Role of antibiotic ligand in nascent peptide ... - Semantic Scholar

Role of antibiotic ligand in nascent peptidedependent ribosome stalling Nora Vázquez-Laslopa,1, Dorota Klepackia, Debbie C. Mulhearna, Haripriya Ramua, Olga Krasnykhb, Scott Franzblaub, and Alexander S. Mankina,1 a Center for Pharmaceutical Biotechnology, University of Illinois, 900 South Ashland Avenue, Chicago, IL 60607; and bInstitute for Tuberculosis Research, College of Pharmacy, University of Illinois, 833 South Wood Street, Chicago, IL 60612

Edited by Charles Yanofsky, Stanford University, Stanford, CA, and approved May 16, 2011 (received for review March 3, 2011)

Specific nascent peptides in the ribosome exit tunnel can elicit translation arrest. Such ribosome stalling is used for regulation of expression of some bacterial and eukaryotic genes. The stalling is sensitive to additional cellular cues, most commonly the binding of specific small-molecular-weight cofactors to the ribosome. The role of cofactors in programmed translation arrest is unknown. By analyzing nascent peptide- and antibiotic-dependent ribosome stalling that controls inducible expression of antibiotic resistance genes in bacteria, we have found that the antibiotic is directly recognized as a part of the translation modulating signal. Even minute structural alterations preclude it from assisting in ribosome stalling, indicating the importance of precise molecular interactions of the drug with the ribosome. One of the sensors that monitor the structure of the antibiotic is the 23S rRNA residue C2610, whose mutation reduces the efficiency of nascent peptide- and antibiotic-dependent ribosome stalling. These findings establish a new paradigm of the role of the cofactor in programmed translation arrest in which a small molecule is recognized along with specific nascent peptide sequences as a composite structure that provokes arrest of translation. A similar mechanism could be used by the ribosome to sense a variety of cellular metabolites. erythromycin ∣ ketolides

T

he ability to monitor the structure of the nascent peptide is a fundamental but poorly understood property of the ribosome. Peptide audit takes place in the nascent peptide exit tunnel through which the newly synthesized proteins leave the ribosome. The tunnel is ca. 100 Å long and 10–20 Å wide. It starts at the peptidyl transferase center (PTC), spans the body of the large ribosomal subunit, and opens at its opposite side. Specific nascent peptide sequences can elicit functional ribosomal response by interacting with the tunnel elements (reviewed in ref. 1). One of the manifestations of such a response is nascent peptide-dependent ribosome stalling, which plays a key role in control of expression of a number of bacterial and eukaryotic genes (2, 3). The sequences that direct ribosome stalling are confined to the C-terminal segments of the nascent peptides, indicating that sensors that interrogate the peptide structure are located in the tunnel segment proximal to the PTC. Several 23S rRNA nucleotides of the upper chamber of the tunnel as well as amino acid residues of proteins L4 and L22 have been shown to be involved in peptide recognition (4–9). Results of biochemical, genetic, and structural analyses argue that the “stalling” nascent peptides establish idiosyncratic contacts with these tunnel sensors (4–8, 10–12). Once the sensors detect the presence of the stalling peptide sequence in the tunnel, the signal is relayed to the PTC active site, possibly via conformational change in the ribosome structure. Inhibition of peptide bond formation causes translation arrest (4, 5, 7, 10, 12). The efficiency of programmed translation arrest that regulates gene expression depends on specific cellular cues. Often, binding of a small-molecular-weight cofactor is required for the formation of the stable stalled ribosome complex (SRC). Thus, ribosome stalling at the last sense codon of the tnaC regulatory

