![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Previous Article | Next Article 
Antimicrobial Agents and Chemotherapy, March 2002, p. 813-820, Vol. 46, No. 3
0066-4804/02/$04.00+0 DOI: 10.1128/AAC.46.3.813-820.2002
Copyright © 2002, American Society for Microbiology. All Rights Reserved.
Antimicrobial Agents Research Group, Division of Immunity and Infection, University of Birmingham, Birmingham, United Kingdom B15 2TT
Received 9 July 2001/ Returned for modification 3 October 2001/ Accepted 11 December 2001
 |
ABSTRACT |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
However, the concentration of new fluoroquinolones accumulated by wild-type or efflux mutant S. pneumoniae isolates and any relationship with activity has not been clearly examined. To date, there have been two publications describing methods for measuring the accumulation of fluoroquinolones by S. pneumoniae (18, 19). However, there are key differences between these two studies: they involve different strains of pneumococci, different culture media, different buffers in the accumulation assay, different fluoroquinolones, and different levels of expression of data, making direct comparisons difficult. Zeller et al. (19) published data mostly for the uncapsulated strain R6, whereas Piddock et al. (18) presented data for two capsulated strains of S. pneumoniae NCTC 7465 (M4) and a clinical isolate. One of the key differences between our earlier work and that of Zeller et al. (19) was the demonstration of an efflux pump for fluoroquinolones in wild-type bacteria (indicated by inhibition by CCCP), which was not observed with the strains examined by Piddock et al. (18). This may have been due to differences in strain or method, or as Zeller et al. showed, the CCCP effect was pH dependent, and this in turn was fluoroquinolone dependent. Zeller et al. (19) also demonstrated that addition of glucose to the accumulation medium led to lower concentrations of fluoroquinolone accumulated, presumably due to energizing an efflux pump(s).
Since 1997, there have been no further reports describing fluoroquinolone accumulation by S. pneumoniae, and anecdotal reports from several laboratories have suggested that numerous problems have been encountered in measuring the accumulation of fluoroquinolones by pneumococci. The problems include unreliable growth kinetics (therefore making comparison of similar cultures at the same stage of growth on different occasions difficult), problems in obtaining a good pellet after centrifugation so that unreproducible numbers of bacterial cells were harvested and used in the accumulation experiment, and variable responses to CCCP.
With this background, we sought to establish a reproducible, straightforward method for growing S. pneumoniae such that the cells could be used in an accumulation assay and to compare the procedures used in this laboratory with those of Zeller et al. (19) to measure accumulation. Because the modified fluorometric accumulation assay is an easy procedure and was used in both previous published studies, we sought to establish all of the parameters for growing and measuring accumulation of fluoroquinolone by capsulated S. pneumoniae (because this is the predominant pathogenic form of this species) compared with uncapsulated R6. The concentrations of fluoroquinolones that had accumulated in the presence and absence of either reserpine or CCCP by each strain were compared with the MIC, hydrophobicity, and molecular size of each agent to determine whether any of these parameters influenced the concentration of drug that had accumulated. In addition, two efflux mutants, R6N and M22, were investigated to determine the substrate profile of the efflux pump protein, PmrA.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Antibiotics and susceptibility determination. The MIC of each antibiotic for each strain was determined by the agar doubling dilution method. All antibiotics were gifts and were made up and used according to the manufacturers' instructions: ciprofloxacin and moxifloxacin, Bayer AG, Leverkusen, Germany; sparfloxacin, Rhone DPC Europe, Paris, France; grepafloxacin, GlaxoWellcome, Uxbridge, United Kingdom; gatifloxacin, Grunenthal GmBH, Aachen,Germany; clinafloxacin, Parke-Davis Warner Lambert, Somerville, N.J.; levofloxacin, Hoechst Marion Roussel, Strasbourg, France; sitafloxacin, Daiichi, Tokyo, Japan; norfloxacin, tetracycline, chloramphenicol, and acriflavine, Sigma, St. Louis, Mo.; and ethidium bromide, BDH, Poole, United Kingdom. Plates containing doubling dilutions of antibiotic were inoculated by transferring 1 µl of the undiluted overnight culture to the surface of the agar with a multipoint inoculator (DenleyTech, Billingshurst, United Kingdom) to give a final inoculum of 106 CFU. All plates were incubated in 5% CO2 at 37°C overnight. The MIC of the antibiotic was defined as the lowest concentration of antibiotic (in micrograms per milliliter of agar) at which no more than 10 colonies were detected; a slight haze of growth was ignored. Reserpine (Sigma) was added to a final concentration of 20 µg/ml, and CCCP was added to a final concentration of 100 µM.
