bases in 16S ribosomal RNA - NCBI - NIH

10 downloads 119 Views 1007KB Size Report
Jun 17, 1991 - Michael Cannon, Julian Davies2 and. Harry F.Noller1. Department ..... Cundliffe,E. (1990) In Hill,W.E., Dahlberg,A.E., Garrett,R.A.,. Moore,P.B. ...
The EMBO Journal vol. 1 0 no. 10 pp. 3099 - 3103, 1991

Interaction of antibiotics with A- and P-site-specific bases in 16S ribosomal RNA

Joanna Woodcock, Danesh Moazed1"3, Michael Cannon, Julian Davies2 and Harry F.Noller1 Department of Biochemistry, Division of Biomolecular Sciences, King's College, London WC2R 2LS, UK, 'Sinsheimer Laboratories, University of California at Santa Cruz, CA 95064, USA and 2Unit6 de Genie Microbiologique, Institut Pasteur, 75724 Paris, France 3Present address: Department of Biochemistry and Biophysics, University of California at San Francisco, San Francisco, CA 94143, USA Communicated by J.Davies

We have studied the interactions of the antibiotics apramycin, kasugamycin, myomycin, neamine and pactamycin with 16S rRNA by chemical probing of drug-ribosome complexes. Kasugamycin and pactamycin, which are believed to affect translational initiation, protect bases in common with P-site-bound tRNA. While kasugamycin protects A794 and G926, and causes enhanced reactivity of C795, pactamycin protects G693 and C795. All four of these bases were previously shown to be protected by P-site tRNA or by edeine, another P-site inhibitor. Apramycin and neamine, which both induce miscoding and inhibit translocation, protect A1408, G1419 and G1494, as was also found earlier for neomycin, gentamicin, kanamycin and paromomycin. A1408 and G1494 were previously shown to be protected by A-site tRNA. Surprisingly, myomycin fails to give strong protection of any bases in 16S rRNA, in spite of having an apparently identical target site and mode of action to streptomycin, which protects several bases in the 915 region. Instead, myomycin gives only weak protection of A1408. These results suggest that the binding site(s) of streptomycin and myomycin have yet to be identified. Key words: antibiotic/chemical probing/rRNA/ribosome

Introduction A wide range of antibiotics act by inhibiting protein synthesis, and the majority of these drugs interact directly with ribosomes (Cundliffe, 1981). In early studies on antibiotic-resistance mutations that affected ribosomes, attention was drawn to the ribosomal proteins as possible target sites for antibiotic interaction. More recently, however, resistance mutations for many antibiotics have been found that result in alterations of ribosomal RNA (rRNA), raising the possibility that such antibiotics may actually interact with rRNA, perhaps exclusively (De Stasio et al., 1988). Indeed, many antibiotics, with the notable exceptions of puromycin (Moazed and Noller, 1987b) and sparsomycin (Moazed and Noller, 1991), protect specific nucleotides in ©) Oxford University Press

16S and/or 23S rRNA from chemical probes when they bind to ribosomes, each producing a characteristic footprint on the rRNA (Moazed and Noller, 1987a,b). The protected nucleotides are, in most cases, identical with or located adjacent to bases that have been directly implicated in those functional processes known to be affected by the antibiotic in question. Accordingly, it has been suggested that the mode(s) of action of such drugs may be to interfere directly with the function of highly conserved sites in rRNA (Moazed and Noller, 1987a,b; Noller et al., 1990). In this study, we report the effects of several additional 30S subunit-specific antibiotics: kasugamycin, pactamycin, apramycin, neamine and myomycin. Kasugamycin and pactamycin are known to inhibit translational initiation (Cohen et al., 1969; Okuyama et al., 1971; Tai,P.-C. et al., 1973), while both apramycin and neamine not only increase the frequency of translational errors but also inhibit translocation (Delcuve et al., 1978; Perzynski et al., 1979). Myomycin has a limited structural resemblance to streptomycin. Nevertheless, these two aminoglycoside-aminocyclitol antibiotics apparently share an identical mode of action, inducing misreading and inhibiting translational initiation in cell-free protein synthesizing systems (Davies et al., 1988). We show that kasugamycin and pactamycin protect P-site-specific bases, which may explain how they block initiation. Apramycin and neamine protect bases in or closely adjacent to tRNA A-site-protected sites on ribosomes, in keeping with their abilities to increase miscoding. Interestingly, myomycin gives only very weak protection of a single base within 16S rRNA, and fails to protect any bases in common with streptomycin, despite the apparently identical inhibitory actions of these two antibiotics.