10496–10501 ∣ PNAS ∣ June 28, 2011 ∣ vol. 108 ∣ no. 26

ORF, which controls expression of the tryptophanase operon, depends on binding of free tryptophan to the ribosome (9, 13). Programmed translation arrest at the arginine attenuator peptide gene in fungi depends on concentration of arginine (14, 15), whereas ribosome stalling at the regulatory ORF of the mammalian gene of S-adenosylmethionine decarboxylase is enhanced at high concentrations of spermidine (16). However, the binding site of the cofactors and their role in establishing the translationarrest state have not been convincingly established in any of these cases. An important class of genes regulated by cofactor-dependent programmed translation arrest are the inducible genes of antibiotic resistance (17, 18). The best-characterized example of these medically relevant genes is ermC, the gene that confers resistance to erythromycin and other macrolide antibiotics. Macrolide drugs (Figs. 1 and 2) bind in the exit tunnel in the vicinity of the PTC and inhibit translation by narrowing the tunnel and provoking peptidyl-tRNA drop-off at the early rounds of translation (19, 20). Erm methyltransferase renders cells resistant to macrolides by dimethylating A2058 of 23S rRNA located in the site of the drug binding (17). Expression of ermC is controlled by a short upstream regulatory ORF ermCL (Fig. 3A) (21, 22). In the absence of the drug, ermC is translationally attenuated, whereas ermCL is constitutively translated. When the inducing macrolide antibiotic is present, translation of ermCL is impeded. A significant fraction of the ribosomes loose peptidyl-tRNA early in translation and dissociate from mRNA (20). Yet some drug-bound ribosomes manage to reach the ninth codon of ermCL. Once the N-terminal 9-amino acid sequence of the ErmCL peptide is polymerized and is lodged in the exit tunnel alongside the bound macrolide molecule, the ribosome stalls at the ninth codon of ermCL. This alters the mRNA secondary structure and activates the expression of the downstream ermC gene. Programmed ribosome stalling requires the sequence of the four C-terminal amino acid residues Ile-Phe-Val-Ile (IFVI) of the ErmCL nascent peptide and is absolutely dependent on the binding of the antibiotic cofactor to the ribosome (7, 23). Mutations at several conserved 23S rRNA residues located in the tunnel adjacent to the critical IFVI sequence of the peptide abolish SRC formation, indicating that these nucleotides are directly involved in sensing the peptide (7, 24). The aperture of the unobstructed exit tunnel is wide enough for the nascent peptide to avoid contacts with rRNA sensors. However, when the antibiotic molecule is bound in the tunnel, the peptide would be compelled to come in direct contact with the rRNA residues involved in nascent peptide recognition. Author contributions: N.V.-L. and A.S.M. designed research; N.V.-L., D.K., D.C.M., and H.R. performed research; O.K. and S.F. contributed new reagents/analytic tools; N.V.-L. and A.S.M. analyzed data; and N.V.-L. and A.S.M. wrote the paper. The authors declare no conflict of interest. This article is a PNAS Direct Submission. Freely available online through the PNAS open access option. 1

To whom correspondence may be addressed. E-mail: [email protected] or [email protected].

This article contains supporting information online at www.pnas.org/lookup/suppl/ doi:10.1073/pnas.1103474108/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1103474108

This observation offered one possible role for the antibiotic cofactor in the mechanism of ribosome stalling—that of a deflector that directs peptides to interact with the tunnel sensors (6, 7). It is unknown, however, whether the necessity of the drug for programmed ribosome stalling is limited to this simple task or whether its purpose expands beyond being a mere space filler. In contrast to other small-molecular-weight coeffectors of ribosome stalling, whose binding sites in the ribosome are largely unknown, the site of binding of macrolides in the ribosome has been characterized with high precision (25–30). When bound to the ribosome, macrolides establish a network of hydrophobic and hydrogen bonding interactions with 23S rRNA residues in the exit tunnel. The macrolactone ring lays flat against the tunnel wall, whereas C5- and C3-linked sugar residues (desosamine and cladinose, respectively, in erythromycin) that play an important role in drug binding and action (27–30) protrude toward the PTC Vázquez-Laslop et al.

(Fig. 1 A and B) and closely approach residues A2058, A2059, and C2610. Importantly, the capacity of macrolide antibiotics to induce programmed ribosome stalling is distinct from their function as protein synthesis inhibitors (31–33). Thus, the lack of C3 cladinose sugar in ketolides, which represent the latest generation of medically important macrolides, dramatically impairs their ribosome stalling capacity and, thus, erm-inducing activity while fully preserving the ability to inhibit protein synthesis and interfere with cell growth (34, 35). However, the general relationships between the structure of the drug and their ability to promote SRC formation remain largely unknown. To gain insights into the role of cofactors in programmed ribosome stalling, we investigated how alterations in the antibiotic structure affect its ability to support SRC formation at ermCL. We further tested how mutations of specific rRNA nucleotides that interact with the drug, but not with the nascent peptide, PNAS ∣