Measurement of the accumulation of quinolones by S. pneumoniae. The modified fluorescence method was used essentially as described by Mortimer and Piddock (12) for measuring the concentration of ciprofloxacin, clinafloxacin, gatifloxacin, grepafloxacin, levofloxacin, moxifloxacin, norfloxacin, ofloxacin, sitafloxacin, and trovafloxacin accumulated by each strain. A starter culture (see Results) was added to 100 ml of BHI broth and grown at 37°C in 5% CO2 to an optical density of 0.7 at 660 nm for M3, M4, and M22 and a density of 0.5 for R6 and R6N. After centrifugation at 10,000 x g at 4°C for 10 min, the pellet was washed at 4°C with 0.1 M sodium phosphate at pH 7.0 (except for clinafloxacin, where phosphate buffer at pH 8.0 was used, and trovafloxacin, where buffer at pH 6.0 was used [see Results]). Following further centrifugation as before, the pellet was concentrated to make a suspension with an A660 equivalent to 20 U. The suspension was allowed to equilibrate at 37°C for 10 min prior to accumulation measurements. For some experiments, 0.2% (wt/vol) glucose was added during this equilibration stage. For others, reserpine and CCCP were also added at this stage (see below and Results). Fluoroquinolones were added at a final concentration of 10 µg/ml (except for trovafloxacin, clinafloxacin, and sitafloxacin, which were used at a final concentration of 20 µg/ml due to their poor fluorescence [see Results]). Because it is known that some fluoroquinolones are unstable in the presence of UV light, all tubes were wrapped in aluminum foil to prevent light from entering. Standard curves of fluorescence were obtained for all agents. Except for clinafloxacin and trovafloxacin, cell lysis was achieved in 0.1 M glycine (pH 3) (12); for clinafloxacin and trovafloxacin, cell lysis was achieved by boiling for 10 min (because fluorescence values were higher after lysis by this method). The concentration of fluoroquinolone in the supernatant was determined by measuring fluorescence at the appropriate excitation and emission wavelength for each agent. The accumulation data were converted and expressed as nanograms of quinolone per milligram (dry weight) of cells. Where used, reserpine was added to a final concentration of 20 µg/ml, and CCCP was added to a final concentration of 100 µM. Control experiments with Klebsiella pneumoniae (NCTC 9633) were performed in parallel to confirm the well-documented phenomenon of apparent enhancement of the amount of fluoroquinolone accumulated and to confirm that the CCCP was active. All experiments were performed in duplicate on at least three separate occasions, and the mean values and standard deviations were determined. To determine whether any differences in data were statistically significant, the values obtained after a 5-min exposure to each agent were compared by using Student's two-tailed t test. A P value of <0.05 was considered significant. The steady-state concentration (SSC) values were plotted against the molecular size of the free form of each agent and the partition coefficient (Papp). The Papp was calculated by determining the concentration of agent in the aqeuous phase (0.1 M sodium phosphate buffer [pH 7]) and organic phase (n-octanol) of each agent as described previously (12).
