The slow dissociation rate of K-1602 contributes to the enhanced inhibitory activity of this novel alkyl–aryl-bearing fluoroketolide

Abstract

Ketolides belong to the latest generation of macrolides and are not only effective against macrolide susceptible bacterial strains but also against some macrolide resistant strains. Here we present data providing insights into the mechanism of action of K-1602, a novel alkyl–aryl-bearing fluoroketolide. According to our data, the K-1602 interacts with the ribosome as a one-step slow binding inhibitor, displaying an association rate constant equal to 0.28 × 10⁴M⁻¹ s⁻¹ and a dissociation rate constant equal to 0.0025 min⁻¹. Both constants contribute to produce an overall inhibition constant K_i equal to 1.49 × 10⁻⁸M, which correlates very well with the superior activity of this compound when compared with many other ketolides or fluoroketolides.

• Antibiotics
• antibiotic resistance
• macrolides
• ribosomes
• slow binding inhibitors

Introduction

Macrolides are members of a large family of protein synthesis inhibitors and are of special interest, because many of them are used in clinical therapy1^,2. Macrolides are no longer seen as general inhibitors of protein synthesis, but rather as specific inhibitors that inhibit translation depending on the sequence of the nascent polypeptide being translated3. According to our current understanding, for the majority of proteins, binding of the drug within the tunnel causes protein synthesis to stop when the nascent peptide chain reaches a length of between 5 and 10 amino acids, which in turn leads to dissociation (or drop-off) of the peptidyl-tRNA from the ribosome4–6.

The widespread usage of macrolides has led to the selection of resistant bacterial strains, providing a strong incentive for the development of newer macrolide drugs that can overcome such resistance mechanisms7^,8. This led to the development of a new generation of macrolides, the ketolides, with the most distinguished members being telithromycin, cethromycin, and solithromycin (Figure 1)9–12. All are semi-synthetic derivatives of erythromycin, where the keto group at the C-3 position of the lactone ring replaces the cladinose sugar that is present in erythromycin. In addition, ketolides possess a cyclic carbamate fused at positions C-11 and C-12 of the lactone ring, as well as a heteroaromatic side chain that is linked via a flexible alkyl-linker to the C11–C12 cyclic carbamate (telithromycin, Figure 1) or via an −O− bridge to the C6 position (cethromycin and K-1602; Figure 1).

Figure 1. Chemical structures of macrolides erythromycin, tylosin, cethromycin, telithromycin, and Kosan-1602.

Recently, we published data concerning new ketolides carrying an alkyl–aryl group via an −O− bridge at the C6 position of the lactone ring and one or two fluoride atoms at the C3 or C13 position of the lactone ring13. Among these compounds, which are all strong inhibitors of protein synthesis in vivo and in vitro, K-1602 (Figure 1) was shown to be the only compound that did not induce expression of the macrolide resistance gene erm(C) and is, therefore, effective against bacteria with macrolide-inducible resistance13. It is well known that a small number of specific short nascent peptides, such as those encoded in the regulatory cistrons of macrolide resistance genes, can induce ribosome stalling, which retains the peptidyl-tRNA, but prevents peptide bond formation with the incoming aminoacyl-tRNA14^,15. In the absence of macrolides, ErmC expression is repressed, because the ribosome-binding site (RBS) and the AUG start codon of the ermC mRNA are sequestered within a stem-loop structure. However, in the presence of sub-inhibitory concentrations of a macrolide, such as erythromycin, ribosomes translating the ErmCL leader peptide become stalled, leading to an alternative stem-loop structure in the mRNA that exposes the RBS and start codon of the ermC gene and thus allows ribosome binding and induction of ErmC expression16^,17. Initially, it was thought that ketolides are not ErmCL inducers; however, it was subsequently shown that they can induce the erm(C) gene although in a narrow range of concentrations and with a low level of induction compared with macrolides18. As mentioned earlier, one exception is the ketolide K-1602, which prompted us to study further its interaction with the ribosome. Following detailed kinetic analysis, we determined the association and dissociation constants of K-1602 with the ribosome. Calculations were based on a previously developed kinetic model19–21, which allowed us to conclude that K-1602 behaves as a slow binding slowly dissociating inhibitor22, thus explaining the unique characteristics and effectiveness of this compound. Slow binding inhibitors represent a large kinetic class, mainly including transition state and reaction intermediate analogs, and as such show promising features for future drug development23.

