Catalytic Properties of Mutant 23 S Ribosomes Resistant to Oxazolidinones*

Kinetic analysis of ribosomal peptidyltransferase activity in a methanolic puromycin reaction with wild type and drug-resistant 23 S RNA mutants was used to probe the structural basis of catalysis and mechanism of resistance to antibiotics. 23 S RNA mutants G2032A and G2447A are resistant to oxazolidinones both in vitro and in vivo with the latter displaying a 5-fold increase in the value of K m for initiator tRNA and a 100-fold decrease in V max in puromycin reaction. Comparison of the K i values for oxazolidinones, chloramphenicol, and sparsomycin revealed partial cross-resistance between oxazolidinones and chloramphenicol; no cross-resistance was observed with sparsomycin, a known inhibitor of the peptidyltransferase A-site. Inhibition of the mutants using a truncated CCA-Phe- X -Biotin fragment as a P-site substrate is similar to that observed with the intact initiator tRNA, indicating that the inhibition is substrate-independent and that the peptidyltransferase itself is the oxazolidinone target. Mapping of all known mutations that confer resistance to these drugs onto the spa-tial structure of the 50 S ribosomal subunit allows for docking of an oxazolidinone into a proposed binding pocket. The model suggests that oxazolidinones bind between the P- and A-loops, partially overlapping with the peptidyltransferase P-site. Thus, kinetic, mutagene-sis, of Initiator fMet-tRNA iMet vitro Puromycin Peptidyltransferase Reaction— The in vitro peptidyltrans- ferase assay was carried out in a 45- (cid:2) l mixture containing the following components: 36.7 m M Tris-HCl (pH 7.4 at 23 °C), 10 m M MgCl 2 , 266.7 m M KCl, 0.01 (cid:2) M of E. coli 70 S ribosomes, 0.01–0.4 (cid:2) M f-[ 35 S]Met-tRNA iMet (20 Ci/mmol), 10–200 (cid:2) M puromycin, 33.3% MeOH, and 0.0083% Me were for various (cid:2) N KOH was added stop the reaction, and the quenched reactions were incubated for min at 37 °C followed by the addition of 600 (cid:2) l of 1 M NaHPO 4 (pH 7.0) and 250 ml of ethyl acetate. The product [ 35 S]formyl-methionyl-puromycin was extracted into the ethyl acetate phase by shaking for 15 min and centrifugation for 5 min. 80 (cid:2) l of the ethyl acetate phase was mixed with 125 (cid:2) l of Microscint-40 to the organic phase and counted in a Packard Topcount-NTX TM (Packard Instrument Co.). Fragment Peptidyltransferase Reaction— Reactions were performed at room temperature and contained the following components in a 60- (cid:2) l volume: 36.7 m M Tris-HCl (pH 7.4 at 23 °C), 10 m M MgCl 2 , 266.7 m M KCl, 0.15 (cid:2) M of E. coli MRE600 70 S ribosomes, various concentrations of CCA-Phe- X -Biotin, 50 (cid:2) M 3 H]puromycin (0.5 33.3% MeOH, and 0.0083% Me SO. The reaction mixtures were stopped by addition of 0.5

Oxazolidinones, the only novel class of antibiotics identified in the last two decades, are the focus of intensive discovery efforts (1)(2)(3)(4)(5)(6)(7)(8)(9). Linezolid, an oxazolidinone, is approved for treatment of infections caused by Gram-positive bacteria that are resistant to other antibiotics. Emerging resistance to all known drugs, including the "last resort" vancomycin family, poses a serious threat to the public health worldwide. Understanding the mechanism of action of oxazolidinones at the molecular level, therefore, has a great importance for the development of the next generation of these novel antibiotics and, ultimately, for the outcome of the ongoing battle against drug-resistant pathogens.