Accumulation of radiolabeled fluoroquinolones. For comparison with the data obtained by the fluorescence method, the accumulation of 3H-levofloxacin (specific activity, 167.5 µCi/µg; Daiichi) and 14C-clinafloxacin was measured (specific activity, 440 µCi/µg; Parke-Davis Warner Lambert) by using the vacuum filtration method and Ultipore filters (16). With a 4% (vol/vol) inoculum, the bacterial strains were grown as before. A 10-ml aliquot of the bacterial culture was preincubated for 10 min at 37°C prior to the addition of fluoroquinolone; 500-µl samples were withdrawn at regular time intervals, and the cells were harvested on the filters, washed with buffer, the filters were dried, and the radioactivity was determined by scintillation counting.
|
RESULTS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
4-fold. The MICs of ciprofloxacin, clinafloxacin, gatifloxacin, and sitafloxacin were also lower in the presence of reserpine, but by only 1 dilution. Efflux mutant M22 was resistant to norfloxacin and ciprofloxacin and less susceptible to ethidium bromide and acriflavine, and reserpine reduced the MICs of these agents. The MICs of ciprofloxacin, clinafloxacin, gatifloxacin, grepafloxacin, moxifloxacin, sitafloxacin, and tetracycline were lowered by 1 dilution.
|
Method of accumulation of fluoroquinolone.
The main difference between the methods used by Zeller et al. (19) and those used by this laboratory was the buffer system used to suspend the cells in which to measure fluoroquinolone accumulation. Comparisons of the buffer system and the effect of glucose (11 mM, 0.2% [wt/vol]) were performed. Zeller et al. (19) showed that glucose enhanced S. pneumoniae efflux and lowered the concentration of fluoroquinolone accumulated by 
3-fold (300%). In the absence of glucose, as in the method used by this laboratory, higher accumulation values were obtained. In the present study, the glucose effect was found to be strain dependent and was much less marked than that seen by Zeller et al. (19). SSC values obtained from experiments with glucose-free buffer were typically within 20% of the values with glucose (Fig. 1). The addition of glucose to the cell suspension in phosphate buffer during the equilibration period resulted in a small but reproducible reduction in the SSCs accumulated by M3, as well as a small increase in the SSCs accumulated by M4. Glucose apparently antagonized the effect of the efflux inhibitors CCCP and reserpine (Fig. 1).
|
To determine whether poor fluorescence of some agents influenced the accumulation data, accumulation of radiolabeled clinafloxacin and levofloxacin was determined, and the data were compared with those obtained with cells lysed in glycine and those lysed by boiling. The radioactive method and the fluorescence method with lysis of the cells achieved by boiling gave very similar data: e.g., M4 accumulated 16.62 ng of 14C-clinafloxacin/mg (dry weight) of cells compared with 16.41 ng/mg (dry weight) of cells when the fluoresence was measured. However, data obtained by lysing the cells in glycine gave one-quarter of the values obtained by the other two procedures.
Accumulation of 10 to 100 µg of levofloxacin per ml by M3 and M4 showed that a linear relationship existed, with a proportional increase in SSC and an increase in the exogenous concentration. Accumulation of fluoroquinolones by M3 and M4 was also measured at 1, 10, 20, 30, and 37°C. Even at 0°C, and in particular for M4 (which produced a lot of capsule), approximately 9 ng/mg (dry weight) of cells was accumulated. As for other bacteria, with an increase in temperature, the SSC increased. Magnesium chloride (20 mM) decreased the concentration of fluoroquinolone accumulated by M3 and M4.
Accumulation of fluoroquinolones by fluoroquinolone-susceptible strains. The accumulation of ciprofloxacin, clinafloxacin, gatifloxacin, grepafloxacin, levofloxacin, moxifloxacin, norfloxacin, ofloxacin, sitafloxacin, and trovafloxacin was determined with M3, M4, and the noncapsulated strain, R6 (Fig. 2). Essentially, the fluoroquinolones fell into two groups based upon the concentration accumulated by the three wild-type strains. The first group (consisting of ciprofloxacin, grepafloxacin, and norfloxacin) accumulated to significantly higher concentrations in all strains than the remainder of the agents. Of these three agents, grepafloxacin acccumulated to the highest concentration (72 to 92 ng/mg [dry weight] of cells), and norfloxacin accumulated to the lowest concentration (35 to 74 ng/mg [dry weight] of cells). The second group (clinafloxacin, gatifloxacin, levofloxacin, moxifloxacin, ofloxacin, sitafloxacin, and trovafloxacin) accumulated to low concentrations in all three strains—between 3 and 30 ng/mg (dry weight) of cells. Compared with M3 and R6, M4 consistently accumulated the lowest concentrations of all fluoroquinolones tested (Fig. 2). Ofloxacin, a racemic mixture of D- and L-isomers, accumulated to higher concentrations for all three strains than levofloxacin (the D-isomer of ofloxacin). These data suggest that the L-isomer is accumulated to a higher concentration than the D-isomer, even though it has poorer antimicrobial activity. Sitafloxacin consistently accumulated to the lowest concentrations in all strains.