Results The different antibiotics were bound to Escherichia coli 70S ribosomes at concentrations where they are known to exert their effects on translation, and probed with the single-strandspecific RNA probes kethoxal and dimethyl sulfate. The sites of chemical modification and protection by the drugs were identified by primer extension. All of the antibiotic-dependent protections are shown in Figure 1; for each of the panels, lane 1 shows the modification pattern for drug-free ribosomes, while the other numbered lanes show the effects of the various antibiotics. Since pactamycin had to be dissolved in ethanol, lane 7 shows the results of probing ribosomes treated with ethanol alone (5%) as a control against possible effects caused by the solvent. Pactamycin and kasugamycin, two inhibitors of translational initiation, both protect bases in 16S rRNA that are also protected by P-site binding of tRNA (Moazed and Noller, 1990). Pactamycin protects G693 at both its Nl and N7 positions (as inferred by protection from both kethoxal and DMS modification), and C795 at N3. Kasugamycin 3099

J.Woodcock et al.

..

...... .,e

.S ~~~~~~~~~~~~~ .~~?'. ~~

g. .

r.I

,d

ir.:

io

.:

Fig. 1. Autoradiographs showing protection of bases in 16S rRNA caused by binding of antibiotics to 70S ribosomes. A and G are dideoxy sequencing lanes. Lane K, unmodified 70S ribosomes. Lanes 1-7 are 70S ribosomes modified in the presence or absence of the various ligands. Lane 1, no antibiotic; lane 2, kasugamycin; lane 3, myomycin; lane 4, apramycin; lane 5, neamine; lane 6, pactamycin; lane 7, ethanol (5 per cent). Chernical modification was done with kethoxal (A and D) and dimethyl sulfate (B, C, E and F); aniline-induced strand scission was performed for DMS-modified samples in B and F, to identify N7 methylation of guanine.

protects A794 at N I and G926 at N I and, in addition, causes enhanced reactivity of C795 at its N3 position. Apramycin and neamine, both of which induce miscoding and inhibit translocation, protect the NI of A1408 and the N7 positions of G1491 and G1494. These nucleotides are positioned at the base of the penultimate stem of the secondary structure of 16S rRNA (Figure 2A) in and around the location of the tRNA A-site footprint (Moazed and

Noller, 1990).

The results obtained with myomycin were very unexpected since this antibiotic is regarded as having a mode of inhibitory action that is identical to that of streptomycin (Davies et al., 1988). Myomycin gives only a very weak, although reproducible, protection of a single base-AI408. We have failed to detect strong protection or enhancement of any bases in 16S rRNA by this drug. The data for protection of bases in 16S rRNA by antibiotics and tRNA are s-immarized in Table I.

Discussion

Figure 2 summarizes the chemical footprinting results for A- and P-site tRNA on 16S rRNA (Moazed and Noller, 1990) and compares them with the corresponding footprints obtained from binding antibiotics in this paper as well as from earlier work (Moazed and Noller, 1987a). The two sets of data show a clear correlation. 3100

-Edeine, pactamycin and kasugamycin, all of which inhibit translational initiation, protect sites that are also protected by P-site tRNA. Since initiation involves binding of initiator tRNA to the 30S P-site, our footprinting results can account in a straightforward way for the mode of action of these three drugs. While edeine protects four of the strongly protected P-site bases (G693, A794, C795 and G926), the other two drugs each affect a different subset. Pactamycin protects G693, although in a manner that is clearly distinguishable from that of edeine. Thus, whereas the latter protects the N I position of G693, pactamycin protects both positions N I and N7. In addition, although pactamycin protects C795, it has no protective effects on either A794 or G926. In contrast, kasugamycin protects A794 and G926, and causes enhanced DMS reactivity at C794, but has no effect on G693. Thus, these three antibiotics that are all known to affect initiation of protein synthesis, a process that involves the function of the ribosomal P site, interact with the 16S rRNA P site in exquisitely different ways. These observations may be attributed to the fact that the three drugs have distinctly different chemical structures (Cundliffe, 1981), which are reflected in inhibitory actions that have evolved independently to affect the same functional target site on ribosomes. In spite of the wide separation of these four P-site bases in the secondary structure, they are believed to be relatively close to each other as a result of the tertiary folding of 16S