June 28, 2011 ∣

vol. 108 ∣

no. 26 ∣

10497

BIOCHEMISTRY

Fig. 1. Differential effect of macrolide antibiotics on nascent peptide-dependent ribosome stalling. (A) Chemical structure of a prototype macrolide antibiotic, erythromycin. (B) Binding of macrolides in the exit tunnel of vacant E. coli ribosome. Position of erythromycin (ERY) (red) was experimentally determined by X-ray crystallography (29). The other drugs were modeled in the macrolide binding site by energy minimization (see SI Appendix). (C, Top) Placement in the exit tunnel (viewed from the PTC down the tunnel) and structures of macrolides ERY, clarithromycin (CLR), telithromycin (TEL), and ITR compounds used in the toeprinting experiments. C3 cladinose sugars of ERY and CLR are circled with a broken line; an arrow indicates the other C3 substitutions. The aperture of the tunnel at the level of macrolide binding is shown as green mesh. The antibiotic molecule in the tunnel is colored salmon with the C3 side chain highlighted in red. (Bottom) Toeprinting analysis of the stalled complex formed on ermCL mRNA on its in vitro translation in the presence of macrolide antibiotics (50 μM) with varying C3 side chains. Ribosome stalling signals on the gel are indicated by arrows. The ninth codon of ermCL located in the P site of the stalled ribosome is boxed. Sequencing lanes are marked.

stalling at the ninth codon of ermCL (Fig. 1C, lane 2). In contrast, the ketolide telithromycin, in which the cladinose side chain is replaced with a keto group, does not support SRC formation (Fig. 1C, lane 4) (7). If the functional role of the cladinose moiety in ribosome stalling is limited to that of a mere space filler, then replacement of this sugar with different bulky side chains, including those whose size is comparable to that of cladinose, should preserve the ability of the drug to promote stalling. To test this possibility, we compared the effect of erythromycin, its O6methylated cousin clarithromycin, and derivatives of clarithromycin with varying C3 side chains on SRC formation in vitro. Footprinting analysis revealed that all of the tested compounds bind in the well-known site of macrolide action, where they protect residues A2058 and A2059 from modification with dimethyl sulfate (26) (Fig. S1A). Computer modeling of the compounds in the erythromycin-binding site showed that their varying C3 side chains would protrude into the space occupied by the cladinose of erythromycin (Fig. 1 B and C). Although all of the compounds readily inhibited in vitro protein synthesis (Fig. S1C), only the cladinose-containing drugs (erythromycin and clarithromycin) promoted SRC formation (Fig. 1C, lanes 2 and 3). In contrast, compounds in which cladinose sugar was replaced with other side chains, including substituents as bulky as cladinose (ITR162, ITR163), could not stall the ribosome at the ninth codon of ermCL ORF (Fig. 1C, lanes 5–8). This result suggested that the role of the drug and, more specifically, of the C3 cladinose sugar in SRC formation is not limited to simply narrowing the tunnel. Rather, it appears that the antibiotic molecule can play a more refined role in generation of SRC. Minor Modifications of the Cladinose Structure in Macrolide Antibiotics Interfere with Ribosome Stalling. A more sophisticated role of

Fig. 2. Modifications of the C3 cladinose of macrolide antibiotics are detrimental for ribosome stalling at ermCL. (A) Effects of C3 cladinose modifications on the ability of antibiotic to induce ribosome stalling at ermCL mRNA, determined by toeprinting assay. (Top) Chemical structures of macrolides with modifications of the C3 sugar. Modifications of the cladinose in the ITR compounds and the inverted C 5″ stereocenter of the oleandrose in oleandomycin (OLN) are indicated by arrows. (Bottom) The toeprint signals corresponding to SRC occupying the ermCL ninth codon (arrows). (B) Overlay of the experimental erythromycin structure in the ribosome exit tunnel (29) and the C3 side chains of the modeled structures of ITR054, ITR074, and oleandomycin. The erythromycin molecule is colored salmon and the C3 side chains of the other three compounds are colored green (ITR074), blue (ITR054), and cyan (OLN).

affect establishment of the translation-arrest state. Our data argue that, together with the sequence of the ErmCL nascent peptide, the structure of the antibiotic cofactor is monitored by the ribosome, that the drug is recognized as a part of the stalling signal, and that the ribosome integrates the response of the tunnel sensors dedicated to recognition of the nascent peptide and the cofactor. Results Substitution of C3 Cladinose with Bulky Side Chains Does Not Restore Ribosome Stalling. The extension inhibition assay, commonly

known as toeprinting (36), makes it possible to directly monitor SRC formation on mRNA (7, 37, 38). When the ermCL ORF is translated in vitro, erythromycin efficiently promotes ribosome 10498 ∣