|
|
A relationship was also found between the MIC of the agent and the SSC. Although it is counterintuitive, the higher the MIC was, the higher the SSC was, such that the most active agents accumulated to the lowest concentrations (correlation coefficients: M3, 0.84; M4, 0.90; R6, 0.53) (data not shown).
Effect of efflux inhibitors on the concentrations of fluoroquinolones accumulated by fluoroquinolone-susceptible strains. Although reserpine gave more reliable data than CCCP, both were used in all experiments to determine the role of an efflux pump(s) in moderating the concentration of fluoroquinolone accumulated by both strains. In all initial experiments with CCCP and reserpine, it was noted that a similar fold increase (where observed) was produced in the presence of either inhibitor. All experiments were performed in the absence of glucose, so that any efflux pumps were not activated and the action of the inhibitor was antagonized or masked.
The responses of the three strains to the efflux pump inhibitor reserpine differed (Fig. 4). For strain M3, there was little or no effect of reserpine upon the accumulation of eight agents (data not shown). Only for clinafloxacin and moxifloxacin was there a significant difference between accumulation in the presence of reserpine and accumulation without reserpine. Of interest, reserpine decreased the accumulation of moxifloxacin, but increased the accumulation of clinafloxacin by this strain (t test values: clinafloxacin, 0.008; moxifloxacin, 0.011). For strains M4 and R6, there was no significant difference between the accumulation of any fluoroquinolone with or without reserpine.
|
|
To determine whether the lack of any significant effect of both efflux inhibitors on the accumulation of several of the fluoroquinolones was an experimental artifact, CCCP and reserpine were added to the cell suspension during equilibration (i.e., 10 min prior to addition of antibiotic (as was done when determining the effect of glucose). However, no enhancing effect was observed.
Finally to ensure that the data were valid, and because Zeller et al. (19) suggested that the CCCP effect was pH dependent, experiments were performed with ciprofloxacin, norfloxacin, ofloxacin, and sitafloxacin at pH 6.5, 7, and 8 and in the presence and absence of reserpine or CCCP for all strains. It was found that the pH effect was minimal and varied from strain to strain and for each agent, and no consistent effect was observed.
Accumulation of fluoroquinolones in the presence and absence of efflux inhibitors by efflux pump mutants. Strain M4 and its efflux mutant M22 both contain a large deletion in pmrA (unpublished data), and therefore the efflux mutant phenotype exhibited by M22 is not due to PmrA overexpression. M22 accumulated less ciprofloxacin, gatifloxacin, and ofloxacin than did M4 (Fig. 4) (t test values: ciprofloxacin, 0.02; gatifloxacin, 0.004; ofloxacin, 0.05). Addition of reserpine significantly increased the concentration of clinafloxacin (t test, 0.035). M22 consistently accumulated a higher concentration of norfloxacin than its isogenic parent strain, M4, and reserpine enhanced this concentration to an even higher level (t test, 0.02). CCCP enhanced the concentration of grepafloxacin accumulated by both M4 and M22 to similar levels. CCCP had no significant effect on the accumulation of any other fluoroquinolone by strain M22 (data not shown).