Oxazolidinones impose their action at the initiation stage of translation (4,5), apparently via inhibition of preinitiation complex formation (9). Our recent finding that oxazolidinones interfere with binding of initiator tRNA to the ribosomal P-site (1), thus inhibiting formation of the first peptide bond, prompted a search for similarities between oxazolidinones and known inhibitors of peptidyltransferase. To address this and other mechanistic questions, we have studied catalytic properties of oxazolidinone-resistant ribosomes and compared the mechanism of oxazolidinone inhibition with the action of known peptidyltransferase inhibitors, such as chloramphenicol and sparsomycin. To further define the oxazolidinone binding site, we have mapped resistant mutations onto the three-dimensional structure of the ribosomal 50 S subunit to reveal an inhibitor binding pocket and docked an oxazolidinone inhibitor therein to show its fitness. Preparation of Oxazolidinone-resistant 23 S RNA Mutant Ribosomes-Oxazolidinone-resistant mutant strains carrying a single mutated copy of ribosomal 23 S RNA were selected and characterized as reported in Refs. 2 and 33 using plasmids with site-directed mutations (10) and a recently described single plasmid-encoded rRNA operon system in Escherichia coli (11). Cells from 4 liters of exponentially growing wild type or mutant E. coli culture (A 590 ϭ 0.6) were pelleted by centrifugation in a Sorvall SLA3000 rotor at 9110 rpm for 15 min at 4°C. Cell pellets were resuspended in an equal volume of buffer A (20 mM HEPES-KOH (pH 7.6 at 0°C), 6 mM MgCl 2 , 30 mM NH 4 Cl, 4 mM ␤-mercaptoethanol, 0.1 mM phenylmethylsulfonyl fluoride, and 0.1 mM benzamidine hydrochloride) and centrifuged at 9110 rpm in SLA3000 rotor for 10 min. The pellets were transferred into a mortar and ground with prechilled alumina followed by the addition of an equal volume of Buffer A supplemented with ␤-mercaptoethanol, phenylmethylsulfonyl fluoride, and benzamidine hydrochloride. After incubation with RNasefree DNase for 30 min. at 0°C, alumina and cell debris were removed by centrifugation at 15,000 for 15 min and 30,000 ϫ g for 30 min, respectively. Supernatant was then layered over an equal volume of 1. Puromycin Peptidyltransferase Reaction-The in vitro peptidyltransferase assay was carried out in a 45-l mixture containing the following components: 36.7 mM Tris-HCl (pH 7.4 at 23°C), 10  Mutant K m and V max Determination-Assuming that the ribosomal peptidyltransferase has an ordered mechanism of action (12), in which the P-site substrate, A, and the A-site substrate, B, associate with the ribosome, E, in an obligate order: the A-site substrate can only bind to the preformed P-substrate-ribosome complex, as shown in Reaction 1,

Reagents and Materials
the corresponding velocity equation would be given by Equation 1.
Determination of apparent V max and K m parameters for fMet-tRNA i Met was performed under the pseudo first-order reaction conditions, in which a fixed saturating concentration of puromycin (200 M) with 70 S ribosomes at 0.005 M has been used, and fMet-tRNA i Met concentration was varied from 0.04 to 0.8 M, depending on the mutant studied. To obtain corresponding puromycin K m values (K B ), a fixed saturating concentration of 0.8 M fMet-tRNA i Met was used for the wild type and all three mutants, whereas the puromycin concentration was varied from 5 to 400 M. Time courses at each set of conditions were performed, and initial rates were determined by linear regression analysis of steadystate kinetics. For sparsomycin, a competitive inhibitor of A-site (13-15), Equation 2 was used.
Based on competitive behavior toward both initiator tRNA (P-site substrate) and puromycin (A-site) (1), Equation 3 was used for oxazolidinones.
It was shown previously that chloramphenicol binds to a different site on the ribosome from the A-site inhibitor sparsomycin (16) and also inhibits sparsomycin-induced binding of the fragment P-site substrate to the ribosomes (17). Competitive mixed or competitive behavior toward puromycin (A-site substrate) (18,19) has also been reported for this inhibitor. Based on these results, it was assumed that chloramphenicol is a competitive inhibitor with respect to both sites, and the data were fitted to the same equation (3) (23); thus, the docking procedure was predominantly based on steric considerations as well as on mutant inhibition data. The oxazolidinone inhibitor XA043 was modeled into a potential binding pocket on the ribosome using the x-ray coordinates with Protein Data Bank entry code 1FFK and Sybyl molecular modeling software. The inhibitor was built and minimized in the gas phase before manually docking it into the pocket near residues where oxazolidinone-resistant mutations occur and the active site K ϩ resides.

Catalytic Properties of 23 S RNA Mutants in Puromycin
Reaction-K m and V max values for both A-and P-site substrates in the puromycin reaction for the drug-resistant 23 S RNA mutants were determined and compared with the wild type ribosomes. For mutant G2447A (position 2482 in H. marismortui), significant changes in both K m and V max values are observed for both substrates (Fig. 1, A and B). The values of K m for initiator tRNA and puromycin were increased by 5-and 2-fold, respectively, and V max decreased 100-fold (Table I). Erythromycin-resistant mutant A2058G (2099 in H. marismortui), which is weakly resistant to oxazolidinones in vivo (2, 33), exhibited a 3-fold increase in K m value for initiator tRNA with no change for that of puromycin. Oxazolidinone-and chloramphenicol-resistant and erythromycin-hypersensitive mutant G2032A (2973) has K m and V max values that are similar to the wild type. Thus, two of the three studied oxazolidinone-resistant mutants (G2447A and A2058G) had decreased affinity for the initiator tRNA (P-site) substrate with none of the mutations affecting the interaction with the A-site substrate puromycin.