the antibiotic in inducing ribosome stalling could involve specific chemical interactions between the drug molecule, especially its cladinose side chain, and either the nascent peptide or the ribosome, as proposed earlier by Weisblum (17). If this is the case, even minimal alterations in the structure of cladinose would be expected to negatively affect the capacity of the antibiotic to induce stalling. To test this hypothesis, we assessed the effects of fairly minor changes in the structure of the C3 sugar on SRC formation (Fig. 2). In ITR054, one of the clarithromycin derivatives tested in these experiments, a small acetyl group is attached at the 4″ hydroxyl of cladinose; in the other derivative, ITR074, this acetyl group is replaced with acetylaniline. In addition, we tested oleandomycin, whose C3 oleandrose is structurally similar to cladinose except for an inverted stereocenter at the C 3″ carbon (several other minor structural differences between oleandomycin and erythromycin, not pertaining to the C3 side chain, are not expected to interfere with stalling). Remarkably, even these minimal modifications of the cladinose sugar significantly reduced the capacity of the drug to promote ribosome stalling at the ninth codon of ermCL. These results emphasize the notion that precise atomic interactions involving the antibiotic are critically important for formation of SRC at the ermCL ORF. Antibiotic Structure Is Recognized by the Ribosome as One of the Stalling Cues. When erythromycin binds to the ribosome, the cladi-

nose sugar is positioned close to the tunnel wall, coming into contact with the 23S rRNA residue C2610. The short distance (3.5 Å) allows for direct interactions between the 3″ methyl group of cladinose and the hydrophobic face of the pyrimidine nitrogen base (Fig. 1B). Such proximity suggests that the presence of cladinose and its spatial placement could be monitored by C2610. We tested the importance of these interactions by analyzing the effects of mutations of C2610 on programmed ribosome stalling at the ermCL ORF. C2610 in the plasmid-borne 23S rRNA gene was substituted with Tor A (the G mutation was not viable), and the mutant ribosomes were expressed as a homogeneous Vázquez-Laslop et al.

Fig. 3. Mutation of C2610 affects drug-dependent SRC formation at ermCL. (A) Structures of the inducible ermCL-ermC operon and the pZα101t reporter. (B) Erythromycin-induced expression of the reporter in E. coli cells carrying wild-type or mutant ribosomes with substitution at the 23S rRNA residue 2610. The paper disk contains 0.5 mg of erythromycin. (C and D) Toeprinting analysis of erythromycin-dependent stalling of wild-type and mutant ribosomes at the ermCL ORF and (E) drug-independent ribosome stalling at the secM ORF. Bands representing stalled ribosomes are indicated by arrows. In D the indicated samples were supplemented with borrelidin (Bor) during translation, which resulted in the ribosome stalling at codon 10 of ermCL (gray arrows and gray box) due to the lack of Thr-tRNA. In E, control samples contained thiostrepton (Ths), which inhibits initiation of translation under our experimental conditions (6). The results of quantification of the toeprint bands (an average of two experiments in C) are shown below the gels, with the error bars representing experimental deviation.

Vázquez-Laslop et al.

(Fig. 3A) (6). Although replacement of C2610 with A had little effect, the C2610U transition significantly reduced erythromycindependent expression of lacZ (Fig. 3B), arguing that the proper interaction between the drug and rRNA residues is required for translation arrest. The reduced inducibility of the reporter in the C2610U mutant was further confirmed in liquid cultures (Fig. S3). To verify that the C2610U mutation indeed affects the ability of the ribosomes to recognize stalling cues, we compared the efficiency of SRC formation by wild-type and mutant ribosomes in vitro. Although the C2610U mutation had no obvious effect on translation activity of the ribosome in the cell-free system (Fig. S2A), it notably reduced erythromycin-dependent ribosome stalling at ermCL (approximately twofold) (Fig. 3C, Left). To ensure that this difference in stalling is not simply due to potential variations in translation capacity of the wild-type and mutant ribosomes, we repeated toeprinting with an internal standard. For that we employed borrelidin, an antibiotic that inhibits threonyl aminoacyl-tRNA synthetase (41). In the presence of borrelidin and erythromycin, the translating ribosomes could be captured either at codon 9 (due to erythromycin-dependent ribosome stalling) or at codon 10 (due to the lack of Thr-tRNA in the cell-free system) (Fig. 3D). The relative intensities of the toeprinting bands corresponding to Ery-dependent stalling at the ninth codon (black arrows) and Ery-independent stalling at the 10th codon (gray arrows) therefore reflect the efficiency of programmed ribosome stalling, irrespective of the general activity of the ribosome preparation. The results of this experiment confirmed that the C2610U mutant ribosomes were less prone to stall in erythromycin-dependent manner at codon 9 of ermCL. The fact that translation can be elongated for at least one codon beyond codon 9 if the ribosome has the C2610U alteration shows that the ErmCL- and erythromycin-mediated stalling of the wild-type ribosome involves, at least in part, arrest of translation elongation that in turn can prevent rapid peptidyltRNA drop-off. Interestingly, the same C2610U mutation had no effect on programmed translation arrest at the secM ORF, which does not require the presence of antibiotic (Fig. 3E). Taken together, these data argue that C2610 could be an antibiotic sensor that specifically responds to the presence of the properly positioned cladinose sugar of the inducing macrolide drug and that this recognition contributes to ribosome stalling at the ermCL ORF. Discussion Previous studies (6, 7) illuminated one possible role of the macrolide cofactor in programmed ribosome stalling: Binding of the antibiotic narrows the exit tunnel and thus ensures that the critical sequence of the nascent peptide interacts with the tunnel sensors responsible for peptide recognition. Our findings argue that functions of the antibiotic in this mechanism are not simply those of a roadblock, but are far more refined. C3 cladinose of the drug molecule is essential for SRC formation. It cannot be replaced with other similarly bulky side chains (Fig. 1). Not even minor modifications in the structure of the C3 sugar are tolerated (Fig. 2), suggesting that not only a mere obstruction of the tunnel by the drug but precise atomic interactions of the antibiotic with the nascent peptide or with the ribosome sensors are required for generating the stalling signal. We identified one such putative drug sensor in the ribosome tunnel. The mutation of C2610 to U, which has no effect on drug-independent ribosome stalling at the secM regulatory ORF, significantly reduces efficiency of the drug- and nascent peptide-dependent ribosome stalling at ermCL (Fig. 3). None of the “trivial” explanations can account for the effect of the C2610U mutation on programmed translation arrest. The mutant ribosomes are fully active (Fig. S2). The mutation does not alter susceptibility of cells to erythromycin. It has no important effect on either drug binding to the ribosome or its dissociation rate (Fig. S4). C2610 is in close PNAS ∣