Materials and methods

l-Phenylalanine, puromycin dihydrochloride (disodium salt), tylosin, erythromycin, GTP, ATP, and tRNA from Escherichia coli strain W were purchased from Sigma Chemical Co. (St. Louis, MO). l-[2,3,4,5,6-³H]Phenyl-alanine was purchased from Amersham Life Science (Piscataway, NJ). Cellulose nitrate filters (type HA, 24-mm diameter, 0.45-μm pore size) were from Millipore Corporation (Billerica, MA). Avian myeloblastosis virus-reverse transcriptase, dNTPs, and dideoxy-NTPs were obtained from Roche Diagnostics GmbH (Mannheim, Germany) and T4 polynucleotide kinase from Takara (Tokyo, Japan). Scintillation cocktail Filter Count was purchased from Perkin-Elmer (Waltham, MA). Ketolide K-1602 was provided by Kosan Bioscience Inc. now Bristol-Myers Squibb (Hayward, CA). Telithromycin was kindly supplied by Sanofi-Aventis Inc. (Bridgewater, NJ) to D.L. Kalpaxis

Biochemical preparations

70 S tight-coupled ribosomes were obtained from E. coli K-12 cells as reported previously24, MF-mRNA was prepared with run-off transcription, and acetyl-phenylalanine-tRNA (Ac-Phe-tRNA) with tritium-labeled phenylalanine were prepared as described previously25. Complex C, i.e., the [70S ribosome ċ MF-mRNA ċ Ac[³H]Phe-tRNA] complex was prepared as described previously21. The initially formed complex C was adsorbed on cellulose nitrate filters and then washed with three 4-ml portions of cold buffer (100 mM Tris-HCl pH 7.2/50 mM KCl/10 mM MgCl₂/6 mM 2-mercaptoethanol). Under these conditions, the majority of the total ribosomes are in the form of complex C with bound donor almost completely at the P site, as titrated with puromycin reaction.

Puromycin reaction

The reaction between complex C and excess of puromycin was carried out at 10 mM Mg²⁺ and 100 mM NH₄⁺, as described previously19. Briefly, complex C reacted with excess of puromycin in the presence or in the absence of macrolides, and the progress of the reaction was analyzed over a wide range of puromycin and macrolide concentrations. When desired, the reaction was terminated, by adding an equal volume of 1 M NaOH. The product, AcPhe-puromycin (P), was extracted in ethyl acetate, and the radioactivity was measured in a liquid scintillation spectrometer. In the kinetic analysis, the product is expressed as a percentage of the isolated radioactivity on the filter (100 × P/N_o). Controls without puromycin were included in each experiment, and the values obtained were subtracted.

Inactivation of complex C by tylosin in the absence and presence of K-1602

Complex C adsorbed on a cellulose nitrate filter was exposed to various concentrations of tylosin in a 2-ml buffer solution (100 mM Tris-HCl pH 7.2/100 mM NH₄Cl/10 mM Mg²⁺ (acetate)/6 mM 2-mercaptoethanol). The reaction was allowed to proceed at 25 °C for the desired time intervals and was stopped by immersing the filter in 15 ml of cold buffer. The remaining active complex C, after washing the cellulose nitrate filter to remove traces of tylosin not specifically bound, was determined by titration with puromycin (2 mM, 2 min at 25 °C) as described previously19. The inactivation of complex C by tylosin was also examined in the presence of K-1602, and the remaining active complex C versus puromycin, named P_t, was monitored as mentioned above.

Antibiotic probing and chemical modification

Aliquots of 70S ribosomes, 50 pmoles per tube, were incubated with or without antibiotic at 37 °C. The incubation took place in buffer containing 80 mM HEPES-KOH (pH 7.8), 20 mM MgCl₂, 100 mM NH₄Cl, and 1.5 mM dithiothreitol for 10 min, followed by 10 min further incubation at 20 °C. After cooling on ice, chemical modification of ribosomes was carried out by using DMS (dimethyl sulfate), as described previously26^,27. The reactions were quenched and then ribosomes precipitated by ethanol. The pellets were resuspended in 50 μL of buffer containing 10 mM Tris-HCl (pH 7.5), 100 mM NH₄Cl, 5 mM EDTA, and 0.5% SDS, and the rRNA extracted with equal volumes of phenol, phenol–chloloform, and last chloroform. The ribosomal RNA was precipitated by ethanol and resuspended in water.