Effect of Peptidyltransferase Inhibitors upon Activity of 23 S RNA Mutants in Puromycin and Fragment Reactions-To define the oxazolidinone binding site, we studied inhibition of oxazolidinone-resistant mutants in two different substrate systems: in the puromycin reaction that employs initiator fMet-tRNA i Met as a P-site substrate and puromycin as an A-site substrate and in the fragment reaction where the truncated version of the P-site substrate, CCA-Phe-Biotin, is utilized, whereas the A-site substrate, puromycin, is the same as in the first system. To compare inhibitory properties of oxazolidinones with those of known peptidyltransferase inhibitors, chloramphenicol and sparsomycin, we determined corresponding IC 50 values in both puromycin and fragment reactions performed by mutant ribosomes.
Inhibition profiles of mutant peptidyltransferases in the puromycin reaction with the oxazolidinones, XA043 and linezolid, are shown in Fig. 2, A and B, respectively. As expected based on their high levels of resistance in vivo (2, 33), oxazolidinoneresistant mutants G2447A and G2032A are not inhibited by oxazolidinones in vitro. The effect of chloramphenicol on peptidyltransferase mutants is shown in Fig. 2C. Oxazolidinoneresistant mutant G2447A is not inhibited by chloramphenicol either, although it is highly sensitive to the A-site inhibitor sparsomycin. In contrast, oxazolidinone-resistant mutant G2032A shows rather a moderate in vitro resistance against chloramphenicol, indicated by less than 4-fold decrease in affinity for this inhibitor. Inhibition of mutant peptidyltransferases with the A-site inhibitor sparsomycin is shown in Fig.  2D. An inhibition constant of ϳ0.4 M determined in this work for sparsomycin was in good agreement with the K i value of ϳ1 M reported previously for this antibiotic with a pentamer fragment substrate by Harris and Pestka (34). Oxazolidinoneresistant mutants are either more sensitive to inhibition by sparsomycin than the wild type (A2058G and G2447A) or 2-fold less sensitive (G2032A), indicating little cross-resistance.
When the fragment P-site substrate was used in place of  initiator fMet-tRNA, the cross-resistance to oxazolidinones (Fig. 3, A and B) and chloramphenicol (Fig. 3C) for mutant G2447A remains with a somewhat lower level of resistance to oxazolidinone XA043. It is noteworthy that an opposite effect, a stimulation of the fragment reaction, was observed for this mutant in the presence of linezolid. The effects of inhibitors on the reactions catalyzed by mutant G2032A were affected by which P-site substrate was used: initiator tRNA leads to less inhibition than the truncated P-site substrate. At the same time, mutations had insignificant, less than 2-fold, effect on the magnitude of inhibition with sparsomycin (Fig. 3D).
To compare inhibitors with different mechanisms of action, K i values were calculated (summarized in Table II). Mutants A2058G and G2032A are ϳ3-4-fold less sensitive toward linezolid in comparison with XA043. Both of these mutants are less affected by chloramphenicol than wild type. Mutant G2447A was not inhibited by chloramphenicol at all, but its sensitivity to oxazolidinones decreased by more than 100-fold. A weakly resistant to oxazolidinones mutant (A2058G) (2, 33) is sensitive toward oxazolidinones in a fragment reaction, while displaying a moderate in vitro resistance when whole initiator tRNA was used as a substrate (Figs. 2 and 3, compare A and B,  respectively). None of the mutations studied displayed a significant level of resistance to an A-site inhibitor sparsomycin. DISCUSSION Mutational studies of the ribosome by genetic means have been a fruitful approach to investigation of the functional role of the different regions of this remarkable ribozyme (24 -26). However, the biochemical effects of ribosomal mutations were  H. marismortui residues are labeled with the corresponding E. coli residues in parentheses. C, docking of oxazolidinone XA043 structure into the binding pocket. After docking the inhibitor XA043 into the putative binding pocket on the ribosome, an 8-Å sphere around the inhibitor was extracted for display. The residues that upon mutation confer oxazolidinone resistance are colored in purple, and the other residues are colored in green, XA043 is colored by atom type, and the K ϩ cation is in orange. H. marismortui residues are labeled with the corresponding E. coli residues in parentheses. The view is from the center of the P-site helix (Fig. 4A) looking outward radially. hindered by an inability to isolate mutant ribosomes free of wild type counterparts because of the presence of multiple copies of both 16 and 23 S RNA in most eubacteria. Development of the single plasmid-encoded rRNA operon system in E. coli (11) made possible engineering, isolation, and enzymatic characterization of mutant ribosomes in vitro. Steady state experiments using purified drug-resistant ribosomal mutants have yielded meaningful catalytic parameters. These, in turn, have allowed initial structure-function relationships at the peptidyltransferase center to be determined, as well as possible mechanisms of action and resistance to antibiotics at the level of the catalyst itself.