June 28, 2011 ∣

vol. 108 ∣

no. 26 ∣

10499

BIOCHEMISTRY

population in the Escherichia coli strain SQ171 lacking chromosomal rRNA alleles (39, 40). Importantly, mutant cells did not show any appreciable growth defect or resistance to erythromycin. The minimal inhibitory concentration of the drug was 32–64 μg∕mL for the wild type and both mutant strains, indicating that the drug readily binds to ribosomes carrying mutations at position 2610. This conclusion was further verified for the most interesting C2610U mutant (see below) by in vitro experiments where efficient binding of erythromycin to the mutant ribosome was revealed by footprinting (Fig. S1B). Furthermore, the drug inhibited translation catalyzed by wild-type and mutant ribosomes equally well (Fig. S2B). To analyze the effect of mutations on ribosome stalling in vivo, we used a previously developed pZα101t reporter in which expression of the 5′-terminal portion of the β-galactosidase lacZ gene encoding α-peptide is controlled by erythromycin-induced ribosome stalling at the ermCL ORF

Fig. 4. Dual function of the antibiotic cofactor in programmed ribosome stalling. (A) First, binding in the ribosome exit tunnel of a macrolide antibiotic (salmon mesh) ensures interaction of the ermCL stalling nascent peptide (cyan) with tunnel sensors, including U1782 and A2062 (blue). (B and C) Second, the antibiotic is recognized as a part of the composite stalling signal in the tunnel. Specific sensors (experimentally identified C2610 and A2503 as well as suspected A2058 and A2059) (red) are positioned to interact specifically with the drug molecule. The nascent peptide (cyan) and antibiotic (salmon) are shown in surface representation. (B) Side view of the peptide– antibiotic complex; (C) top view from the PTC down the tunnel.

contact with the antibiotic molecule but can hardly interact with the nascent peptide (Fig. 4). Thus, we believe that together with tunnel sensors, e.g., A2062 and U1782, that do interact with the ErmCL nascent peptide (7, 24) (Fig. S5), C2610 participates in recognizing the stalling cues by monitoring the structure of the antibiotic molecule. An additional line of evidence points to C2610 as a sensor of the drug structure. The mutation of the neighboring rRNA residue U2609 to C increases sensitivity of E. coli to cladinose-containing macrolides but not ketolides (42). Because the base of U2609 is separated by more than 7 Å from the C3 cladinose of the macrolide (29), the effect of the mutation must be allosteric, most likely via its immediate neighbor C2610. In turn, U2609 could be part of a signal relay pathway that links the tunnel antibiotic sensor to the peptidyl transferase active site (10). This pathway may supplement the one operating on the opposite wall of the tunnel that directly responds to the nascent peptide structure (6, 7, 10). C2610 is likely not the only sensor exploited by the ribosome for discerning stalling cues from the antibiotic structure. We have shown that mutations of a conserved adenine at position 2503 10500 ∣