Antibiotics and 16S rRNA

A

p

Fig. 2. Locations of antibiotic-protected bases in 16S rRNA (Moazed and Noller, 1987a; this work), compared with bases protected by tRNA bound to the ribosomal A- and P-sites (Moazed and Noller, 1990). Vertical arrows indicate enhancement of reactivity. On the left are shown tRNAprotected bases, and on the right are the corresponding antibiotic protections. Apr, apramycin; Ede, edeine; Hyg, hygromycin; Ksg, kasugamycin; Myo, myomycin; Nea, neamine; Neos, neomycin and related aminoglycosides; Pct, pactamycin.

rRNA in the ribosome. Figure 3 shows the location of the 690, 790 and 930 region stems in the three-dimensional model of 16S rRNA proposed by Stern et al. (1988b). In the latter study, it was proposed that these three stems surround the cleft of the 30S subunit, where the anticodon stem-loop of P-site tRNA is bound during its interaction with mRNA. Direct evidence for the proximity of two of these loops comes from intramolecular crosslinking of 16S rRNA in 30S ribosomal subunits between bases located approximately at positions 695 and 794 or 799, by treatment with bis-(2-chloroethyl)methylamine (Atmadja et al., 1986). If the protection of bases by edeine, pactamycin and kasugamycin is the result of direct contact between these drugs and the protected bases, then our data can be taken as evidence supporting the mutual proximity of these three structural features: because of the relatively small molecular dimensions of the antibiotics, positions 693, 794-795 and 926 must be relatively close to each other for any two of the RNA sites to be simultaneously in contact with the same

drug molecule. In the model of Stern et al. (1988b), the distances are 24 A (G693/A794) and 25 A (G693/G926 and A794/G926), respectively. Within the uncertainty of the model, these distances are consistent with possible direct contact between the antibiotics and the protected bases. Apramycin and neamine, which both cause miscoding and inhibit translocation (Perzynski et al., 1979; Delcuve et al., 1987), give footprints in the same tightly constrained region of the 16S rRNA secondary structure that is located at the base of the penultimate stem. This highly conserved region is the site where certain bases are protected by the anticodon stem-loop region of A-site-bound tRNA (Moazed and Noller, 1986, 1990), accounting for the miscoding effects of these drugs. Their footprints are, therefore, similar to those of the neomycin-related antibiotics neomycin, kanamycin, gentamicin and paromomycin, and of hygromycin (Moazed and Noller, 1987a), whose effects are also to cause miscoding and inhibit translocation. A simple explanation for their miscoding effects is that they perturb 3101

J.Woodcock et al. Table I. Protection of bases in 16S rRNA by antibiotics and tRNA Ligand

Protected bases G693

A794

C795

G926

+ + +

+ + -

+ + +

+ +

+

E

+

P-site tRNA Edeine Pactamycin

(N 1,N7) Kasugamycin

-

G1491

G1494

(N7)

(N7)

4

-

+

+ + + +

+ + + + +

A1408 A-site tRNA Neomycins Hygromycin Apramycin Neamine Myomycin

E + i 4-

Symbols indicate (+) protection or (±) weak protection from or (E) enhancement of chemical modification of the designated bases. Ni and N7 refer to the site of modification of guanine bases that is protected. Data are from Moazed and Noller (1986, 1990) (A-site and P-site tRNA); Moazed and Noller (1987a) (edeine, neomycins and hygromycin), and this paper (pactamycin, kasugamycin, apramycin, neamine and myomycin).

G 693

>

A 794

Fig. 3. Model for the folding of 16S rRNA in 30S ribosomal subunits (Stem et al., 1988b), showing the locations of bases that are protected in common by P-site-bound tRNA and by edeine, pactamycin and kasugamycin (cf. Table I). Residues G693, A794, C795 and G926 are all protected by tRNA as well as by edeine. Pactamycin protects G693 and C795, while kasugamycin protects A794 and G926, and enhances C795.

the site in the 30S ribosomal subunit where anticodon -codon recognition takes place, perhaps by strengthening nonspecific interactions between A-site tRNA and the ribosome. Such a mechanism might also help to explain their inhibitory effect on translocation. The results obtained for myomycin are surprising and intriguing. This antibiotic shares certain structural features with kasugamycin, streptothricin and streptomycin, but the inhibitory actions of myomycin are apparently identical to