Primer extension

The modifications in the 23S rRNA were monitored by primer extension analysis using avian myeloblastosis virus-reverse transcriptase and 5′-labeled primer. DNA primers (18–20 nt) were labeled at the 5′-end using T4 polynucleotide kinase and [γ-³²P]ATP. The cDNA products of the primer extension reactions were separated on 6% polyacrylamide/7 M urea sequencing gels. Gels were scanned with a PhosphorImager-type Fujifilm (FLA-3000; Berthold, Oak Ridge, TN) and analyzed with ImageQuant software AIDA (Raytest, Wilmington, NC). The stop positions during the cDNA synthesis were identified by reference to dideoxy sequencing reactions performed in parallel on 23S rRNA26.

Results

Puromycin reaction in the presence and absence of K-1602

Puromycin mimics the A-site bound 3′-aminoacyl end of aminoacyl-tRNA and acts as a pseudosubstrate. This reaction is a model system for testing peptidyltransferase (PTase) activity and takes place according to the kinetic Scheme 1 where C is the complex of 70S E. coli ribosome programmed with MF-mRNA and bearing AcPhe-tRNA at the P site, S is puromycin, P is the product AcPhe-puromycin and C′ is the ribosome without bound donor. Since the reaction buffer is free of unbound Ac-Phe-RNA, C′ is unable to be reconverted to C and react for a second cycle with puromycin. In the presence of excess puromycin, the reaction follows pseudo-first-order kinetics. According to the integrated first-order rate law, the relationship (1) holds19, where C_o is the initial total reactive complex C at zero time, k_obs is the apparent rate constant of product formation, and t is the reaction time. The rate constant k_obs is related to puromycin concentration by the following equation: (2) which allows the determination of both and k₃ and K_S constants by double reciprocal plotting28. In Figure 2, we can see the progress of puromycin reaction in the absence and in the presence of K-1602 and/or tylosin. As shown, K-1602 does not cause any inhibition in product formation, while in the presence of tylosin, there is strong inhibition and, after 2 min, the logarithmic plot reaches a plateau (Figure 2A). This is in agreement with previous observations according to which19, tylosin, a 16-membered ring macrolide with a long disaccharide extension of the lactone ring (Figure 1) reaches the PTase area and inhibits the peptide bond formation, whereas other macrolides with monosaccharide extensions, such as erythromycin or K-1602 (Figure 1), do not reach the PTase area and, therefore, do not inhibit peptidyltransferase activity. As expected, the simultaneous presence of K-1602 relieves the inhibitory effect of tylosin, and at high concentrations of K-1602 becomes completely nullified. This is clearly shown in Figure 2(B), where tylosin strongly inactivates complex C versus puromycin. Nevertheless, inactivation of tylosin is partially relieved by K-1602 and almost abolished at higher K-1602 concentrations. This latter finding indicates that the binding site of K-1602 overlaps with Tylosin as expected based on the similarity of binding sites of 14- and 16-membered macrolides observed on the ribosome by X-ray crystallography29. The competition of tylosin with K-1602 for a common or overlapping binding site on the ribosome is also supported by our chemical footprinting data: as shown in Figure 2(C), erythromycin, telithromycin, tylosin, and K-1602 strongly protect bases A2058 and A2059 of 23S rRNA, which are both located within the ribosomal tunnel and comprise part of the macrolide binding site.

Figure 2. (A) Progress curve of puromycin reaction, in the absence of antibiotic (▴) and in the presence of 1 μM K-1602 (•) and tylosin 3 μM (○). (B) Puromycin reaction in the absence of antibiotic (bar 1) and in the presence of 5 μM K-1602 (bar 2), 5 μM Tylosin (bar 3), and 5 μM Tylosin in the simultaneous presence of K-1602 either at 5 μM (bar 4) or 20 μM (bar 5). (C) Protection of bases in domain V of 23S rRNA from DMS modification by bound antibiotics. Lanes are marked from the left as follows: U, A, G, C sequencing lanes. Lane 1, untreated 23S rRNA; lane 2, 23S rRNA modified by DMS in the absence of antibiotic; lanes 3–6, ribosomes were first preincubated for 10 min at 37 °C with erythromycin, telithromycin, K-1602 and tylosin, and then modified by DMS. All antibiotics were used at a final concentration equal to 50 μM.