Differences in the concentration of fluoroquinolone accumulated by R6 and R6N and the effect of CCCP and reserpine were found to be highly dependent upon the growth phase of these strains. In rigorously controlled experiments, there was a significant difference (t test value, 0.023) between the concentrations of norfloxacin accumulated by R6N and strain R6 (Fig. 5). The apparent difference in the levels of accumulation of ciprofloxacin between R6 and R6N was not confirmed statistically (t test value, 0.351). For R6N, the effect of reserpine and CCCP was found to only occur at the early- to mid-logarithmic phase; despite this, these efflux inhibitors had little or no significant effect upon the concentration of any other fluoroquinolone accumulated by strain R6N.
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
As might have been predicted, the more hydrophobic molecules accumulated to lower concentrations than the more hydrophilic agents. For all strains tested, the most hydrophilic molecules, norfloxacin and ciprofloxacin, were accumulated the most readily, especially by M3. One notable exception was grepafloxacin. This agent is strongly hydrophobic, yet was readily accumulated by both strains to high concentrations. This may suggest that accumulation by passive diffusion across the cell membrane may not be the only way in which S. pneumoniae is able to accumulate fluoroquinolones.
The relationship between the molecular size of the free form of each agent and SSC was such that the compounds with the lowest molecular sizes, notably ciprofloxacin and norfloxacin, were accumulated by both strains to the highest concentrations. Fluoroquinolones with high molecular weights, such as trovafloxacin and moxifloxacin, were accumulated to the lowest concentrations; however, grepafloxacin was again a notable exception in this respect. It has a high molecular weight, similar to that of trovafloxacin and moxifloxacin, but the concentration accumulated was as high for grepafloxacin as for norfloxacin and ciprofloxacin.
Active efflux as a mechanism of low-level fluoroquinolone resistance in pneumococci has been proposed by several groups (1, 4, 18, 19). We were interested to investigate if any of the wild-type strains in our study showed evidence of efflux, as is the case for wild-type Pseudomonas aeruginosa. The efflux inhibitors reserpine and CCCP were used in a series of experiments measuring fluoroquinolone accumulation by S. pneumoniae. CCCP will only inhibit efflux by proteins that use the proton motive force for energy. For pumps that use ATP, such as ABC transporters, CCCP will have no effect. Such transporters have already been described for S. pneumoniae (14). Experiments with fluoroquinolones and CCCP can also be complicated in their interpretation due to a pH effect on the fluoroquinolone (altering its charged state), and the charged state can influence the rate of accumulation by the bacterial cell (13). Zeller et al. (19) previously demonstrated a CCCP effect for strain R6 when accumulation of several quinolones was measured; however, this effect appeared to be pH dependent as well as agent dependent. No difference was found in the concentration accumulated or in the effect of CCCP between pH 6.5 and 8.0 for ciprofloxacin, confirming the data of the present study. Zeller et al. (19) showed enhanced accumulation of pefloxacin and sparfloxacin at pH 6.5 with a CCCP effect, which was abolished at pH 8.0. Some of the agents examined in the present study fluoresce poorly (as does sparfloxacin), and we postulate that pH may affect fluorescence and hence the CCCP effect. For some agents, the action of CCCP was not as expected, and instead of enhancing the concentration of fluoroquinolone accumulated (due to inhibition of an efflux pump), the opposite effect was observed (i.e., a reduction in the SSC). These data suggest that CCCP itself may be a substrate for one or more efflux pumps in S. pneumoniae and that there is competition between the fluoroquinolone and CCCP. Further experiments are in progress.