alleviated SRC formation at ermCL (6). Because A2503 is shielded from the nascent peptide by the macrolide molecule bound in the ribosome tunnel and can hardly participate in direct interaction with the ErmCL nascent peptide, we proposed that A2503 could be part of a signal relay pathway linking the exit tunnel sensors to the peptidyl transferase active site. In view of our findings and given the proximity of A2503 to the drug, it is also possible that, similar to C2610, A2503 serves as a tunnel sensor that recognizes the presence and structure of antibiotic. This model does not necessarily contradict involvement of A2503 in signal relay: Interaction with the drug may be required for placing A2503 in an orientation suitable for relaying the stalling signal to the PTC. Two other adenines, A2058 and A2059, also emerge as good candidates for the role of drug sensors. These residues form a hydrophobic cleft that interacts with the C5 desosamine sugar of erythromycin (27–30). The nascent peptide could alter the antibiotic pose and, as a consequence, placement of rRNA nucleotides that interact with the drug. Even minor changes in the position of A2058 or A2059 could affect functions of the peptidyl transferase active site (12), which would allow these residues to serve as additional triggers of the stalling response. Our finding that recognition of the antibiotic structure is required for SRC formation at ermCL principally changes the general view of how the stalling mechanism operates. In the previous model, the only entity thought to be recognized by the ribosome was the nascent peptide. The drug played a supportive role, presenting the critical peptide sequence to the tunnel sensors (Fig. 4A). The facts that the fine structure of the inducing antibiotic is critical for stalling and that the identities of rRNA nucleotides that contact the drug affect the efficiency of SRC formation argue that the ribosome actually recognizes the composite anatomy of the ErmCL nascent peptide and the inducing antibiotic (Fig. 4 B and C). Several tunnel sensors, e.g., A2062 and U1782, are committed to monitoring the structure of the peptide (Fig. 4 and Fig. S5). The other sensors, C2610, A2503, and possibly A2058 and A2059, are dedicated to recognition of the antibiotic molecule (Fig. 4 B and C). Generation of the arrest state of the ribosome may require integration of the stalling signals coming from the sensors interacting with both the small cofactor and the nascent peptide. This model illuminates a more general scenario of how the ribosome could monitor the presence and abundance of various small molecules in the cell. The rugged surface of the tunnel contains many cavities that could serve as binding sites for a number of putative stalling cofactors (43). Translation of the proper nascent peptide would provide the missing components of the composite signal and pause or even arrest ribosome progression. Alternatively, the nascent peptide in the tunnel could assist in organizing the binding site for a small molecule. Once bound to such a transiently formed binding site, the low-molecularweight cofactor can help to modulate the strength of the translation-arrest signal. The ensuing ribosomal pausing or stalling could be used to regulate translation in response to various cellular cues. Materials and Methods Antibiotics and Chemicals. Erythromycin was from Sigma, clarithromycin from USP, borrelidin from A. G. Scientific, ITR compounds 054, 074, 156, 162, 163, and 166 were synthesized as previously reported (44), telithromycin was from CEMPRA Pharmaceuticals, and oleandomycin was kindly provided by B. Weisblum (University of Wisconsin, Madison, WI). All of the chemicals were from Fisher Scientific and the enzymes were from Fermentas. Cell-Free Translation and rRNA Mutagenesis. The final concentration of all macrolide antibiotics in cell-free translation reactions or of borrelidin (Fig. 3D) was 50 μM. Inhibitory activity of antibiotics on cell-free translation was evaluated using the E. coli S30 extract system (Promega) by monitor-


ing translation of the firefly luciferase as directed by the manufacturer. Toeprinting analysis was carried out in the cell-free translation PURESYSTEM (Postgenomics Institute) (45) or in its Δ1 version supplemented with mutant ribosomes as described (6, 7). Ribosomes from mutant cells were prepared according to ref. 46. The pAM55 plasmid (to be described elsewhere), a simplified version of pLK35 (47), was used to introduce mutations at the 23S rRNA residue 2610 using the QuikChange Lightning Multi Site mutagenesis kit (Stratagene). The mutant plasmids were eventually transformed into E. coli strain

SQK15 (6), where a homogeneous population of mutant ribosomes was obtained as described (48). Modeling the ErmCL nascent peptide and antibiotics in the ribosome is described in SI Materials and Methods.