3102

those displayed by streptomycin. Both drugs affect translational accuracy and inhibit the initiation of protein synthesis (Davies et al., 1988). Myomycin and streptomycin phenotypically suppress the same range of nonsense and missense mutations of E. coli. Furthermore, single amino acid changes in ribosomal protein S12 can cause cross resistance to both drugs (Davies et al., 1988). These close similarities provide strong evidence that myomycin and streptomycin share the same ribosomal target site-a conclusion reinforced by the studies of Montandon et al. (1986) which demonstrate that a C to U base change at position 912 in E. coli 16S rRNA gives streptomycin resistant strains that are cross resistant to myomycin (Davies et al., 1988). Nevertheless, although the primary sites protected within 16S rRNA by streptomycin are in the 915 region, we find that myomycin protects no bases within this region (data not shown). Instead, it gives very weak protection of A1408-a result which suggests that myomycin, like neomycin and related aminoglycosides, might perturb A-site codon - anticodon interaction in some way. It may be that structural differences between streptomycin and myomycin, although allowing an identical inhibitory action, cause the drugs to bind to the same functional domain on ribosomes but at different contact points. However, we note that the protection of bases 911-915 by streptomycin is incomplete, even when its binding to the ribosome is saturated (D.Moazed, unpublished; T.Powers and H.F.Noller, submitted). Similarly, protection of A1408 by myomycin is also incomplete at high drug concentrations, suggesting that the observed protections by both drugs are indirect, and not due to direct contact with 16S rRNA. If this is the case, we conclude that the true binding sites for streptomycin and myomycin have yet to be identified. An important unsolved problem is whether the observed protection of bases by antibiotics and tRNA (Table I) results from direct ligand -rRNA contact, or from ligand-induced conformational changes in 16S rRNA. All four of the previously reported protections by edeine mimicked four Psite tRNA protections (Moazed and Noller, 1987a, 1990), a result that could be interpreted in terms of a conformational change induced by either edeine or tRNA. However, the results of the present study argue against such a simple interpretation. Thus, two of the four sites (G693 and C795) are protected by pactamycin, whereas this drug has no effect on the reactivity of the other two sites; therefore, there would have to be two conformational changes. The finding that, unlike for tRNA and edeine, pactamycin protects not only the NI position of G693 but also the N7 position, introduces further complications, and would require that the putative conformational change is somehow different in the case of this drug. Furthermore, pactamycin protects C795 without affecting the adjacent base A794 (which is protected by kasugamycin, edeine and tRNA). Here again, the results are more simply accounted for in terms of direct contact than by multiple, independent conformational changes. Such an interpretation is consistent with our current understanding of 16S rRNA structure. Another question, raised implicitly by these studies, is the role of ribosomal proteins in the interactions of antibiotics with ribosomes. Antibiotic resistance is well known to be conferred by mutations in r-proteins. In addition to the possibility that the proteins may contribute all or part of certain drug-binding sites, their role(s) could, alternatively, be in the modulation of the detailed conformation of rRNA

Antibiotics and 16S rRNA

(Allen and Noller, 1989; Stern et al. 1989). In at least one case, that of thiostrepton (Cundliffe, 1990), it has been shown that the rRNA itself is capable of binding the drug in the absence of protein. It remains to be seen how far this will extend to the many other antibiotics that interact with ribosomes.

Stern,S., Moazed,D. and Noller,H.F. (1988a) Methods Enzymol., 164, 481 -489. Stem,S., Weiser,B. and Noller,H.F. (1988b) J. Mol. Biol., 204, 447-481. Stern,S., Powers,T., Changchien,L.-M. and Noller,H.F. (1989) Science,

244, 783-790.

