the other samples. This is important for quantifying antibiotic-induced reactivity changes. It also enables one to establish whether antibiotic binding induces conformational changes in the rRNA.

• 4. The base-specific reactions stop buffers are added to all samples, as indicated in the original protocols (15) (see Note 8 ). Ribosomal proteins, ribonucleases, and excess reagent are eliminated by phenol extraction (see Note 9 ). After phenol has been distilled it is saturated with buffer and can be stored frozen for long periods or at 4°C for up to 2 weeks. The RNA is precipitated by 2.5 vol ethanol (see Note 10 ). For the extreme halophile ribosomes, the salt pellet can be extracted with phenol and precipitated with ethanol to remove the high salt. The enzymatic reactions are stopped by phenol extraction.

• 5. Hydroxyl radical probing is performed by placing 30-µg ribosomes in 25 µL modification buffer. To this is added the freshly prepared hydroxyl radical mixture consisting of 5 µL 50 mM Fe(NH₄)_2·(SO4)_2·6H₂O, 5 µL 100 mM EDTA, 5 µL 250 mM Na ascorbate, and 5 µL 2.5% H₂O₂ per reaction. The reactions are performed for 4 min on ice and stopped by precipitation with 0.3M Na acetate, pH 6.0, and 2.5 vol cold ethanol.

23.3.4 Primer Extension

Modified sites are identified by primer extension with reverse transcriptase.

• 1. Oligonucleotide primers (17–23 mer) are labeled at their 5′ ends using [γ³²P]-ATP and polynucleotide kinase as described earlier (15) (see Note 11 ). Priming sites are spaced at 150–200 nucleotide intervals along the rRNA.

• 2. When a drug site has been approximately localized, annealing a primer approx 50 nucleotides 3′ to the site is optimal for further analysis (see Note 12 ).

• 3. Primers are purified in 20% polyacrylamide gels containing 7M urea Bands are extracted from the gel and eluted in 90–120 µL H₂O by incubating at 37°C for 20 h.

• 4. After the extension, reactions are stopped by precipitating with ethanol, and samples are centrifuged at 13,000 rpm for 10 min at 4°C. After washing with 80% ethanol, samples are dried in a Speed-Vat centrifuge and resuspended in loading buffer (3–9 µL), so that 1 µL produces about 100 cps. The loading buffer consists of a formamide solution: 80% formamide, 1 mM EDTA, 0.05% xylene cyanol, 0.05% bromophenol blue, and 10 mM NaOH to enhance DNA denaturation and degrade any excess RNA.

• 5. After denaturing for 3 min at 95°C, 1 µL of each sample is loaded onto a 5–6% polyacrylamlde gel containing 8M urea (see Note 13 ). Samples are run at 60 W for 1–1.5 h.

• 6. Polyacrylamide gels can be made wedge-shaped, or with a salt gradient, to obtain bands which are evenly spaced, especially at the bottom of the gel. The salt gradient works best for short gels (e.g., 20 × 40–50 cm).

• 7. Pre-electrophorese to remove persulfate ions (from the ammonium persulfate) and to warm up the gel.

• 8. Dry the pellet briefly (1–2 mm) to eliminate ethanol, before dissolving in loading buffer. Otherwise, the samples will sit poorly in the wells when loaded and the sample may be partially aggregated.

• 9. NaOH is included in the loading buffer because it facilitates DNA denaturatlon and degrades any excess RNA.

• 10. After denaturation, place the samples on ice and then load within 10 min to minimize sample renaturation.

• 11. Load the same number of counts in each well approx 100 cpm for the best results.

• 12. Three microliters is the minimum volume of loading buffer in which labeled RNA should be dissolved to avoid losing the sample in the syringe or by evaporation.

• 13. Thin gels (0.2 mm) are washed with 3% acetic acid (2 × 15 min), dried (after the run to give high resolution) and autoradiographed for 12–24 h.

• 14. The band’s intensities can be evaluated visually or by microdensitometry (see Notes 14 – 18 ).

23.3.5 Sequences

Dideoxy sequencing lanes A, C, G, and T are run in parallel with the modified samples.

• 1. Templates can be obtained from phenol extractions of approx 25 µg ribosomes.

• 2. After the extraction (15), samples are resuspended in 15 µL of TE buffer (10 mM Tris HCl, pH 7.5, 0.1 mM EDTA). The primer was annealed to 1 µL template (15) in a volume of 6 µL and extension was performed by adding 1 µL of the appropriate ddNTP (initial concentrations ddATP—0.33 mM; ddCTP—0.22 mM; ddGTP—0.22 mM, and ddTTP—0.33 mM) and 3 µL of extension mix, consisting of 0.4 µL 25X reverse transcriptase buffer (1.25M Tris HCI, pH 8.4, 250 mM MgCl₂, 50 mM DTT), 0.8 µL 10 × chase (2.5 mM of each dNTP), 1.8 µL H₂O and 1 unit reverse transcriptase (15).

23.4 Notes

23.4.1 Preparation of Active Ribosomes

• 1. Ribosomes dissociate, and progressively lose activity, on freezing and thawing.

• 2. It is important to use freshly prepared S100, or an aliquot that has only been frozen and thawed once, in the activity assay.