Competition of K-1602 and tylosin for binding on the ribosomal complex C

The exposure of the complex C in a mixture of tylosin and K-1602 is a classical competition experiment and the progress as well as the final products depend on the initial concentrations of each antibiotic and the specific values of kinetic constants. The final products of complex C with tylosin or complex C with K-1602 are kinetically easily discriminated, because the complex with bound tylosin is inactive towards puromycin, whereas the complex with bound K-1602 is fully active. In other words, K-1602, although not inhibiting the puromycin reaction, offers protection in complex C from tylosin inactivation. This type of competition can follow two alternative mechanisms, as depicted in Schemes 2 and 3.

Kinetic equations concerning Schemes 2 and 3 have been derived in previous publications19^,20, and the main functions are presented in Table 1. Plotting our data to these equations gives us the opportunity to distinguish, which of the two mechanisms is followed by K-1602. As shown in Figure 3(A), inactivation of complex C follows first-order kinetics and from the slope of each plot, an apparent rate constant F for the inactivation of complex C can be determined. The linearity of F versus the concentration of K-1602 (Figure 3B) supports the model of Scheme 3, according to which K-1602 interacts with complex C via a one-step mechanism. Otherwise, a hyperbolic equation would be expected according to Equation (4) of Table 1. The linearity of F versus K-1602 was verified with two different tylosin concentrations, namely 3 and 6 μM, respectively. To further support this model, [C*I]_∞ versus tylosin concentration was plotted (Figure 4A). The orthogonal hyperbola of the plot is again in agreement with Scheme 3, while the linearity of the double reciprocal plot, 1/[C*I]_∞ versus 1/[I] supports the same mechanism (Figure 4B). Moreover, the plots of F versus the concentration of K-1602 allow us to calculate the association rate constant k₆, since the slope of this plot is equal to k₆K_Tyl/(K_Tyl+[I]) according to Equation (1) of Table 1.

Figure 3. (A) Progress curve of complex C inactivation by tylosin in the presence of K-1602. Complex C absorbed on a cellulose nitrate filter was exposed to a solution containing 3 μM tylosin plus K-1602 either 2 μM (•) or 8 μM (○). (B) Function of the apparent rate constant F versus K-1602 concentration in the simultaneous presence of tylosin. Tylosin concentration was either 3 μM (Δ) or 6 μM (•). The apparent rate constant F was calculated from the slope of the plots as (A).

Figure 4. (A) Inactivation of complex C by tylosin in the presence of 4 μM K-1602. (C*I)_∞ represents the inactivated complex C by tylosin at infinite time. (B) Variation of 1/(C*I)_∞ as a function of reverse of tylosin concentration. The K-1602 concentration remains constant and equal to 4 μM as in (A). (C*I) is presented as the ratio of total (C).

Figure 5. Determination of the dissociation rate constant (k₇) of C*A complex. Drug–ribosome complex (C*A) was prepared in the presence of 5 μM K-1602, and absorbed on a cellulose nitrate filter. Then after dilution, it was exposed to 10 μM tylosin for various time intervals, after which the remaining activity was titrated with puromycin. The k₇ value was estimated from the slope of the linear time plot.

Scheme 1. Kinetic model of puromycin reaction.

Scheme 2. Two step kinetic model of competition between K-1602 (A) and tylosin (I) for binding on the ribosomal complex (C).

Scheme 3. One step kinetic model of competition between K-1602 (A) and tylosin (I) for binding on the ribosomal complex (C).

Table 1. Kinetic equations for one-step and two-steps mechanisms of drug binding to ribosomes.

Function	One step mechanism	Two step mechanism
F
(C*I)∞

F represents the apparent rate constant of ribosomal complex C interaction with both antibiotics tylosin and K-1602, and (C*I)_∞ represents the product AcPhe-puromycin formed at infinite time of exposing the complex in defined antibiotic concentrations.

Determination of the dissociation rate constant k₇ of the complex

The K-1602-ribosome complex was prepared in the presence of 2 × 10⁻⁶M antibiotic, and was absorbed on a cellulose nitrate filter. Under conditions of excess antibiotic, the macrolide binding site is completely occupied by the drug and the complex is totally active towards puromycin. Then, the complex was exposed to a near-infinite dilution of K-1602 but in the presence of 1 × 10⁻⁵M tylosin for various time intervals, after which the remaining activity of the complex was titrated with puromycin. This means that complex C*A was dissociated to complex C through a k₇ dissociation rate constant but was unable to be recycled due to the infinite dilution of K-1602. Subsequently, with the addition of excess of tylosin, the regenerated complex C was immediately inactivated to C*I. Therefore, the k₇ value can be estimated from the slope of the linear time plot of C*A titration by puromycin (Figure 5). This was found to be equal to 0.0025 min⁻¹. By dividing the value of k₇ to the value of k₆, we estimated the overall dissociation constant K_A, which is equal to 14.9 nM, a value that is very close to the value previously calculated by equilibrium binding studies13.