Tai,P.-C., Wallace,B.J. and Davis,B.D. (1973) Biochemistry, 12, 616-620. Received on May 7, 1991; revised on June 17, 1991

Materials and methods Antibiotics were obtained from the following sources: kasugamycin, Sigma; pactamycin, The Drug Synthesis and Chemistry Branch, Division of Cancer Treatment, National Cancer Institute; apramycin, Lilly Research Laboratories, Indianapolis; neamine, Upjohn Company, Kalamazoo, MI; myomycin, Parke-Davis, Ann Arbor, MI. Ribosomes were prepared from E. coli MRE600 as described previously (Moazed and Noller, 1989). Antibiotics were bound to 100 pmol of 70S ribosomes in 100 Id of 80 mM potassium cacodylate (pH 7.2) 20 mM MgCl2, 100 mM NH4Cl, 1 mM DTT, 0.5 mM EDTA for 30 min at 37°C and then for 10 min on ice. Antibiotic concentrations were: kasugamycin, 200 1M; myomycin, 50 zM; apramycin, 100 /tM; neamine, 100 /M; pactamycin 100 itM. Chemical modification was performed by addition of dimethyl sulfate or kethoxal followed by incubation at 37°C for 10 min, as described previously (Moazed and Noller, 1987a). The rRNA was then isolated by phenol extraction, precipitated with ethanol and the sites of chemical modification identified by primer extension as detailed previously (Stem et al., 1988a).

Acknowledgements J.W. and M.C. gratefully acknowledge the financial support of the Science and Engineering Research Council, the Nuffield Foundation and the Wellcome Trust. D.M. and H.F.N. were supported by NIH grant No. GM-17129. We thank Lilly Research Laboratories, the Drug Synthesis and Chemistry Branch of the National Cancer Institute, Upjohn Laboratories, Parke-Davis, Inc., E.De Stasio and A.E.Dahlberg for gifts of antibiotics. We thank Bryn Weiser for computer graphics drawings.

References Allen,P.N. and Noller,H.F. (1989) J. Mol. Biol., 208, 457-468. Atmadja,J., Stiege,W., Zobawa,M., Greuer,B., Osswald,M. and Brimacombe,R. (1986) Nucleic Acids Res., 14, 659-673. Bollen,A. (1978) Biochem. J., 174, 1-7. Cohen,L.B., Herner,A.E. and Goldberg.I.H. (1969) Biochemistnr, 8, 1312-1326. Cundliffe,E. (1981) In Gale,E.F., Cundliffe,E.* Reynolds,P.E.. Richmond,M.H. and Waring,M.J. (eds), The Molecular Basis of Antibiotic

Action. John Wiley and Sons, New York, pp. 402 -457. Cundliffe,E. (1990) In Hill,W.E., Dahlberg,A.E., Garrett,R.A., Moore,P.B., Schlessinger,D. and Warner,J.R. (eds), The Ribosome, Structure, Function & Evolution. American Society for Microbiology, Washington, DC, pp. 479-490. Davies,J., Cannon,M. and Mauer,M.B. (1988) J. Antibiotics, 41, 366-372. Delcuve,G., Cabezon,T., Herzog,A., Cannon,M. and Bollen,A. (1978) Biochem. J., 174, 1-7. De Stasio,E.A., Goringer,H.U., Tapprich,W.E. and Dahlberg,A.E. (1988) In Tuite,M.F., Picard,M. and Bolotin-Fukuhara,M. (eds), Genetics of

Translation: New Approaches. Springer-Verlag, pp. 17-42. Moazed,D. and Noller,H.F. (1986) Cell, 47, 985-994. Moazed,D. and Noller,H.F. (1987a) Nature, 327, 389-394. Moazed,D. and Noller,H.F. (1987b) Biochimie, 69, 879-884. Moazed,D. and Noller,H.F. (1989) Cell, 87, 585-597. Moazed,D. and Noller,H.F. (1990) J. Mol. Biol., 211, 135-145. Moazed,D. and Noller,H.F. (1991) Proc. Natl. Acad. Sci. USA, in press. Montandon,P.E., Wagner,R. and Stutz,E. (1986) EMBO J., 5, 3705-3708. Noller,H.F., Moazed,D., Stem,S., Powers,T., Allen,P.N., Robertson,J.M., Weiser,B. and Triman,K. (1990) In Hill,W.E., Dahlberg,A.E., Garrett,R.A., Moore,P.B., Schlessinger,D. and Warner,J.R. (eds), The Ribosome. Structure, Function & Evolution. American Society for Microbiology, Washington, DC, pp. 73-92.

Okuyama,A., Machiyama,N., Kinoshita,T. and Tanaka,N. (1971) Biochemll. Biophys. Res. Commun., 43, 196-199. Perzynski,S., Cannon,M., Cundliffe,E., Chahwala,S.B. and Davies,J. (1979) Eur. J. Biochem., 99, 623-628.

3103