1. Tenson T, Ehrenberg M (2002) Regulatory nascent peptides in the ribosomal tunnel. Cell 108:591–594. 2. Morris DR, Geballe AP (2000) Upstream open reading frames as regulators of mRNA translation. Mol Cell Biol 20:8635–8642. 3. Ito K, Chiba S, Pogliano K (2010) Divergent stalling sequences sense and control cellular physiology. Biochem Biophys Res Commun 393:1–5. 4. Cruz-Vera LR, Rajagopal S, Squires C, Yanofsky C (2005) Features of ribosome-peptidyltRNA interactions essential for tryptophan induction of tna operon expression. Mol Cell 19:333–343. 5. Nakatogawa H, Ito K (2002) The ribosomal exit tunnel functions as a discriminating gate. Cell 108:629–636. 6. Vazquez-Laslop N, Ramu H, Klepacki D, Mankin AS (2010) The key role of a conserved and modified rRNA residue in the ribosomal response to the nascent peptide. EMBO J 29:3108–3117. 7. Vazquez-Laslop N, Thum C, Mankin AS (2008) Molecular mechanism of drugdependent ribosome stalling. Mol Cell 30:190–202. 8. Lawrence M, Lindahl L, Zengel JM (2008) Effects on translation pausing of alterations in protein and RNA components of the ribosome exit tunnel. J Bacteriol 190:5862–5869. 9. Cruz-Vera LR, New A, Squires C, Yanofsky C (2007) Ribosomal features essential for tna operon induction: Tryptophan binding at the peptidyl transferase center. J Bacteriol 189:3140–3146. 10. Seidelt B, et al. (2009) Structural insight into nascent polypeptide chain-mediated translational stalling. Science 326:1412–1415. 11. Yang R, Cruz-Vera LR, Yanofsky C (2009) 23S rRNA nucleotides in the peptidyl transferase center are essential for tryptophanase operon induction. J Bacteriol 191:3445–3450. 12. Fulle S, Gohlke H (2009) Statics of the ribosomal exit tunnel: Implications for cotranslational peptide folding, elongation regulation, and antibiotics binding. J Mol Biol 387:502–517. 13. Gong F, Yanofsky C (2002) Instruction of translating ribosome by nascent peptide. Science 297:1864–1867. 14. Luo ZL, Sachs MS (1996) Role of an upstream open reading frame in mediating arginine-specific translational control in Neurospora crassa. J Bacteriol 178:2172–2177. 15. Gaba A, Wang Z, Krishnamoorthy T, Hinnebusch AG, Sachs MS (2001) Physical evidence for distinct mechanisms of translational control by upstream open reading frames. EMBO J 20:6453–6463. 16. Raney A, Baron AC, Mize GJ, Law GL, Morris DR (2000) In vitro translation of the upstream open reading frame in the mammalian mRNA encoding S-adenosylmethionine decarboxylase. J Biol Chem 275:24444–24450. 17. Weisblum B (1995) Erythromycin resistance by ribosome modification. Antimicrob Agents Chemother 39:577–585. 18. Lovett PS (1990) Translational attenuation as the regulator of inducible cat genes. J Bacteriol 172:1–6. 19. Menninger JR (1985) Functional consequences of binding macrolides to ribosomes. J Antimicrob Chemother 16(Suppl A):23–34. 20. Tenson T, Lovmar M, Ehrenberg M (2003) The mechanism of action of macrolides, lincosamides and streptogramin B reveals the nascent peptide exit path in the ribosome. J Mol Biol 330:1005–1014. 21. Horinouchi S, Weisblum B (1980) Posttranscriptional modification of mRNA conformation: Mechanism that regulates erythromycin-induced resistance. Proc Natl Acad Sci USA 77:7079–7083. 22. Gryczan TJ, Grandi G, Hahn J, Grandi R, Dubnau D (1980) Conformational alteration of mRNA structure and the posttranscriptional regulation of erythromycin-induced drug resistance. Nucleic Acids Res 8:6081–6097. 23. Mayford M, Weisblum B (1989) ermC leader peptide. Amino acid sequence critical for induction by translational attenuation. J Mol Biol 206:69–79. 24. Vázquez-Laslop N, Ramu H, Mankin AS (2011) Nascent peptide-mediated ribosome stalling promoted by antibiotics. Ribosomes: Structure, Function, and Dynamics, eds MV Rodnina, R Green, and W Wintermeyer (Springer-Verlag, Vienna).