We find that 23 S RNA G2032A and G2447A mutations result in decreased affinity for the P-site substrate initiator fMet-tRNA and insensitivity toward oxazolidinones and chloramphenicol. Nucleotide 2447 belongs to the peptidyltransferase region that has been shown to be involved in interactions with the acyl moiety of initiator tRNA (P-site) (27)(28)(29)(30). We also demonstrate that the affinity for the A-site substrate is not affected by the mutations that confer resistance to oxazolidinones. According to our inhibition results (Figs. 2 and 3), chloramphenicol, an inhibitor of both P-and A-sites (16,17), displays partial cross-resistance with oxazolidinones, whereas a specific inhibitor of the A-site, sparsomycin, does not. Thus, this suggests that oxazolidinones inhibit the methanolic peptidyltransferase reaction primarily via interference with the Psite substrate binding. In addition, we have shown previously that oxazolidinones are competitive with initiator tRNA (P-site substrate) inhibitors (1). The accumulated evidence leads us to conclude that oxazolidinones are inhibitors of the peptidyltransferase P-site.
Truncation of the P-site substrate did not eliminate in vitro resistance toward oxazolidinones and chloramphenicol. This can be interpreted in terms of the binding site for these drugs being located on the 50 S part of the P-site in a close proximity to the CCA-fMet binding site. We suggest that mutations at positions G2447 and G2032 interfere with the binding of the isoacceptor part of the initiator tRNA and at the same time contribute to formation of the oxazolidinone binding site. Mutant A2058G showed no resistance toward oxazolidinones in the fragment reaction, whereas a moderate decrease in affinity for oxazolidinone linezolid was observed in the pulomycin reaction. This can be interpreted in terms of the possible structural changes in the P-site induced by this mutation, in particular in binding of the non-isoacceptor portion of the initiator tRNA. A remarkable observation that a catalysis of the fragment reaction by mutant G2447A is stimulated in the presence of linezolid, but inhibited by XA043, suggests that inhibition with oxazolidinones might occur via conformational change in the mutant active center that distorts the substrate(s) binding site, leading to a dichotomous outcome depending on the structure of the inhibitory molecule bound.
To further define the oxazolidinone binding site, mutations that lead to oxazolidinone resistance (2,20,21,33) were mapped onto the crystal structure (23) of the peptidyltransferase 50 S subunit (Fig. 4). Although the mutated nucleotides are relatively dispersed in the primary structure of 23 S rRNA, these mutations map into a confined space between the P-and A-loops in the three-dimensional structure, which partially overlaps with the P-site. As seen in Fig. 4A, the mutations are clustered near the central channel of the ribosome that permits substrate entry to the P-site. There is a cavity of sufficient size to accept the inhibitor XA043 as shown in Fig. 4C, which contains a potassium ion that might be required for catalytic activity (30). Little reorganization of the nucleotide residues is necessary for inhibitor access to the binding site. Although the binding site model is preliminary and cannot predict the exact conformations of the binding pocket and the inhibitor, especially in light of potential structural differences between archaeal and eubacterial ribosomes, the model offers an explanation of the inhibitory action against P-site activity by oxazolidinones.
The functional role of the nucleotide G2447 is of a particular interest due to the ongoing controversy about the mechanism of catalysis by ribosomal peptidyltransferase. The low catalytic efficiency and markedly decreased affinity for the P-site substrate of mutant G2447A suggests an important contribution to the catalysis by this nucleotide. At the same time, it raises questions about the proposed general acid-base mechanism hypothesis (30,31). If nucleotide G2447 plays an important role in the pK a shift of the catalytic residue A2051, one would expect this nucleotide to be essential both in vivo and in vitro, which has not been observed (3,10,32). Thus, there is a clear need for further mechanistic studies of this essential ribozyme.
We conclude that oxazolidinones inhibit binding of the initiator tRNA to the peptidyltransferase P-site on 50 S, preventing the formation of the first peptide bond. One would also predict that clinical resistance to this class of drugs would evolve by introducing mutations that would alter the fine structure of the P-site, as was observed in the case of two oxazolidinone-resistant mutants, selected in the laboratory. One would also predict that frequency of resistance acquired through this target-based mechanism should be quite low due to redundancy of rRNA operons in bacteria. Thus, this key ribozyme, encoded by multiple rDNA genes, presents an excellent drug discovery target for the development of novel antibiotics, the clinical usefulness of which would not be eroded by single point mutations.