Reserpine had a clear effect on susceptibility of the wild-type strains M3 and M4 and reduced the MIC of fluoroquinolones by one to two tubes (two- to fourfold). Therefore, it was expected that the MIC synergy would be mirrored by similar enhancements of the concentration of fluoroquinolone accumulated. However, reserpine, like CCCP, had little or no effect upon accumulation of most fluoroquinolones by all strains, including R6N, which overexpresses PmrA. There are two explanations for these data: (i) most fluoroquinolones are not effluxed by any efflux pump proteins of S. pneumoniae, or (ii) the low concentrations of most fluoroquinolones accumulated are due to much of the intracellular agent being effluxed, but by pump proteins insensitive to the concentration of reserpine used in the present study. Due to the amino acid sequence similarity between Bmr and the predicted amino acid sequence of PmrA (35% identity), it has been assumed that reserpine inhibits Pmr in a similar manner to Bmr, and this gives rise to the synergistic effect observed in MIC studies between many fluoroquinolones and reserpine. Reserpine interacts directly with the Bmr protein at amino acids phenylalanine 143, valine 286, and phenylalanine 306 (10). Although these amino acids are distant from each other, when the protein is folded into its tertiary structure, it is proposed that these amino acids form part of a reserpine binding pocket (10). It has been found that if valine 286 is substituted for with another amino acid, the affinity of reserpine is increased or decreased. Substitution of valine 286 with the larger residue, leucine, results in a fourfold reduction in sensitivity of Bmr to reserpine. Leucine 286 is found in Staphylococcus aureus NorA, and it was suggested that a higher concentration of reserpine would be needed to give an inhibitory effect. Substitution of amino acids at the other loci also affected reserpine sensitivity and can also give resistance to fluoroquinolones and other agents. It is clear that alignment of PmrA with NorA and Bmr indicates that while there is overall homology between the amino acid sequences of the three proteins, there are differences in the putative reserpine binding site. NorA contains two of the three amino acids important in binding reserpine in Bmr, whereas PmrA does not contain any of these residues. PmrA contains tyrosine 143 and glycine 306; these amino acids may not be employed in reserpine binding. PmrA also contains the larger residue, tryptophan 286, which may significantly reduce the sensitivity of PmrA to reserpine. These data, coupled with the lack of effect of reserpine in the accumulation experiments with R6N, suggest that the synergistic effect seen in the MIC experiments with this strain is not due to interaction between a reserpine binding pocket in PmrA analogous to the pocket in Bmr. It would seem likely that reserpine either interacts elsewhere in PmrA or with another target in the bacterial cell, possibly another efflux pump protein, although the accumulation data do not support the latter hypothesis.
The effects of reserpine and CCCP upon norfloxacin accumulation by R6N were highly dependent upon the phase of growth that the cells were harvested. Gill has recently suggested that expression of pmrA is dependent upon growth phase (M. J. Gill, Proc. 7th Int. Symp. New Quinolones, abstr. 4.4, 2001). Therefore, the amount of antibiotic effluxed via PmrA will also depend upon the growth phase.
For the efflux mutants M22 and R6N, the data obtained suggest that each strain has increased expression of a different pump. Genetic data for R6N demonstrate that this is PmrA (7). M22 had greater resistance to norfloxacin and ciprofloxacin and decreased susceptibility to several other fluoroquinolones than did R6N. Except for norfloxacin, this was associated with less fluoroquinolone accumulated. Although counterintuitive, M22 accumulated more norfloxacin than the parent strain, M4. Since little or no expression of pmrA was detected in M4 or M22 (L. J. V. Piddock et al., unpublished data), the identity of the pump(s) in M22 is currently being sought. It is likely that S. pneumoniae has multiple efflux pumps that can export antibiotics as in other bacteria. However, it appears that the only fluoroquinolone effluxed by PmrA is norfloxacin, and the pump postulated to be overexpressed or have an altered substrate profile in M22 has a broader substrate profile, including ciprofloxacin and certain other fluoroquinolones. Finally, until other efflux pump proteins are identified in S. pneumoniae, and inhibitors of these identified the contribution of efflux to any inherent resistance to fluoroquinolones and other antibiotics cannot be determined.
|
ACKNOWLEDGMENTS |
|---|
We are grateful to Martin Gill and Nigel Brenwald for the gift of S. pneumoniae R6 and R6N and for helpful discussions during preliminary work with their strains. We are also grateful to Helen Kent for performing preliminary experiments and initial work on method development.
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Clin. Vaccine Immunol. | Clin. Microbiol. Rev. |
|---|---|
| J. Clin. Microbiol. | ALL ASM JOURNALS |