Discussion

The increasing incidence of antibiotic resistance and the toxicity associated with some of the available compounds constitute a formidable challenge for further research looking for new, safe, and effective antibiotics. Taking into account that the ribosome is a particularly versatile target for drug development30 as well as the serious side effects of telithromycin in clinical applications31, the further exploitation of the ribosome as a target for new ketolides seems to be an essential road. Such a new compound is K-1602, which could be a useful weapon in the endless war against macrolide-resistant pathogens in the future. Its advantages are known from a previous publication13 but additional data were produced in the present study. According to them, K-1602 binds to the ribosome like a slow binding-slowly reversible inhibitor, following a one-step mechanism22. Its slow association constant (0.28 × 10⁴M⁻¹ s⁻¹) classifies K-1602 in the class of slow binding inhibitors, which includes many other macrolides tested in our lab, such as erythromycin, tylosin, and telithromycin32^,33. In the case of K-1602, we failed to detect a rapidly formed encounter complex CA, although a fast initial step cannot be easily excluded, because sometimes it cannot be detected under the experimental conditions and the antibiotic concentrations used34^,35. The slow reversibility of the complex C*A supports the notion that strong binding forces may keep K-1602 bound tightly to the ribosome. Comparing K-1602 with erythromycin19, we conclude that although the association rate constants do not differ substantially, the dissociation rate constant of K-1602 is significantly lower. A probable scenario when erythromycin or K-1602 and peptidyl-tRNA are simultaneously bound to a stalled ribosome is that the ribosome–drug complex could dissociate before drop-off of the peptidyl-tRNA. Subsequently, the ribosome could become refractory to the drug and protein chain elongation can be resumed. However, the rate constant for drop-off of an oligopeptidyl-tRNA calculated by Lovmar et al.36 is much higher than 0.0025 min⁻¹, which implies that K-1602 at saturation concentrations completely blocks the peptide-chain elongation process. On the contrary, erythromycin has a dissociation rate constant equal to 0.06 min⁻¹37 about 30 times higher than K-1602, a fact rendering the scenario of erythromycin dissociation before oligo peptidyl-tRNA drop-off, more probable. It seems that the additional binding energy is coming from the aryl–alkyl side chain of K-1602. In contrast, telithromycin possessing also an alkyl–aryl side chain linked to the C11–C12 cyclic carbamate is characterized by a different dissociation constant33. Therefore, the type of alkyl–aryl side chain and the lactone-ring position, at which this chain is attached, can influence the tightness of the drug–ribosome complex, as observed when K-1602 is compared with erythromycin and other ketolides. In addition, the fixation of a drug on the ribosome depends on the stereochemistry of the drug binding pocket in the ribosome, which may also differ among the bacterial species38^,39. From the standpoint of a drug design for pharmaceutical applications, a slow rate of dissociation is desirable as it may be expected to result in a longer lived C*I complex, while a slow rate of association should be avoided since it delays the time required for inhibition at a given concentration40. From this point of view, whenever a drug interaction with ribosomes is investigated, an appropriate experimental system is required to reveal possible late events of interaction, and additional kinetic constants are necessary to completely characterize the drug potency.

Conclusions

The extremely low dissociation constant k₇ of K-1602 means that this novel compound is bound to the ribosome in an almost covalent-like fashion. Together with the tight binding affinity of K-1602, this may explain its inability to induce the macrolide resistance erm(C) gene and, therefore, demonstrates the exciting potential of K-1602 as a therapeutically useful ketolide to combat macrolide-resistant bacteria.

Declaration of interest

The authors report no conflicts of interest.

Acknowledgements

We thank Kosan Biosciences Inc. for providing the new ketolide K-1602 and Dr Daniel Wilson for critical reading of the manuscript. This work was partially supported by the Research Committee of Patras University, program “K. Karatheodori” (grant C.573 to G.P.D.), and the IKYDA program for short-term fellowship to M.K. and M.S.