23.4.2 Preparation of Antibiotic-Ribosome Complexes

• 3. Antibiotics are available from various sources but especially from pharmaceutical companies and private collections. The former, at least, will generally provide information about the level of purity. If there is any doubt, and a longer series of experiments is planned, it is wise to analyze the batch by, for example, mass spectrometry, before proceeding. Some antibiotics (for example, thiostrepton) generally contain inactive conformer that will probably not show up in the analyses, and this may increase with the age of the sample.

• 4. Dissolving drugs can be difficult. Although some such as erythromycin, clindamycin and licomycin are available in water-soluble salt form from drug stores, many others are hydrophobic and exhibit low solubility in aqueous solution. These can generally be dissolved, and stored, in dimethyl sulfoxide (DMSO) or in aqueous mixtures with DMSO. The following are recommended: 100% DMSO—bruceantin, carbomycin, chloramphenicol, griseoviridin, micro-coccin, T2 toxin, thiostrepton, tylosin, virginiamycin M₁; 50% DMSO (v/v)—anisomycin and narciclasine, and water—amicetin, anthelmycin, blasticidin S, and sparsomycin.

23.4.3 Chemical Modification

• 5. Tris-buffer is unsuitable for modification with the base-specific chemicals since the Tris ion may react with the reagents, also the buffering capacity of Tris is strongly temperature dependent HEPES is generally the preferred buffer

• 6. DMS should be stored dry, and at 4°C when not in use, it reacts with water to produce sulfuric acid and methanol. For stronger modification, the incubation time should be increased. If more DMS is added the sulfuric acid generated may cause a drop in pH and consequent rearrangement of the RNA structure and/or destabilization of the antibiotic-ribosome complex. Anything that has been in direct contact with DMS should be placed in 10M NaOH, including the remains from the first phenol extraction.

• 7. Kethoxal gives best results when it has not been lyophilized or frozen. Since it modifies arginine side chains slowly, this may help to inactivate some ribo-nucleases and help to reduce the levels of background cuts on the autoradiograms but it may also interfere with protein-RNA interactions on the ribosome. To dissolve kethoxal, add approx 35 mg to 200 µL absolute EtOH. Shake to dissolve and then add 800 µL water. Precipitation generally occurs and the sample turns white. It should be shaken until the precipitate has redissolved (it may take several hours). After kethoxal modification, add Na borate at pH 7.5 or lower to stabilize the kethoxal-base adduct. Then adjust to 0.3M Na acetate at pH 6.0.

• 8. Dissociation of the antibiotic-ribosome complex may occur if the modification is continued for too long.

• 9. If it is necessary to store the samples between the modification step and the phenol extraction, samples should be stored on dry ice, since significant reaction may occur at −20°C.

• 10. Precipitate rRNA with at least 2.5 vol cold ethanol.

23.4.4 Primer Extension

• 11. Polynucleotide kinase is inhibited by NH₄ ⁺ ions and they should be avoided when labeling the primer. For labeling, use a primer γ[³²P]-ATP ratio of 1 pmol 1 pmol.

• 12. The primer is the limiting reagent in the RNA-primer annealing step. One can use either a labeled or an unlabeled primer. Gel resolution is normally best if a [³²P]-labeled primer is used. However, unlabeled samples last longer. If [³⁵S]-labeled nucleotide triphosphates are used, information may be lost because films are less sensitive to the weaker β particles emitted by [³⁵S]. When an unlabeled primer is used, it is possible that degraded RNA will act as a primer and anneal to the RNA template. Thus, when reverse transcriptase is added, cDNA is synthesized both from the oligodeoxynucleotide primer and the annealed RNA and will produce artefact bands.

23.4.5 Sequencing

• 13. Acrylamide solutions will keep for up to 6 months, if deionized and protected from light. Acrylamide breaks down in solution (as does bis-acrylamide) to acrylic acid, by deamination, which affects the mobility of molecules through the gel matrix. Breakdown is catalyzed by light and alkali. The acrylamide solution should be stored at pH 7.0 or lower, and be degassed, prior to use, to facilitate polymerization which is inhibited by oxygen.

23.4.6 General Notes

• 14. During protein biosynthesis, ribosomes generate different functional conformers (23). Antibiotics may recognize one or more of these conformers and, also, alter the equilibrium between them on vacant ribosomes. This, together with the low binding affinities of many of the drugs, may explain the nonreproducible effects that are occasionally observed. Therefore, it is important to repeat the experiments several times using antibiotic complexes with different batches of ribosomes. It is also advisable to test a range of drug concentrations (as in Fig. 1 ) in order to reinforce the specificity of the observed reactivity changes.

• 15. Some thought should be given to whether the drug under investigation is likely to be modified directly by the chemical reagents. Extensive information on the active sites of the drugs is available in earlier reviews (3–5).

• 16. The whole rRNA sequence should be scanned, since many long-range interactions occur in the large rRNA structures and distant nucleotides in the sequence may be close neighbors in the ribosomal tertiary structure.

• 17. The probing data obtained is necessarily minimal, since many nucleotide positions cannot be read owing to the occurrence of control bands on the autoradiograms (see Fig. 1 ). These may arise from cuts incurred during the preparation, or treatment of the ribosomes, or from posttranscriptional modifications or stable rRNA structures, which cause pausing of the reverse transcriptase.

• 18. It has been shown recently that aminoglycoside drugs that affect the decoding site on 16S rRNA (6) can interact directly with a synthetic RNA fragment corresponding to this region (14). This raises the possibility that other ribosomal antibiotics can be mapped directly onto RNA fragments by the above methods.