25. Ettayebi M, Prasad SM, Morgan EA (1985) Chloramphenicol-erythromycin resistance mutations in a 23S rRNA gene of Escherichia coli. J Bacteriol 162:551–557. 26. Moazed D, Noller HF (1987) Chloramphenicol, erythromycin, carbomycin and vernamycin B protect overlapping sites in the peptidyl transferase region of 23S ribosomal RNA. Biochimie 69:879–884. 27. Schlunzen F, et al. (2001) Structural basis for the interaction of antibiotics with the peptidyl transferase centre in eubacteria. Nature 413:814–821. 28. Tu D, Blaha G, Moore PB, Steitz TA (2005) Structures of MLSBK antibiotics bound to mutated large ribosomal subunits provide a structural explanation for resistance. Cell 121:257–270. 29. Dunkle JA, Xiong L, Mankin AS, Cate JH (2010) Structures of the Escherichia coli ribosome with antibiotics bound near the peptidyl transferase center explain spectra of drug action. Proc Natl Acad Sci USA 107:17152–17157. 30. Bulkley D, Innis CA, Blaha G, Steitz TA (2010) Revisiting the structures of several antibiotics bound to the bacterial ribosome. Proc Natl Acad Sci USA 107:17158–17163. 31. Pestka S, et al. (1976) Induction of erythromycin resistance in Staphyloccus aureus by erythromycin derivatives. Antimicrob Agents Chemother 9:128–130. 32. Mayford M, Weisblum B (1990) The ermC leader peptide: Amino acid alterations leading to differential efficiency of induction by macrolide-lincosamide-streptogramin B antibiotics. J Bacteriol 172:3772–3779. 33. Allen NE (1977) Macrolide resistance in Staphylococcus aureus: Inducers of macrolide resistance. Antimicrob Agents Chemother 11:669–674. 34. Rosato A, Vicarini H, Bonnefoy A, Chantot JF, Leclercq R (1998) A new ketolide, HMR 3004, active against streptococci inducibly resistant to erythromycin. Antimicrob Agents Chemother 42:1392–1396. 35. Bryskier A (2000) Ketolides-telithromycin, an example of a new class of antibacterial agents. Clin Microbiol Infect 6:661–669. 36. Hartz D, McPheeters DS, Traut R, Gold L (1988) Extension inhibition analysis of translation initiation complexes. Methods Enzymol 164:419–425. 37. Sachs MS, et al. (2002) Toeprint analysis of the positioning of translation apparatus components at initiation and termination codons of fungal mRNAs. Methods 26:105–114. 38. Muto H, Nakatogawa H, Ito K (2006) Genetically encoded but nonpolypeptide prolyltRNA functions in the A site for SecM-mediated ribosomal stall. Mol Cell 22:545–552. 39. Asai T, Zaporojets D, Squires C, Squires CL (1999) An Escherichia coli strain with all chromosomal rRNA operons inactivated: Complete exchange of rRNA genes between bacteria. Proc Natl Acad Sci USA 96:1971–1976. 40. Bollenbach T, Quan S, Chait R, Kishony R (2009) Nonoptimal microbial response to antibiotics underlies suppressive drug interactions. Cell 139:707–718. 41. Hütter R, Poralla K, Zachau HG, Zahner H (1966) Metabolic products of microorganisms. 5l. On the mechanism of action of borrelidin-inhibition of the threonine incorporation in sRNA. Biochem Z 344:190–196. 42. Garza-Ramos G, Xiong L, Zhong P, Mankin A (2001) Binding site of macrolide antibiotics on the ribosome: New resistance mutation identifies a specific interaction of ketolides with rRNA. J Bacteriol 183:6898–6907. 43. Voss NR, Gerstein M, Steitz TA, Moore PB (2006) The geometry of the ribosomal polypeptide exit tunnel. J Mol Biol 360:893–906. 44. Zhu ZJ, et al. (2008) Structure-activity relationships of macrolides against Mycobacterium tuberculosis. Tuberculosis 88(Suppl 1):S49–63. 45. Shimizu Y, et al. (2001) Cell-free translation reconstituted with purified components. Nat Biotechnol 19(8):751–755. 46. Ohashi H, Shimizu Y, Ying BW, Ueda T (2007) Efficient protein selection based on ribosome display system with purified components. Biochem Biophys Res Commun 352:270–276. 47. Douthwaite S, Powers T, Lee JY, Noller HF (1989) Defining the structural requirements for a helix in 23S ribosomal RNA that confers erythromycin resistance. J Mol Biol 209:655–665. 48. Zaporojets D, French S, Squires CL (2003) Products transcribed from rearranged rrn genes of Escherichia coli can assemble to form functional ribosomes. J Bacteriol 185:6921–6927.


PNAS ∣

June 28, 2011 ∣

vol. 108 ∣

no. 26 ∣

10501

BIOCHEMISTRY

ACKNOWLEDGMENTS. We thank Yijia Luo for help with some experiments, Shannon Sparenga for editing the manuscript, and Michael Johnson for support of D.C.M. This work was supported by Grant MCB-0824739 from the National Science Foundation.